Acid buffering and sulfate retention in
SO
2
-impacted Andosols
–
a multi-methodological approach
Von Diplom-Agrarbiologin Andrea Franziska Herre
Fakultät VI – Planen Bauen Umwelt
der Technischen Universität Berlin
zur Erlangung des akademischen Grades
Doktor der Naturwissenschaften
-Dr. rer. nat.-
genehmigte Dissertation
Promotionsausschuss:
Vorsitzender: Prof. Dr. Gerd Wessolek
Gutachter: Prof. Dr. Martin Kaupenjohann
Gutachter: apl. Prof. Dr. Jörg Prietzel
Gutachter: PD Dr. habil Friederike Lang
Tag der wissenschaftlichen Aussprache: 27. März 2009
Berlin 2009
D 83
V
TABLE OF CONTENTS
Table of Contents....................................................................................................V
List of Figures ........................................................................................................XI
List of Tables........................................................................................................ XV
Abstract ............................................................................................................. XVIII
Zusammenfassung...............................................................................................XX
1 General Introduction.........................................................................................1
1.1 Acid buffering in soils............................................................................................. 1
1.2 Sulfate retention in SO
2
-affected soils .................................................................. 3
1.3 Sulfate retention by allophane ............................................................................... 4
1.4 Identification of BAS phases in soils..................................................................... 5
1.5 Research objectives................................................................................................ 6
2 Mechanisms of acid buffering and formation of secondary minerals in
vitric Andosols..................................................................................................9
2.1 Summary................................................................................................................... 9
2.2 Introduction............................................................................................................ 10
2.3 Materials and Methods.......................................................................................... 11
2.3.1 Study sites .................................................................................................. 12
2.3.1.1 Poás Volcano............................................................................... 12
2.3.1.2 Masaya Volcano.......................................................................... 12
2.3.2 Estimation of S-deposition, soil sampling and description.......................... 13
2.3.2.1 Poás Volcano............................................................................... 13
2.3.2.2 Masaya Volcano.......................................................................... 15
2.3.3 Sample preparation..................................................................................... 15
2.3.4 Short-term acidification experiments at pH 3.............................................. 17
2.3.5 Fourier-transformed diffuse reflectance infrared spectroscopy.................. 18
2.3.6 Soil saturation extracts................................................................................ 18
2.3.7 Microscopic methods..................................................................................... 19
2.4 Results and Discussion ........................................................................................ 20
2.4.1 Mechanisms of H
+
buffering at Poás........................................................... 20
2.4.1.1 Short-term buffering under cation removal.................................. 20
2.4.1.2 Short-term buffering under cation accumulation ......................... 24
2.4.1.3 Relevance for the acid buffering under field conditions............... 25
2.4.2 Mechanisms of H
+
buffering at Masaya...................................................... 28
2.4.2.1 Short-term buffering under cation removal.................................. 28
2.4.2.2 Short-term buffering under cation accumulation ......................... 29
VI
2.4.2.3 Relevance for the acid buffering under field conditions...............29
2.4.3 Formation of secondary minerals................................................................30
2.5 Conclusions............................................................................................................33
3 Simultaneous thermal analysis for the identification of S-pools in soils –
possibilities and limitations............................................................................35
3.1 Summary.................................................................................................................35
3.2 Introduction............................................................................................................ 36
3.3 Materials and Methods..........................................................................................37
3.3.1 Experimental approach...............................................................................37
3.3.2 Sulfur containing reference compounds .....................................................38
3.3.3 Matrix components for mixtures..................................................................38
3.3.4 Synthesis of sample Ref
ads
(dominantly sulfate adsorbed to allophane).... 39
3.3.5 Pure samples and mixtures of reference compounds with matrix
components for STA-MS analysis...............................................................40
3.3.6 Simultaneous thermal analysis coupled with mass spectrometry
(STA-MS) ....................................................................................................40
3.4 Results.................................................................................................................... 42
3.4.1 Characterization of the synthesized S containing reference compounds
by X-ray diffraction......................................................................................42
3.4.2 Mass spectrometry signals of STA-MS analysis.........................................43
3.4.2.1 Pure S containing reference compounds ....................................43
3.4.2.2 Mixtures of S containing reference compounds with matrix
components....................................................................................44
3.4.2.2.1 Mixtures with allophane .......................................................44
3.4.2.2.2 Mixtures with SiO
2
................................................................45
3.4.2.2.3 Mixtures with Al
2
O
3
...............................................................46
3.4.2.3 Mass spectrometry signals of sample Ref
ads
...............................47
3.5 Discussion..............................................................................................................48
3.5.1 Is a differentiation between estersulfate phases, alunite and basaluminite
possible by STA-MS analysis? ................................................................... 48
3.5.2 Can “sulfate adsorbed to allophane” be distinguished from BAS and
estersulfate phases by STA-MS? ............................................................... 49
3.6 Conclusions............................................................................................................50
4 In situ precipitation of basic aluminum sulfate phases at the allophane-
water interface.................................................................................................51
4.1 Summary.................................................................................................................51
4.2 Introduction............................................................................................................ 52
4.3 Materials and Methods..........................................................................................54
VII
4.3.1 Experimental approach............................................................................... 54
4.3.2 Experimental sites....................................................................................... 55
4.3.2.1 Masaya Volcano.......................................................................... 55
4.3.2.2 Poás Volcano............................................................................... 55
4.3.3 Synthesis of glass, allophane and S-containing reference compounds..... 56
4.3.3.1 Synthesis of glass........................................................................ 56
4.3.3.2 Synthesis of allophane ................................................................ 57
4.3.3.3 Synthesis of S-containing reference compounds........................ 57
4.3.4 Field exposure of glass, allophane and ion exchange resins..................... 58
4.3.5 Resin analysis............................................................................................. 59
4.3.6 Total carbon and sulfur analysis................................................................. 59
4.3.7 Scanning electron microscopy with energy dispersive X-ray analysis
(SEM-EDX) ................................................................................................. 59
4.3.8 Simultaneous thermal analysis coupled with mass spectrometry
(STA-MS).................................................................................................... 60
4.3.9 Sulfur K-edge X-ray absorption near edge spectroscopy (S K-edge
XANES)....................................................................................................... 60
4.3.10 Attenuated total reflectance Fourier-transformed infrared spectroscopy
(ATR-FTIR) ................................................................................................. 61
4.3.11 Wet chemical extraction methods............................................................... 62
4.3.11.1 Sulfate extractability .................................................................... 62
4.3.11.2 Oxalate extraction........................................................................ 62
4.4 Results.................................................................................................................... 63
4.4.1 Sulfate and aluminum fluxes in soils........................................................... 63
4.4.2 Total carbon and sulfur contents of glass and allophane ........................... 64
4.4.3 SEM-EDX of glass and allophane............................................................... 65
4.4.3.1 SEM-EDX of glass samples ........................................................ 65
4.4.3.2 SEM-EDX of allophane................................................................ 67
4.4.4 STA-MS....................................................................................................... 70
4.4.5 S K-edge XANES........................................................................................ 72
4.4.6 ATR-FTIR.................................................................................................... 74
4.4.7 Sulfate extractability.................................................................................... 74
4.4.8 Oxalate extraction of allophane .................................................................. 75
4.5 Discussion.............................................................................................................. 77
4.5.1 Effects of field exposure on allophane and glass characteristics............... 77
4.5.2 Do BAS precipitates preferentially form in situ at the allophane-water
interface? .................................................................................................... 79
4.6 Conclusions ........................................................................................................... 81
VIII
5 Sulfate retention by allophane – adsorption or precipitation?
Part 1. Isotherms........................................................................................83
5.1 Summary.................................................................................................................83
5.2 Introduction............................................................................................................ 84
5.3 Materials and Methods..........................................................................................86
5.3.1 Experimental approach...............................................................................86
5.3.2 Synthesis of allophane................................................................................87
5.3.3 Synthesis of sulfate-containing reference compounds...............................88
5.3.4 Sorption isotherms at pH 4.0, 4.5 and 5.0 .................................................. 88
5.3.5 Simultaneous thermal analysis coupled with mass spectrometry
(STA-MS) ....................................................................................................89
5.3.6 Sulfur K-edge X-ray absorption near edge spectroscopy (S K-edge
XANES).......................................................................................................90
5.3.7 Scanning electron microscopy with energy-dispersive X-ray analysis
(SEM-EDX)..................................................................................................91
5.3.8 Sulfate extractability....................................................................................91
5.3.9 Porosity measurements by N
2
adsorption................................................... 91
5.4 Results and Discussion ........................................................................................92
5.4.1 Analysis of sulfate, aluminum and silica concentrations and of acid
consumption................................................................................................92
5.4.2 STA-MS.......................................................................................................95
5.4.3 S K-edge XANES........................................................................................97
5.4.4 SEM-EDX.................................................................................................... 99
5.4.5 Sulfate extractability....................................................................................99
5.4.6 Porosity measurements ............................................................................100
5.4.7 Influence of sulfate concentration and pH on sulfate retention processes101
5.4.8 Ecological implications..............................................................................105
5.5 Conclusions..........................................................................................................106
6 Sulfate retention by allohpane – adsorption of precipitation?
Part 2. Kinetics..............................................................................................107
6.1 Summary...............................................................................................................107
6.2 Introduction..........................................................................................................108
6.3 Materials and Methods........................................................................................110
6.3.1 Experimental approach.............................................................................110
6.3.2 Synthesis of allophane..............................................................................111
6.3.3 Synthesis of sulfate-containing reference compounds.............................112
6.3.4 Retention experiments..............................................................................112
IX
6.3.5 Simultaneous thermal analysis coupled with mass spectrometry
(STA-MS).................................................................................................. 114
6.3.6 Sulfur K-edge X-ray absorption near edge spectroscopy (S K-edge
XANES)..................................................................................................... 115
6.3.7 Scanning electron microscopy with energy dispersive X-ray analysis
(SEM-EDX) ............................................................................................... 116
6.3.8 Attenuated total reflectance Fourier-transformed infrared spectroscopy
(ATR-FTIR) ............................................................................................... 116
6.3.9 Wet chemical extraction methods............................................................. 117
6.3.9.1 Sulfate extractability .................................................................. 117
6.3.9.2 Oxalate extraction...................................................................... 118
6.4 Results and Discussion ...................................................................................... 118
6.4.1 Analysis of sulfate, aluminum and silica concentrations and of acid
consumption.............................................................................................. 118
6.4.1.1 Sulfate retention kinetics at pH 4.5............................................ 118
6.4.1.2 Sulfate retention kinetics at pH 5.0............................................ 120
6.4.2 ATR-FTIR.................................................................................................. 122
6.4.3 Oxalate extraction..................................................................................... 123
6.4.4 STA-MS..................................................................................................... 124
6.4.4.1 Differential scanning calorimetry ............................................... 124
6.4.4.2 Mass spectrometry signals........................................................ 124
6.4.5 S K-edge XANES...................................................................................... 126
6.4.6 SEM-EDX.................................................................................................. 126
6.4.7 Sulfate extractability.................................................................................. 127
6.4.8 Allophane dissolution / transformation during the pH 4.5 experiment...... 128
6.4.9 Kinetics of sulfate retention processes and consequences for the sulfate
extractability.............................................................................................. 129
6.5 Conclusions ......................................................................................................... 131
7 Extended Summary, General Conclusions and Outlook............................133
7.1 Extended Summary............................................................................................. 133
7.1.1 Which processes are relevant for the effective acid neutralization in SO
2
-
impacted Andosols and which consequences do these processes have for
secondary mineral formation (chapter 2)?................................................ 134
7.1.2 Do soil hydrological conditions influence acid buffering processes
(chapter 2)?............................................................................................... 134
7.1.3 Evaluation of the suitability of STA-MS for the differentiation between BAS
phases, organically bound estersulfates and adsorbed sulfate
(chapter 3)................................................................................................. 135
X
7.1.4 Are BAS phases preferentially formed in situ at the allophane-water
interface (chapter 4)?................................................................................ 135
7.1.5 Effects of sulfate concentration and pH on the relative contribution of
adsorption and BAS precipitation to the total sulfate retention by allophane
(chapter 5).................................................................................................135
7.1.6 How do the relative contributions of adsorption and BAS precipitation to the
total sulfate retention by allophane change with time (chapter 6)?........... 136
7.2 General Discussion and Conclusions...............................................................136
7.2.1 Acid buffering in SO
2
-impacted Andosols of the (sub)tropics – a
comparison with soils of heavily acid-affected sites in the temperate
regions.......................................................................................................136
7.2.2 Suitability of different methods for BAS identification ............................... 138
7.2.3 Contribution of adsorption processes and BAS precipitation to the total
sulfate retention in SO
2
-impacted Andosols .............................................140
7.2.4 Ecological implications of BAS formation in SO
2
-impacted Andosols ...... 141
7.3 Outlook..................................................................................................................142
8 References.....................................................................................................145
Acknowledgements.............................................................................................159
Curriculum vitae ..................................................................................................161
Appendix..............................................................................................................A-1
XI
LIST OF FIGURES
Figure 2.1 Aerial photographs of the Poás Volcano with sampling sites P1 - P5 (a)
and the Masaya Volcano with sampling sites M1 and M2 (b) (Arrows start from
the active craters and indicate the main wind direction).
Figure 2.2 Acid neutralizing capacity, released DOC, Al, Si, Ca and Na (mean
values and standard errors) during the resin experiment for Poas (P1 - P5) and
Masaya (M1 & M2) samples.
Figure 2.3 Acid neutralizing capacity, released DOC, Al, Si, Ca and Na (mean
values and standard errors) during the pH
stat
titration experiment for Poas (P1 -
P5) and Masaya (M1 & M2) samples.
Figure 2.4 Scanning electron microscope images of thin sections of grains from
the 2AC horizons of sample P5 (4a, 4b) and sample P1 (4c, 4d) showing repre-
sentative plagioclase, magnetite and pyroxene crystals in a glass matrix. Larger
mineral crystals have been denominated with P for plagioclase, Py for pyroxene
and M for magnetite.
Figure 2.5 Depth profiles of saturation indices (SI) for amorphous SiO
2
in soil satu-
ration extracts (calculated with the geochemical model Visual Minteq) in profiles
P1 - P5 of Poás Volcano and M1 - M2 of Masaya Volcano. (Dotted lines repre-
sent the lower and upper limits of the SI, indicating an equilibrium between soluti-
on concentrations and amorphous SiO
2
).
Figure 2.6 Scanning electron microscope image of a sand grain (a) with EDX ana-
lyses for points A (b) and B (c). The grain originates from the 2AC horizon of pro-
file P1, Poás.
Figure 3.1 XRD patterns of the synthesized S containing reference compounds
(a) amorphous basaluminite (b) K-alunite and (c) Na-alunite
Figure 3.2 SO and H
2
O mass spectrometry signals during STA-MS analysis of the
pure reference compounds basaluminite, N-alunite, K-alunite and Na-
dodecylsulfate (S
org
).
Figure 3.3 SO and H
2
O mass spectrometry signals during STA-MS analysis of the
reference compounds basaluminite, N-alunite, and Na-dodecylsulfate (S
org
) in
mixture with allophane. In the case of S
org
SO signals of two mixtures with diffe-
XII
rent mixing ratios are shown (dashed line: mixing ratio 19:1, solid line: mixing ra-
tio 107:1).
Figure 3.4 SO and H
2
O mass spectrometry signals during STA-MS analysis of the
reference compounds basaluminite, N-alunite, and Na-dodecylsulfate (S
org
) in
mixture with SiO
2
.
Figure 3.5 SO and H
2
O mass spectrometry signals during STA-MS analysis of the
reference compounds basaluminite, N-alunite, and Na-dodecylsulfate (S
org
) in
mixture with Al
2
O
3
.
Figure 3.6 SO and H
2
O mass spectrometry signals during STA-MS analysis of
sample Ref
ads
.
Figure 4.1 Sulfur and aluminum fluxes (in kg ha
-1
) in layers 1-4 of resin boxes from
Masaya (left column) and Poás (right column) sites over the whole experimental
period of 18 months
Figure 4.2 Scanning electron microscopy images of pristine glass (a,b) and glass
that had been exposed in soils close to Masaya (c,d) and Poás (e,f). Arrows in
Figure 4.2e indicate S-rich and SiO
2
-rich particles, circles indicate C-rich, Fe-
containing phases.
Figure 4.3 Scanning electron microscopy images of pristine allophane (a,b) and
allophane that had been exposed in soils close to Masaya (c,d) and Poás (e,f).
Arrows indicate C-rich, Fe-containing phases.
Figure 4.4 Scanning electron microscopy images (backscattered electrons) and
mapping of energy dispersive X-ray analysis of S, Al and Si (in this sequence
from top to bottom) of allophane samples after field exposure close to Masaya
(left column) and Poás (right column)
Figure 4.5 Scanning electron microscopy image of S-rich particle with crystalline
aspect on the surface of an allophane aggregate after exposure in soil close to
Poás Volcano.
Figure 4.6 SO-signals of STA-MS analysis of (a) basaluminite, (b) Na-alunite,
(c) Na-dodecylsulfate, (d) Ref
ads
, (e) allophane sample exposed at Masaya and
(f) allophane sample exposed at Poás. Reference compounds were analyzed in
physical mixture with allophane. For Masaya and Poás samples zoomed sket-
ches of SO signals up to 600°C are additionally shown.
XIII
Figure 4.7 (a) S K-edge X-ray absorption near edge spectra of K-alunite, Na-alunite,
basaluminite, Na-dodecylsulfate, Ref
ads
and representative allophane samples
which had been exposed in soils close to Masaya and Poás. (b) magnified sketch
of superimposed spectra of allophane samples from Poás and Masaya shows
that Poás samples contain organic S-containing phases.
Figure 4.8 ATR-FTIR spectra of pristine allophane (solid line) and allophane after
field exposure (dashed lines) at Masaya (a) and Poás (b). Deviations of field ex-
posed allophane from pristine allophane can be explained by (i) S retention and
(ii) release of polymeric silica.
Figure 4.9 Sulfate extractability of the reference phases basaluminite, K-alunite, Na-
alunite, Na-dodecylsulfate and Ref
ads
and of allophane samples after exposure in
soils close to Masaya and Poás (mean values +/- standard errors).
Figure 4.10 Oxalate extractable Si, Al and Fe of pristine allophane and allophane
after field exposure at Masaya and Poás. Fe concentrations in oxalate extracts of
pristine allophane and Masaya samples were below the detection limit.
Figure 5.1 Sulfate retention, Al and Si net release and acid consumption for iso-
therms at pH 4.0 (black), pH 4.5 (grey) and pH 5.0 (white). Graphs show mean
values ± standard errors.
Figure 5.2 Saturation indices for the phases (a) basaluminite and (b) Al(OH)
3
for
isotherms at pH 4.0 (black), pH 4.5 (grey) and pH 5.0 (white). For the calculation
of SI
Al(OH)3
the logK = 10.8 was taken from Nordstrom et al, 1990. Graphs show
mean values ± standard errors.
Figure 5.3 Mass spectrometry signals of m/z=48 (SO) during STA-MS analysis of
amorphous basaluminite, Ref
ads
and of samples I15 of the isotherms at pH 4.0,
4.5 and 5.0.
Figure 5.4 S K-edge XANES spectra of amorphous basaluminite, Ref
ads
and of
representative measurements for samples I15 of the isotherms at pH 4.0, 4.5 and
5.0.
Figure 6.1 Sulfate retention, Al and Si net release and acid consumption during the
kinetic experiment at pH 4.5. Graphs show mean values ± standard errors. In the
case of Al a second and smaller graph zooms into Al release up to 200 hours.
XIV
Figure 6.2 Sulfate retention, Al and Si net release and acid consumption during the
kinetic experiment at pH 5.0. Graphs show mean values ± standard errors. In the
case of Al a second and smaller graph zooms into Al release up to 50 hours.
Figure 6.3 ATR-FTIR spectra of pure allophane and solid sample aliquots taken
from the kinetic experiment after 48 hours, 2 weeks, 2 months and 6 months.
Figure 6.4 Mass spectrometry signals of m/z=48 (SO) during STA-MS analysis
amorphous basaluminite, Ref
ads
and samples taken after 24 hours, 2 weeks, 1
month, 2 months and 6 months from the kinetic experiment.
Figure 6.5 S K-edge XANES spectra of amorphous basaluminite, Ref
ads
and sample
aliquots taken after 2 hours, 8 hours, 2 weeks and 2 months of the kinetic expe-
riment.
Figure 6.6 Scanning electron microscopy images of (a) pristine allophane and (b) a
sample aliquot taken after 2 months of the kinetic experiment
XV
LIST OF TABLES
Table 2.1 Literature reports on acid buffering mechanisms in Andosol subtypes.
Table 2.2 Sulfation rates (mg SO
2
m
-2
day
-1
) at soil profile sampling sites at Poás
and Masaya.
Table 2.3 Soil sample characteristics at Poás (P1 - P5) and Masaya (M1 - M2).
Table 2.4 Amounts of total released Al (Al
t
), Ca and Al released by plagioclase
dissolution (Ca
plag
, Al
plag
) and Al released from Al-organic complexes (Al
org
) in the
resin and pH
stat
titration experiments of Poás samples.
Table 2.5 Comparison of the acid neutralizing capacity in the resin experiment
attributable to protonation of variable charge on organic matter (ANC
prot
) and to
plagioclase dissolution (ANC
plag
, as the difference between total ANC and
ANC
prot
) with estimated acid deposition rates at the Poás Volcano.
Table 2.6 Comparison of the acid neutralizing capacity in the pH
stat
experiment
attributable to protonation of variable charge (ANC
prot
) and to mineral dissolution
(ANC
diss
) with estimated acid deposition rates
a)
at Masaya Volcano.
Table 3.1 Sample numbering, composition, mixing ratios (matrix:reference com-
pound) and sample weights for STA-MS analysis.
Table 4.1 Carbon and sulfur concentrations of glass and allophane samples before
(pristine) and after exposure in soils close to Masaya and Poás Volcanoes.
Table 5.1 Precipitated sulfate (% of totally retained sulfate) calculated by fitting
Gaussian peaks to SO signals of STA-MS as a function of pH and sulfate con-
centration of added solutions in the isotherm experiment.
Table 5.2 Precipitated sulfate (% of totally retained sulfate, mean values ± standard
error, in parenthesis: minimum and maximum values) determined by linear com-
bination fitting of S K-edge XANES spectra as a function of pH and sulfate con-
centration of added solutions in the isotherm experiments. Amorphous basalumi-
nite and Ref
ads
were applied as endmembers in LCF.
Table 5.3 Mesopore volumes (MePV), micropore volumes (MiPV) and average
micropore diameters (AMiPD) of pristine allophane and samples I1 and I15 from
the pH 4.5 and pH 4.0 experiment. Values are given as means ± standard error.
XVI
Table 6.1 Oxalate extractable aluminum and silica and molar Al:Si ratios from solid
sample aliquots taken after 24 hours, 1 month, 2 months and 6 months of the ki-
netic experiment.
XVII
XVIII
ABSTRACT
Large deposition of sulfuric acid to soils in close vicinity of active volcanoes affects
the dynamics of sulfur and other nutrients and can have toxic effects on animals,
plants and microorganisms. Both acid buffering and sulfur retention processes are
hardly investigated for Andosols of the (sub-)tropics, however. The objectives of my
study were (i) to estimate the contributions of different processes to the effective
acid neutralization capacity and (ii) to determine the relevance of adsorption and
basic aluminum sulfate (BAS) precipitation for the total sulfate retention in SO
2
-
impacted Andosols. Acid buffering was studied in short-term pH
stat
laboratory ex-
periments with samples taken at two sites which differ in climatic conditions and
soils (Masaya: wet-dry cycles, vitric Andosols developing towards the sil-andic
subtype; Poás: perhumid, vitric Andosols developing towards the alu-andic subtype).
Additionally, samples from a transect of increasing acid input at Poás were analyzed
by infrared spectroscopy (IR) and scanning electron microscopy with energy disper-
sive X-ray analysis (SEM-EDX) to identify the processes which are relevant for the
long-term acid buffering under field conditions. For the identification of S retention
processes a combination of macroscopic field and laboratory experiments with
synthetic allophane and glass were combined with microscopic and spectroscopic
techniques (SEM-EDX; S K-edge X-ray absorption near edge spectroscopy,
XANES; simultaneous thermal analysis coupled with mass spectrometry, STA-MS;
IR). The suitability of STA-MS to differentiate S pools in samples was evaluated
before by the analysis of synthetic reference compounds in pure form and in mix-
tures with simple matrix components. Thereby I could show that organically bound
sulfur, basaluminite, alunite and adsorbed sulfate can be distinguished from each
other in pure form by STA-MS. However, matrix components distort STA-MS signals
thus limiting this method to samples of simple and known matrix composition.
The protonation of variable charge of organic matter and mineral / glass dissolution
dominate the short-term acid buffering at pH 3.0 in vitric Andosols. Protonation of
variable charge is of larger relevance in Poás than in Masaya samples due to larger
contents of organic matter at Poás. Accumulation of base cations and aluminum in
solution hamper mineral weathering. This is likely the case during the dry season at
Masaya leading to a decrease in the effective acid neutralization capacity for this
site. Under field conditions, protonation of variable charge is effectively reversed by
Abstract
XIX
cation and aluminum release from mineral / glass weathering which is thus the
dominant long-term acid buffering process at both study sites.
After an 18 months field experiment close to Masaya and Poás Volcanoes no
measurable in situ BAS formation could be detected on surfaces of glass particles.
Contrastingly, BAS contributed with 11% and 19% to the total sulfur retention by
synthetic allophane exposed in soils close to Masaya and Poás Volcanoes, respec-
tively. The presence of BAS in allophane samples exposed at Poás showed that
large organic matter contents and perhumid climatic conditions do not impede BAS
formation.
Laboratory pH
stat
experiments showed that decreasing pH, increasing sulfate load
and increasing time tend to increase the relevance of BAS for S retention by allo-
phane but the effects were small and adsorption processes dominated with 86% to
98% the total sulfate retention. The observed slow increase in sulfate adsorption
within 6 months could be attributed to an increased availability of existing adsorption
sites due to silica depolymerization and desorption and to the formation of new
adsorption sites due to an increase in defect sites of the allophane spherules.
Although BAS precipitation increases the S retention capacity of SO
2
-impacted
allophane containing Andosols, no measurable effect on the S extractability could be
identified in laboratory experiments. Whether in situ formed BAS decrease the sul-
fate extractability deserves further investigation. Additionally, the elucidation of the
process underlying BAS precipitation, i.e. surface precipitation or solution supersatu-
ration in void spaces, would help to predict the effects of further parameters – as
e.g. the concentrations of fluoride, DOC or phosphate or the allophane aggregation
degree – on BAS formation.
XX
ZUSAMMENFASSUNG
Die Deposition von Schwefelsäure auf Böden in der Umgebung aktiver Vulkane
beeinflusst den S-Haushalt sowie die Dynamik anderer Nährelemente und kann
toxisch auf Tiere, Pflanzen und Mikroorganismen wirken. Die Prozesse der Säure-
pufferung und der S-Retention sind für vulkanische Böden der (Sub)tropen kaum
untersucht. Die Ziele meiner Arbeit waren deshalb, (i) die Relevanz verschiedener
Prozesse für die Säurepufferung abzuschätzen und (ii) die Bedeutung der Sulfatad-
sorption und der Fällung von basischen Aluminium-Hydroxosulfaten (BAS) für die S-
Retention in SO
2
-beeinflussten vulkanischen Böden zu bestimmen. Pufferprozesse
wurden mithilfe von Kurzzeit-pH
stat
-Versuchen an Bodenproben zweier klimatisch
und bodenkundlich unterschiedlicher Standorte identifiziert (Masaya: wechselfeucht,
Präsenz von Allophan; Poás: perhumid, Dominanz von Al-Humus-Komplexen).
Zusätzlich wurden für den Standort Poás Bodenproben entlang eines Transektes
abnehmender Säuredeposition mithilfe von IR-Spektroskopie (IR) und Rasterelekt-
ronenmikroskopie (SEM-EDX) untersucht, um die Relevanz der zuvor identifizierten
Pufferprozesse für die Säureneutralisation unter Freilandbedingungen zu prüfen.
Für die Untersuchung der S-Retentionsprozesse habe ich makroskopische Freiland-
und Laborexperimente mit synthetischem Allophan und Glas mit spektroskopischen
und mikroskopischen Methoden kombiniert (SEM-EDX; Röntgenabsorptionsspekt-
roskopie, S K-edge XANES; Thermische Analyse gekoppelt mit Massenspektrosko-
pie, STA-MS; IR). Die Eignung von STA-MS für die Identifizierung verschiedener S-
Pools in Böden habe ich zuvor anhand der Analyse synthetischer Referenzsubstan-
zen in Reinform und in Mischung mit Mineralen (Allophan, SiO
2
, Al
2
O
3
) getestet.
Dabei zeigte sich, dass mit dieser Methode Basaluminit, Alunit, Estersulfate und an
Allophan adsorbiertes Sulfat in Reinform voneinander unterschieden werden kön-
nen, die Methode jedoch auf Proben sehr einfacher und bekannter Matrixzusam-
mensetzung beschränkt ist.
Die Protonierung funktioneller Gruppen der organischen Substanz und die Auflö-
sung von silikatischen Mineralen und Glas dominieren die schnelle Säurepufferung
in vitric Andosolen. Aufgrund höherer Gehalte an organischer Substanz ist deren
Protonierung für den Standort Poás bedeutender als am Masaya. Die Anreicherung
von basischen Kationen und Aluminium in der Bodenlösung hemmt die Mineralver-
witterung, was insbesondere während der Trockenzeit am Standort Masaya zu einer
Zusammenfassung
XXI
Erniedrigung der Säurepufferkapazität führen kann. Unter Freilandbedingungen wird
die kurzzeitige Protonierung variabler Ladungen durch die Freisetzung von basi-
schen Kationen und Aluminium bei der Mineral- oder Glasverwitterung rückgängig
gemacht, so dass Verwitterungsprozesse die Langzeitpufferung von Säureeinträgen
an diesen Standorten dominieren.
In situ gebildete BAS konnten nach 18-monatiger Exposition von synthetischem
Allophan und Glas in Böden nahe der Vulkane Masaya und Poás nur an Al-
lophanproben identifiziert werden, in denen sie mit 11% (Masaya) und 19% (Poás)
zur Gesamt-S-Retention beitrugen. Perhumide Bedingungen und hohe Gehalte an
organischer Substanz, wie sie am Standort Poás vorliegen, verhindern somit die
BAS-Bildung nicht.
Obwohl pH-Erniedrigung, höhere Sulfatkonzentrationen und längere Versuchszeiten
in pH
stat
-Laborversuchen mit synthetischem Allophan tendenziell zu einer Zunahme
von BAS führten, waren die Effekte als gering einzustufen und Adsorption dominier-
te die Sulfatretention mit 86% - 98%. Eine langsame Zunahme der Sulfatadsorption
bei pH 4.5 im Laufe von 6 Monaten konnte ich auf eine Freilegung existierender
Adsorptionsplätze (durch die Depolymerisierung und Desorption von Si) und die
Bildung neuer Adsorptionsplätze (durch die Zunahme der Allophan-Fehlstellen)
zurückführen.
Die Fällung von BAS erhöht die Sulfatretentionskapazität SO
2
-beeinflusster vulkani-
scher Böden, die Sulfatextrahierbarkeit wird dadurch zumindest in Laborversuchen
jedoch nicht beeinflusst. Ob in situ gebildete BAS die Sulfatextrahierbarkeit erniedri-
gen, bedarf weiterer Untersuchungen. Zusätzlich sollte der Mechanismus, der der
BAS-Bildung zugrunde liegt (Bildung von Oberflächenpräzipitaten oder Lösungs-
übersättigung in Mikro- und Mesoporen), aufgeklärt werden. Dadurch könnte der
Einfluss weiterer Faktoren auf die BAS-Bildung, wie z.B. der Präsenz von Fluorid,
DOC oder Phosphat oder des Allophan-Aggregierungsgrades, abgeschätzt werden.
1
1 GENERAL INTRODUCTION
1.1 ACID BUFFERING IN SOILS
Atmospheric acid deposition and the consecutive acidification of soils affect nutrient
cycles and may have toxic effects on soil organisms, plants and aquatic species in
streams and lakes due to an increase in concentrations of e.g. available aluminum
(Ulrich et al., 1980; Van Breemen et al., 1982; 1984; Falkengren-Grerup, 1989).
Acid inputs to soils are buffered by different mechanisms and the knowledge of
buffering processes and kinetics is important for the predictability of soil acidification.
First studies on acid rain and acid buffering in soils date back until the 19
th
century
(Rogers and Rogers, 1848; Smith, 1852, 1872), major advances regarding this topic
were made in the 80ies of the 20
th
century in the frame of intensive research on the
causes of forest dieback in Central Europe and North America (e.g. Van Breemen et
al., 1982, 1983, 1984; Paces, 1985; Ulrich, 1986; Reuss and Johnson, 1986;
Matzner, 1989; Mulder et al., 1989; Ulrich and Sumner, 1990). Being now textbook
knowledge, carbonate, silicate, sesquioxide and clay mineral dissolution as well as
the protonation of variable charge and specific anion adsorption were identified as
processes responsible for acid buffering in soils (Ulrich, 1986; Reuss & Walthall,
1990; Bruggenwert et al., 1991). The relevance of each process depends on the pH
and the soil mineralogy.
Contrasting to the vast knowledge on acid buffering in soils of the temperate re-
gions, only little is known about acid buffering in volcanic soils, which are located in
the (sub)tropics, although these soils can be subjected to large acid deposition due
to volcanic activity (Delmelle et al., 2001). Generally, the nature of buffering proc-
esses in volcanic soils of the (sub)tropics is expected to be the same as in inten-
sively studied soils of the temperate regions. This has been confirmed by a number
of authors who identified specific processes of acid buffering in volcanic soils as the
protonation of variable charge and the dissolution of primary silicate minerals, allo-
phane or sesquioxides (Parnell, 1986; Parnell & Burke, 1990; Takamatsu et al.,
1992; Dahlgren & Saigusa, 1994; Takahashi et al., 1995; Baba et al., 1995; Baba &
Okazaki, 2000; Hayashi & Okazaki, 2002, 2003). Nevertheless, the relevance of
each buffering process for the overall acid buffering has not been investigated by
Chapter 1
2
these authors and cannot be deduced from studies of acid impacted sites of the
temperate regions as physical conditions close to active volcanoes of the
(sub)tropics differ substantially:
(i) Acid deposition to soils close to active volcanoes reaches values up to
109 kmol H
+
ha
-1
year
-1
(Delmelle et al., 2001) while the largest acid deposition
rates in Central Europe in the 70
ies
were around 6 kmol H
+
ha
-1
year
-1
(Van
Breemen, 1991).
(ii) The geogenic mineralogy of Andosols is dominated by the presence of glass
and small phenocrysts of crystalline minerals while primary minerals in soils of
the temperate regions are mostly crystalline with crystallites being of compara-
tively larger size. This may have a strong influence on the rates of mineral
weathering.
(iii) Climatic conditions in soils are different for Andosols close to active volcanoes
of the (sub)tropics when compared with formerly studied acid-impacted sites
of the temperate region. Besides temperature effects, the hydrologic regime
may affect buffering processes: during the dry season of the subtropics con-
centrations of basic cations and aluminum - released due to rapid protonation
of organic matter - increase and may thereby hamper mineral weathering due
to product limitation.
The contribution of different pools to the acid neutralization capacity of Andosols has
only been investigated by Delfosse et al. (2005a) for vitric and eutric Andosols close
to Masaya Volcano. They determined the total acid neutralization capacity by a
modified equation based on Van Breemen et al. (1984) as the sum of the total re-
serve of bases (Ca, Mg, Na, K) plus total aluminum plus iron contained in Fe-
bearing silicates minus the reserve of sulfur and phosphorus in soils. Additionally,
they determined the amount of exchangeable cations and based on the size of the
pool of exchangeable and non-exchangeable cations they inferred the relevance of
ion exchange and mineral weathering for the overall acid neutralization. Neverthe-
less, the presence of a potentially acid-neutralizing pool not necessarily means that
this pool is active. The effective acid neutralization capacity, which includes all
active pools under given acidification conditions (e.g. amount of acid input, reaction
time), is much smaller and may be constituted by different pools than the total acid
neutralization capacity. The contribution of different processes to the effective acid
Introduction
3
neutralization capacity of Andosols has not been quantified so far. Based on sulfate
retention experiments conducted by Delfosse et al. (2006) it can be inferred that
sulfate retention does not contribute significantly to the effective acid neutralization
in vitric Andosols. The role of silicate and sesquioxide dissolution and of protonation
of variable charge for the effective acid neutralization is still unclear.
1.2 SULFATE RETENTION IN SO
2
-AFFECTED SOILS
Anthropogenic and natural acid emissions contain large amounts of sulfur which is
an essential nutrient for microorganisms, plants and animals. At heavily affected
sites in Central Europe sulfur deposition rates reached up to 80 kg S ha
-1
year
-1
in
the 70
ies
(Reuss & Johnson, 1986) while rates up to 1 450 kg S ha
-1
year
-1
have
been measured close to Masaya Volcano a few years ago (Delmelle et al., 2001).
As a consequence of sulfur deposition concentrations of S in soils increased (Priet-
zel, 1992; Delfosse et al., 2005a). Besides the direct effects on S cycling, sulfur
deposition can influence other nutrient cycles, e.g. due to increasing cation leaching
as a consequence of increased sulfate concentrations in soil solution (Johnson and
Cole, 1977, 1980).
Deposited sulfate may be incorporated into soil organic matter, it may be adsorbed
to variable charge minerals like iron or aluminum sesquioxides or allophane or it
may be included in inorganic precipitates of the basic aluminum sulfate group (BAS)
(Johnson et al., 1982; Prietzel 1992; Mayer et al., 2001; Prietzel et al., 2004). Espe-
cially for acid-impacted sites, where large deposition of sulfate combines with alumi-
num release due to acid buffering, the question whether sulfate is mainly adsorbed
or incorporated into BAS phases has been discussed extensively and controversely
for soils of the temperate regions. While Van Breemen (1973), Adams & Rawajfih
(1977), Nordstrom (1982), Khanna et al. (1987), Courchesne & Hendershot (1990),
Prietzel & Feger (1991), Evans (1991), Prietzel (1992) and Mayer et al. (2001)
suggested that BAS phases are present in S-affected soils, sulfate dynamics in soils
of the Solling area and the Hartz mountains could not be modeled by precipita-
tion/dissolution of BAS phases (Alewell et al., 1995; Lükewille et al., 1995). Similarly,
Kaiser & Kaupenjohann (1998) described that BAS phases most probably do not
control sulfate release from acid forest soils. Bigham and Nordstrom (2000) claim
that BAS phases may form when the pH is 5 or larger but not under more acidic
conditions. This is contradicted by the fact that BAS phases were found in extremely
acid hydrothermal systems of active volcanoes (Hochstein & Browne, 2000).
Chapter 1
4
Basic aluminum sulfate phases in soils may have varying compositions. Applying
equilibrium modeling Van Breemen (1973) described that solution compositions of
acid sulfate soils followed the solubility line of jurbanite and not of alunite or K-alum.
Nordstrom (1982) proposed that alunite and jurbanite are the stable BAS phases
under less and more acidic conditions, respectively, but that basaluminite is formed
as a meta-stable phase instead of alunite. According to the same author, alunogen
is only formed under the most extreme acid conditions of acid sulfate soils or acid
mine waters. Bigham and Nordstrom (2000) amplified the number of BAS phases
and stated that basaluminite, zaherite, aluminite, jurbanite and alunogen are precipi-
tated in this sequence with decreasing pH.
For volcanic soils, a few studies gave hints on BAS precipitation (Wolt et al., 1992;
Delmelle et al., 2003; Delfosse et al., 2005b) and basaluminite- and aluminite-like
phases have been visualized by transmission electron microscopy coupled with
energy dispersive X-ray analysis (TEM-EDX) (Delfosse et al., 2005c). These authors
also suggested that BAS phases may preferentially form during sulfate retention at
the allophane-water interface but no unambiguous evidence was given for this.
1.3 SULFATE RETENTION BY ALLOPHANE
Up to now, only macroscopic experiments have been conducted to study sulfate
retention by allophane. Overall, these studies showed that sulfate retention by allo-
phane increases (i) with increasing sulfate load without reaching a retention maxi-
mum despite large sulfate loads and (ii) with decreasing pH (Gebhardt and Cole-
man, 1974; Rajan, 1979; Padilla et al., 2002; Pigna & Violante, 2003; Jara et al.,
2006, Ishiguro et al., 2006; Delfosse et al., 2006). Only very few and even contradic-
tory results have been presented on sulfate retention kinetics: while the amount of
retained sulfate did not increase in studies of Rajan (1979) from 3 hours up to 52
hours, Jara et al. (2006) reported that sulfate was retained by a combination of fast
and slow processes with no apparent equilibrium within the experimental period of
two weeks.
Most studies on sulfate retention by allophane concentrate on sulfate adsorption
without regarding the possibility of BAS formation. Sulfate is adsorbed to aluminol
sites of allophane and Padilla et al. (2002) and Abidin et al. (2007) stated that defect
sites of the hollow spherules present preferential sites for anion adsorption. Sulfate
binds to allophane via inner-sphere and outer-sphere complexation with inner-
sphere complexation increasing with decreasing pH (Padilla et al., 2002). Based on
Introduction
5
the macroscopic data different types of inner-sphere complexes (e.g. mono-
/bidentate) have been discussed (Rajan, 1979).
The observation that sulfate retention by allophane did not reach a maximum de-
spite partly large sulfate loads suggests that BAS precipitates may form at the allo-
phane-water interface. Similarly as for sulfate retention, precipitate formation has
also been discussed for phosphate and fluoride retention by allophane, during which
taranakite and cryolite have been suggested to form, respectively (Wada, 1989).
Only few studies addressed the formation of BAS precipitates and gave contradic-
tory results: Rajan (1979) concluded that no BAS precipitates form during sulfate
retention by allophane at pH 5. Contrastingly, Delfosse et al. (2006) and Ishiguro et
al. (2006) who analyzed sulfate retention by Andosol samples under various pH
conditions and similar sulfate addition as Rajan (1979), interpreted biphasic iso-
therms – a Langmuir-type isotherm at low sulfate loads and linearly or exponentially
increasing sulfate retention at high sulfate loads – as a shift from sulfate adsorption
to sulfate precipitation.
1.4 IDENTIFICATION OF BAS PHASES IN SOILS
The controversial discussion on the presence or absence of BAS phases in soils or
at the allophane-water interface is largely attributable to difficulties in an unambigu-
ous BAS identification. This task is rather complex as BAS phases in the mentioned
environments are of small size and amorphous. Consequently, traditional methods
of mineral identification like e.g. X-ray diffraction fail to identify them.
Up to now, the identification of BAS phases in soils or model experiments has been
almost exclusively based on selective extraction procedures applied to solid sam-
ples or on geochemical equilibrium modeling of solution data. Solutions of NH
4
F are
expected to extract BAS phases and adsorbed sulfate (Prietzel and Hirsch, 1998),
KH
2
PO
4
is applied to extract adsorbed sulfate solely and consequently, the differ-
ence between NH
4
F extractable sulfate and KH
2
PO
4
extractable sulfate has been
used to estimate the amount of BAS phases in soils (Prietzel and Hirsch, 1998;
Delfosse et al., 2005b). However, selective extraction procedures for the identifica-
tion of different S phases in soils are not selective sensus strictus (Prietzel et al.,
2003) and thus, they do not allow an unambiguous BAS identification. Similarly,
geochemical equilibrium modeling of solution data may give hints on thermodynami-
cally possible phases in soils or model systems without proofing the existence of
such phases. This has also been emphasized by Sposito (1986) and Ford et al.
Chapter 1
6
(2001) who demonstrated that macroscopic date alone can lead to misinterpreta-
tions of results and that a combination of macroscopic, microscopic and spectro-
scopic tools is required for an unambiguous phase or process identification. Del-
fosse et al. (2005c) successfully identified BAS phases in volcanic soils by
transmission electron microscopy coupled with energy dispersive X-ray analysis
(TEM-EDX). However, sample pretreatment for BAS identification by TEM-EDX may
alter BAS phases as described by Delfosse et al. (2005c). Additionally, the absence
of BAS phases in TEM images not necessarily means that no BAS phases are
present as they may represent not more than a “few needles in a haystack”, thus
being difficult to visualize.
Extended X-ray absorption fine structure spectroscopy (EXAFS) and X-ray absorp-
tion near edge spectroscopy (XANES) both have the potential to differentiate be-
tween adsorbed and amorphous, precipitated species (Fendorf et al., 1994; Schulze
& Bertsch, 1995). However, EXAFS signals of the low atomic weight element S are
only very weak and cannot be interpreted meaningfully (personal communication
Dr. Jörg Göttlicher, Institute for Synchrotron Radiation, Forschungszentrum
Karlsruhe). Therefore, EXAFS spectroscopy can be regarded as non-suitable for the
differentiation between adsorbed and precipitated sulfate phases in Andosols.
XANES has been applied to differentiate between adsorbed and mineral forms of
phosphate (Hesterberg et al., 1999), thus it may also have a potential for BAS identi-
fication in soils.
Additionally, it may be hypothesized that adsorbed and precipitated species differ in
their binding and thus in their thermal energies. If this was the case, simultaneous
thermal analysis coupled with mass spectrometry (STA-MS) may also potentially
allow differentiating between adsorbed and precipitated sulfate. This method has not
been tested so far for this purpose but it has been applied repeatedly in soil science
to identify pools of organic carbon with different thermal stabilities (e.g. Plante et al.,
2005; Dorodnikov et al., 2007).
1.5 RESEARCH OBJECTIVES
The main objectives of my study were (i) to estimate the contributions of different
processes to the effective acid neutralization capacity and (ii) to determine the rele-
vance of adsorption processes and BAS precipitation for the total sulfate retention in
SO
2
-impacted Andosols. For the investigation of S retention processes, I combined
Introduction
7
macroscopic experiments at the field and laboratory scale with the analysis of solid
samples by microscopic, spectroscopic and thermal techniques.
The relevance of different processes for acid buffering in Andosols is largely un-
known. Additionally it is unclear, if and in which way climatic conditions may influ-
ence acid buffering in Andosols. These questions are addressed in chapter 2.
Therefore, I estimated the contribution of different processes to the effective acid
neutralization capacity in short-term laboratory experiments with soil samples from
Masaya and Poás Volcanoes simulating different leaching conditions and analyzing
samples from a transect of decreasing acid input. Additionally, the effects of acid
buffering on secondary mineral formation are addressed.
To be able to estimate the contribution of adsorption processes and BAS precipita-
tion to the total S retention, I evaluated different methods for their suitability to differ-
entiate between adsorbed sulfate and sulfate included in different BAS phases. As a
co-author I contributed to the evaluation of the suitability of XANES (Prietzel et al.,
2008). This method showed to be powerful for the differentiation of adsorbed sulfate
and sulfate contained in BAS phases. However, estersulfate compounds may be
confounded with BAS. Additionally, I tested the method of simultaneous thermal
analysis coupled with mass spectrometry (STA-MS) for its ability to distinguish
between different sulfate-species and the results are presented in chapter 3. For
this test, I synthesized reference compounds and analyzed them by STA-MS in pure
form and in mixture with different matrix components.
In chapter 4 I analyzed whether BAS phases are preferentially formed in situ at the
allophane-water or at the glass-water interface. I did this by analyzing the solid
phases for the presence of BAS after 18 months of exposure in soils close to the
SO
2
-emitting volcanoes Masaya and Poás. Additionally, both allophane and glass
were analyzed for potential changes caused by field exposure.
In chapter 5 and 6 I investigated the effects of sulfate concentration, pH and time
on the relative contribution of adsorption processes and BAS precipitation to the
total sulfate retention by allophane. To do this, I conducted pH
stat
experiments with
synthetic allophane and the solid experimental products were analyzed by a combi-
nation of wet-chemical methods, SEM-EDX, STA-MS and S K-edge XANES. For the
Chapter 1
8
long-term experiment, allophane was additionally analyzed for potential structural
changes.
9
2 MECHANISMS OF ACID BUFFERING AND FORMATION OF SECONDARY
MINERALS IN VITRIC ANDOSOLS
2.1 SUMMARY
Andosols in the vicinity of active volcanoes receive large inputs of SO
2
and HCl. I
studied (i) the mechanisms of acid buffering, (ii) the effect of cation removal on the
short-term acid neutralization capacity and (iii) the consequences of acid buffering
for secondary mineral formation in vitric Andosols around the Central American
volcanoes Poás and Masaya. Two types of short-term (24 h) acidification experi-
ments at pH 3 were conducted to simulate an open system in which leaching pre-
vails (extraction with protonated cation exchange resin) and a closed system with no
leaching (pH
stat
titration with cation accumulation). Long-term buffering under field
conditions (mean soil pH: 4.6) and its effect on secondary mineral formation were
studied by analysis of samples from a transect of decreasing acid input by IR spec-
troscopy, microscopic methods and geochemical equilibrium modeling. In Poás
samples the main short-term buffering mechanisms at pH 3 are plagioclase dissolu-
tion and protonation of organic matter. Long-term acid buffering under field condi-
tions led to weathering of plagioclase crystals but did not result in protonated car-
boxyl groups. In Masaya samples mineral and/or glass dissolution are the dominant
acid buffering mechanisms in laboratory experiments and under field conditions. For
both sites, cation accumulation during pH 3 acidification experiments led to a de-
crease of the effective acid neutralization capacity. Due to different climatic condi-
tions, Al is precipitated as basaluminite at Masaya while it seems to be susceptible
to leaching at Poás. Acid buffering resulted in the formation of amorphous silica at
both sites.
Chapter 2
10
2.2 INTRODUCTION
Input of acids and subsequent effects on soils have been intensively studied in
temperate regions of the northern hemisphere (Van Breemen, 1991). Acid deposi-
tion in this area has been mainly of anthropogenic origin and reached up to 6 kmol
H
+
ha
-1
year
-1
in the 1980s in strongly polluted forested areas of western and central
Europe (Van Breemen, 1991). In contrast, few studies exist on the impact of acid
emissions from active volcanoes on surrounding soils, where estimated H
+
deposi-
tion rates can reach up to 109 kmol H
+
ha
-1
year
-1
(Delmelle et al., 2001). While
anthropogenic acid emissions are dominated by HNO
3
and H
2
SO
4
, active volcanoes
release predominantly H
2
SO
4
or HCl with minor amounts of HF.
According to Bruggenwert et al. (1990) and Ulrich (1990), different acid buffering
mechanisms with specific buffer capacities and rates are distinguishable in soils. In
the temperate regions, protonation of variable charge with subsequent liberation of
basic cations and dissolution of sesquioxides are the major short-term acid buffering
mechanisms (Van Breemen et al., 1983). In Andosols various buffering mechanisms
can be relevant, depending on acid deposition rates, the pH of the soil solution and
the Andosol subtype. Besides, leaching conditions are expected to influence acid
buffering as the accumulation of cations in soil solution can lead to product limitation
and thus hamper buffering processes. Changes in buffer mechanisms and rates
may occur over short periods of time, as volcanic emissions show a strong temporal
variability, causing large pH fluctuations in surrounding Andosols (Parnell, 1986).
Field and laboratory experiments on the effects of acid input in different Andosol
subtypes suggest the protonation of allophane and humic substances as well as the
dissolution of allophane, iron oxides and primary silicate minerals or glass as buffer-
ing mechanisms (Table 2.1). As a consequence, cations and Si are released into the
soil solution and can be leached or precipitated as secondary mineral phases.
Studies on the effects of large acid inputs on the formation of secondary minerals in
Andosols are scarce. Delfosse et al. (2005c) reported the formation of basic alumi-
num sulfate minerals due to large S deposition and Al release by volcanic glass
weathering in soils around the Masaya Volcano in Nicaragua. Secondary phases
can precipitate on weathering mineral surfaces. This can have a protecting effect
against further dissolution (Courchesne et al. 1996). The formation of secondary
phases on weathering mineral surfaces as a consequence of acid buffering in An-
dosols has not been studied so far.
Acid buffering
11
Table 2.1 Literature reports on acid buffering mechanisms in Andosol subtypes.
Andosol subtype Buffering mechanism Reference
Vitric Andosol Protonation of variable charge Takamatsu et al. 1992
Sesquioxide dissolution Takamatsu et al. 1992
Silicate mineral weathering Delfosse et al. 2005a
Sil-andic Andosol Protonation of variable charge on
allophane (and organic matter)
Parnell 1986
Dahlgren & Saigusa 1994
Takahashi et al. 1995
Delfosse et al. 2005a
Dissolution of allophane or amor-
phous Al
Baba & Okazaki 2000
Takahashi et al. 1995
Alu-andic Andosol Protonation of variable charge on
organic matter
Dahlgren & Saigusa 1994
Takahashi et al. 1995
Therefore, the aims of my study were (i) to identify the mechanisms of acid buffering
in vitric Andosols, (ii) to test the effect of cation removal on the contribution of differ-
ent buffering mechanisms to the effective short-term acid neutralization capacity and
(iii) to analyze the effects of acidification on the formation of secondary minerals.
2.3 MATERIALS AND METHODS
In short-term-laboratory-acidification experiments (with and without cation removal) I
analyzed the fast buffering processes at pH 3, which are important for the rapid
response of Andosols to peaks of acid input. The relevance of these processes for
acid buffering under field conditions was studied by the analysis by infrared spec-
troscopy and microscopic methods of soil samples from a transect of decreasing
acid input. Saturation indices of potential secondary minerals were calculated by
geochemical equilibrium modeling of soil saturation extracts. Additionally, I analyzed
mineral grains of one study site by microscopic methods in order to elucidate
whether potential secondary phases form on weathering mineral surfaces.
My study sites were the volcanoes Poás (Costa Rica) and Masaya (Nicaragua).
Both emit large amounts of acid-generating gases; their tephra are of similar minera-
logical composition. However, climatic conditions differ considerably between the
two settings.
Chapter 2
12
2.3.1 Study sites
2.3.1.1 Poás Volcano
The Poás Volcano (10°12´N, 84°14´W, Figure 2.1a) is a composite stratovolcano
situated in the central Costa Rican volcanic range and rising to 2708 m above sea
level. Mean annual temperatures vary from 8 to 12 °C, and mean annual rainfall
from 3800 to 4600 mm, with monthly averages ranging from 120 mm (December to
April) to 420 mm (May - November) (Rowe et al., 1992; Martinez et al., 2000). Poás
Volcano has shown different types of volcanic activity ranging from phreatic to
phreato-magmatic eruptions in 1953 - 1955, 1910 and 1834, with minor eruptions
and fumarolic activity between these eruptive cycles (Alvarado Induni, 2005). During
eruptions volcanic ash of basaltic to andesitic composition has been ejected to-
gether with variable amounts of S-rich sediments from the extremely acid crater lake
(Prosser & Carr, 1987). Lake sediments comprise a substantial amount of tephra
from the 1910 eruption, making this ash layer easily recognizable in the field by its
whitish-greyish colour. Basaltic-andesitic tephra contain plagioclase, pyroxene and
magnetite phenocrysts in a glass matrix (Rowe & Brantley, 1993).
Since 1955 the Poás Volcano has shown fumarolic activity of varying magnitude,
emitting SO
2
, HCl and HF, with SO
2
being the dominant acid-generating component
of the gas plume (Rowe et al., 1992). Measured SO
2
emission rates range from 8 to
700 t SO
2
day
-1
(Casadevall et al., 1984; Andres et al., 1992; Zimmer et al., 2004).
This leads to large acid deposition onto soils in the vicinity, with measured pH val-
ues in rain water ranging from 2.9 to 5.8 at Cerro Pelón, an exposed hill site 2.4 km
downwind of the active crater (Martinez et al., 2000).
Natural vegetation at Poás is a cloud forest. With increasing acid input, the tree
species Clusia adorata and Scheflera rodriguesiana dominate over other species.
Also, in some places the cloud forest has been replaced by myrtle stands (Pernettia
coriacea, Vaccinium poasanum).
2.3.1.2 Masaya Volcano
The Masaya Volcano (11°59´N, 86°10´W, rising to 560 m above sea level, Figure
2.1b) is a shield volcano of basaltic composition with tephra containing glass, pla-
gioclase, olivine and augite (Walker et al., 1993). It also is located in the Central
American Volcanic Belt. Because of its lower elevation in comparison to the Poás
Acid buffering
13
Volcano, the climate is markedly different, with a mean annual temperature of 26.3
°C and a mean annual precipitation of 1379 mm of which approximately 90% falls
during the rainy season (May - October) (UNA & ISRIC, 1994). Apart from its erup-
tive activity, the Masaya Volcano has also shown cycles of fumarolic emissions, with
SO
2
fluxes in the actual degassing period ranging from 390 to 1850 t SO
2
day
-1
(Delmelle et al., 2002). Also, substantial amounts of halogens (HCl and HF) are
emitted (Delmelle et al., 2001). This leads to acid deposition ranging from 1 to 30
mg H
+
m
-2
day
-1
(Delmelle et al., 2001) and pH values of rainwater between 2.5 and
5 (Johnson & Parnell, 1986). As a consequence of the acid input, the soil pH and
thus the effective cation exchange capacity are lower in strongly acid-affected sites
compared with control sites, while sulfate and fluoride contents in soils show a
marked increase with increasing acid input (Delmelle et al., 2003; Delfosse et al.,
2005a).
Figure 2.1 Aerial photographs of the Poás Volcano with sampling sites P1 - P5 (a) and the
Masaya Volcano with sampling sites M1 and M2 (b) (Arrows start from the active craters and
indicate the main wind direction).
2.3.2 Estimation of S-deposition, soil sampling and description
2.3.2.1 Poás Volcano
In December 2002 five sampling sites were selected (Figure 2.1a). Sulfur input was
estimated at four of the five sites during a period of three weeks, using sulfation
plates according to the method described in Delmelle et al. (2001). Briefly, Petri
dishes filled with a lead oxide-containing paste are exposed in the field where SO
2
reacts with PbO to form PbSO
4
, which is subsequently extracted in the laboratory.
The amount of trapped sulfate is determined by ion chromatography. Absorption of
SO
2
on sulfation plates depends on the SO
2
air concentration and wind speed,
a
b
Chapter 2
14
therefore sulfation rates do not equal SO
2
deposition rates on plants or soils. Never-
theless, Delmelle et al. (2001) showed that for the Masaya Volcano sulfation plates
give a rough estimate of S deposition. Table 2.2 shows that sulfation rates at the
sampling sites around Poás Volcano decrease from profile one (P1) to five (P5).
Based on S deposition rates on samplers, H
+
deposition rates were estimated.
In November 2003 soil profiles were dug at the sites P1 to P5 to a depth of 80 - 100
cm and described and classified according to WRB standard guidelines (WRB,
1998). All selected profiles contain ash horizons from the last two bigger eruptions
(1910 and 1953-1955); this guarantees that all soil profiles have developed from the
same parent material.
Carbon concentrations reflect the stratification of the profiles with buried A horizons
between ash layers (Table 2.3). Soil pH values range from 3.5 to 4.8 and do not
show a clear effect of acid deposition along the transect. The same holds true for
exchangeable cations. Iron concretions and mottling, as observed in the field, and
the increase in oxalate-extractable iron in deeper horizons (23 - 80 g Fe
o
kg
-1
in mBs
horizons) indicate iron translocation. Low oxalate-extractable Si (Si
o
) concentrations
could be due to the dissolution of Si from crystalline mineral surfaces (Dahlgren,
1994); the presence of allophane in these soils is unlikely. The ratios of oxalate-
extractable Al (Al
o
) to pyrophosphate-extractable Al (Al
p
) equaled one. The latter
suggests the dominance of Al-organic complexes. The contents of volcanic glass
and the contents of Al
o
+ 0.5 Fe
o
are larger than 10 and 0.4%, respectively, there-
fore the soils have been classified as vitric Andosols.
Table 2.2 Sulfation rates (mg SO
2
m
-2
day
-1
) at soil profile sampling sites at Poás and Ma-
saya.
Poás P1 P2 P3 P4 P5
Masaya
a)
820
M1
nd
M2
275
166
9
~600 ~400
a)
data taken from map published in Delfosse et al. (2005a)
nd: not determined
Acid buffering
15
2.3.2.2 Masaya Volcano
At the Masaya Volcano two sites (M1 & M2, Figure 2.1b) were selected in May 2004
for soil sampling, based on the results of Delmelle et al. (2001) and Delfosse et al.
(2005a) on effects of acid input in vitric Andosols in the vicinity of the volcano. I
selected the most (“vitric 2” in Delfosse et al., 2005a) and least (“vitric 9” in Delfosse
et al., 2005a) acid-affected sites of their study. Sulfation rates on S samplers at
these sites were taken from the map published in Delfosse et al. (2005a) and are
reported in Table 2.2.
The Masaya soils are stratified and show a clear influence of acid deposition on pH
values and the effective cation exchange capacity (ECEC) (Table 2.3) as described
by Delmelle et al. (2003). Oxalate-extractable Si and Al contents indicate the pres-
ence of allophane, reflecting the development towards the sil-andic subtype. How-
ever, Masaya soils have also been classified as vitric Andosols because of glass
contents and Al
o
+ 0.5 Fe
o
larger than 10% and 0.4%, respectively.
2.3.3 Sample preparation
Apart from the microscopic observations, all analyses were carried out on the entire
soil samples. Therefore, coarse fragments, which constituted between 10 and 30%
of the Masaya samples, were crushed gently with a porcelain mortar to a grain size
smaller than 2 mm before analysis. Poás samples contained no components
coarser than 2 mm. Field-moist samples of Poás soils were used for the short-term
acidification experiments and the soil saturation extracts, whereas air-dried samples
of Masaya soils were used as most of them were already dry at the moment of
sampling.
16
Table 2.3 Soil sample characteristics at Poás (P1 - P5) and Masaya (M1 - M2).
----------- cations exchangeable by 1M NH
4
Cl -----------
Profile Horizon Depth pH
a)
C
org
b)
Fe
o
c)
Si
o
c)
Al
o
c)
Al
p
c)
ECEC
d)
Al Ca Mg K Na
/ cm / % / g kg
-
1
/ g kg
-
1
/ g kg
-
1
/ g kg
-
1
-------------------------- / mmol
c
kg
-
1
--------------------------
P1 Ah 0-6 4.6 0.9 6.0 0.74 1.6 0.9 8.8 5.7 2.2 0.3 0.2 0.4
2AC 6-9 4.8 1.0 11.5 0.78 1.0 0.9 7.4 5.1 1.7 0.2 0.2 0.2
3Ah 9-17 4.0 3.4 2.5 0.11 1.3 1.5 17.8 13.9 1.9 1.3 0.3 0.4
P2 L 19-9 nd nd nd nd nd nd nd nd nd nd nd nd
Ofh 9-0 nd nd nd nd nd nd nd nd nd nd nd nd
Ah 0-4 4.3 4.1 3.2 0.25 1.6 1.6 36.3 21.1 9.6 3.5 0.6 1.5
2AC 4-9 4.7 1.0 1.8 0.67 1.9 1.8 8.6 6.5 1.3 0.2 0.2 0.4
3Ah 9-15 4.2 10.4 6.1 0.22 5.8 5.9 29.7 25.9 2.6 0.4 0.5 0.3
P3 L 12-8 nd nd nd nd nd nd nd nd nd nd nd nd
Ofh 8-0 3.6 36.9 1.7 0.13 4.0 3.9 44.8 24.0 12.8 3.5 3.9 0.6
Ah 0-4 3.9 15.7 3.3 0.16 2.5 2.8 37.4 21.6 7.8 5.4 1.7 0.9
2AC 4-11 4.7 1.0 1.8 0.78 2.3 2.1 7.2 5.5 1.0 0.2 0.2 0.3
3Ah 11-25 4.3 8.0 3.5 0.14 4.2 4.3 21.7 19.1 1.5 0.4 0.4 0.3
P4 L 5-3 nd nd nd nd nd nd nd nd nd nd nd nd
Ofh 3-0 3.6 nd nd nd nd nd nd nd nd nd nd nd
2AC 0-7 4.6 3.5 2.1 0.49 2.6 2.7 19.3 14.6 2.9 0.9 0.4 0.5
3Ah 7-11 4.8 11.9 4.8 0.37 5.9 5.9 23.2 19.8 2.0 0.6 0.4 0.4
P5 L 33-29 nd nd nd nd nd nd nd nd nd nd nd nd
Ofh 29-0 4.1 34.1 2.2 0.20 4.3 3.7 49.4 8.6 28.1 8.6 3.2 0.9
2AC 0-3 4.8 3.5 5.9 0.66 6.7 6.6 7.0 3.7 1.6 0.2 0.3 1.2
3Ah 3-9 4.7 7.5 1.8 0.28 7.0 7.0 9.2 4.7 2.4 1.1 0.4 0.6
M1 A 0-9 4.1 1.8 10.4 2.4 4.6 2.0 22.7 15.5 2.7 0.8 1.8 1.9
2BC 9-14 4.4 2.2 11.2 7.2 12.8 3.2 21.6 15.8 2.9 0.5 1.7 0.7
3A 14-24 4.5 2.6 15.7 14.6 24.3 4.2 26.8 19.1 3.8 0.8 1.5 1.6
M2 A 0-6 4.5 5.8 11.3 2.4 4.9 2.1 163.5 12.7 121.1 18.3 10.3 1.1
2BC 6-15 4.8 1.8 11.1 3.1 5.6 2.1 45.7 10.4 27.4 3.0 3.5 1.4
3A 15-21 4.9 3.2 17.5 7.9 12.3 2.6 116.7 4.7 89.7 12.5 9.0 0.8
a)
measured in soil saturation extracts
b)
organic carbon content (CN analyser Elementar Vario ELIII)
c)
oxalate-extractable Fe (Fe
o
), Si (Si
o
), Al (Al
o
) and pyrophosphate-
extractable Al (Al
p
) according to Dahlgren (1994)
d)
effective cation exchange capacity nd: not determined
16
Acid buffering
17
2.3.4 Short-term acidification experiments at pH 3
The short-term acidification experiments at pH 3 were conducted on samples of
horizons representing the last volcanic eruption. At Poás, this corresponds to the
1953 - 1955 ash layer (horizon 2AC), and at Masaya to the horizon 2BC. Short-term
buffering mechanisms were analyzed by two methods with three replicates for each
analysis. In method 1, batch experiments containing protonated cation exchange
resin as a sink for released cations were conducted. This approach simulates an
open system and may be taken as a model for soils where leaching prevails. In
method 2, pH
stat
titrations were carried out during which released cations remain in
solution. This system simulates conditions under which leaching and plant uptake of
cations are low and where product limitation of buffering processes may occur, as
described by Kaupenjohann & Wilcke (1995). For both methods, the short-term acid
neutralization capacity (ANC) was determined as the sum of the released cations
and will be termed hereafter as ANC
res
for method 1 and ANC
titr
for method 2.
(i) Method 1: Batch experiments using H
+
-saturated cation exchange resin
I used a cation exchange resin extraction as described by Schwarz et al. (1999).
Briefly, samples corresponding to 2 g of air dried soil were weighed in 100 ml PE
bottles and 50 ml of deionized water and 3 g of H
+
-saturated resin (Amberjet
1200H, Rohm & Haas) were added. The resin had been washed previously with
deionized water and dried and packed into Nylon bags (pore size 150 µm). After
24 hours the supernatant was filtered through membrane filters (cellulose ace-
tate, 0.45 µm), Si and DOC were determined by flame atomic absorption spec-
troscopy and a TOC analyzer, respectively, and sulfate by capillary electropho-
resis (HP
3D
CE, Agilent) using a buffer solution contained in the Agilent anion
analysis kit (product number: 5063-6511). At the end of the extraction, the pH of
supernatant solutions ranged from 2.9 to 3.5. Cations adsorbed to the resin were
extracted by shaking the resin bags 3 times for 30 minutes with 100 ml of 2 M
HNO
3
each. Between each extraction the resin bags were washed with deion-
ized water. The concentrations of Al, Ca, Mg, K and Na in the extracts were de-
termined by flame atomic absorption spectroscopy (Perkin Elmer 1100 B).
Chapter 2
18
(ii) Method 2: pH
stat
titration
Samples were titrated at a constant pH of 3 during 24 hours using automatic ti-
trators (Mettler Toledo DL50 Graphix), which recorded acid consumption during
the duration of the experiment. Therefore samples corresponding to 2 g of air-
dried soil were weighed into reaction vessels and 50 ml of 50 mM KCl were
added. Subsequently, samples were stirred for 15 minutes before starting the ti-
tration with 0.02 M HCl. After 24 hours, supernatants were filtered through
membrane filters (cellulose acetate, 0.45 µm) and the concentrations of Al, Si,
Ca, Mg and Na were determined by flame atomic absorption spectroscopy
(Perkin Elmer 1100 B). The concentration of dissolved organic carbon (DOC)
was determined by a TOC analyzer (Shimadzu TOC-5050A). The amount of ti-
trated H
+
at pH 3 equaled for all analyzed samples the sum of extracted cations.
2.3.5 Fourier-transformed diffuse reflectance infrared spectroscopy
For the analysis of protonated and dissociated carboxyl groups of the soil organic
matter, infrared spectra were recorded with a Bio-Rad FTS 135 FT-IR spectrometer
in the diffuse reflectance mode (DRIFT). Spectra were recorded in the range of 400
- 4000 cm
-1
with a 1 cm
-1
spectral resolution and 16 replicate scans for each sample.
In order to standardize the water content, all samples were stored for 24 hours in a
desiccator before analysis. The band between 1710 and 1740 cm
-1
is assigned to
protonated carboxyl groups and the band between 1600 and 1640 cm
-1
to dissoci-
ated carboxyl groups. Therefore, by calculating the ratio of the two band heights the
degree of protonation of carboxylic groups can be determined.
2.3.6 Soil saturation extracts
Soil saturation extracts (SSEs) were obtained as described by Rhoades (1982) with
minor modifications. Briefly, amounts of samples corresponding to 150 g air-dried
soil were weighed into 500 ml PE bottles and deionized water was added under
mixing to obtain a uniformly saturated soil-water paste. At this point soil samples
glistened and flowed slightly when the bottles were tipped. After equilibration for 24
hours at 4°C in the dark, SSEs were obtained by filtering the samples with suction
through membrane filters (cellulose acetate, 0.45 µm), which had been rinsed previ-
ously with 200 ml of deionized water. After determination of pH, the filtrates were
ultracentrifuged at 370 000 g for 30 minutes. The supernatant solutions were ana-
Acid buffering
19
lyzed for Al, Si and Fe by flame atomic absorption spectroscopy. Calcium, Mg and
Na were determined by capillary electrophoresis (HP
3D
CE, Agilent) according to
Beck & Engelhardt (1992), sulfate, chloride and nitrate by capillary electrophoresis
as described for ANC
res
and DOC by a TOC analyzer (Shimadzu TOC-5050A). Total
fluoride in SSEs was quantified by using an ion-selective electrode after addition of
TISAB buffer to samples according to Adriano & Doner (1982). A preliminary ex-
periment on selected samples showed no differences in SSE composition when the
equilibration time was increased from 24 hours to 6 weeks.
Aluminum species in SSEs and saturation indices (SI) for secondary Al or Si con-
taining minerals were estimated using the geochemical model Visual Minteq 2.31
(Allison et al., 1991) using the equilibrium constants included therein. Saturation
indices between -0.1 and +0.1 were taken as an indication that the soil solution
might be in equilibrium with a “hypothetical” secondary phase.
2.3.7 Microscopic methods
Grains of the sand fraction were isolated for microscopic analyses from samples of
the 2AC horizons of Poás soil profiles by wet sieving (mesh size: 63 µm). Subse-
quently, adherent fine material was removed by washing with deionized water and
ultrasonic cleaning until the supernatant was clear. Grains were dried at 60 °C, an
aliquot of each sample was embedded in epoxy resin, and thin sections with a thick-
ness of 25-30 µm were prepared as uncovered specimens.
Thin sections were observed under a petrographic microscope (Reichert Jung,
Polyvar) for their mineralogical composition and for weathering features on minerals.
Additionally, they were observed using a FEI XL 30 field-emission environmental
scanning electron microscope (ESEM-FEG) and analyzed for the average mineral
composition by energy-dispersive X-ray analysis (EDX). The EDX analyses were
performed in the low-vacuum mode, applying 20 keV, and the mineral composition
was calculated as an average of 8 measurements.
For the analysis of potential secondary minerals on primary mineral surfaces, entire
mineral grains were mounted on a double-adhesive carbon tape, coated by carbon
and analysed by SEM (Hitachi S-2700) with EDX, applying 20 keV.
Chapter 2
20
2.4 RESULTS AND DISCUSSION
2.4.1 Mechanisms of H
+
buffering at Poás
2.4.1.1 Short-term buffering under cation removal
The ANC
res
in 2AC horizons increased with decreasing SO
2
inputs (P1 P5), with a
concomitant increase in the release of DOC, Al, Si, Ca and Na (Figure 2.2). The
amount of Al released corresponded to 70 and 82% of the ANC
res
for P1 and P5,
respectively, with intermediate portions for P2 - P4. Released Ca corresponded to 8
- 16% of ANC
res
and released Na to 0.8 - 1.1%. Similarly to Na, released K and Mg
(data not shown) contributed negligibly to ANC
res
(0.2-0.4 and 0.6-1.5%, respec-
tively).
ANCres / mmolc kg-1
0
200
400
600
800
1000
DOC / g kg-1
0
2
4
6
8
P1P2P3P4P5 M1M2
Si / mmol kg-1
0
20
40
60
80
100
120
140
Al / mmolc kg-1
0
200
400
600
800
Ca / mmolc kg-1
0
20
40
60
80
100
Na / mmolc kg-1
0
1
2
3
4
5
6
P1P2P3P4P5 M1M2
P1P2P3P4P5 M1M2 P1P2P3P4P5 M1M2
P1P2P3P4P5 M1M2 P1P2P3P4P5 M1M2
Figure 2.2 Acid neutralizing capacity, released DOC, Al, Si, Ca and Na (mean values and
standard errors) during the resin experiment for Poas (P1 - P5) and Masaya (M1 & M2)
samples.
Acid buffering
21
Vance and David (1991) showed that DOC can be mobilized in soils by acidification
due to protonation of variable charge of organic matter, causing the liberation of
bridging cations like Ca or Al. As the exchangeable cation composition in Poás soils
is dominated by Al (Table 2.3), I assume that DOC was mainly mobilized from Al-
organic complexes. Van der Salm et al. (2000) showed that various pools of Al-
organic complexes with differing Al-binding strength exist in soils. Ammonium chlo-
ride is supposed to extract the most labile pool while the cation exchange resin
additionally removes Al from more stable complexes. This explains the larger contri-
bution of Al to the ANC
res
compared to the contribution of Al
NH4Cl
to the ECEC.
Nevertheless, the large amounts of released Al and the release of Si, Ca, Mg, K and
Na indicate additional sources for ANC
res
. Possible buffering substances may be
volcanic glass, plagioclase or pyroxene which constitute the mineral fraction of the
2AC horizon, as determined by X-ray diffraction analyses. Besides, amorphous Al
hydroxide, if present, could dissolve during the ANC determination. Nevertheless,
based on the fact that soil saturation extracts were clearly undersaturated with
respect to amorphous Al(OH)
3
(see below) I discarded this possibility. The fact that
the amounts of mobilized Ca were 9 to 20 times larger than the corresponding Mg
amounts rules out pyroxene as the main source of mobilized cations. Strong positive
correlations between Ca and Si (r
2
= 0.98, p < 0.05) and Ca and Na (r
2
= 0.96, p <
0.05) and no significant correlation between these three elements and K point to
plagioclase as the dominant weathering mineral. This finding agrees with results of
Rowe & Brantley (1993) who reported larger dissolution rates of plagioclase in
comparison to glass in a flank aquifer at Poás Volcano.
I determined the plagioclase composition in Poás samples by energy-dispersive X-
ray analysis on thin sections as anorthite68, resulting in a composition of
Na
0.32
Ca
0.68
Al
1.68
Si
2.32
O
8
. Based on the Al:Ca ratio of 2.47, calculated from the aver-
age composition, I estimated the amounts of Al released by plagioclase weathering
(Al
plag
) from the amounts of Ca released, assuming congruent mineral dissolution.
The amount of Ca released by plagioclase weathering was taken as the difference
between total released Ca and exchangeable Ca. Furthermore, I calculated the
amount of Al released from Al-organic complexes (Al
org
) as the difference between
total released Al (Al
tot
) and Al
plag
, assuming that contributions of pyroxene and glass
dissolution are negligible. This assumption was based on the fact that the amounts
Chapter 2
22
of Mg and K released during the resin experiment were very low in comparison to Al
and Ca release.
The amounts of Al
org
(Table 2.4) were smaller than the contents of pyrophosphate-
extractable Al, indicating that not all organically complexed Al was released in the 24
hour resin extraction at pH 3. For P1 20% of pyrophosphate-extractable Al was
released while for the other samples between 50 and 60% was mobilized. The
results show an increase of Al
org
from 20.2 to 398.8 mmol
c
kg
-1
from P1 to P5, corre-
sponding to 15 - 54% of the total ANC
res
. Thus, soil profiles receiving larger acid
inputs from Poás Volcano released less organically complexed Al than control sites
during the short-term acidification experiment at pH 3. This was caused by decreas-
ing contents of organically complexed Al from P5 to P1, as reflected by a decrease
of pyrophosphate-extractable Al in the 2AC horizon from P5 to P1 (Table 2.3). De-
creasing amounts of organically complexed Al with increasing acidification have
been reported by Takamatsu et al. (1992) for volcanogenous Regosols in Japan and
by Berggren et al. (1998, and references therein) for soils in temperate regions.
Berggren et al. (1998) mention the protonation of variable charge on organic matter
as the cause of decreasing amounts of organically complexed Al. Whether this is the
case for the Poás samples is discussed below in the section on the relevance of the
buffering mechanisms under field conditions.
My calculations on the contribution of the protonation of variable charge and mineral
dissolution on acid buffering are based on congruent dissolution of homogeneous
plagioclase crystals during the laboratory experiment. Incongruent dissolution of
plagioclase (Muir & Nesbitt, 1997) or preferential dissolution of Ca rich plagioclase
patches over Na rich patches as described by Shotyk and Nesbitt (1992) for anor-
thite60, would lead to a lower Al:Ca ratio in the resin extracts compared to the mean
ratio determined by ESEM-EDX in plagioclase phenocrysts on thin sections and
thus to an overestimation of Al
plag
. Amrhein and Suarez (1992) analyzed the dissolu-
tion of anorthite in laboratory experiments in the absence and presence of exchange
resin and they found incongruent anorthite dissolution and the formation of a
leached layer in the absence of exchange resin, while dissolution in the presence of
exchange resin was congruent. Based on these results, I assumed that plagioclase
dissolved congruently during the determination of the ANC
res
. Nevertheless, plagio-
clase crystals in thin sections showed rhythmic zoning which could be due to het-
erogeneities in the Ca and Na content of the different zones. Therefore, a preferen-
tial dissolution of Ca-rich zones over Na-rich zones, as described by Shotyk and
Acid buffering
23
Nesbitt (1992) for anorthite60, cannot be ruled out. The fact that the Ca:Na ratio in
resin extracts (5.3) was larger than the Ca:Na ratio of the average plagioclase com-
position as determined by ESEM-EDX strengthens this possibility. If I assume con-
gruent dissolution of Ca-rich zones and calculate the mean plagioclase composition
based on the Ca:Na ratio in the resin extracts, I obtain a mean plagioclase composi-
tion of Na
0.16
Ca
0.84
Al
1.84
Si
2.16
O
8
and therefore a Al:Ca ratio of 2.19. This ratio is
smaller than the Al:Ca ratio of 2.47 determined by ESEM-EDX on thin sections by
13% and therefore Al
plag
as reported in Table 2.4 could be overestimated by up to
13%.
Table 2.4 Amounts of total released Al (Al
t
), Ca and Al released by plagioclase dissolution
(Ca
plag
, Al
plag
) and Al released from Al-organic complexes (Al
org
) in the resin and pH
stat
titra-
tion experiments of Poás samples
Sample Al
t
Ca
plag
Al
plag
Al
org
---------------------------------/ mmol
c
kg
-1
---------------------------------
resin experiment
P1 95.8 20.4 75.6 20.2
P2 226.3 33.9 125.6 100.7
P3 322.9 47.5 176.0 146.9
P4 290.4 27.8 103.0 187.4
P5 604.1 55.4 205.3 398.8
pH
stat
titration
P1 55.1 18.1 67.1 -12.0
P2 70.0 3.6 13.3 56.7
P3 94.4 3.8 14.1 80.3
P4 90.6 4.0 14.8 75.8
P5 197.5 4.0 14.8 182.7
On the other hand, the starting material of the resin experiment could already con-
tain a leached layer formed under field conditions. Such a leached layer, depleted in
Ca and Na as described by Shotyk and Nesbitt (1992) for plagioclase weathered
under acid conditions, could dissolve during the resin experiment at pH 3. If this is
the case, Al
plag
would be underestimated. Scanning electron microscopy with EDX
on plagioclase in thin sections of Poás samples did not show an altered or leached
surface layer. Nevertheless, leached layers on plagioclase in literature reports do
Chapter 2
24
not exceed a few hundred nanometers (e.g. Shotyk & Nesbitt, 1992), which is too
small to be detected with SEM-EDX, and therefore an underestimation of Al
plag
cannot be ruled out. For sample P1, the total amount of released Al during the resin
experiment was 27% larger than Al
plag
(Table 2.4), indicating that at least for this
sample Al
plag
cannot be underestimated by more than 27%. As the site P1 receives
larger acid inputs in the field than the sites P2-P5, any existing leached layer in the
starting material for the ANC
res
determination would be expected to be thickest in the
P1 sample, compared to P2-P5. Thus a possible Al
plag
underestimation should be
largest in P1 and should therefore not exceed 27%.
2.4.1.2 Short-term buffering under cation accumulation
The ANC
titr
and the release of DOC, Si and cations were substantially lower than in
the resin experiments (Figure 2.3). The ANC
titr
and the amounts of released Al were
highest in P5 but in contrast to the resin extraction no significant differences could
be observed between the other profiles. Dissolved organic carbon release followed
the same pattern as observed in the resin experiment but with significantly lower
amounts of released DOC during pH
stat
titrations. Calcium release was highest in P1
with no significant differences between the other horizons.
As described for the resin experiment I calculated the amount of Al released from
plagioclase weathering (Al
plag
) and from Al-organic complexes (Al
org
) during pH
stat
titration based on the amounts of released Ca (Table 2.4). The estimated amount of
Al originating from plagioclase dissolution was highest in P1 (67 mmol
c
kg
-1
) and for
this sample nearly equaled the amount of Al
plag
released during the resin experiment
(76 mmol
c
kg
-1
). In P2 - P5 Al
plag
was substantially lower for the pH
stat
titrations than
the amounts released in the resin experiment. Aluminum mobilized from Al-organic
complexes increased from P2 to P5, corresponding to 72 - 86% of the ANC
titr
. In the
case of P1 there was no statistical difference between Al
plag
and Al
tot
and therefore
the negative Al
org
value was not statistically different from zero.
The observation that in P2 - P5 substantially lower amounts of Al
plag
were released
in the pH
stat
titration compared to the resin experiment could be explained by an
inhibition of plagioclase dissolution during pH
stat
titrations. This could be caused by
an increase in Al solution concentrations due to the comparatively faster release of
Al from Al-organic complexes. The negative influence of high Al solution concentra-
tions on the dissolution of aluminosilicate minerals has been described by Oelkers et
Acid buffering
25
al. (1994). The inhibition of plagioclase dissolution in Poás samples was lowest in
P1, where negligible amounts of Al
org
were released.
ANCtitr / mmolc kg-1
0
200
400
600
800
1000
DOC / g kg-1
0
2
4
6
8
Al / mmolc kg-1
0
200
400
600
800
Si / mmol kg-1
0
20
40
60
80
100
120
140
Ca / mmolc kg-1
0
20
40
60
80
100
Na / mmolc kg-1
0
1
2
3
4
5
6
P1P2P3P4P5 M1M2 P1P2P3P4P5 M1M2
P1P2P3P4P5 M1M2 P1P2P3P4P5 M1M2
P1P2P3P4P5 M1M2 P1P2P3P4P5 M1M2
Figure 2.3 Acid neutralising capacity, released DOC, Al, Si, Ca and Na (mean values and
standard errors) during the pH
stat
titration experiment for Poas (P1 - P5) and Masaya (M1 &
M2) samples.
2.4.1.3 Relevance for the acid buffering under field conditions
Because of the perhumid climate, highly permeable soils and presumably large ion
uptakes by the dense vegetation at Poás, I supposed that the resin-containing sys-
tem resembles soil-solution conditions in the field much better than the closed sys-
tem of pH
stat
titrations. Based on this assumption I propose that both the protonation
of organic carbon-based variable charge and plagioclase weathering are the main
Chapter 2
26
short-term acid buffering processes in Poás soils, with different contributions from
each process, depending on the distance from the active caldera. A comparison of
ANCs from P1 to P5 with estimates of acid input shows that plagioclase weathering
prevails at P1, while at P3 to P5 ANC based on protonation of variable charge of
organic matter (ANC
prot
= Al
org
) is large enough to be able to buffer acid inputs at pH
3 for time spans from 8 to nearly 400 years (Table 2.5). For P2 the comparison
cannot be made, as sulfation rates were not measured. Thus, if the processes iden-
tified in the short-term acidification experiments at pH 3 were relevant for the long-
term acid buffering under present soil pH conditions, I expect a high degree of pro-
tonation of organic matter and of plagioclase weathering in sample P1 but hardly
weathered plagioclase and a high Al saturation of organic matter in P5.
Table 2.5 Comparison of the acid neutralizing capacity in the resin experiment attributable to
protonation of variable charge on organic matter (ANC
prot
) and to plagioclase dissolution
(ANC
plag
, as the difference between total ANC and ANC
prot
) with estimated acid deposition
rates at the Poás Volcano.
Sample ANC
prota)
ANC
plaga)
Acid deposition rate
/ kmol H
+
ha
-1
/ kmol H
+
ha
-1
/ kmol H
+
ha
-1
year
-1
P1 20.2 115.8 94
P2 100.7 209.9 nd
P3 146.9 268.1 19
P4 187.4 184.1 9
P5 398.8 335.9 1
a)
I assumed a bulk density of 1 g cm
-3
and a horizon thickness of 10 cm
nd: not determined
Infrared spectra of samples P1 - P5 showed a band at 1620 cm
-1
, caused by disso-
ciated carboxyl groups, but no band between 1710 and 1740 cm
-1
was observed,
indicating the absence of protonated carboxyl groups in Poás samples. Based on
this, I postulate that under field conditions the protonation of variable charge on
organic matter may rapidly buffer peaks in acid input but that, consecutively, cations
(especially Al) released by mineral weathering re-exchange with H
+
on soil organic
matter. Consequently, buffering does not result in protonated organic matter at the
study sites under present soil pH conditions. Thus, the observed decrease in the
concentration of organically complexed Al with increasing acid input from P5 to P1
Acid buffering
27
(Table 2.3) was not caused by different degrees of H and Al saturation of organic
matter but may have been caused by differences in the Al complexation capacity. A
lower complexation capacity at sites with larger H
+
deposition may be due to a lower
humus content and differences in organic matter quality. The latter can be caused
by different vegetation types. At Poás, sites P1 and P4 are dominated by myrtle
plants while sites P2, P3 and P5 are forest sites. The fact that the ratio of pyrophos-
phate-extractable Al to total organic matter was 0.04 and 0.03 for sites P1 and P4,
respectively, while sites P2, P3 and P5 had ratios of 0.08 - 0.09, reflects the differ-
ences in Al complexation capacity of organic matter between the vegetation types.
This difference, in connection with larger carbon contents at sites P4 and P5 than at
P1 - P3, seems to lead to the observed trend in organically complexed Al. The large
Al
p
:C
t
ratios at sites P2, P3 and P5 of 0.08 - 0.09 corresponded to an Al complexa-
tion capacity of organic matter of 20 - 23 mmol
c
g
-1
C and were in the upper range of
the potential acidity of soil organic matter as reported by Stevenson (1994).
The observations on thin sections confirmed the importance of plagioclase dissolu-
tion for the long-term acid buffering in Poás soils under field conditions. Petrographic
and scanning electron microscopy showed plagioclase crystals with weathering
channels and rather pristine crystals (Figure 2.4). The number and extension of the
weathering channels tended to be larger in P1 than in the other samples; neverthe-
less, quantification by image-analysis was not done.
2.4.2 Mechanisms of H
+
buffering at Masaya
2.4.2.1 Short-term buffering under cation removal
The ANC
res
was larger in the 2BC horizon from M2 (low acid input) than from M1
(high acid input) (Figure 2.2). Correspondingly, the amounts of released DOC, Al, Si,
Ca, Mg, K and Na were larger for M2 than for M1. Compared to Poás samples,
dissolved organic carbon in M2 was about half the amount released from P2 - P4
and one fifth of the amount released from P5; for M1 the amounts of released DOC
were even lower, suggesting that H
+
buffering due to protonation of variable charge
of organic matter is of comparably less importance in Masaya soils. In addition, dry
samples of Masaya soils were analyzed, which could have caused a liberation of
hydrophilic DOC from microbial biomass residues at the moment of wetting (Kaiser
et al., 2001). Therefore it is likely that DOC release due to protonation of variable
charge on organic matter was even lower than the measured total DOC release. For
Chapter 2
28
the low contribution of Al-organic complexes to ANC
res
I supposed that for Masaya
samples the sum of NH
4
Cl-extractable cations (ECEC) represents a quite good
estimate of the fraction of ANC
res
based on protonation of variable charge (ANC
prot
).
Figure 2.4 Scanning electron microscope images of thin sections of grains from the 2AC
horizons of sample P5 (4a, 4b) and sample P1 (4c, 4d) showing representative plagioclase,
magnetite and pyroxene crystals in a glass matrix. Larger mineral crystals have been de-
nominated with P for plagioclase, Py for pyroxene and M for magnetite.
The amounts of cations released during the resin experiment (Figure 2.2) were
much larger than the amounts of exchangeable cations (Table 2.3), indicating that
protonation of variable charge cannot be the only source of H
+
buffering. Additional
buffering mechanisms may be the dissolution of allophane, primary silicate minerals
(plagioclase and pyroxene, identified by X-ray diffraction), glass or basic aluminum
sulfate minerals that have been identified in Masaya soils by Delfosse et al. (2005c).
In an attempt to elucidate which of the mentioned minerals may contribute to the
ANC, I estimated the amount of Al mobilized through dissolution (Al
diss
) by subtract-
ing the amount of NH
4
Cl-exchangeable Al (Al
NH4Cl
) from the total amount of Al re-
leased during the resin experiment and calculated thereafter the ratio of Al
diss
:Si.
The ratio was 3.0 for M1 and 1.9 for M2. The comparison of these ratios with (Al
o
-
a
b
c
d
Acid buffering
29
Al
p
)/Si
o
ratios of allophane in 2BC horizons of Masaya soils (M1: 1.4, M2: 1.1) indi-
cates that allophane cannot be the only source of released Al. Besides, Si
o
in M1
was about double the amount of Si
o
in M2 (Table 2.3), while Si released in the resin
experiment showed the opposite behavior. The latter suggests, especially for M2, an
additional source for Si - and concomitant cation - release. This could be the disso-
lution of any of the primary silicate minerals (plagioclase or pyroxene) or glass but
with the present data I was not able to decide which one was dissolved. Parallel to
Poás results, plagioclase could be the main dissolving mineral. Nevertheless, as
dissolution rates depend on mineral composition and the spatial distribution in the
sample and as Poás and Masaya samples possibly differ in these characteristics, a
conclusion on Masaya samples cannot be drawn from the results of Poás samples.
Aluminum could also be released by the dissolution of basic aluminum sulfate min-
erals (BAS). Nevertheless, the amounts of released sulfate during the resin experi-
ment (4.6 and 5.7 mmol
c
kg
-1
for M1 and M2, respectively) corresponded to 28 and
34 mmol
c
Al kg
-1
from the dissolution of basaluminite, which accounted for only 5 -
7% of Al
diss
.
2.4.2.2 Short-term buffering under cation accumulation
Similarly to Poás, the ANC
titr
of Masaya soils was lower than the ANC
res
and no
differences in ANC
titr
could be observed between M1 and M2. As described for the
results of the resin experiments, part of the released cations could be due to proto-
nation of variable charge on organic matter. The kinetics of cation exchange are
faster than those of mineral dissolution. Therefore, the results could be explained by
the rapid release of cations from the exchange sites, leading to an increase in cation
solution concentrations and thereby to product limitation of mineral weathering.
2.4.2.3 Relevance for the acid buffering under field conditions
The climate at Masaya is much drier than that at Poás, with potential evapo-
transpiration exceeding rainfall during the dry season (UNA & ISRIC, 1994). Also,
infiltration of rain water is hampered by indurated soil horizons, locally called
“talpetate”. Therefore, I suggest that for the dry season (and the beginning and end
of the rainy season) the closed pH
stat
system resembles soil solution conditions
better than the open, resin-containing system. Based on this assumption I conclude
that protonation of variable charge and mineral or glass dissolution are the short-
Chapter 2
30
term acid buffering processes during the dry season but that mineral / glass dissolu-
tion is hampered by the increase in cation solution concentrations due to the release
from exchange sites. During the rainy season, part of the dissolution products may
be leached, thus increasing mineral dissolution, and consequently temporarily in-
creasing the effective acid neutralizing capacity of Masaya soils.
Acid deposition rates around Masaya have been measured by Delmelle et al. (2001)
during the dry season. The comparison of these deposition rates with the ANC
attributable to the protonation of variable charge (ANC
prot
= ECEC) and the fraction
of the ANC
titr
attributable to mineral dissolution (ANC
diss
, as the difference between
total ANC
titr
and ANC
prot
), indicates that the dissolution of minerals plays a dominant
role in the long-term acid buffering in Masaya soils (Table 2.6). This is confirmed by
the fact that the IR spectra of Masaya samples did not show the absorption band
between 1710 and 1740 cm
-1
attributable to protonated carboxyl groups. Neverthe-
less, mineral dissolution can be hampered by product limitation, thus leading under
continuing acid input to a decrease in soil pH, as described by Delmelle et al.
(2003). Also, ion accumulation in soil solution favours the formation of secondary
minerals.
Table 2.6 Comparison of acid neutralizing capacity in the pH
stat
experiment attributable to
protonation of variable charge (ANC
prot
) and to mineral dissolution (ANC
diss
) with estimated
acid deposition rates
a)
at Masaya Volcano.
Sample ANC
protb)
ANC
dissb)
Acid deposition rate
/ kmol H
+
ha
-1
/ kmol H
+
ha
-1
/ kmol H
+
ha
-1
year
-1
M1 21.6 151.8 ~90
M2 45.7 126.3 ~50
a)
data taken from the map published in Delmelle et al. (2001)
b)
I assumed a bulk density of 1 g cm
-3
and a horizon thickness of 10 cm
2.4.3 Formation of secondary minerals
In Poás samples, all SSEs were undersaturated with respect to amorphous Al(OH)
3
and basaluminite (Al
4
SO
4
(OH)
10.
4H
2
O), with saturation indices ranging from -0.4 to -
4.4 and -0.4 to -22, respectively. Therefore I suppose that Al solution concentrations
may be controlled by Al-organic complexes, as described by Farmer (1990) for
Podzol Bs horizons and by Skyllberg (1999) for Podzol O and E horizons. Aluminum
Acid buffering
31
released from these complexes by acidification may be leached by percolating
water. Results of the geochemical speciation indicate [Al(SO
4
)
+
]
aq
as the dominant Al
species in solution. Therefore aluminum mobilized during acidification may be
leached as [Al(SO
4
)
+
]
aq
, leading to crypto-podzolisation as described by Duchaufour
and Souchier (1965, cited in Ulrich, 1990) for soils affected by acid deposition. The
question of whether leached Al is immobilized in deeper horizons cannot be an-
swered by the analytical results for the depth profiles, as soils around Poás Volcano
are stratified.
In the case of Si, saturation indices from -0.7 to 0.13 point to an equilibrium of SSEs
with amorphous SiO
2
in part of the samples (Figure 2.5). Additionally, saturation
indices for amorphous SiO
2
tended to increase from P5 (low acid load) to P1 (high
acid load), suggesting that larger acid inputs at P1 cause an increase in plagioclase
weathering, which in turn leads to higher Si solution concentrations. This hypothesis
is strengthened by a significant correlation between Si and sulfate in SSEs (r
2
= 0.62
with n = 53, p < 0.05), as SO
2
is the main source of acid input at Poás Volcano.
Scanning electron microscope images (SEM) of sand grains from 2AC horizons
showed dominantly bare grain surfaces and to a lesser extent grains with areas of
different morphology, as shown in Figure 2.6a. Energy-dispersive X-ray analyses of
the differing areas revealed zones with Si enrichment (Figure 2.6b) above zones
constituted by Si, Al, Fe and basic cations (Figure 2.6c). The areas enriched in Si
could be interpreted as SiO
2
secondary phases or as a residual Si accumulation on
mineral surfaces due to cation leaching (formation of a leached layer). Nevertheless,
the comparison of the thickness of the SiO
2
-enriched area in Figure 2.6a (approxi-
mately 5 µm) with published data on leached layer dimensions, not exceeding a few
hundred nanometers (e.g. Hamilton et al., 2000; Shotyk & Nesbitt, 1992; Muir &
Nesbitt, 1997), renders this possibility unlikely. With SEM-EDX analyses it is not
possible to decide whether SiO
2
-enriched zones are constituted by amorphous or by
crystalline SiO
2
like disordered cristobalite, traces of which I detected by X-ray dif-
fraction analyses. In order to elucidate whether amorphous SiO
2
phases are present
in Poás samples, I examined mineral grains with petrographic microscopy by the
immersion method with glycerine. Volcanic glass, cristobalite and amorphous SiO
2
are all isotropic phases. Glass and cristobalite have a refractive index larger than
the 1.48 of glycerine, while amorphous SiO
2
has a refractive index smaller than 1.48
(Sedov et al., 2003). Based on this fact, the Becke test makes it possible to distin-
Chapter 2
32
guish between glass / cristobalite and amorphous SiO
2
and showed in the Poás
samples that amorphous SiO
2
was present.
saturation index
-0.6 -0.4 -0.2 0.0 0.2
depth (cm)
-100
-80
-60
-40
-20
0
P1
saturation index
-0.6 -0.4 -0.2 0.0 0.2
P2
depth (cm)
-100
-80
-60
-40
-20
0
P3
depth (cm)
-100
-80
-60
-40
-20
0
P5
P4
depth (cm)
-100
-80
-60
-40
-20
0
M1 M2
Figure 2.5 Depth profiles of saturation indices (SI) for amorphous SiO
2
in soil saturation
extracts (calculated with the geochemical model Visual Minteq) in profiles P1 - P5 of Poás
Volcano and M1 - M2 of Masaya Volcano. (Dotted lines represent the lower and upper limits
of the SI, indicating an equilibrium between solution concentrations and amorphous SiO
2
).
Acid buffering
33
These findings agree with the results of Jongmans et al. (1996), who described the
formation of silica-rich crusts at soil surfaces on the slopes of Costa Rican active
volcanoes, and Poetsch (2004) and Usai and Dalrymple (2003), who report the
formation of silica-rich pedofeatures in volcanic soils of Mexico and Italy. Similarly,
the formation of amorphous Si precipitates on mineral grains has been reported by
Veerhoff and Brümmer (1993) for extremely acid soils in the temperate region.
The SSEs of all Masaya samples were undersaturated with respect to amorphous
Al(OH)
3
. Contrastingly, slightly positive saturation indices for basaluminite in some
of the samples suggest that Al solution concentrations may be controlled by this
phase. This agrees with the results of Delfosse et al. (2005c), who deduced the
existence of basic aluminum sulfate minerals in Masaya soils from SEM-EDX analy-
ses. Saturation indices for SiO
2
in SSEs ranged from -0.3 to 0.03 in Masaya soils
(Figure 2.5), suggesting the existence of silica precipitates as also found in the Poás
samples.
Figure 2.6 Scanning electron microscope image of a sand grain (a) with EDX analyses for
points A (b) and B (c). The grain originates from the 2AC horizon of profile P1, Poás.
2.5 CONCLUSIONS
Short-term decreases in soil pH as a consequence of large peaks in acid input at
Poás Volcano lead to the protonation of variable charge on organic matter and to
plagioclase dissolution. Due to differences in organic matter quantity and quality, the
contribution of organic matter to the short-term acid neutralizing capacity is larger at
control sites than at sites with high acid deposition, where plagioclase dissolution
dominates. Under field conditions with soil pH values of 4.5 - 5, the protonation of
variable charge is of little importance for the acid buffering, as indicated by the
keV
0 1 2 3 4 5
counts
0
2000
4000
6000
8000 Si
Al
keV
0 1 2 3 4 5
counts
0
2000
4000
6000
8000
Si
Al
Na KCa
a
b
c
Chapter 2
34
absence of protonated carboxyl groups in all Poás samples. Plagioclase dissolution
is the main buffering process under field conditions. Mobilized Al seems to be sus-
ceptible to leaching in the form of [Al(SO
4
)
+
]
aq
, as soil saturation extracts show un-
dersaturation with respect to any possible Al-containing secondary phase. Con-
versely, increased plagioclase dissolution due to large acid input leads to an
increase in Si solution concentrations and the consecutive precipitation of amor-
phous SiO
2
on mineral surfaces.
At Masaya Volcano, short-term acid buffering at pH 3 and buffering under field
conditions with a mean soil pH value of 4.6 are governed by mineral and/or glass
dissolution. Climatic conditions and restricted infiltration are supposed to favor the
accumulation of Si and cations in solution. This in turn hampers mineral weathering
due to product limitation and favors the formation of secondary minerals such as
basic aluminum sulfates and amorphous silica. As a consequence of product limita-
tion, weathering rates can get too low to buffer acid deposition effectively, leading to
a pH decrease in Masaya soils.
My results imply that the relevance of different buffering processes changes as a
response to highly fluctuating acid deposition rates on soils in the vicinity of active
volcanoes. Also, the quantity and quality of acid buffering depend on whether reac-
tion products are removed by leaching or not. Without leaching, the rapid release of
cations due to protonation of variable charge seems to hamper mineral weathering,
leading to a decrease of the effective acid neutralization capacity.
35
3 SIMULTANEOUS THERMAL ANALYSIS FOR THE IDENTIFICATION OF S-
POOLS IN SOILS – POSSIBILITIES AND LIMITATIONS
3.1 SUMMARY
The differentiation of basic aluminum sulfate phases (BAS), adsorbed sulfate and
estersulfate compounds in soils is of ecological relevance. However, existing meth-
ods (selective extraction procedures, S K-edge X-ray absorption near edge structure
spectroscopy and scanning-electron microscopy with energy dispersive X-ray analy-
sis) are limited in their suitability for the differentiation of these phases. Therefore,
possibilities and limitations of simultaneous thermal analysis (differential scanning
calorimetry and thermogravimetry) coupled with mass spectrometry (STA-MS) as a
fingerprint method for the identification of S pools in soils were tested. Reference
compounds (basaluminite, K-alunite, Na-alunite, Na-dodecylsulfate and sulfate
adsorbed to allophane) were analyzed by STA-MS in pure form and in physical
mixture with simple matrix components (allophane, SiO
2
and Al
2
O
3
). Based on MS
signals of SO and H
2
O pure reference compounds can be distinguished from each
other. However, the addition of matrix components partly leads to overlapping MS
signals limiting the suitability of STA-MS for the identification of S pools to samples
of simple and known matrix composition. Similarly, the CO
2
release pattern of Na-
dodecylsulfate is affected by the addition of matrix components. In the light of these
results existing measurements of thermal stabilities of organic carbon pools in soils
of varying mineralogy may have to be re-evaluated based on the analysis of organic
reference compounds in mixture with various soil minerals.
Chapter 3
36
3.2 INTRODUCTION
Sulfur is retained in aerated soils due to (i) sulfate adsorption to variable charge
minerals, (ii) the precipitation of basic aluminum sulfate phases (BAS) and (iii) by
incorporation into organic matter. Large sulfur inputs into soils can be caused by
anthropogenic combustion processes (e.g. Kopacek et al., 2001) or locally by natu-
ral sources as e.g. active volcanoes (Delmelle et al., 2003). The effect of large sulfur
inputs on sulfur retention processes in soils is a matter of ongoing scientific debate
(Delfosse et al., 2005b; Prietzel et al., 2008). Process understanding, however, is
the basis for the prediction of sulfate dynamics as (i) the products of different proc-
esses are expected to differ with respect to their formation and dissolution kinetics
and (ii) adsorption is limited by the number of available sites while inorganic precipi-
tation or incorporation into organic substances can be infinite sulfate sinks.
Generally, the identification of BAS phases in soils is complicated as these phases
are very fine grained and amorphous so that classical methods as e.g. X-ray diffrac-
tion (XRD) fail to identify them. Until recently, adsorbed sulfate, BAS phases and
organically bound sulfate were distinguished from each other by applying different
dissolution procedures to soils (e.g. Prietzel et al., 2001; Delfosse et al., 2005b).
Nevertheless, such procedures are operationally defined and not selective sensus
strictus and thus, they do not allow an unambiguous differentiation between different
forms of sulfate binding in soils (Prietzel et al., 2003). Very recently, Prietzel et al.
(2008) showed that BAS phases in soils can be distinguished from adsorbed sulfate
by S K-edge X-ray absorption near edge structure spectroscopy (S K-edge XANES).
However, XANES does not allow distinguishing between BAS phases and organi-
cally bound sulfate as e.g. estersulfate compounds (Prietzel et al., 2008). Addition-
ally, due to small differences in XANES spectra of different BAS phases as e.g.
basaluminite and alunite, the quantification of precipitated and adsorbed sulfate in
the absence of organic substances can at best be semi-quantitative unless the exact
nature of precipitated phases is known. Although basaluminite and alunite can be
distinguished in pure form from each other by differences in their XANES spectra,
the differentiation is limited when these phases are only present in small amounts in
the analyzed samples. Nevertheless, the differentiation between basaluminite and
alunite is of ecological relevance as both phases differ in their formation and dissolu-
tion kinetics (Nordstrom, 1982).
STA-MS
37
To my knowledge, simultaneous thermal analysis (STA) has not been tested so far
for its ability to identify and potentially (semi-)quantify different S pools in soils.
However, the combination of differential thermal analysis (DTA), differential scan-
ning calorimetry (DSC) and / or thermogravimetry (TG) has been applied in numer-
ous studies in soil science for the identification of organic carbon pools of different
thermal stabilities (e.g. Leinweber and Schulten, 1992; Friedrich et al., 1996; Siewert
C., 2001; Lopez-Capel et al., 2005; Plante et al., 2005; Dorodnikov et al., 2007). In
analogy to these studies, I hypothesized that a differentiation between the before
mentioned sulfate pools in soils may be possible due to differences in their thermal
stabilities. Thus, I investigated whether STA (DSC/TG) coupled with mass spec-
trometry (MS) allows the differentiation between organically bound sulfate, the BAS
phases basaluminite and alunite and adsorbed sulfate. I tested this for pure com-
pounds and for mixtures of the pure compounds with simple matrix components.
3.3 MATERIALS AND METHODS
3.3.1 Experimental approach
The sulfur containing reference compounds (i) amorphous basaluminite, (ii) K-
alunite, (iii) Na-alunite and (iv) the estersulfate compound Na-dodecylsulfate were
analyzed by STA-MS. All reference compounds were analyzed in pure form. Addi-
tionally, amorphous basaluminite, Na-alunite and Na-dodecylsulfate were analyzed
in mixtures with allophane, Al
2
O
3
or SiO
2
. Allophane, Al
2
O
3
and SiO
2
were chosen as
model compounds for secondary minerals of sulfate-retaining volcanic soils. By
comparison of the mass spectrometry signals of S-containing gases and H
2
O I
analyzed if the mentioned S containing reference compounds can be distinguished
from each other. Additionally, a sample was generated in the laboratory which pre-
sumably contains dominantly sulfate adsorbed to allophane (Ref
ads
) and STA-MS
signals of this sample were compared with signals of the reference compounds in
mixtures with allophane in order to evaluate if Ref
ads
has a STA-MS signal different
from the signals of the reference compounds thus allowing the differentiation be-
tween adsorbed and precipitated or organically bound sulfate. Besides, on a se-
lected sample mixture the influence of the mixing ratio on STA-MS signals has been
studied.
Chapter 3
38
3.3.2 Sulfur containing reference compounds
Sodium dodecylsulfate (C
12
H
25
NaO
4
S for biochemistry and tenside testing) was
purchased from Merck (Darmstadt, Germany). The other sulfur containing reference
compounds were synthesized in the laboratory by titration of an Al
2
(SO
4
)
3
solution
with an alkali solution up to a molar OH/Al ratio of 2.0 similar as described by Adams
and Rawajfih (1977) and Prietzel and Hirsch (1998). Amorphous basaluminite and
Na-alunite were synthesized by adding 40 mL 1 M NaOH (Titrisol
) dropwise to
200 mL 0.05 M Al
2
(SO
4
)
3
*16H
2
O under constant stirring. Potassium alunite was
synthesized by adding 1M KOH instead of NaOH. Precipitates formed in the synthe-
sis of amorphous basaluminite were washed immediately after precipitation while
precipitates of the synthesis of Na-alunite and K-alunite were allowed to age at
50±0.5 °C for 16 weeks before washing. Washing was done by repeated addition of
doubly deionized water until the conductivity was lower than 100 µS cm
-1
followed
by three washings with acetone. Precipitates were allowed to dry for two days at
40 °C. They were then finely ground in an agate mortar and stored in polyethylene
bottles in a desiccator. For each reference compound 10 replicates were conducted
and the obtained products were pooled at the end of the synthesis procedure. The
synthesis products were characterized by XRD analysis (Philips PW 3710, Cu-Kα
radiation).
3.3.3 Matrix components for mixtures
Allophane was synthesized according to Ohashi et al. (2002) with minor modifica-
tions. Briefly, 225 g of 100 mM Na
4
SiO
4
and 300 g of 100 mM AlCl
3
* 6H
2
O were
mixed rapidly, stirred for one hour and centrifuged for two hours at 7600 g. After
decanting the clear supernatant, 200 mL of doubly deionized H
2
O were added, the
allophane precursors were resuspended and the suspension was boiled for 48
hours under a reflux condenser. Remaining salts were eliminated by dialysis (Spec-
tra/Por
Dialysis membrane, MWCO: 12-14,000, Spectrum Laboratories, Inc., Ran-
cho Dominguez, USA) until the electrical conductivity of the suspension was lower
than 10 µS cm
-1
. The suspension was centrifuged for four hours at 7600 g and the
clear supernatant was decanted. By the described procedure an allophane gel was
obtained which was re-suspended in H
2
O and freeze-dried after shock-freezing the
suspension by dropwise injection into liquid N
2
(-196°C). The allophane was charac-
terized by X-ray diffraction (Philips PW 3710), differential scanning calorimetry -
STA-MS
39
thermogravimetry coupled with mass spectrometry (Netzsch 409 PC – Pfeiffer
Thermostar MS), mid and far FT infrared spectroscopy (Thermo Nicolet Nexus),
oxalate extraction according to Schwertmann (1964), and transmission electron
microscopy (Philips CM 12 TEM, FEI Company, Oregon, USA). By the described
synthesis procedure allophane with the formula 1.25 SiO
2
* Al
2
O
3
* 3.2 H
2
O and a
Al:Si ratio of 1.6 was obtained. All other mentioned analyses gave results character-
istic for allophane.
The matrix compounds SiO
2
(SiO
2
60 for column chromatography, 70-230 mesh,
ASTM) and Al
2
O
3
(type 506-C-I, 150 mesh ASTM) were purchased from Merck
(Darmstadt, Germany) and Aldrich (St. Louis, USA), respectively.
3.3.4 Synthesis of sample Ref
ads
(dominantly sulfate adsorbed to allophane)
The question on the relevance of adsorption and precipitation processes for sulfate
retention in volcanic soils has not been answered so far. The same holds true for
sulfate retention by the mineral allophane. Consequently, the reference compound
“sulfate adsorbed to allophane” cannot be synthesized without ambiguity as precipi-
tated phases may also form during sulfate retention by allophane. I selected synthe-
sis conditions such (rather high pH resulting in low aluminum solution concentra-
tions) that a formation of precipitated phases is unlikely. The potential presence of
such phases in Ref
ads
will be discussed in section 3.5.2.
Freeze-dried allophane (0.80 g) was weighed into a reaction vessel (120mL, PP),
40 mL of 100 mM KCl were added and the suspension was stirred for 4.5 hours.
Subsequently, 40 mL of 6 mM K
2
SO
4
(in 100 mM KCl) were added and the sample
was stirred for further 15.5 hours. After a total of 20 hours the suspension was fil-
tered through a membrane filter (cellulose nitrate, 0.45 µm, Sartorius AG, Göttingen,
Germany) and the visibly clear filtrate was centrifuged for one hour at 159 000 g
(Beckman Coulter Optima
TM
L-90K, Fullerton, USA). The filter with the solid product
was washed by filtration of 40 mL doubly deionized H
2
O, frozen at –20° C, freeze-
dried and the dry solid product stored in a desiccator. The filtrate and an aliquot of
the K
2
SO
4
solution were analyzed for sulfate by anion chromatography (DX-120,
Dionex, Sunnyvale, USA, eluent: 3.5 mM NaCO
3
/ 1 mM NaHCO
3
), and for Al and Si
by flame atomic absorption spectroscopy (Perkin Elmer 1100 B, Waltham, USA).
The pH was not controlled and was 5.2 at the beginning and 5.7 at the end of the
experiment. The Al solution concentration in this sample was 26 µg L
-1
at the end of
the experiment.
Chapter 3
40
3.3.5 Pure samples and mixtures of reference compounds with matrix compo-
nents for STA-MS analysis
Pure S containing reference compounds and mixtures of these compounds with the
before mentioned matrix components were analyzed by STA-MS without any sam-
ple pretreatment. Sample numbering, mixing ratios and sample weights for analysis
are shown in Table 3.1.
3.3.6 Simultaneous thermal analysis coupled with mass spectrometry (STA-
MS)
All samples were weighed in Pt/Rh crucibles, tapped by a Pt/Rh lid with a small
orifice for gas release and analyzed with a STA 449C Jupiter (Netzsch, Selb, Ger-
many) equipped with a thermogravimetry / differential scanning calorimetry
(TG/DSC) sample holder. An empty Pt/Rh crucible was used as reference. Thermo-
gravimetric and differential scanning calorimetry signals were recorded and released
gases were detected by a coupled quadrupole mass spectrometer (QMS 403C
Aëolos, Netzsch, Selb, Germany) which was connected by a heated capillary to the
STA 449C to prevent condensation during gas transport to the MS. Signals of
masses 18, 44, 48, 64, 80 and 96 were recorded corresponding to H
2
O, CO
2
, SO,
SO
2
, S
2
O/SO
3
and S
2
O
2
, respectively. Water vapor and sulfur containing gases were
analyzed as previous studies for alunites showed that such phases show patterns of
characteristic H
2
O and S-release (Frost et al., 2006). Carbon dioxide was analyzed
for the interpretation of STA-MS signals of Na-dodecylsulfate. After a 10 minutes
isothermal segment at 35 °C, samples were heated under an atmosphere of air / N
2
(gas fluxes 80 / 20 mL min
-1
, respectively) up to 1400 °C with a heating rate of 10 °C
min
-1
. Measurement signals were analyzed with the software package Proteus
Version 4.8.1 (Netzsch, Selb, Germany). For the differentiation between S contain-
ing reference compounds mass spectrometry signals of H
2
O and S containing gases
were evaluated as these signals are more sensitive to specific differences between
samples than DSC or TG signals. Mass spectrometry signals were smoothed by the
Golay Savitzky algorithm.
STA-MS
41
Table 3.1 Sample numbering, composition, mixing ratios (matrix:reference compound) and
sample weights for STA-MS analysis.
Sample number Reference compound Mixing ratio Sample weight (mg)
Pure reference compounds
1 basaluminite - 25
2 Na-alunite - 27
3 K-alunite - 101
4 Na-Dodecylsulfate - 26
5 Ref
ads
- 97
Mixtures with allophane
6 basaluminite 3:1 100
7 Na-alunite 3:1 101
8 Na-Dodecylsulfate 19:1 100
9 Na-Dodecylsulfate 107:1 100
Mixtures with SiO
2
10 basaluminite 3:1 84
11 Na-alunite 3:1 97
12 Na-Dodecylsulfate 3:1 100
Mixtures with Al
2
O
3
13 basaluminite 3:1 100
14 Na-alunite 3:1 101
15 Na-Dodecylsulfate 3:1 102
3.4 RESULTS
3.4.1 Characterization of the synthesized S containing reference compounds
by X-ray diffraction
For amorphous basaluminite two broad humps could be observed in the XRD pat-
tern reflecting the amorphous nature of the precipitated phase (Figure 3.1). No
reflexes of crystalline phases were present. Total elemental analysis resulted in Al
and S concentrations typical for basaluminite. For K-alunite the typical reflexes of a
well crystallized K-alunite could be observed without any hints on the presence of
further phases. Contrastingly, for Na-alunite XRD patterns showed that the sample
contains well crystallized Na-alunite and additionally traces of a poorly crystallized
basaluminite-like phase as indicated by the reflexes as shown in Figure 3.1.
Chapter 3
42
2Theta
0 10 20 30 40 50 60 70
Intensity
0
200
400
600
800
1000
2Theta
0 10 20 30 40 50 60 70
Intensity
0
1000
2000
3000
4000
5000
KA
KA
KA
KA
KA KA
KA
KAKA
KA KA
KAKAKA
2Theta
0 10 20 30 40 50 60 70
Intensity
0
1000
2000
3000
4000
NA
NA
NA
NA
NA
NA NA NA NA
NA
Bas Bas
Figure 3.1 XRD patterns of the synthesized S containing reference compounds (a) amor-
phous basaluminite (b) K-alunite and (c) Na-alunite
3.4.2 Mass spectrometry signals of STA-MS analysis
3.4.2.1 Pure S containing reference compounds
For all analyzed pure reference compounds as well as for sample mixtures de-
scribed in section 3.4.2.2 SO was the dominant S-containing gas that could be
detected. Besides, traces of SO
2
were present showing an identical dependence on
sample temperature as SO. Sulfur trioxide, S
2
O or S
2
O
2
could not be detected.
Gases detected by MS not necessarily reflect the nature of gases released by the
samples as gases can be transformed during ionization in the mass spectrometer
(cracking pattern). In the following only SO signals will be presented in combination
with H
2
O signals, for Na-dodecylsulfate additionally CO
2
signals are discussed. As
the sensitivity of the mass spectrometer in STA-MS analysis generally fluctuates
over measurement periods of days and weeks, signals of the analyzed samples
cannot be interpreted quantitatively by comparing peak heights. Thus, only peak
positions and peak height ratios within each measurement were interpreted. Mass
a
b
c
STA-MS
43
spectrometry signals have been partially rescaled for better visual comparability of
signals from different samples.
All reference compounds showed several steps of water release (Figure 3.2). In the
case of alunites the H
2
O release up to ~200 °C may be attributed to the re lease of
adsorbed water while the H
2
O release at higher temperatures most probably is due
to dehydroxylation processes (Frost et al., 2006). A first flush of sulfur release from
these samples between 700 °C and 1000 °C can be att ributed to the decomposition
of a dehydrated and dehydroxylated alunite-derived compound releasing a sulfur-
containing gas under the simultaneous formation of Al
2
O
3
and K
2
SO
4
/ Na
2
SO
4
(Frost et al., 2006). The thereby formed sulfate salts decompose further at tempera-
tures larger than 1300 °C as can be observed in Figure 3.2. For basaluminite a
similar sequence of processes may be postulated with dehydration, dehydroxylation
and sulfur release from an Al-S-containing compound. Dehydroxylation tempera-
tures seemed to be lower for basaluminite compared with the alunitic compounds.
Sulfur is released completely in one stage around 990 °C which was affirmed by the
fact that the sulfur content of the basaluminite sample after STA-MS analysis was
below the detection limit of 0.2 g kg
-1
. Traces of basaluminite, which are present in
the Na-alunite sample according to XRD analysis, could not be identified based on
the SO MS signals. This may be attributed most probably to a masking effect due to
a large and gently decreasing SO release from Na-alunite. The organic compound
Na-dodecylsulfate released H
2
O and SO between 200 °C and 400 °C due to the
oxidation of the organic material with a simultaneous CO
2
release (data not shown),
for SO an additional release could be observed between 1200 °C and 1400 °C
which may be attributed to the formation and consecutive decomposition of an
alkali-sulfate compound during STA-MS as described in the case of alunites.
Chapter 3
44
temperature (°C)
0 400 800 1200
S
org
K-alunite
Na-alunite
basaluminite
Ion current (m/z=18, H
2
O) / a.u.
temperature (°C)
0 400 800 1200
S
org
K-alunite
Na-alunite
basaluminite
Ion current (m/z=48, SO) / a.u.
Figure 3.2 SO and H
2
O mass spectrometry signals during STA-MS analysis of the pure
reference compounds basaluminite, N-alunite, K-alunite and Na-dodecylsulfate (S
org
).
3.4.2.2 Mixtures of S containing reference compounds with matrix components
3.4.2.2.1 Mixtures with allophane: When allophane was added an additional water
release peak could be observed around 180 °C (Figure 3.3) which is attributable to
dehydration and dehydroxylation processes of allophane (Wada, 1989). The large
H
2
O release from allophane partly masked the H
2
O release from the S containing
reference compounds, especially in the case of basaluminite and to a lower extent
for Na-dodecylsulfate and Na-alunite. Nevertheless, for Na-alunite the water release
signal due to dehydroxylation processes around 500 °C is only little influenced by
the allophane water release signal and can thus be identified unambiguously.
Not surprisingly, from pure allophane no sulfur release could be detected (signal not
shown). However, the S release patterns of the S containing reference compounds
were at least partially influenced by the admixture of allophane (Figure 3.3). In the
case of Na-alunite the main sulfur release has shifted from approximately 780 °C to
880 °C and a second maximum of SO release can be observed between 1000 °C
and 1200 °C. The SO maximum between 1200 °C and 140 0 °C which could be
observed in the analysis of the pure compound and which was attributed to the
decomposition of Na
2
SO
4
(Frost et al., 2006) disappeared when allophane was
added. The SO release from basaluminite was only very slightly affected by the
addition of allophane with a small increase in the SO release in the signal shoulder
previous to the main SO peak. Contrastingly, the SO signal of Na-dodecylsulfate
was influenced remarkably by the addition of allophane. With an allophane:S
org
STA-MS
45
mixing ratio of 19:1, the main SO release between 200 °C and 400 °C was still
observable but disappeared nearly completely when the mixing ratio was 107:1. In
contrast, SO was released around 500 °C and 900 °C which could not be observed
for pure S
org
. Compared with the SO signal of the pure compound, the SO peak
between 1200 °C and 1400 °C disappeared completely or shifted to lower tempera-
tures for mixing ratios of 19:1 and 107:1, respectively.
temperature (°C)
0 400 800 1200
Sorg+allo (1:19)
Sorg
Na-alunite+allo
Na-alunite
basaluminite+allo
basaluminite
Ion current (m/z=18, H
2
O) / a.u.
temperature (°C)
0 400 800 1200
Sorg+allo
Sorg
Na-alunite+allo
Na-alunite
basaluminite+allo
basaluminite
Ion current (m/z=48, SO) / a.u.
Figure 3.3 SO and H
2
O mass spectrometry signals during STA-MS analysis of the reference
compounds basaluminite, N-alunite, and Na-dodecylsulfate (S
org
) in mixture with allophane.
In the case of S
org
SO signals of two mixtures with different mixing ratios are shown (dashed
line: mixing ratio 19:1, solid line: mixing ratio 107:1).
3.4.2.2.2 Mixtures with SiO
2
: Similarly as described for the mixtures with allophane,
an additional H
2
O peak could be observed in the temperature range 30 °C – 200 °C
due to the addition of SiO
2
(Figure 3.4). The water release from basaluminite and
Na-alunite was not affected by SiO
2
addition and for Na-alunite the signal attribut-
able to dehydroxylation processes around 500 °C cou ld be identified unambigu-
ously. In the case of Na-dodecylsulfate the addition of SiO
2
influenced the shape
and position of the main water release signal slightly.
For all S containing reference compounds the addition of SiO
2
affected the release
of SO only to a minor extent (Figure 3.4). The basaluminite signal showed no sub-
stantial change. In the case of Na-alunite the SO signal around 700 °C was not
affected while the signal between 1200 °C and 1400 °C shifted to lower tempera-
Chapter 3
46
tures. The latter was also observed for Na-dodecylsulfate, for which additionally the
shape of the main SO release peak was changed due to SiO
2
addition.
temperature (°C)
0 400 800 1200
S
org
+SiO
2
S
org
Na-alunite+SiO
2
Na-alunite
basaluminite+SiO
2
basaluminite
Ion current (m/z=18, H
2
O) / a.u.
temperature (°C)
0 400 800 1200
S
org
+SiO
2
S
org
Na-alunite+SiO
2
Na-alunite
basaluminite+SiO
2
basaluminite
Ion current (m/z=48, SO) / a.u.
Figure 3.4 SO and H
2
O mass spectrometry signals during STA-MS analysis of the reference
compounds basaluminite, N-alunite, and Na-dodecylsulfate (S
org
) in mixture with SiO
2
.
3.4.2.2.3 Mixtures with Al
2
O
3
: Water was released from Al
2
O
3
in the temperature
range between 80 °C and 400 °C. Additionally, a sma ll H
2
O release around 540 °C
could be observed. Water release from basaluminite and Na-alunite was not af-
fected substantially while the shape and position of the main H
2
O release signal
from Na-dodecylsulfate changed due to the addition of Al
2
O
3
(Figure 3.5).
The addition of Al
2
O
3
had a substantial influence on the SO release pattern from all
S containing reference compounds (Figure 3.5). In the case of basaluminite the
main peak broadened substantially. For Na-alunite the first peak of SO release
shifted from approximately 780 °C to 880 °C – 980 ° C with a substantial peak
broadening and the SO release between 1300 °C and 1 400 °C which was attributed
to the decomposition of Na
2
SO
4
shifted towards lower temperatures. The latter could
also be observed for Na-dodecylsulfate for which an additional SO peak appeared
around 900 °C and the shape of the peak between 200 °C and 400 °C also
changed.
STA-MS
47
temperature (°C)
0 400 800 1200
S
org
+Al
2
O
3
S
org
Na-alunite+Al
2
O
3
Na-alunite
basaluminite+Al
2
O
3
basaluminite
Ion current (m/z=18, H2O) / a.u.
temperature (°C)
0 400 800 1200
S
org
+Al
2
O
3
S
org
Na-alunite+Al
2
O
3
Na-alunite
basaluminite+Al
2
O
3
basaluminite
Ion current (m/z=48, SO) / a.u.
Figure 3.5 SO and H
2
O mass spectrometry signals during STA-MS analysis of the reference
compounds basaluminite, N-alunite, and Na-dodecylsulfate (S
org
) in mixture with Al
2
O
3
.
3.4.2.3 Mass spectrometry signals of sample Ref
ads
:
The sample Ref
ads
showed a water release which is typical for dehydration and
dehydroxylation processes of allophane (Figure 3.6). Sulfur monoxide release from
Ref
ads
occurred in two steps with a main peak around 880 °C and a peak shoulder
around 950 °C. The position of the latter coincides with the position of the main SO
release from amorphous basaluminite. With the aim to estimate the contributions of
the main peak and the peak shoulder to the overall SO signal I fitted two Gaussian
peaks to the temperature region 700 – 1000 °C of the SO signals with the software
package PeakFit
TM
Version 4 (Systat Software Inc., San Jose, USA). Peak centers
and peak widths were allowed to float without restrictions as the exact peak position
and peak width in STA-MS analysis depend slightly on the contents of the analyzed
phases (Smykatz-Kloss, 1974). By this procedure, for the main peak and the peak
shoulder contributions of 98% and 2% to the overall SO signal were obtained, re-
spectively.
Chapter 3
48
temperature (°C)
0 400 800 1200
ion current / a.u.
SO
H2O
Figure 3.6 SO and H
2
O mass spectrometry signals during STA-MS analysis of sample
Ref
ads
.
3.5 DISCUSSION
3.5.1 Is a differentiation between estersulfate phases, alunite and basaluminite
possible by STA-MS analysis?
While amorphous basaluminite, Na-alunite, K-alunite and Na-dodecylsulfate in pure
form can be distinguished unambiguously from each other based on their character-
istic patterns of H
2
O and SO release, the differentiation between these S containing
phases is complicated when matrix components are admixed. Matrix components
influence SO and H
2
O release patterns. This influence is more pronounced for SO
than for H
2
O and may be explained by a number of different reactions. Transforma-
tion reactions which do not occur when the pure S compounds are heated may be
enabled due to surface catalysis with the matrix components acting as catalyzing
agents. Such catalysis reactions have been postulated also by Emmerich and
Smykatz-Kloss (2002) for the thermal analysis of organic molecules in the presence
of talc. Additionally, gases which evolve during any of the occurring reactions may
be captured by the solid phases due to adsorption processes or under the formation
of new solid phases. Such phase transformations during thermal analysis have been
reported for alunites in the presence of excess amounts of potassium and calcium
salts. The presence of these salts leads to a suppression in the S release around
800 °C from alunites due to the formation of K
2
SO
4
and CaSO
4
which have a higher
thermal stability compared with alunites (Cipriani et al., 1997).
STA-MS
49
Overall, my results imply that a differentiation between alunitic phases, basaluminite
and estersulfate compounds with STA-MS is very limited. While these phases may
be distinguished from each other in pure form, their identification in complex sam-
ples as soils, which present mixtures of a number of different matrix components, is
most likely impossible. As an intermediate, in very simple systems of known matrix
composition a differentiation between the before mentioned S-pools may be possi-
ble. Such simple systems are often applied as model systems for directed investiga-
tions of specific scientific questions. Thus, in such cases STA-MS may be helpful for
the differentiation of S pools.
For the analyzed simple systems I conclude that in the presence of allophane ba-
saluminite and alunite can be identified unambiguously by the combined interpreta-
tion of SO and H
2
O MS signals while the MS signals of alunite may mask the signals
of S
org
. If only SiO
2
is present as admixture, all phases can be distinguished from
each other, while with Al
2
O
3
as singular admixture SO signals of all reference com-
pounds overlap making a differentiation in this case impossible. In the presence of
Al
2
O
3
only alunitic compounds can be identified unambiguously by their characteris-
tic dehydroxylation around 500 °C in combination with SO release at higher tem-
peratures. Thus, especially sulfate adsorbing phases seem to influence the sulfur
release signals.
3.5.2 Can “sulfate adsorbed to allophane” be distinguished from BAS and
estersulfate phases by STA-MS?
The comparison of the SO release pattern of Ref
ads
with SO signals of the S contain-
ing reference compounds in mixture with allophane shows that SO release from
adsorbed sulfate can be masked in the case that organic estersulfate compounds or
alunitic phases are present. In case that the presence of organic or alunitic phases
can be ruled out based on the absence of CO
2
release and the absence of H
2
O
release around 500 °C, adsorbed sulfate may be distinguished from basaluminite-
like phases by differences in the temperature dependence of the main SO release
(880 °C vs. 950 °C).
Nevertheless, it has to be emphasized that characteristic “fingerprints” during ther-
mal analysis cannot be readily interpreted in terms of reaction mechanisms
(Langier-Kuzniarowa, 2002). This can be exemplarily seen for the SO signal of
sample Ref
ads
which shows a peak shoulder besides the main peak and the ques-
tion on the source of this two-step SO release arises. On one side the peak shoulder
Chapter 3
50
may be explained by the presence of a small portion of a basaluminite-like phase in
Ref
ads
accounting with 2% to the total SO-release. As described in section 3.3.4 the
formation of such a phase during the synthesis of Ref
ads
cannot be ruled out. Never-
theless, the biphasic SO release from Ref
ads
could hypothetically also be explained if
only adsorbed sulfate would be present and if Ref
ads
would decompose into two
different solid phases during thermal analysis with each solid phase retaining part of
the sulfur. Subsequently, sulfur may be released in different heating stages accord-
ing to the thermal stabilities of the phases which form during STA-MS analysis.
Additionally, the biphasic SO release may also be explained by the presence of
different adsorption complexes, i.e. inner sphere and outer sphere complexes al-
though this is unlikely as water is removed in an early stage of STA-MS analysis
resulting most probably in the transformation of outer-sphere into inner-sphere
complexes. Such a transformation has been observed for sulfate adsorbed on goe-
thite (Wijnja and Schulthess, 2000) as a consequence of goethite air drying.
By applying high-temperature XRD I attempted to elucidate the processes which are
responsible for the biphasic SO release from Ref
ads
. Nevertheless, this attempt
failed as no crystalline S-containing phases could be detected. Thus, both the pres-
ence of traces of a basaluminite-like phase in Ref
ads
and/or the transformation of
“sulfate adsorbed to allophane” into two S-containing solid phases of different ther-
mal stability during STA-MS analysis may be responsible for the biphasic SO re-
lease.
3.6. CONCLUSIONS
The sulfate containing phases estersulfate, basaluminite and alunite can be distin-
guished from each other by STA-MS when they are analyzed in pure form. Never-
theless, the admixture of inorganic components as allophane, SiO
2
or Al
2
O
3
affect
water and SO release patterns substantially, thus limiting the applicability of STA-
MS for the differentiation of S-pools to very simple systems of known composition.
My results imply that the identification of S-pools in complex samples as soils by
STA-MS is most probably impossible. These results additionally suggest that the
suitability of thermal methods for the identification of soil carbon pools of different
thermal stabilities might be re-evaluated by the systematic analysis of model com-
pound mixtures.
51
4 IN SITU PRECIPITATION OF BASIC ALUMINUM SULFATE PHASES AT THE
ALLOPHANE-WATER INTERFACE
4.1 SUMMARY
The precipitation of basic aluminum sulfate phases (BAS) contributes to the sulfur
retention in Andosols close to SO
2
-emitting volcanoes. To test whether allophane-
water interfaces are preferential sites for in situ BAS formation I exposed synthetic
allophane and glass for 18 months in vitric Andosols close to the active volcanoes
Masaya and Poás. These sites differ in their climatic conditions (wet-dry cycles at
Masaya, perhumid at Poás) and their soil organic matter contents which are larger
at Poás. Samples were analyzed for the presence of BAS phases by a combination
of S K-edge X-ray absorption near edge spectroscopy, simultaneous thermal analy-
sis coupled with mass spectrometry, scanning electron microscopy coupled with
energy dispersive X-ray analysis and sulfate extraction. While no BAS phases could
be detected on glass surfaces, sulfate containing precipitates contributed with 11%
and 19% to the total sulfur content of 6.0 g kg
-1
and 11.7 g kg
-1
of allophane ex-
posed at Masaya and Poás, respectively. Continued leaching under the perhumid
climatic conditions and the competition between solid or dissolved organic carbon
and sulfate for dissolved aluminum thus do not impede BAS formation at Poás. The
larger BAS contribution to the total sulfur concentration of allophane exposed at
Poás compared with Masaya may be attributed to (i) deposition of alunitic aerosol
particles originating from the hydrothermal crater lake, (ii) larger sulfate fluxes in soil
solution and/or (iii) a smaller sulfate adsorption capacity due to the adsorption of
dissolved organic carbon by allophane. Organic carbon coatings on allophane sur-
faces stabilized allophane against oxalate dissolution indicating that the quantifica-
tion of allophane based on oxalate extraction data can lead to largely erroneous
results.
Chapter 4
52
4.2 INTRODUCTION
Soils in the vicinity of active, SO
2
-emitting volcanoes store large amounts of sulfate.
The processes by which sulfate is retained in these soils are a matter of ongoing
scientific debate. In recent studies, the question whether sulfate is mainly adsorbed
to variable charge minerals like allophane and iron (hydr)oxides or if it is additionally
incorporated into precipitates of the group of basic aluminum sulfate phases (BAS)
like e.g. basaluminite, alunite or aluminite has been discussed (Delfosse et al,
2005a; Prietzel et al., 2008). The knowledge on the relevance of both processes is
important for the prediction of sulfate dynamics as (i) the products are expected to
differ in their formation and dissolution kinetics and (ii) adsorption is limited by the
number of available adsorption sites while precipitation can be an infinite sulfate
sink.
First hints on the presence of BAS phases in SO
2
impacted volcanic soils were
given by geochemical equilibrium calculations and selective dissolution experiments
(Wolt et al., 1992; Delfosse et al., 2005b). Transmission electron microscopy cou-
pled with energy dispersive X-ray fluorescence (Delfosse et al., 2005c) and S K-
edge XANES analysis (Prietzel et al., 2008) of soil samples from Masaya Volcano,
one of the largest natural SO
2
sources of the world, affirmed the presence of BAS
precipitates in these soils. Based on the observation that BAS-like precipitates are
located in close vicinity to allophane particles in transmission electron micrographs,
Delfosse et al. (2005c) suggested that BAS phases may preferentially form at the
allophane water interface. The authors hypothesized that surface precipitation may
occur, i.e. that two-dimensional sulfate adsorption complexes may develop into
three-dimensional BAS precipitates on the surface of allophane-rich aggregates.
This process has been described in detail by Ford et al. (2001) and it may be one
potential mechanism to explain a preferential BAS formation at the allophane-water
interface. Additionally, surfaces may act as physical nuclei for BAS precipitation. In
this case the efficacy of surfaces for preferential BAS formation depends only on the
size of the available surface area irrespective of the presence or absence of sites for
sulfate adsorption. Allophane aggregates have specifically large surface areas
(Bartoli et al., 2007) and may thus have a high potential to act as nucleation sites.
This may be enhanced by the fact that a large percentage of the allophane surface
area is due to internal surfaces in micro- and mesopores of allophane aggregates
(Paterson, 1977; Bartoli et al., 2007). Such void spaces may present preferential
In situ BAS formation
53
sites for BAS precipitation as ion exchange between the void spaces and the freely
draining soil solution may be diffusion limited allowing for BAS saturation in void
spaces while the freely draining solution may be undersaturated.
The mentioned processes make it plausible that BAS precipitates are preferentially
formed at the allophane-water interface. Nevertheless, volcanic glass and mineral
phenocrysts as e.g. plagioclase microcrystallites are the dominant aluminum
sources for BAS formation in volcanic soils (Delfosse et al., 2005a; Herre et al.,
2007). Thus, BAS precipitates may also form on glass or plagioclase surfaces de-
pending on the composition of the surrounding microenvironment. Gislason and
Eugster (1987) showed that alteration minerals did form on glass particles, espe-
cially in voids which developed during glass dissolution.
Identification of the small and mostly amorphous BAS phases in soils is rather com-
plex and a number of different methods has been applied. Delfosse et al. (2005c)
used transmission electron microscopy coupled with energy dispersive X-ray analy-
sis (TEM-EDX). Nevertheless, by own experience identifying BAS precipitates by
electron microscopy in complex samples as soils is much like searching a needle in
a haystack. Delfosse et al. (2005c) tried to overcome this problem by analyzing not
samples as a whole but selecting the S-rich clay fraction. However, sample pre-
treatment risks dissolution of BAS precipitates as discussed by Delfosse et al.
(2005c).
Limitations for BAS identification by TEM-EDX may be overcome in two ways: (i) the
installation of pure mineral phases in the field with their posterior analysis and/or
(ii) the use of other methods for BAS identification for which no sample pretreat-
ments are necessary. Prietzel et al. (2008) could show recently that S K-edge X-ray
absorption near edge fine structure spectroscopy (S K-edge XANES) is a powerful
tool to identify BAS phases. Nevertheless, by this method BAS phases may be
confounded with estersulfate phases. Besides, similarly as with electron microscopy,
S K-edge XANES probes rather small sample aliquots. Thus the question arises
whether XANES results are representative for the whole sample. This limitation is
overcome by the method of simultaneous thermal analysis coupled with mass spec-
trometry (STA-MS) in which a larger sample aliquot is analyzed without sample
pretreatment (chapter 3). Additionally, STA-MS has the potential to distinguish
between different BAS phases even when they are present in rather small amounts.
Nevertheless, the sample matrix composition can distort STA-MS signals substan-
tially limiting this method to very simple systems of known composition. As a third
Chapter 4
54
and more indirect method, sulfate extractability with the application of different ex-
tractants has been used to quantify the contribution of different S-containing phases
(as e.g. BAS, adsorbed sulfate, C-bound sulfate, estersulfate compounds and oth-
ers) to S retention in soils (e.g. Erkenberg et al., 1996; Prietzel et al., 2001; Delfosse
et al, 2005a). This method probes the bulk sample. Nevertheless, selective extrac-
tion methods are not selective sensus strictus (Prietzel et al., 2003) and may at best
give an estimate of different S-containing phases when applied for S speciation in
soils. Thus, all mentioned methods for BAS identification have both potentials and
limitations and a combination of two or more methods may be necessary for an
unambiguous BAS identification in soils.
The primary goal was to investigate whether BAS precipitates preferentially form in
situ at the allophane-water interface compared to glass surfaces. I did this by field
exposure of synthetic allophane and glass at two SO
2
-impacted volcanic sites. The
sites were chosen such that they differ in climatic conditions and the soil organic
carbon content as these characteristics may influence BAS precipitation: climatic
conditions influence leaching conditions in soils and thus the saturation state of the
soil solution while organic carbon compounds compete with sulfate for adsorption
sites and for dissolved aluminum. As sulfur retention and BAS formation can addi-
tionally be influenced by changes in the allophane and glass characteristics, I inves-
tigated changes in the characteristics of these phases caused by field exposure.
4.3 MATERIALS AND METHODS
4.3.1 Experimental approach
I conducted an 18 months field experiment during which I exposed synthetic glass
and allophane in separate devices at two acid impacted volcanic sites. Study sites
were close to Masaya Volcano in Nicaragua and Poás Volcano in Costa Rica. Soils
of these sites have been described previously (Herre et al., 2007); in the present
study additionally sulfate and aluminum fluxes in soils were quantified by installation
of ion exchange resins. Both glass and allophane were analyzed before and after
exposure for the presence of sulfur by total elemental analysis and by scanning
electron microscopy coupled with energy dispersive X-ray analysis (SEM-EDX).
Sulfur-containing samples were additionally analyzed for the presence of BAS
phases by simultaneous thermal analysis coupled with mass spectrometry (STA-
MS), S K-edge X-ray absorption near edge structure spectroscopy (XANES), and for
In situ BAS formation
55
their sulfate extractability. Sample signals were compared with those of synthetic S-
containing reference compounds. Additionally, allophane samples were analyzed for
potential mineralogical changes. Therefore, samples were characterized by simulta-
neous thermal analysis (differential scanning calorimetry / thermogravimetry,
DSC/TG), attenuated total reflectance Fourier-transformed infrared spectroscopy
(ATR-FTIR) and oxalate extraction. Potential glass transformations were analyzed
by SEM-EDX.
4.3.2 Experimental sites
4.3.2.1 Masaya Volcano
Masaya Volcano is a shield volcano with tephra of basaltic composition. It is located
in Nicaragua elevating to 560 m above sea level. The mean annual temperature and
mean annual precipitation are 26.3 °C and 1379 mm, respectively with 90% of the
precipitation falling during the rainy season from May to October (UNA & ISRIC,
1994). Masaya Volcano is one of the largest natural SO
2
sources of the world with
SO
2
fluxes of the actual degassing period ranging from 390 to 1850 t SO
2
day
-1
(Delmelle et al., 2002). Besides, substantial amounts of halogens (HCl and HF) are
emitted (Delmelle et al., 2001).
Soils in the close vicinity of Masaya Volcano classify as Vitric Andosols. The pres-
ence of allophane reflects the development towards the sil-andic subtype. A more
detailed description of the soils can be found in Delmelle et al. (2003) and Herre et
al. (2007). For the field experiment of the present study the site named “M1” in Herre
et al. (2007) has been selected. This site is located approximately 2 – 3 km down-
wind of the active crater and is affected by large acid inputs. For soil samples taken
at this site, transmission electron microscopy coupled with energy dispersive X-ray
analysis (Delfosse et al., 2005b) and S K-edge XANES analysis (Prietzel et al.,
2008) revealed the presence of BAS precipitates.
4.3.2.2 Poás Volcano
Poás Volcano is a composite stratovolcano in central Costa Rica with an elevation
of 2708 m above sea level. Mean annual temperatures vary from 8 to 12 °C and
mean annual rainfall from 3800 to 4600 mm, with monthly averages ranging from
120 mm (December to April) to 420 mm (May to November) (Rowe et al., 1992;
Martinez et al., 2000). Poás Volcano has shown different types of volcanic activity
Chapter 4
56
ranging from phreatic to phreato-magmatic eruptions and fumarolic activity between
the eruptive cycles (Alvarado Induni, 2005). Tephra are of basaltic to andesitic
composition and contain variable amounts of S-rich sediments from the extremely
acid crater lake (Prosser & Carr, 1987). Since 1955 Poás Volcano has shown fu-
marolic activity of varying magnitude, emitting SO
2
, HCl and HF with SO
2
being the
dominant acid-generating gas of the plume (Rowe et al., 1992). Measured acid
emission rates range from 8 to 700 t SO
2
day
-1
(Casadevall et al., 1984; Andres et
al., 1992; Zimmer et al., 2004).
Soils in the vicinity of Poás Volcano are classified as Vitric Andosols of the alu-andic
subtype. For the field experiment of the present study the site named “P2” in Herre
et al. (2007) has been selected. This site is situated approximately 2 km downwind
of the active crater and is affected by the acid plume.
4.3.3 Synthesis of glass, allophane and S-containing reference compounds
4.3.3.1 Synthesis of glass
With the aim to obtain pure glass of basaltic-andesitic composition, I synthesized
glass by melting and rapid cooling of volcanic ash from the 1953-1955 eruption of
Poás Volcano. The ash sample had been taken from the middle-bottom of a thick
ash layer which looked macroscopically unaltered. The volcanic ash sample was
oven-heated in a Pt10/Rh crucible to 1600 °C, maintained for 2 hours at this tem-
perature and rapidly cooled by large-area contact with a highly heat-conductive
metal plate. The obtained glass was crushed in a jaw crusher (Fritsch GmbH, Ger-
many) and broken up in a stainless steel disk mill (Fritsch GmbH, Germany) until
glass fragments passed a sieve with mesh size of 350 µm. Subsequently, adherent
fine particles were removed by washing the glass repeatedly with doubly deionized
H
2
O and ultrasonic treatment until the supernatant was clear. By the whole proce-
dure a glass containing 54% SiO
2
, 18% Al
2
O
3
, 9.5% Fe
2
O
3
, 8.4% CaO, 4.7% MgO,
3% Na
2
O, 1% K
2
O and 0.8% TiO
2
was obtained as determined by X-ray fluores-
cence analysis. X-ray diffraction, petrographic microscopy and microprobe analysis
for the elements Si, Al, Ca, Mg, Na and Fe confirmed the amorphous nature and
homogeneous elemental distribution of the obtained glass.
In situ BAS formation
57
4.3.3.2 Synthesis of allophane
Allophane was synthesized according to Ohashi et al. (2002) with minor modifica-
tions. Briefly, 225 g of 100 mM Na
4
SiO
4
and 300 g of 100 mM AlCl
3
* 6H
2
O were
mixed rapidly, stirred for one hour and centrifuged for two hours at 7600 g. After
decanting the clear supernatant, 200 mL of doubly deionized H
2
O were added, the
allophane precursors were resuspended and the suspension was boiled for 48
hours under a reflux condenser. Remaining salts were eliminated by dialysis (Spec-
tra/Por
Dialysis membrane, MWCO: 12-14,000, Spectrum Laboratories, Inc., Ran-
cho Dominguez, USA) until the electrical conductivity was lower than 10 µS cm
1
.
The suspension was centrifuged for four hours at 7600 g and the clear supernatant
was decanted. By the described procedure an allophane gel is obtained which was
re-suspended in H
2
O and oven-dried at 40 °C. During oven-drying allophane forms
rather large aggregates which were softly broken up in an agate mortar until passing
a sieve with 200 µm mesh aperture.
The allophane was characterized by X-ray diffraction (Philips PW 3710), differential
scanning calorimetry - thermogravimetry coupled with mass spectrometry (Netzsch
409 PC – Pfeiffer Thermostar MS), mid and far FT infrared spectroscopy (Thermo
Nicolet Nexus) and oxalate extraction according to Schwertmann (1964). By the
described procedure allophane with the formula 1.25 SiO
2
* Al
2
O
3
* 3.2 H
2
O and a
Al:Si ratio of 1.6 was obtained. All other mentioned analyses gave results character-
istic for this type of allophane.
4.3.3.3 Synthesis of S-containing reference compounds
The BAS phases basaluminite, Na-alunite and K-alunite were synthesized according
to Prietzel and Hirsch (1998) with minor modifications. The exact procedure has
been described in detail chapter 3. Therein, also the synthesis of a reference com-
pound for “sulfate adsorbed to allophane” is described. Sodium-dodecylsulfate
(C
12
H
25
NaO
4
S for biochemistry and tenside testing) as reference compound for
organically-bound estersulfate compounds was purchased from Merck (Darmstadt,
Germany).
Chapter 4
58
4.3.4 Field exposure of glass, allophane and ion exchange resins
Sulfate and aluminum fluxes in soils of the study sites were determined by installa-
tion of ion exchange resins in “self-integrating accumulators (SIAs)” as patented and
described by Bischoff (2007). With this method the materials are exposed in verti-
cally installed plastic tubes in the soils. The installation method as detailed below
guarantees (i) that the structure of the overlying soil is not disturbed and (ii) that
solutes pass the installed materials irrespective of soil solution transport by matrix or
preferential flow. Additionally, with the resins a per-area elemental flux is measured
directly without the need of water budget modeling. At each site eight replicate tubes
were installed. For allophane and glass five replicates each were installed at each
site using a modified tube technique as detailed below.
A Nylon grid with a mesh size of 150 µm was attached at the bottom of each tube
(diameter 10 cm). All materials were sandwiched between a 3 cm quartz sand layer
at the bottom and a 1 cm quartz sand layer at the top of the tubes. Allophane and
glass were installed in tubes of 4.5 cm height. For each allophane tube, 8 g of the
mineral were mixed with 72 g of quartz sand. In order to be able to separate allo-
phane and quartz sand by sieving at the end of the experiment, the quartz sand had
been sieved previously to contain only particles larger than 250 µm. Glass grains
were installed in pure form and were separated from the underlying and overlying
quartz layers by introduction of a 10 cm diameter Nylon grid disk with a mesh size of
20 µm to guarantee the complete separation of glass and quartz at the end of the
experiment. For each resin tube, 400 g of a strongly acidic/basic mixed bed ion
exchange resin were mixed with 900 g quartz sand and the mixture, also sand-
wiched between the 3 cm and 1 cm quartz sand layers, was installed in boxes of
15 cm height. All quartz sand used had been washed previously for 24 hours in
0.01 M HCl with subsequent removal of the acid by repeated washings with deion-
ized water. Before installation, the sand has been dried at 65 °C.
At Masaya tubes were installed in the soil below the Ah horizon (upper limit of tubes
at 12 cm depths) and at Poás below the Ofh horizon (upper limit of tubes at 9 cm
depths). For tube installation in the soils a trench was dug at each site. Small tun-
nels were excavated perpendicular to the trench wall and tubes were installed verti-
cally in these tunnels assuring good contact between the tube material and the
overlying soil by pushing the tubes carefully against the overlying soil. Remaining
holes at the bottom and side of the tubes were refilled by soil material. Finally, the
In situ BAS formation
59
trench was refilled by soil material. Allophane, glass and resins remained for 18
months in the soils (June 2005 – December 2006). After this period the tubes were
taken out and the upper quartz sand layer was removed from each tube. The allo-
phane/quartz and glass layers were sampled as a whole while the resin/quartz
layers of each tube were sampled in 4 sublayers, named further sublayer 1 to 4 from
top to bottom of the tubes. All 4 sublayers were weighed and analyzed separately in
the laboratory. This procedure is applied to check whether sulfate was quantitatively
retained (indicated by the absence of sulfate in the resin sublayers in the middle of
the tube) or if it did break through (indicated by the presence of sulfate in all 4 resin
sublayers).
4.3.5 Resin analysis
For sulfate and aluminum analysis 5 g of resin/quartz mixtures were weighed in PE
bottles and 50 mL of 2 M HNO
3
were added. Samples were shaken for one hour
and the supernatant was decanted through a Nylon grid of 20 µm mesh size.
Resin/quartz mixtures which have not been exposed in the field were analyzed as
blanks. Sulfate and aluminum concentrations were quantified by inductively coupled
plasma optical emission spectroscopy (Vista Pro, Varian, Australia).
4.3.6 Total carbon and sulfur analysis
Total contents of carbon and sulfur were determined for allophane and glass sam-
ples by a CNS analyzer (vario EL III, Elementar, Hanau, Germany) equipped with a
thermal conductivity detector for carbon and a UV detector for sulfur quantification.
Depending on the expected C and S concentrations 1 – 80 mg of the samples were
weighed in tin capsules and approximately double the amount of WoO
3
was added
to prevent the formation of non-volatile S-phases.
4.3.7 Scanning electron microscopy coupled with energy dispersive X-ray
analysis (SEM-EDX)
Samples were mounted on a double adhesive carbon tape, coated with a thin film of
carbon or gold and analyzed by SEM-EDX (Hitachi S-4000, Maidenhead, UK).
Analyses were done applying accelerating voltages between 5 and 20 kV. For part
of the allophane samples elemental mapping (S, Al, Si and Fe) was done by SEM-
Chapter 4
60
EDX (Hitachi S-2700, Maidenhead, UK) to check whether sulfur is distributed homo-
geneously in allophane samples or localized in S hot-spots.
4.3.8 Simultaneous thermal analysis coupled with mass spectrometry (STA-
MS)
According to the method described in chapter 3 all samples were weighed in Pt/Rh
crucibles, tapped by a Pt/Rh lid with a small orifice for gas release and analyzed
with a STA 449C Jupiter (Netzsch, Selb, Germany) equipped with a thermogravim-
etry / differential scanning calorimetry (TG/DSD) sample holder. Released gases
were detected by a coupled quadrupole mass spectrometer (QMS 403C Aëolos,
Netzsch, Selb, Germany) which was connected by a heated capillary to the STA
449C to prevent condensation during gas transport to the MS. Signals of masses
18, 48, 64, 80 and 96 were recorded corresponding to H
2
O, SO, SO
2
, S
2
O/SO
3
and
S
2
O
2
, respectively. After a 10 minutes isothermal segment at 35 °C, samples were
heated under an atmosphere of air / N
2
(gas fluxes 80 / 20 mL min
-1
, respectively)
up to 1400 °C with a heating rate of 10 °C min
-1
. Measurement signals were ana-
lyzed with the software package Proteus Version 4.8.1 (Netzsch, Selb, Germany)
and mass spectrometry signals were smoothed by the Golay Savitzky algorithm.
4.3.9 Sulfur K-edge X-ray absorption near edge spectroscopy (S K-edge
XANES)
By sulfur K-edge XANES spectroscopy I determined the percentage of adsorbed
and precipitated sulfate as described by Prietzel et al. (2008). Sulfur K-edge XANES
spectra were acquired with the scanning transmission X-ray microscope at the ID-21
beamline at the European Synchrotron Radiation Facility (ESRF, Grenoble, France)
described by Barrett et al. (2000). The method and the instrumental setup have
been described in detail in Prietzel et al. (2007). The monochromaticity of the beam
and the energy scan were ensured by a fixed exit double crystal Si <111> mono-
chromator located upstream the microscope, which offers an energy resolution of
0.5 eV necessary to resolve the XANES structures. The X-ray transmission and
fluorescence signals were recorded simultaneously with a Si photodiode mounted
downstream the sample and an energy-dispersive silicon drift diode (XFLASH 2001,
Röntec, Berlin, Germany), respectively.
In situ BAS formation
61
I obtained XANES spectra by irradiating samples that had been sandwiched be-
tween two Ultralene® foils (thickness 4 µm, Spex-Certiprep Comp., Metuchen, USA)
with X-rays in the energy range between 2450 and 2530 eV with stepwise energy
increments of 0.2 eV and a dwell time of 1 s. Because for some compounds it was
not possible to acquire a transmission signal with the described sample mounting
technique, the fluorescence spectra of all samples were evaluated in my study.
Comparisons between the transmission and fluorescence spectra of selected sam-
ples and reference compounds assured that self-absorption did not constitute a
problem during XANES measurements. For each spectrum, the signals of 10 scans
were compiled. The energy calibration was done with pure CaSO
4
(white line:
2482.5 eV). I analyzed all reference compounds and samples by recording XANES
spectra in a non-focused mode (with the zone plate removed), the size of the beam
being determined by a 200 µm pinhole. In this mode, the integrating signal of an
area of 0.13 mm² of the sample was analyzed. Additionally, one allophane sample
which had been exposed in soils at the Masaya site was analyzed by spatially re-
solved S K-edge XANES. The sample mounting technique and the measurement
details for spatially resolved XANES are described in Thieme et al. (2006) and
Prietzel et al. (2003), respectively.
Spectra were analyzed using the software package WinXAS 3.1 (Ressler, 1998). All
spectra were baseline-corrected and normalized to the edge jump. The contributions
of adsorbed and precipitated sulfate to the total sulfate content were estimated
applying linear combination fitting (LCF). As predictor components the spectra of
adsorbed sulfate, amorphous basaluminite, K-alunite or Na-alunite were taken. As
predictor component for organically bound sulfate the spectrum of Na-
Dodecylsulfate was used. Additionally, for all spectra the full width at half maximum
(FWHM) of the white line at 2482.5 eV was determined.
4.3.10 Attenuated total reflectance Fourier-transformed infrared spectroscopy
(ATR-FTIR)
With the aim to detect potential allophane transformations during its exposure at
Masaya and Poás Volcanoes FTIR spectra were recorded by a Bruker IFS66/S
(Bruker Optics, Ettlingen, Germany) equipped with a “Golden Gate” attenuated total
reflectance micro-unit (Specac, London, UK) with a single reflection diamond crystal.
Spectra were recorded from 4000 – 400 cm
-1
with a resolution of 4 cm
-1
and 64
scans per sample. Therefore, samples were deposited on the diamond crystal and a
Chapter 4
62
defined pressure was applied on the samples by tightening the sample screw with a
constant torque of 90 cNm with a dynamometric key. All spectra were background
and baseline corrected. As background signal the IR spectrum of the empty ATR-
diamond in air was used. ATR-FTIR spectra of allophane differ from IR spectra
obtained by the more commonly applied KBr technique and this may be attributed
most probably to artifacts caused by the KBr technique (chapter 6). Thus, for a
comparison of ATR-FTIR signals of structural vibrations of my samples with litera-
ture data my data should be shifted by 30-40 cm
-1
to larger wavenumbers.
4.3.11 Wet chemical extraction methods
4.3.11.1 Sulfate extractability
For samples, which contained sulfur after field exposure, I determined the fraction of
sulfate exchangeable with phosphate after 1 hour, 8, 24, 48 and 168 hours of ex-
traction. Therefore, sample aliquots corresponding to 40 µg S were weighed into 25
mL PE bottles and 5 mL of 0.016 M KH
2
PO
4
were added. Suspensions were shaken
for 1 hour, 8, 24, 48 or 168 hours on a rotary shaker, filtered through 0.2 µm mem-
brane filters (Minisart RC 15, Sartorius, Göttingen, Germany) and the filtrates were
analyzed for sulfate by capillary electrophoresis (HP 3D CE, Agilent) using a buffer
solution contained in the Agilent anion analysis kit (product number: 5063-6511).
4.3.11.2 Oxalate extraction
To detect potential allophane modifications or changes in the allophane extractabil-
ity, oxalate extractable Al, Si and Fe were determined by shaking sample aliquots
for 2, 4 and 168 hours with ammonium oxalate / oxalic acid solution at pH 3 accord-
ing to Schwertmann (1964) and a solid:solution ratio of 1:1000. After 2, 4 or 168
hours, samples were filtrated and the concentrations of Al, Si and Fe were deter-
mined by flame atomic absorption spectroscopy (Perkin Elmer 1100 B). For quality
control, in each analysis batch aliquots of synthetic allophane were used as in-
house reference material. In case that solid residuals could be observed on the filter,
these residuals were analyzed by ATR-FTIR with measurement parameters as
described in section 4.3.10.
In situ BAS formation
63
4.4. RESULTS
4.4.1 Sulfate and aluminum fluxes in soils
Depth profiling of sulfate and aluminum fluxes in resin boxes (Figure 4.1) showed
that for both study sites no sulfate and only little aluminum could be detected in
sublayer 3. The absence of sulfate in sublayer 3 indicates that percolating sulfate
did not break through the resin boxes but was retained quantitatively in sublayers 1
and 2. The presence of aluminum in sublayer 3 may indicate aluminum break-
through but aluminum release from field-weathered impurities in the sand-resin-
mixture is more probably the cause for the observed as sublayers 2 and 3 of both
study sites release very similar aluminum amounts despite differences in the Al
contents of sublayers 1. The presence of sulfate and aluminum in sublayer 4 can be
explained by (i) the upward movement of soil solution due to capillary rise or (ii)
diffusion of ions into the tubes with the resin representing an active ion sink. While
for the Masaya site upward water movement is probable, for the Poás site upward
water movement is expected to be very low due to the perhumid climatic conditions.
This makes it likely that the presence of sulfate and aluminum in sublayer 4 is at
least partially caused by ion diffusion into the tubes. Consequently, for the calcula-
tion of ion fluxes sublayer 4 has not been taken into account. In the case of sulfate,
ion contents of sublayers 1 and 2 have been summed up while aluminum fluxes
were determined based solely on the contents of sublayers 1.
By this procedure sulfate fluxes of 64 ± 15 (6 – 142) kg S ha
-1
and 515 ± 147 (130 –
1 392) kg S ha
-1
and aluminum fluxes of 44 ± 9 (9 – 94) kg Al ha
-1
and 26 ± 3 (13 –
37) kg Al ha
-1
are obtained for the experimental period of 18 months for the Masaya
and Poás site, respectively (mean values ± standard errors with minima-maxima in
parenthesis). Fluxes may be overestimated as ions most probably also diffuse from
the top end of the tubes into sublayer 1. Nevertheless, as diffusion effects decrease
with increasing ion load on resins, this effect is expected to be lower for sublayer 1
compared to sublayer 4 when summed up for the whole experimental period. Thus,
subtracting the ion fluxes measured for sublayers 4 from the above reported sulfate
and aluminum fluxes results in an estimate of the minimal total ion fluxes in percolat-
ing water of 48 kg S ha
-1
and 432 kg S ha
-1
and 40 kg Al ha
-1
and 14 kg Al ha
-1
over
the experimental period for the Masaya and Poás site, respectively. The sulfate
fluxes are markedly lower than aboveground S deposition rates which have been
measured with sulfation plates in February-March 1999 and December 2002 at the
Chapter 4
64
Masaya and Poás site, respectively (Delmelle et al., 2001; Herre et al., 2007). Dif-
ferences may be caused by temporal variations of volcanic S emissions, by S reten-
tion and release processes in the soil horizon overlying the resin tubes or by meth-
odological differences between resin tubes and sulfation plates.
Masaya
S (kg ha
-1
)
0 150 300 450 600
1
2
3
4
Poás
S (kg ha
-1
)
0 150 300 450 600
1
2
3
4
Al (kg ha
-1
)
0 10 20 30 40 50 60
Al (kg ha
-1
)
0 10 20 30 40 50 60
1
2
3
4
1
2
3
4
Figure 4.1 Sulfur and aluminum fluxes (in kg ha
-1
) in layers 1-4 of resin boxes from Masaya
(left column) and Poás (right column) sites over the whole experimental period of 18 months
4.4.2 Total carbon and sulfur contents of glass and allophane
Due to exposure in the field, glass samples of the Poás site showed a slight in-
crease in C concentrations while C concentrations of glass samples from Masaya
were below the detection limit (Table 4.1). Carbon contents on allophane increased
for both sites with the increase being larger for Poás samples. While sulfur concen-
trations of glass samples did not increase significantly due to field exposure, mean
sulfur concentrations of allophane samples were 0.6 and 1.17% after sample expo-
sure at Masaya and Poás, respectively.
In situ BAS formation
65
Table 4.1 Carbon and sulfur concentrations of glass and allophane samples before (pristine)
and after exposure in soils close to Masaya and Poás Volcanoes.
C (%) S (%)
Glass samples
pristine 0.04±0.007a 0.002±0.0009a
Masaya 0.05±0.004a 0.002±0.0005a
Poás 0.09±0.012b 0.003±0.0002a
Allophane samples
pristine 0.16±0.002a <0.001
Masaya 0.28±0.03b 0.60±0.07a
Poás 1.40±0.18c 1.17±0.03b
Means were compared by the unpaired t-test for glass and allophane samples separately.
Values in the same column and for each mineral that are followed by the same letter are not
statistically different with p<0.05.
4.4.3 SEM-EDX of glass and allophane
4.4.3.1 SEM-EDX of glass samples
Scanning electron micrographs of carbon coated, pristine glass samples showed
features of conchoidal fracturing (Figure 4.2a) and the presence of very few isolated
holes (latter not shown). With larger magnification a certain roughness of the surface
got visible (Figure 4.2b). EDX analyses at different surface locations confirmed the
homogeneous elemental distribution in glass samples (data not shown). After expo-
sure in soils close to Masaya Volcano (Figure 4.2c,d) the number of holes increased
but elemental composition of the etch pits, as determined by EDX analyses, did not
differ from the elemental composition of unaltered glass surfaces (data not shown).
No further features could be detected on glass samples from the Masaya site.
Chapter 4
66
Figure 4.2 Scanning electron microscopy images of pristine glass (a,b) and glass that had
been exposed in soils close to Masaya (c,d) and Poás (e,f). Arrows in Figure 4.2e indicate S-
rich and SiO
2
-rich particles, circles indicate C-rich, Fe-containing phases.
Glass samples which had been exposed in soils close to Poás Volcano showed
different features on their surfaces: on one hand, small particles with dimensions of
a few micrometers can be seen, which appear in a rather bright grey tone in the
backscattered electron microscope image (exemplarily marked by arrows in Figure
4.2e), these particles seem to lie on the glass surface. On the other hand, features
of comparatively dark grey tone and showing no sharp delimitation can be observed
f e
c d
a b
In situ BAS formation
67
(exemplarily marked by circles in Figure 4.2e), which are shown at larger magnifica-
tion in Figure 4.2f.
Energy dispersive X-ray analyses of the bright grey particles reveal the presence of
both (i) SiO
2
rich particles and (ii) particles enriched in Al, S and having traces of Na
and K. SiO
2
–rich particles may have formed in situ but the fact that particles seem to
lie loosely on the glass surface makes the possibility of quartz sand translocation
from the quartz sand layer on top of glass samples more probable. Particles en-
riched in Al, S, Na and K have a rectangular shape indicating a rather crystalline
nature and show elemental compositions similar to BAS precipitates of the alunite
type. Whether these particles have formed in situ will be discussed further in section
4.5.2.
Energy dispersive X-ray measurements of the dark grey structures on glass sur-
faces showed the presence of carbon and iron, additionally to the elements con-
tained in the glass. As samples were coated for SEM-EDX analysis with carbon, at
least part of the observed carbon may stem from the coating. However, SEM-EDX
analysis of Au sputtered sample aliquots confirmed that the dark grey structures on
glass surfaces dominantly consist of C, O and Fe.
4.4.3.2 SEM-EDX of allophane
Scanning electron micrographs of Au sputtered pristine allophane show surfaces of
large aggregates as well as the presence of small aggregates with elemental com-
position typical for the synthetic allophane (Figure 4.3a, b). After exposure in soils
close to Masaya Volcano the sample morphology looks unaltered (Figure 4.3c, d).
With EDX Al, Si, O and S can be detected, elemental mapping with 500-fold magni-
fication (Figure 4.4 left side) shows that all elements including S are homogeneously
distributed in the samples. Allophane samples, which have been exposed in soils
close to Poás Volcano, show phases of irregular shape on allophane surfaces not
easily to distinguish at first glance due to the rough allophane surface morphology
(arrows in Figure 4.3e). Additionally, small aggregates are connected with each
other by chain-like structures (arrows in Figure 4.3f). Energy dispersive X-ray analy-
ses of both overlying and chain-like structures show the presence of C, O and Fe.
As theses structures are rather small compared with the EDX resolution, signals
contain also a contribution of the surrounding and underlying allophane with Al, Si,
O and S contributing to the EDX signal. Elemental mapping of allophane samples
from Poás showed that all elements are homogeneously distributed with a few local
Chapter 4
68
exceptions enriched in (i) SiO
2
or (ii) S (Figure 4.4 right side). Visualization of these
sites shows small particles with dimensions of a few micrometers which seem to lie
loosely on the surfaces of allophane aggregates. Similar as described before for the
glass surfaces S-rich particles on allophane aggregates show a rather crystalline
aspect (Figure 4.5).
Figure 4.3 Scanning electron microscopy images of pristine allophane (a,b) and allophane
that had been exposed in soils close to Masaya (c,d) and Poás (e,f). Arrows indicate C-rich,
Fe-containing phases.
a b
c d
e f
In situ BAS formation
69
Figure 4.4 Scanning electron microscopy images (backscattered electrons) and mapping of
energy dispersive X-ray analysis of S, Al and Si (in this sequence from top to bottom) of
allophane samples after field exposure close to Masaya (left column) and Poás (right col-
umn)
20
µ
µµ
µ
m
20
µ
µµ
µ
m
20
µ
µµ
µ
m
30
µ
µµ
µ
m
30
µ
µµ
µ
m
30
µ
µµ
µ
m
Chapter 4
70
Figure 4.5 Scanning electron microscopy image of S-rich particle with crystalline aspect on
the surface of an allophane aggregate after exposure in soil close to Poás Volcano.
4.4.4 STA-MS
As no measurable quantities of sulfur could be detected in glass samples, only
allophane samples were analyzed by STA-MS. In these samples alunitic com-
pounds are absent or contribute only to little extent to the SO release signals as in
H
2
O mass spectrometry signals the characteristic peak of alunite dehydroxylation
processes around 550 °C was absent. Contrastingly, the presence of organic S-
containing compounds is indicated by SO release around 200 °C for both allophane
samples from the Masaya and Poás site (inlet in Figure 4.6e,f). A comparison of
mass spectrometric SO signals for the temperature range 500 – 1400 °C with sig-
nals of BAS reference compounds, which had been analyzed in physical mixtures
with pristine allophane as described in chapter 3, shows that SO signals of samples
cannot be explained solely by the presence of adsorbed sulfate, basaluminite,
alunite or organically bound sulfate in field-exposed samples (Figure 4.6). Similarly
as described in chapter 3 this may be explained by (i) the presence of BAS com-
pounds different from the analyzed reference compounds as e.g. aluminite or jur-
banite or (ii) a matrix induced change in SO release patterns of BAS reference
compounds caused by a transformation of allophane into newly forming phases of
unknown composition. With the present data it is not possible to differentiate be-
tween both effects. Thus, by STA-MS only the presence of organically bound sulfur
could be detected while a differentiation between further S-containing phases in
allophane samples was not possible.
In situ BAS formation
71
Differential scanning calorimetry signals of pristine allophane and samples which
had been exposed in soils close to Masaya and Poás Volcanoes show the typical
features of allophane, i.e. a broad endothermic peak between 30 and 350 °C corre-
sponding to water loss and a sharp exothermic peak around 1000 °C corresponding
to the transformation of allophane into mullite and/or Al
2
O
3
. For samples which had
been exposed at Poás additionally a broad exothermic peak corresponding to the
oxidation of organic matter could be identified.
Temperature (°C)
0 200 400 600 800 1000 1200
1400
200 400 600
200 400 600
a
b
c
d
e
f
Figure 4.6 SO-signals of STA-MS analysis of (a) basaluminite, (b) Na-alunite, (c) Na-
dodecylsulfate, (d) Ref
ads
, (e) allophane sample exposed at Masaya and (f) allophane sam-
ple exposed at Poás. Reference compounds were analyzed in physical mixture with allo-
phane. For Masaya and Poás samples zoomed sketches of SO signals up to 600°C are
additionally shown.
The exothermic peaks corresponding to allophane transformation into mullite appear
at mean temperatures of 993 °C, 986 °C and 1002 °C for pristine allophane, Masaya
and Poás samples, respectively. The shift towards lower and higher temperatures
for Masaya and Poás samples is statistically significant. A shift to lower tempera-
Chapter 4
72
tures as shown by Masaya samples has been interpreted as a destabilization of
allophane due to an increase in defect sites of hollow spherules (chapter 6). Never-
theless, it is unclear which processes may lead to an increase in the recrystallization
temperature of allophane as observed for Poás samples.
4.4.5 S K-edge XANES
Similarly as with STA-MS, only allophane samples were analyzed by S K-edge
XANES. Generally, shapes of sample spectra look similar to spectra of adsorbed
sulfate, basaluminite or Na-Dodecylsulfate (Figure 4.7a). Alunitic phases may be
present in mixture with these phases. Besides the large sulfate signal with the white
line around 2.4825 keV and electron backscattering features around 2.499 keV, for
Poás samples additionally traces of sulfite phases and of organically bound S other
than estersulfate phases (e.g. organic sulfide, thiol-S, tiophene-S, sulfoxide) could
be identified due to S K-edge XANES signals in the region 2.472 keV – 2.479 keV
(Figure 4.7b) by comparison with data of reference compounds taken from Prietzel
et al. (2003). The exact nature of these compounds could not be determined as they
are only present in traces resulting in very small and rather undefined features in the
XANES spectra. Nevertheless, these features could be identified in all samples
which had been exposed at Poás. Masaya samples did not show differences in this
spectral region in comparison with C-free samples from laboratory experiments on
sulfate retention by allophane. Thus, no hints on organic S compounds could be
identified for Masaya samples. Hints on sulfone and sulfonate phases could not be
found in any sample but may be masked by the very large and overlapping sulfate
signal at 2.4825 keV.
For linear combination fitting only adsorbed sulfate, amorphous basaluminite, alunite
and estersulfate were selected as endmembers, as XANES spectra showed that the
other before-mentioned phases are only present in traces. Linear combination fits
using adsorbed sulfate and amorphous basaluminite as endmembers resulted in
contributions of precipitated sulfate to the total sulfate load ranging from 1% to 18%
and from 7% to 27% for Masaya and Poás samples respectively. Corresponding
mean percentages ± standard errors were 11 ± 2% and 19 ± 1% for Masaya and
Poás samples, respectively. Substituting basaluminite as endmember by alunite
resulted in only very slightly lower percentages of precipitated sulfate. Running LCF
with adsorbed sulfate and Na-Dodecylsulfate as endmembers resulted in 10 ± 2%
and 17 ± 1% of sulfate bound in estersulfate compounds. All spectra could be fitted
In situ BAS formation
73
well by LCF with fit residuals being lower than 5%. No substantial differences in fit
quality were observed when using different endmembers for precipitated sulfate or
estersulfate. Consequently, different BAS precipitates in my samples or the pres-
ence of organically bound sulfate cannot be distinguished from each other by apply-
ing the “best-fit” criterion.
Energy / keV
2.47 2.48 2.49 2.50 2.51 2.52
Fluorescence / a.u.
K-alunite
Na-alunite
basaluminite
Na-dodecylsulfate
allophane+sulfate
Masaya
Poás
a
Energy / keV
2.460 2.465 2.470 2.475 2.480
Fluorescence / a.u.
Masaya
Poás
b
Figure 4.7 (a) S K-edge X-ray absorption near edge spectra of K-alunite, Na-alunite, ba-
saluminite, Na-dodecylsulfate, Ref
ads
and representative allophane samples which had been
exposed in soils close to Masaya and Poás. (b) magnified sketch of superimposed spectra of
allophane samples from Poás and Masaya shows that Poás samples contain organic S-
containing phases.
Prietzel et al. (2008) reported that white lines in XANES spectra of estersulfate
compounds have a rather large full width at half maximum (FWHM = 4.6) compared
with BAS phases (FWHM = 3.0) or adsorbed sulfate (FWHM = 1.8). The FWHM of
samples of the present study varied between 1.8 and 2.0. In former experiments I
studied sulfate adsorption by allophane in laboratory experiments (chapter 5). In
these systems organically-bound estersulfate phases were absent, and white lines
of samples had FWHM between 1.8 and 2.2. The similarity of FWHM´s of samples
of the present study and FWHM´s of C-free samples suggests that estersulfate
phases are of no or of negligible relevance in allophane samples which had been
exposed in soils close to Masaya and Poás Volcanoes. This will be further dis-
cussed in section 4.5.2.
For one allophane sample, which had been exposed at the Masaya site, additionally
µ
-X-ray fluorescence and
µ
-XANES spectra were acquired. Elemental mapping with
Chapter 4
74
µ
-X-ray fluorescence showed that S, Al and Si were heterogeneously distributed in
the sample. Unfortunately, spatially resolved XANES spectra could not be evaluated
as LCF fitting of
µ
-XANES spectra resulted in rather large fit residuals due to a poor
signal to noise ratio of sample spectra.
4.4.6 ATR-FTIR
The infrared spectrum of pure allophane showed the typical features of allophane
which are caused by vibrations of structural and adsorbed water, Si-O bonds and Al-
O bonds in the spectral regions 3800-2800 cm
-1
, 1640-1630 cm
-1
, 1200-800 cm
-1
and 800-400 cm
-1
. Spectra of allophane samples which had been exposed in soils
close to Masaya and Poás Volcanoes also show these features with small devia-
tions from the spectrum of pristine allophane in the region 1250 – 750 cm
-1
(Figure
4.8). As a consequence of field exposure attenuated total IR reflectance increased
around 1100 cm
-1
. As all main sulfate bands both of adsorbed and precipitated
sulfate species are located in this region (e.g. Wijnja and Schulthess, 2000 and own
measurements of reference phases) the increase in signal intensity most probably
reflects the increase in sulfate retention. According to the larger increase in S con-
centrations, this band is more intensive for Poás samples than for Masaya samples.
Additionally, for all samples the intensity between 1040 – 930 cm
-1
decreases when
normalized to the maximum intensity between 1250 and 750 cm
-1
. A similar de-
crease in the intensity of the ATR-FTIR signal could be observed for allophane
samples during a laboratory experiment on sulfate retention kinetics and has been
attributed to a decrease in Si-O bonds of allophane, especially of polymerized Si
(chapter 6). Thereby it is unclear if this decrease can be attributed to the effect of
elevated H
+
or sulfate concentrations.
4.4.7 Sulfate extractability
Sulfate extractability from inorganic reference compounds decreases in the order
adsorbed sulfate (Ref
ads
) > amorphous basaluminite > Na-alunite >~ K-alunite > Na-
dodecylsulfate (Figure 4.9). Extractable sulfate tends to increase for all precipitated
phases with time although the absolute increase is rather low for the alunitic phases.
For Masaya samples the percentages of sulfur extractable as sulfate by exchange
with phosphate are lower than the percentages observed for Ref
ads
but larger than
the percentages observed for precipitated reference phases. Sulfate extractability of
In situ
BAS formation
75
Masaya samples increases from 1 hour up to 8 hours and remains then unchanged
until 168 hours of extraction. For Poás samples, an increase in sulfate extractability
from 1 hour up to 24 hours can be observed with no further significant change up to
168 hours. Percentages of extractable sulfate of Poás samples lie in between those
measured for basaluminite and alunitic phases, respectively.
wave number (cm
-1
)
800900100011001200
ATR-Intensity / a.u.
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0.30
wave number (cm
-1
)
800900100011001200
ATR-Intensity / a.u.
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0.30
Figure 4.8 ATR-FTIR spectra of pristine allophane (solid line) and allophane after field expo-
sure (dashed lines) at Masaya (a) and Poás (b). Deviations of field exposed allophane from
pristine allophane can be explained by (i) S retention and (ii) release of polymeric silica.
extraction time (hours)
0 20 40 60 80 100 120 140 160 180
extractable S (%)
0
20
40
60
80
100
120
basaluminite
K-alunite
Na-alunite
Na-dodecylsulfate
Ref
ads
Masaya
Poás
Figure 4.9 Sulfate extractability of the reference phases basaluminite, K-alunite, Na-alunite,
Na-dodecylsulfate and Ref
ads
and of allophane samples after exposure in soils close to
Masaya and Poás (mean values +/- standard errors).
4.4.8 Oxalate extraction of allophane
Pristine allophane and allophane samples which had been exposed in soils in the
vicinity of Masaya Volcano dissolved completely after 2 hours of extraction and no
a
b
Chapter 4
76
difference in Al:Si ratios could be observed between these samples (Figure 4.10).
Contrastingly, for allophane samples which had been exposed at the Poás site
dissolution was incomplete after 2 hours of extraction and Al and Si concentrations
increased up to 1 week of extraction. Concentrations of extracted Al and Si were
highly correlated for samples from the Poás site (r
2
= 0.99, p<0.05). After 1 week no
solid residuals were visible and the same amount of Al was extracted as for pristine
allophane while Si concentrations were slightly lower than those measured for pris-
tine allophane. Consequently, the Al:Si ratios of the 1-week extraction were slightly
larger for allophane samples from the Poás site (1.67) compared with pristine and
Masaya allophane samples (1.54). Concentrations of oxalate extractable Fe of
pristine allophane and of samples from Masaya were below the detection limit of
0.28 g kg
-1
while for Poás samples the mean concentrations of extracted Fe in-
creased from 0.35 g kg
-1
to 0.52 g kg
-1
from 2 hours up to 1 week of extraction. The
increase in Fe release was significantly correlated with the increase in Si and Al
release (r
2
= 0.45 and 0.44, respectively, p<0.05).
extraction time (hours)
0 20 40 60 80 100120140160180
Si (g kg
-1
)
0
40
80
120
160
200
Masaya
Poás
pristine
extraction time (hours)
0 20 40 60 80 100120140160180
Al (g kg
-1
)
0
50
100
150
200
250
Masaya
Poás
pristine
extraction time (hours)
0 20 40 60 80 100120140160180
Fe (g kg
-1
)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Poás
Figure 4.10 Oxalate extractable Si, Al and Fe of pristine allophane and allophane after field
exposure at Masaya and Poás. Fe concentrations in oxalate extracts of pristine allophane
and Masaya samples were below the detection limit.
In situ
BAS formation
77
The solid residuals of Poás samples after 2 hours of oxalate extraction were ana-
lyzed by ATR-FTIR spectroscopy and the results were compared with spectra of
untreated samples. The overall shape of spectra did change only slightly due to
oxalate extraction: a band between 1700 and 1650 cm
-1
appeared which can be
attributed to the presence of carboxyl groups of oxalate, and the band at 1635 cm
-1
attributable to adsorbed water decreased substantially. The shape of the bands
attributable to vibrations of Si-O and Al-O bonds did not change but the ratio of
these two signals (Si-O/Al-O ratio) increased slightly due to oxalate extraction. This
may indicate the presence of a Si-O-rich phase which is only less rapidly extractable
by oxalate. Nevertheless, initially incongruent allophane dissolution by oxalate with a
preferential release of Al also may explain the observed.
4.5 DISCUSSION
4.5.1 Effects of field exposure on allophane and glass characteristics
Results from infrared spectroscopy, differential scanning calorimetry and oxalate
extraction indicate that allophane underwent structural modifications during field
exposure. In samples of both study sites the relative amount of polymerized silica
decreased. The same effect has been observed during long-term laboratory experi-
ments on sulfate retention by allophane (chapter 6). For samples from Masaya a
decrease of the re-crystallization temperature of allophane could be observed as it
has been also measured for samples from laboratory experiments. This decrease
can be attributed to allophane destabilization caused by an increase in defect sites
in the allophane structure (chapter 6). It remains unclear which structural changes in
allophane samples exposed at Poás may lead to an increase in the re-crystallization
temperature.
Contrasting to changes in IR and DSC signals of allophane samples which had been
exposed at Masaya the oxalate extractability of these samples was not altered by
field exposure. However, for Poás samples the oxalate extractability of allophane
changed substantially. The results with extraction times of 2 hours and 4 hours –
times which are generally applied for allophane quantification in soil science (Dahl-
gren, 1994) – suggest that a rather large part of allophane underwent structural
modifications. Nevertheless, ATR-FTIR spectra of solid residuals after 2 hours of
oxalate extraction are very similar to spectra of untreated, field-exposed allophane
Chapter 4
78
and oxalate extractability of the samples increased substantially up to 1 week of
extraction reaching Si
o
and Al
o
values similar to pristine allophane. Thus, changes in
oxalate extractability most probably cannot be attributed to modifications of the
allophane structure.
Oxalate extractable iron for Poás samples increased from 2 hours of extraction up to
1 week. Scanning electron microscopy with EDX analysis showed that iron is asso-
ciated in these samples with carbon-rich structures. Thus the increase in oxalate
extractable iron with increasing extraction time suggests that C-rich structures are
slowly desorbed during oxalate extraction. Smeck and Novak (1994) and Ochs
(1996) reported that adsorbed humic substances can slow down clay mineral disso-
lution, especially if humic substances are attached to mineral surfaces by multi-site
binding as observed by SEM for allophane samples from the present study. Conse-
quently, C-rich structures on allophane aggregate surfaces are most probably re-
sponsible for the substantial decrease in oxalate extractability with extraction times
of 2 hours and 4 hours. The presence of C-coatings may also explain the apparent
discrepancy that synthetic field-exposed allophane was stable during the 18 months
field experiment at Poás although allophane is not formed naturally at this site.
Meijer et al. (2007) analyzed a number of volcanic soil samples for the amount and
quality of allophane by sequential oxalate extraction and infrared spectroscopy. The
authors did not find clear relations between the oxalate extractability of amorphous
phases and the nature of these phases as determined by infrared spectroscopy.
This discrepancy may be partially explained by my findings that organic coatings
can slow down allophane dissolution kinetics. Thus, oxalate extraction may lead to
largely erroneous results when used for allophane quantification in soils.
Binuclear surface complexes of anions on oxide surfaces may also decrease oxide
extractability (Stumm, 1997). Additionally, it may be possible that the formation of
BAS precipitates on allophane surfaces may have a shielding effect. Nevertheless,
the fact that allophane samples from Masaya as well as allophane samples from
laboratory experiments on sulfate retention did not show such a substantial de-
crease in oxalate extractability makes it very unlikely that sulfate adsorption or BAS
formation may be responsible for the observed.
Glass samples of both sites underwent dissolution processes as indicated by the
presence of etch pits and a decrease in particle size (unpublished data). Possibly
leached layers formed on glass surfaces as described by Hamilton et al. (2000) but
In situ
BAS formation
79
the methods applied in the present study are not able to detect such layers. Similar
as observed for allophane samples, surfaces of glass samples from the Poás site
were partially coated by C-rich and Fe-containing structures. In analogy to alumi-
num-organic precipitates (Scheel et al., 2008) these structures may be iron-organic
precipitates or ferrihydrite-organic matter coprecipitates as they have been de-
scribed by Mikutta et al. (2008) and Eusterhues et al. (2008).
4.5.2 Do BAS precipitates preferentially form in situ at the allophane-water
interface?
For glass samples exposed at the Masaya site, no hints on the presence of BAS
phases could be detected. Contrastingly, on glass samples from the Poás site,
alunite-like particles could be visualized by SEM-EDX. These particles have dimen-
sions of a few micrometers, lie on the surface of glass particles and do have a rather
crystalline aspect when observed with larger magnifications. Similar particles could
also be observed by SEM-EDX in allophane samples from the Poás site but not for
the Masaya site. Size and aspect of the observed particles suggest that they did not
form
in situ
. It is much more likely that these particles stem from the hydrothermal
system of the acid crater lake of Poás Volcano. Sediments of this lake are S-rich
and analysis of tephra from phreatic eruptions with substantial amounts of lake
sediments revealed the presence of alunite-like phases (own unpublished data).
Alunite and jarosite formation during acidic alteration of volcanic rocks by hydro-
thermal fluids has also been described by Hochstein & Browne (2000). The fact that
alunite-like particles could be found in the mineral tubes although tubes are only
overlain by the Ofh-horizon and not by any tephra layer suggests that S-rich aerosol
particles contribute to the S-load of soils around Poás during degassing periods.
This may also explain the very large S fluxes in Poás soils as aerosols entering
resin tubes are at least partly dissolved during resin extraction with 2 M HNO
3
.
Overall, I conclude that no hints on
in situ
BAS formation on glass surfaces could be
found for both study sites.
Contrastingly, S K-edge XANES analysis and a rather low sulfate extractability by
exchange with phosphate suggest the presence of BAS phases in allophane sam-
ples. This is in accordance with TEM observations of Delfosse et al. (2005c), in
which BAS phases are located in close vicinity to allophane-rich aggregates, and
with Delfosse et al. (2006) and Ishiguro et al. (2006) who interpreted the results of
macroscopic sulfate retention experiments by postulating BAS formation at high
Chapter 4
80
sulfate loads on allophane. In the present study, both XANES analysis and sulfate
extractability may misestimate the contribution of BAS precipitates to sulfate reten-
tion as detailed in the following. As described earlier, in XANES analysis BAS
phases may be confounded with estersulfate phases (Prietzel et al., 2008). Never-
theless, besides the hints of the narrow FWHM´s of XANES´s white lines, there is
further evidence that estersulfate phases are of negligible importance for Masaya
and Poás samples: generally, C:S ratios in soils are around 200:1. With carbon
concentrations of 0.28% and 1.40% for allophane samples exposed at Masaya and
Poás, respectively, it is unlikely that organically bound S contributes substantially to
S contents of 0.60% and 1.17%. Additionally, Erkenberg et al. (1996) showed that
C-bonded S is of major importance than estersulfate phases in SO
2
impacted soils
in Central Europe. In the present study only negligible and no signals for C-bonded
S could be detected by XANES for allophane samples exposed at Poás and Ma-
saya, respectively. Thus I conclude that XANES measurements give a rather good
estimate of the contribution of BAS phases to S retention by allophane under field
conditions.
Sulfate which can be easily exchanged by phosphate has been interpreted as ad-
sorbed sulfate. The difference between the total S content and S extractable by
phosphate consequently has been attributed to the presence of organically bound S,
sulfide, sulfite or BAS phases (Prietzel et al., 2003). Organically bound S and sulfite
is of negligible relevance for S retention on allophane at Masaya and Poás and in
XANES analysis no hints on sulfide compounds could be found. Consequently, the
difference between total S and S extractable by phosphate would be interpreted as
the amount of BAS phases following the classical interpretation of sulfate extraction
schemes (e.g. Prietzel et al., 2003; Delfosse et al., 2005b). Thus, the results of
sulfate extractability after 1 week of extraction would indicate that BAS phases
contribute with 27% and 77% to the sulfate retention by allophane exposed at Ma-
saya and Poás, respectively. These percentages of BAS phases are larger than
those determined by XANES analysis, especially in the case of Poás samples.
Mikutta et al. (2006a) described that phosphate desorbability from goethite may
decrease substantially when polygalacturonate, a network-forming model compound
for root exudates, was sorbed to goethite surfaces after phosphate retention. Simi-
larly, C-coatings on allophane aggregates in the present study may limit the desorb-
ability of adsorbed sulfate. This leads to an underestimation of adsorbed sulfate and
thus to an overestimation of BAS phases. Consequently, the contribution of BAS
In situ
BAS formation
81
phases to the total S content in soils close to Masaya may be much smaller than the
reported values of 10 – 50%, which have been determined by Delfosse et al.
(2005b,c) based on selective extraction procedures. This effect may be more pro-
nounced for the C-rich Eutric Andosols compared with Vitric Andosols in the Masaya
region.
Overall, my results show that BAS phases are preferentially formed at the allo-
phane-water interface compared with glass surfaces. XANES results indicate that
the relevance of BAS precipitates for S retention is larger for Poás than for Masaya
samples. This may partly be due to the presence of alunitic aerosol particles in Poás
samples. Additionally, (i) larger sulfate fluxes in soils and/or (ii) larger concentrations
of adsorbed organic carbon compounds, which reduce the capacity for sulfate ad-
sorption, may result in a larger
in situ
BAS formation at the allophane-water interface
at Poás compared with Masaya. Climatic conditions and the competition of solid and
dissolved organic carbon thus do not impede BAS precipitation at Poás.
With exception of aerosol particles in Poás samples no BAS precipitates could be
visualized in allophane samples by SEM-EDX. This may be explained on one side
by the presence of very small precipitates or by the formation of thin BAS-films on
allophane surfaces with dimensions of only a few atomic layers not detectable by
SEM-EDX. Delfosse et al. (2005c) observed structures with compositions similar to
BAS phases having diameters of approximately 100 nm in Andosol samples from
Masaya Volcano by transmission electron microscopy. In the present study, XANES
indicated that BAS contribute with maximal 19% to the total sulfur content of 1.17%
at the largest sulfate load of allophane samples from Poás. Thus, the total amount of
BAS in allophane samples is rather low and BAS with sizes as described by Del-
fosse et al. (2005c) may represent “needles in a haystack” in my samples and may
therefore not be visible by SEM-EDX although being present. Besides, BAS precipi-
tates may have preferentially formed in micro- and mesopore voids of allophane
aggregates thus being not detectable by a surface-sensitive method as SEM.
4.6 CONCLUSIONS
The presence of allophane in acid impacted volcanic soils not only increases the
sulfate retention capacity due to sulfate adsorption but additionally due to a prefer-
ential BAS formation at the allophane-water interface when compared with surfaces
of glass particles. Basic aluminum sulfate formation accounts with 11% and 19% to
the total S retention by allophane during the 18 months field experiment in vitric
Chapter 4
82
Andosols close to Masaya and Poás Volcanoes, respectively. A larger BAS contri-
bution to the total S concentration of samples exposed at Poás may be caused by
(i) a reduction in the sulfate adsorption capacity due to the adsorption of dissolved
organic carbon, (ii) larger sulfate fluxes in soils or (iii) the deposition of aerosol
particles from the hydrothermal system of the acid crater lake. Intensive leaching
and the competition of organic compounds and sulfate for dissolved aluminum do
not impede BAS formation at Poás.
In situ
formed BAS phases could not be visual-
ized by scanning electron microscopy suggesting that these phases are only very
small or form thin films on aggregate surfaces. Besides, BAS phases may preferen-
tially precipitate in intra-aggregate void spaces.
On glass and allophane samples from the Poás site, C-rich coatings and networks
could be identified after field exposure. Allophane dissolution kinetics and the de-
sorbability of adsorbed sulfate are lowered substantially by the presence of the C-
rich structures. These results underline the importance of applying other than selec-
tive extraction procedures for correct phase quantifications in soils.
83
5 SULFATE RETENTION BY ALLOPHANE – ADSORPTION OR
PRECIPITATION? PART 1. ISOTHERMS
5.1 SUMMARY
Allophane-water interfaces are preferential sites for the formation of basic aluminum
sulfate phases (BAS) in SO
2
-impacted Andosols. I investigated the effects of sulfate
load and pH on the relevance of BAS formation for the total sulfate retention by
allophane. Short-term pH
stat
experiments at pH 4.0, 4.5 and 5.0 were conducted with
synthetic allophane and sulfate concentrations ranging from 0 to 15 mM. The re-
lease of OH
-
was recorded and solution concentrations of sulfate, aluminum and
silica were determined after 20 hours. I characterized the dried solid experimental
products by a combination of S K-edge X-ray absorption near edge spectroscopy
(XANES), simultaneous thermal analysis coupled with mass spectrometry (STA-
MS), scanning electron microscopy with energy dispersive X-ray analysis (SEM-
EDX), gas adsorption measurements and for their sulfate extractability. Comparable
amounts of BAS were determined by XANES and STA-MS but BAS phases could
not be visualized by SEM-EDX. Precipitates already formed at low sulfate loads at
the allophane-water interface. Despite an increasing contribution of BAS to the total
sulfate retention with decreasing pH and increasing sulfate concentrations, the
overall effects were small and BAS contributed with maximal 14% to the total short-
term sulfate retention by allophane. The precipitation of BAS phases had no meas-
urable effect on the sulfate extractability. However, it led to a decrease of the allo-
phane micro- and mesoporosity thus having potential ecological implications for the
retention or release of other ions by allophane.
Chapter 5
84
5.2 INTRODUCTION
Basic aluminum sulfate phases (BAS) in acid impacted volcanic soils form rather at
the allophane-water interface than on surfaces of glass particles (Delfosse et al.,
2005c; chapter 4 of this study). The precipitation of BAS increases the sulfate reten-
tion capacity of volcanic soils. Additionally, the sulfate dynamics in the soils may be
influenced as precipitated sulfate may differ from adsorbed sulfate in its retention
and release kinetics. Thus, it is of ecological relevance to know the factors govern-
ing sulfate adsorption and BAS formation in volcanic soils.
Hitherto existing laboratory studies on sulfate retention by allophane or allophane-
containing Andosols uniformly show that the total sulfate retention increases (i) with
increasing sulfate addition and (ii) decreasing pH (e.g. Padilla et al., 2002). How-
ever, contradictory statements are given on the question whether BAS form and the
effects of sulfate concentration and pH on BAS formation are unclear. In studies with
increasing sulfate solution concentrations retention did not approach a finite maxi-
mum despite large sulfate additions (Gebhardt and Coleman, 1974; Rajan, 1979;
Padilla et al., 2002; Jara et al., 2006, Ishiguro et al., 2006; Delfosse et al., 2006).
This may be indicative for BAS formation. However, based on macroscopic data
Rajan (1979) concluded that no BAS form when sulfate is retained by allophane at
pH 5. Contrastingly, Delfosse et al. (2006) and Ishiguro et al. (2006) who analyzed
sulfate retention by Andosol samples under various pH conditions and similar sulfate
addition as Rajan (1979), interpreted biphasic isotherms – a Langmuir-type isotherm
at low sulfate loads and linearly or exponentially increasing sulfate retention at high
sulfate loads – as a shift from sulfate adsorption to sulfate precipitation. Such a shift
has been termed the “adsorption-precipitation-continuum” by McBride (2000). The
author applied the term primarily to adsorption-surface precipitation processes.
Nevertheless, it may be applied irrespective if precipitates form at mineral surfaces
or due to solution oversaturation. In the adsorption-precipitation-continuum the
relevance of precipitate formation increases gradually with increasing saturation of
available sites for sulfate adsorption.
Increasing sulfate retention with decreasing pH can be attributed to the protonation
of hydroxyl groups at the allophane-water interface, leading to a larger number of
adsorption sites and at least partially to an increase of inner-sphere-complex forma-
tion. Spectroscopic evidence for this process is missing for allophane but may be
deduced from information gained for sulfate retention by aluminum and iron
Sulfate retention isotherms
85
(hydr)oxides (Peak et al., 1999; Wijnja and Schulthess, 2000). Additionally, BAS
precipitation may be a potential source of increasing sulfate retention as allophane
dissolution increases with decreasing pH supplying aluminum for potential BAS
formation. Whether this is the case has not been investigated so far.
All hitherto existing studies on sulfate retention by allophane are purely macroscopic
experiments. However, processes cannot be deduced unambiguously from macro-
scopic data alone but a combination of macroscopic and microscopic or spectro-
scopic tools is required (Sposito, 1986; Ford et al., 2001). Thus, it is unclear whether
adsorption and precipitation processes occur simultaneously during sulfate retention
by allophane or if the relevance of precipitation increases markedly after saturation
of available sites for sulfate adsorption and with decreasing pH. For the identification
and quantification of BAS in soils, different methods have been applied. Prietzel et
al. (2008) differentiated between adsorbed and precipitated sulfate by sulfur K-edge
X-ray absorption near edge spectroscopy (S K-edge XANES) as spectra of the
mentioned S species differ in the height ratio of the white line to the post-edge oscil-
lations (WL/[PEF-1] ratio). However, during testing of the suitability of XANES for the
differentiation between adsorbed and precipitated sulfate phases by Prietzel et al.
(2008) no distinction was made between outer and inner sphere surface complexes
as no separate reference compounds for both species could be synthesized. If and
to what extent this limitation of XANES may affect the quantification of BAS by this
method is unclear.
Another method to discriminate between BAS, adsorbed sulfate and estersulfate
compounds is simultaneous thermal analysis coupled with mass spectrometry (STA-
MS) as described in chapter 3. In STA-MS samples are heated at a given rate under
a controlled atmosphere up to a predetermined temperature. Heat fluxes from or to
the sample (DSC signal) and mass losses (TG signal) are recorded by a thermobal-
ance, released gases are identified by a coupled mass spectrometer. Basic alumi-
num sulfate phases and adsorbed sulfate can be distinguished from each other
based on the patterns of sulfur monoxide and water release; inner- and outer-sphere
complexes of sulfate are expected to result in identical STA-MS signals because
outer-sphere complexes are expected to transform at an early stage of STA-MS
analysis into inner-sphere complexes (chapter 3). However, STA-MS is restricted to
samples of rather simple and known matrix composition as matrix components can
distort STA-MS signals substantially (chapter 3).
Chapter 5
86
By scanning electron microscopy coupled with energy dispersive X-ray analysis
BAS precipitates can potentially be visualized (Delfosse et al., 2005c). Nevertheless,
detection of small precipitate clusters or very thin films of BAS is limited by SEM-
EDX. Besides, precipitates which do not form on allophane aggregate surfaces but
in intra-aggregate void spaces are not detected by SEM-EDX unless a sample
pretreatment as e.g. the preparation of thin sections is applied.
Hints on (i) thin BAS films on allophane aggregate surfaces or on (ii) intra-aggregate
BAS precipitation may be obtained by gas adsorption analysis as precipitate forma-
tion on surfaces or in microvoids is expected to influence the micro- and mesoporos-
ity of allophane. Gas adsorption measurements have been applied in various stud-
ies to identify pore clogging (e.g. Kaiser & Guggenberger, 2003; Mikutta et al.,
2006b).
Selective extraction methods have been applied numerously for the quantification of
different S-species in soils (e.g. Erkenberg et al., 1996; Prietzel et al., 2001; Del-
fosse et al, 2005b). Nevertheless, such extraction procedures are not selective
sensus strictus and can at best be semi-quantitative (Prietzel et al., 2003; chapter 4
of this study). However, such procedures allow to investigate whether BAS forma-
tion – as determined by XANES, STA-MS or SEM-EDX – affects sulfate extractabil-
ity.
In this work I combine macroscopic experiments with microscopic and spectroscopic
approaches to analyze the influence of sulfate load and pH on sulfate retention
mechanisms by allophane. Additionally, I discuss the ecological implications of the
obtained results.
5.3 MATERIALS AND METHODS
5.3.1 Experimental approach
I conducted laboratory sulfate retention experiments on synthetic allophane at dif-
ferent pH
stat
(pH 4.0, 4.5 and 5.0). Sulfate concentrations of added solutions ranged
from 0 to approximately 15 mM. The pH values chosen span the pH range observed
for equilibrium solutions of volcanic soils impacted by large SO
2
input (Herre et al.,
2007).
For all experiments the equilibrium solution concentrations of sulfate, Al and Si and
the amounts of consumed acid were determined. Selected samples of the solid
Sulfate retention isotherms
87
experimental products were analyzed (i) by simultaneous thermal analysis coupled
with mass spectrometry (STA-MS), (ii) by sulfur K-edge X-ray absorption near edge
fine structure spectroscopy (S K-edge XANES) (iii) by scanning electron microscopy
with energy-dispersive X-ray analysis (SEM-EDX) and (iv) for their sulfate extracta-
bility. Synthetic sulfate-containing standard compounds were also characterized by
all of the mentioned methods and the results were compared with those of the solid
experimental products. Additionally, selected samples of the retention experiments
were analyzed by N
2
adsorption to detect potential changes in the porosity of allo-
phane aggregates.
5.3.2 Synthesis of allophane
Allophane was synthesized according to Ohashi et al. (2002) with minor modifica-
tions. Briefly, 225 g of 100 mM Na
4
SiO
4
and 300 g of 100 mM AlCl
3
* 6H
2
O were
mixed rapidly, stirred for one hour and centrifuged for two hours at 7600 g. After
decanting the clear supernatant, 200 mL of doubly deionized H
2
O were added, the
emerged allophane precursors were resuspended and the suspension was boiled
for 48 hours under a reflux condenser. Remaining salts were eliminated by dialysis
(Spectra/Por
Dialysis membrane, MWCO: 12-14.000, Spectrum Laboratories, Inc.,
Rancho Dominguez, USA) until the electrical conductivity of the suspension was
lower than 10 µS cm
-1
. The suspension was centrifuged for four hours at 7600 g and
the clear supernatant was decanted. By the described procedure an allophane gel
was obtained which was re-suspended in H
2
O and freeze-dried after shock-freezing
the suspension by dropwise injection into liquid N
2
(-196 °C). The allophane was
characterized by X-ray diffraction (Philips PW 3710, Cu K
α
-radiation), differential
scanning calorimetry - thermogravimetry coupled with mass spectrometry (Netzsch
409 PC – Pfeiffer Thermostar MS), mid and far FT infrared spectroscopy (Thermo
Nicolet Nexus), oxalate extraction according to Schwertmann (1964) and transmis-
sion electron microscopy (Philips CM 12 TEM, FEI Company, Oregon, USA). By the
described procedure allophane with the formula 1.25 SiO
2
* Al
2
O
3
* 3.2 H
2
O and a
Al:Si ratio of 1.6 was obtained. All other mentioned analyses gave results character-
istic for this type of allophane.
Chapter 5
88
5.3.3 Synthesis of sulfate-containing reference compounds
I synthesized one standard reference compound for sulfate adsorbed to allophane
(further named Ref
ads
) and two standard reference compounds for precipitated
sulfate of the aluminum hydroxy sulfate group: (i) amorphous basaluminite and
(ii) K-alunite. As reference compound for adsorbed sulfate I used a sample which I
generated in a similar way as described in section 5.3.4. Details can be found in
chapter 3. The reference mineral compounds of the aluminum hydroxy sulfate group
were synthesized, aged at 50 °C (K-alunite) and dried according to Prietzel and
Hirsch (1998). The products were characterized by X-ray diffraction (Philips PW
3710, Cu K
α
-radiation) and the Al and K contents were determined by atomic ab-
sorption spectrometry (Perkin Elmer 1100 B, Waltham, USA) after wet acid digestion
(HNO
3
conc.), the S content was quantified by a CNS analyzer (Elementar vario EL
III). Elemental compositions and XRD reflexes of the synthesized reference minerals
gave characteristic properties which were comparable to the results reported by
Prietzel and Hirsch (1998).
5.3.4 Sorption isotherms at pH 4.0, 4.5 and 5.0
During all sorption experiments samples were maintained at a preset pH under
constant stirring using automatic titrators (DL50 Graphix, Mettler Toledo) equipped
with combined glass electrodes with ceramic diaphragms (DG111-SC, Mettler
Toledo) and a Teflon blade stirrer. The titrators recorded acid consumption during
the duration of the experiment. They were calibrated before each run using buffer
solutions of pH 4 and pH 7 (WTW, Weilheim, Germany, traceability to NIST/PTB)
and the calibration was controlled after each run. All samples were run in triplicate.
In three sets of experiments samples were maintained at a constant pH of 4.0, 4.5
or 5.0 ± 0.05 during 20 hours and different amounts of sulfate were added. Freeze-
dried allophane (0.80 g) was weighed into 120 mL reaction vessels (PP), 40 mL of
0.1 M KCl were added and the suspensions were stirred for 10 minutes. Afterwards,
the pH was maintained constant for 4.5 hours on the preset value by automatic
addition of 0.1 M HCl (Titrisol
). Mean volumes of acid additions during this period
were 0.6, 3.1 and 9.7 mL for pH 5.0, 4.5 and 4.0, respectively. Subsequently, 40 mL
of 0, 0.2, 1, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 24 or 30 mM K
2
SO
4
(in 0.1 M KCl) were
added and the pH was maintained constant for further 15.5 hours. Corresponding to
Sulfate retention isotherms
89
sample numbering from lowest to highest sulfate addition, the samples will further
be named I1 – I15 with “I” standing for “isotherm”.
After a total of 20 hours the suspensions were filtered through membrane filters
(cellulose nitrate, 0.45 µm, Sartorius AG, Göttingen, Germany) and the visibly clear
filtrates were centrifuged for one hour at 159 000 g (Beckman Coulter Optima
TM
L-
90K, Fullerton, USA) to guarantee the sedimentation of allophane particles larger
than 10 nm. Preliminary particle size measurements (Coulter LS230 laser grain
sizer, Beckman Coulter Inc., Fullerton, USA) on freeze-dried allophane samples
showed that the mean aggregate diameter was 18 µm with 99% of the particles
being larger than 1 µm. The absence of visible pellets after ultracentrifugation addi-
tionally indicated that particles were already effectively separated by filtration. Filters
with solid products were rinsed with 40 mL doubly deionized H
2
O, frozen at –20 °C,
freeze-dried and the dry solid products were stored in a desiccator. Centrifuged
solutions and aliquots of the K
2
SO
4
solutions were analyzed for sulfate by anion
chromatography (DX-120, Dionex, Sunnyvale, USA, eluent: 3.5 mM NaCO
3
/ 1 mM
NaHCO
3
), and for Al and Si by flame atomic absorption spectroscopy (Perkin Elmer
1100 B, Waltham, USA). The amount of retained sulfate was calculated as the
difference between the amount of sulfate added and the amount of sulfate remaining
after centrifugation. On a selection of samples the concentrations of sulfur in the
experimental products were also analyzed by a CNS analyzer (vario EL III, Elemen-
tar, Hanau, Germany) and the results agreed with the calculated results within +/-
20%, with deviations showing no systematic differences. Saturation indices for the
minerals basaluminite, Al(OH)
3
and amorphous SiO
2
were calculated with the geo-
chemical equilibrium model Visual Minteq ver. 2.50 (Allison et al., 1991).
5.3.5 Simultaneous thermal analysis coupled with mass spectrometry (STA-
MS)
All samples were weighed in Pt/Rh crucibles, tapped by a Pt/Rh lid with a small
orifice for gas release and analyzed with a STA 449C Jupiter (Netzsch, Selb, Ger-
many) equipped with a TG/DSC sample holder. Released gases were detected by a
coupled quadrupole mass spectrometer (QMS 403C Aëolos, Netzsch, Selb, Ger-
many) which was connected by a heated capillary to the STA 449C to prevent con-
densation during gas transport to the MS. Signals of masses 18 and 48 were re-
corded corresponding to H
2
O and SO, respectively. After a 10 minutes isothermal
segment at 35 °C, samples were heated under an atmo sphere of synthetic air / N
2
Chapter 5
90
(gas fluxes 80 / 20 mL min
-1
, respectively) up to 1400 °C with a heating rate of 10 °C
min
-1
.
Measurement signals were analyzed with the software package Proteus Version
4.8.1 (Netzsch, Selb, Germany). Mass spectrometry signals for SO were smoothed
by the Golay Savitzky algorithm. In case that SO signals indicated that adsorbed
sulfate and basaluminite were the only S-containing phases, two Gaussian peaks
were fitted to the mass spectrometric data in the temperature region 700 – 1000 °C
using the program PeakFit
TM
Version 4 (Systat Software Inc., San Jose, USA). Peak
centers and peak widths were allowed to float without restrictions as the exact peak
position and peak width in STA-MS analysis depend slightly on the contents of the
analyzed phases (Smykatz-Kloss, 1974).
5.3.6 Sulfur K-edge X-ray absorption near edge spectroscopy (S K-edge
XANES)
Sulfur K-edge XANES spectra were acquired with the scanning transmission X-ray
microscope at the ID-21 beamline at the European Synchrotron Radiation Facility
(ESRF, Grenoble, France) described by Barrett et al. (2000). The method and the
instrumental setup have been described in detail by Prietzel et al. (2007). The
monochromaticity of the beam and the energy scan were ensured by a fixed exit
double crystal Si <111> monochromator located upstream the microscope, which
offers an energy resolution of 0.5 eV necessary to resolve the XANES structures.
The X-ray transmission and fluorescence signals were recorded simultaneously with
a Si photodiode mounted downstream the sample and an energy-dispersive silicon
drift diode (XFLASH 2001, Röntec, Berlin, Germany), respectively.
I obtained XANES spectra by irradiating samples that had been sandwiched be-
tween two Ultralene® foils (thickness 4 µm, Spex-Certiprep Comp., Metuchen, USA)
with X-rays in the energy range between 2450 and 2530 eV with stepwise energy
increments of 0.2 eV and a dwell time of 1 s. For each sample at least 3 spectra
were recorded at different sample positions. For some compounds it was not possi-
ble to acquire a transmission signal with the described sample mounting technique.
Thus, I analyzed the fluorescence spectra of all samples of this study. Comparisons
between the transmission and fluorescence spectra of selected samples and refer-
ence compounds assured that self-absorption was negligible during XANES meas-
urements. For each spectrum, the signals of 10 scans were compiled. The energy
calibration was done with pure CaSO
4
(white line: 2482.5 eV). I analyzed all refer-
Sulfate retention isotherms
91
ence compounds and samples by recording XANES spectra in a non-focused mode
(with the zone plate removed), the size of the beam being determined by a 200 µm
pinhole. In this mode, the integrating signal of an area of 0.13 mm² of the sample
was analyzed.
Spectra were analyzed using the software package WinXAS 3.1 (Ressler, 1998). All
spectra were baseline-corrected and normalized to the edge jump. The contributions
of adsorbed and precipitated sulfate to the total sulfate content were estimated
applying linear combination fitting (LCF). As predictor component for precipitated
sulfate the spectra of amorphous basaluminite and K-alunite were taken conducting
a LCF run for each reference compound separately and not as mixtures. As predic-
tor component for sulfate adsorbed on allophane I used sample Ref
ads
. The posi-
tions of white lines in LCF were restricted to shift not more than 0.5 eV.
5.3.7 Scanning electron microscopy with energy-dispersive X-ray analysis
(SEM-EDX)
For SEM-EDX, samples were mounted on a double-adhesive carbon tape and
coated with a thin film of gold or carbon. Samples were observed by high resolution
SEM (Hitachi S-4000, Maidenhead, UK) with magnifications up to 54 000 fold apply-
ing an accelerating voltage of 20 kV.
5.3.8 Sulfate extractability
Sample aliquots corresponding to 40 µg S were weighed into 25 mL PE bottles and
5 mL of 0.016 M KH
2
PO
4
were added. For samples with high sulfate concentrations,
sample weight and solution volume were doubled to avoid very small sample
weights which can have negative effects on measurement precision. Suspensions
were shaken for 1 hour on a rotary shaker, filtered through 0.2 µm membrane filters
(Minisart RC 15, Sartorius, Göttingen, Germany) and the filtrates were analyzed for
sulfate by capillary electrophoresis (HP 3D CE, Agilent) using a buffer solution
contained in the Agilent anion analysis kit (product number: 5063-6511).
5.3.9 Porosity measurements by N
2
adsorption
Micro- and mesoporosity of samples were determined with a Quantachrome Auto-
sorb-1 (Quantachrome, Syosset, NY) using N
2
as adsorbate. Therefore, 100 mg of
sample were degassed under vacuum until the pressure increase rate by vapor
Chapter 5
92
release was less than 1.3 Pa min
-1
within a 1-minute test interval. Helium was used
as a backfill gas. For N
2
adsorption and desorption isotherms 54 and 25 measure-
ment points from 0.00004 to 0.995 and from 0.995 to 0.05 P/P
0
were included,
respectively. Micropore (< 2 nm) volumes and average micropore diameters were
calculated by the method of Dubinin-Radushkevic (DR method; Gregg and Sing,
1982), mesopore (2 – 50 nm) volumes were obtained based on desorption data
using the Barrett-Joyner-Hallenda method (BJH, Barrett et al., 1951). Differentiation
between small (2-5 nm), medium (5-10 nm) and large mesopores (10 – 50 nm) was
achieved by linear interpolation of BJH desorption data.
5.4 RESULTS AND DISCUSSION
5.4.1 Analysis of sulfate, aluminum and silica concentrations and of acid con-
sumption
Sulfate retention on allophane increased with increasing sulfate load and decreasing
pH (Figure 5.1) as it has been described by Padilla et al. (2002). Sorption isotherms
showed a biphasic pattern and could not be described by the classical Freundlich or
Langmuir models. Similar curves as the observed ones can be obtained by model-
ing cation or anion sorption by a combination of adsorption and surface precipitation
processes as described by Farley et al. (1985). Accordingly, Ishiguro et al. (2006)
and Delfosse et al. (2006) have interpreted such biphasic curves, which they ob-
tained for sulfate retention on volcanic soil samples, as indicative for the combina-
tion of adsorption and precipitation with precipitation becoming relevant after satura-
tion of available sites for sulfate adsorption. Nevertheless, caution has to be taken
as both in my study and the work done by Ishiguro et al. (2006) and Delfosse et al.
(2006) ionic strength was not maintained strictly constant but increases with increas-
ing sulfate load. This may also cause an increase in the sorption capacity (Celi et
al., 2000). The authors showed that a gradual increase in ionic strength can result in
non-gradual steps in ion retention and thus may be the cause of the observed bi-
phasic isotherms of the present study.
The net release of Al and Si and the acid consumption increase with decreasing pH
(Figure 5.1) reflecting an increase in allophane dissolution from pH 5 to pH 4. Based
on the Si solution concentrations I calculated the percentage of dissolved allophane
as 0.6 – 1.4%, 1.3 – 2.2% and 3.0 – 5.3% for pH 5.0, pH 4.5 and pH 4.0, respec-
Sulfate retention isotherms
93
tively. As solutions of samples at pH 5.0 and pH 4.5 are undersaturated with respect
to the precipitation of amorphous SiO
2
and at least part of the released silica may
stem from the pool of adsorbed silica the percentages of allophane dissolution given
above are expected to reflect the maximum allophane dissolution. In contrast, solu-
tions at pH 4.0 are slightly oversaturated with respect to amorphous SiO
2
(maximal
saturation index: 0.1), thus SiO
2
may precipitate and allophane dissolution might be
larger than proposed by Si concentrations.
SO
4
(mM)
0 2 4 6 8 10 12 14
SO
4sorb
(mmol g
-1
)
0.0
0.1
0.2
0.3
0.4
0.5
SO
4
(mM)
0 2 4 6 8 10 12 14
Al (mmol g
-1
)
0.0
0.2
0.4
0.6
0.8
SO
4
(mM)
0 2 4 6 8 10 12 14
Si (mmol g
-1
)
0.0
0.1
0.2
0.3
0.4
SO
4
(mM)
0 2 4 6 8 10 12 14
H
+
(mmol g
-1
)
0.0
0.4
0.8
1.2
1.6
Figure 5.1 Sulfate retention, Al and Si net release and acid consumption for isotherms at pH
4.0 (black), pH 4.5 (grey) and pH 5.0 (white). Graphs show mean values
±
standard errors.
For each of the investigated pH conditions the relation between Al release and
sulfate load is different: at pH 5.0 the net release of Al is independent of the sulfate
concentration, while at pH 4.5 and at pH 4.0 the net release of Al decreases and
increases with increasing sulfate load, respectively. The observed amount of re-
leased Al results from the combination of allophane dissolution and Al consuming
precipitation processes. Constant or decreasing Al concentrations with increasing
Chapter 5
94
sulfate concentration as observed for pH 5.0 and 4.5 may indicate Al consuming
precipitation as the formation of BAS or Al(OH)
3
but a lower allophane dissolution at
high sulfate loads may likewise be caused by the formation of surface adsorption
complexes as it has been described by Stumm (1997) for the adsorption of binuclear
surface complexes to oxide surfaces.
Saturation indices for basaluminite (SI
Bas
, Figure 5.2) indicate oversaturation for all
samples of the pH 4.5 and pH 5.0 isotherms. They increased from sample I1 up to
sample I5 and decreased then up to sample I15. For pH 4.0 SI
Bas
was negative for
samples I1 – I6 and positive for samples I7 – I15 with SI
Bas
tending to decrease
slightly from samples I12 – I15. Decreases in SI
Bas
may be interpreted as precipita-
tion of a basaluminite phase. This may be especially the case for pH 5.0 and 4.5
samples, for which saturation indices larger than 2 allow for the rather rapid precipi-
tation of amorphous phases (McBride, 2000). For Al(OH)
3
different equilibrium con-
stants exist in the literature, Visual Minteq includes two of them (log K
1
= 10.8,
Nordstrom et al., 1990 and log K
2
= 8.29, Gustafsson et al., 1998) and the saturation
index for the samples of this study vary substantially with the chosen equilibrium
constant. When choosing log K
1
the saturation indices were negative for all samples
and decrease from samples I1 – I15 (Figure 5.2). When choosing log K
2
saturation
indices for Al(OH)
3
were positive for all samples of the pH 4.5 and pH 5.0 isotherms
while at pH 4.0 SI
Al(OH)
3
was positive for samples I1 – I7 and negative for samples I8
– I15 (data not graphed). Thus, a precipitation of Al(OH)
3
may not be ruled out
based on the macroscopic data.
Silica release and acid consumption increase with increasing sulfate load for all
investigated pH conditions but the slope of this increase is larger at pH 4.0 com-
pared to pH 4.5 and 5.0. Released Si can originate from the pool of Si adsorbed to
allophane or of structural Si. A larger slope of Si release for pH 4.0 compared to pH
5.0 and pH 4.5 may indicate that sulfate at this pH is more competitive with ad-
sorbed silicate or that the rather low pH and large sulfate concentrations lead to a
destabilization and thus increased dissolution of allophane. The latter argumentation
would be in accordance with increasing Al concentrations and increasing acid con-
sumption at pH 4.0.
Summarizing I conclude that the biphasic shape of sorption isotherms may be ex-
plained by the precipitation of BAS phases with precipitation processes getting
especially important after saturation of available sites for sulfate adsorption. Con-
stant or decreasing aluminum concentrations at pH 5.0 and pH 4.5 and the trends in
Sulfate retention isotherms
95
saturation indices for basaluminite also fit into this model. Nevertheless, I emphasize
that the macroscopic data cannot be interpreted without ambiguity as it has been
pointed out by various authors (e.g. Sposito, 1986; Ford et al., 2001). Accordingly,
the obtained macroscopic results can be explained equally well by a combination of
the other mentioned processes of sulfate adsorption, allophane dissolution and
Al(OH)
3
precipitation as well as further processes as e.g. the formation of aqueous
complexes (e.g. [AlSO
4+
]
aq
, Al(OH)
x3-x
).
SO
4
(mM)
0 2 4 6 8 10 12 14
SI basaluminite
-3
-2
-1
0
1
2
3
4
SO
4
(mM)
0 2 4 6 8 10 12 14
SI Al(OH)
3
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
Figure 5.2 Saturation indices for the phases (a) basaluminite and (b) Al(OH)
3
for isotherms at
pH 4.0 (black), pH 4.5 (grey) and pH 5.0 (white). For the calculation of SI
Al(OH)3
the
logK = 10.8 was taken from Nordstrom et al, 1990. Graphs show mean values
±
standard
errors.
5.4.2 STA-MS
For STA-MS analysis samples obtained by the addition of sulfate solutions with
concentrations of 2, 6, 18 and 30 mM (I4, I6, I12 and I15) were selected from all
isotherms. Depending on the pH, the mentioned concentrations of added solutions
correspond to sulfate solution concentrations of 0.1 – 0.2, 0.8 – 1.3, 4.5 – 6.2 and
9.0 – 11.6 mM at the end of the experiment, respectively. In STA-MS analysis no
water release from dehydroxylation processes of alunitic phases, as described by in
chapter 3, could be observed around 500 °C (data not shown). Thus, alunitic phases
did not form during the experiments. Sulfur monoxide mass spectrometry signals of
basaluminite and Ref
ads
as well as SO signals of samples I15 from all isotherms are
plotted in Figure 5.3. Sulfur monoxide mass spectrometry signals of samples I4, I6
and I12 show a comparable shape as samples I15. As described in chapter 3 the
Chapter 5
96
biphasic SO release from Ref
ads
can be attributed to sulfur release from adsorbed
sulfate around 880 °C and from amorphous basaluminite around 950 °C. The same
holds true for samples of the present study. Thus, STA-MS results indicate that in
addition to sulfate adsorption the reaction of sulfate with allophane leads to the
formation of a basaluminite-like BAS precipitate.
Temperature (°C)
400 600 800 1000
1200
ion current (m/z=48, SO) / a.u.
pH 5.0, I15
pH 4.5, I15
pH 4.0, I15
Ref
ads
basaluminite
Figure 5.3 Mass spectrometry signals of m/z=48 (SO) during STA-MS analysis of amorphous
basaluminite, Ref
ads
and of samples I15 of the isotherms at pH 4.0, 4.5 and 5.0.
In order to quantify the contribution of adsorbed sulfate and basaluminite to the area
of the main SO signal I fitted two Gaussian functions to the sample signals. The
resulting peak maxima attributed to adsorbed and precipitated sulfate varied be-
tween 854 – 883 °C and 945 – 958 °C, respectively. The results in Table 5.1 show
that BAS phases already formed at rather low sulfate loads (samples I4 and I6). The
percentage of precipitated sulfate tends to increase with increasing sulfate load with
the increase being largest from samples I4 to I6 and only little further change from
samples I6 to samples I12 and I15 in the case of pH 4.5 and pH 5.0 samples. For
samples of the pH 4.0 isotherm only a slight increase in the percentage of precipi-
tated sulfate could be observed from sample I4 to I15. For all analyzed samples the
relevance of precipitated phases increases with decreasing pH.
Sulfate retention isotherms
97
Table 5.1 Precipitated sulfate (% of totally retained sulfate) calculated by fitting Gaussian
peaks to SO signals of STA-MS as a function of pH and sulfate concentration of added
solutions in the isotherm experiment.
Sample number I4 I6 I12 I15
pH 5.0 5 8 8 9
pH 4.5 7 11 11 12
pH 4.0 12 12 13 14
5.4.3 S K-edge XANES
As in the case of STA-MS, from each isotherm samples obtained by the addition of
sulfate solutions with concentrations of 1, 6, 18 and 30 mM (samples I4, I6, I12 and
I15) were selected for XANES analyses. Exemplarily, spectra of samples I15 of all
isotherms are shown in Figure 5.4 together with spectra of basaluminite and Ref
ads
.
During linear combination fitting, spectra of samples with the lowest sulfate load
resulted too noisy to be adequately fitted by the combination of the spectra of refer-
ence compounds. Fit residuals for these samples reached up to 11% and fits were
therefore discarded. Residuals for all other samples were lower than 5%. Residuals
in LCF did not vary substantially when different endmembers of precipitated sulfate
were selected and consequently, potentially different BAS precipitates in my sam-
ples cannot be distinguished from each other by applying the “best-fit” criterion. For
the short-term isotherm experiments amorphous basaluminite is the most likely
forming precipitate and the presence of a basaluminite-like phase has been pro-
posed by STA-MS analyses, already. Therefore the results obtained with amor-
phous basaluminite as endmember for precipitated sulfate are shown in Table 5.2
and discussed in the following. Selecting K-alunite as potential endmember for
precipitated sulfate in LCF resulted in slightly lower contributions of precipitated
sulfate, but differences are small and do not affect the overall observed trends.
For pH 5.0 the percentage of precipitated phases tends to increase with increasing
sulfate load with the difference between sample I6 and sample I15 being statistically
significant (Kruskal-Wallis, p<0.1). For pH 4.5 all samples have similar percentages
of precipitated phases whereas at pH 4.0 a significant increase of precipitated
phases can be observed from sample I6 to sample I12 with no further changes with
increasing sulfate load. As a consequence, at low sulfate concentrations (samples
Chapter 5
98
I6) the percentage of precipitated phases is highest at pH 4.5 with no differences
between pH 4.0 and pH 5.0, while for samples I12 and I15 the percentage of precipi-
tated phases is highest at pH 4.0 with no differences between pH 5.0 and pH 4.5.
Energy / keV
2.47 2.48 2.49 2.50 2.51 2.52
pH 5.0, I15
Fluorescence / a.u.
pH 4.5, I15
pH 4.0, I15
Ref
ads
basaluminite
Figure 5.4 S K-edge XANES spectra of amorphous basaluminite, Ref
ads
and of representa-
tive measurements for samples I15 of the isotherms at pH 4.0, 4.5 and 5.0.
Table 5.2 Precipitated sulfate (% of totally retained sulfate, mean values
±
standard error, in
parenthesis: minimum and maximum values) determined by linear combination fitting of S K-
edge XANES spectra as a function of pH and sulfate concentration of added solutions in the
isotherm experiments. Amorphous basaluminite and Ref
ads
were applied as endmembers in
LCF.
Sample number I6 I12 I15
pH 5.0 3±1
(-1 – 5)
11±2
(2 – 14)
13±2
(8 – 15)
pH 4.5 10±1
(8 – 10)
13±1
(11 – 13)
13±1
(10 – 14)
pH 4.0 2±2
(-2 – 6)
20±1
(17 – 23)
20±1
(18 – 21)
Sulfate retention isotherms
99
These results are at least partly contradicting the results obtained by STA-MS
analysis. Differences and overall consequences will be addressed in section 5.4.7 in
which the results of all methods contribute to a combined discussion about the
effects of sulfate concentration and pH on the formation of BAS.
5.4.4 SEM-EDX
Exemplarily, samples with the largest sulfate load from isotherms at pH 4.0 and pH
4.5 were observed by scanning electron microscopy with energy dispersive X-ray
analysis. For both samples nearly all observed surface sections showed structures
comparable to those of pristine allophane and repeated EDX measurements gave
spectra of constant ratios of Al, Si and S. Even the very few observed sample loca-
tions with apparently differing morphology also gave EDX results with the same ratio
of Al, Si, O and S. Thus, no structures indicative for BAS precipitates could be iden-
tified by SEM-EDX.
5.4.5 Sulfate extractability
From the reference compounds amorphous basaluminite and K-alunite 19 ± 0.2 and
1 ± 0.5% of sulfate could be extracted by KH
2
PO
4
within 1 hour (mean values ±
standard errors) and these results are roughly in accordance with data reported by
Prietzel and Hirsch (1998) for synthetic BAS phases. More than 90% of sulfate could
be extracted from the sample Ref
ads
reflecting the dominantly adsorbed nature of
sulfate in this sample. Nevertheless, the sulfate extractability of BAS and the de-
pendence of sulfate extractability of BAS on solid:solution ratios in preliminary ex-
periments (data not shown) do not leave any doubt that a strictly selective extraction
of adsorbed sulfate is not possible by exchange with phosphate. A potential de-
crease in sulfate extractability of samples may therefore just give hints on the pres-
ence of precipitated phases but should be interpreted primarily in terms of effects of
sulfate precipitation – as detected by XANES or STA-MS – on the fraction of easily
exchangeable sulfate.
Sulfate extractable within 1 hour by 0.016 M KH
2
PO
4
ranged between 81 and 117%
for isotherm samples showing no systematic behavior with sulfate load or experi-
mental pH. The large percentage of extractable sulfate points towards a dominance
of adsorbed sulfate and no effect of BAS formation – as determined by XANES and
STA-MS – on sulfate extractability could be detected. Partially, this may be due to
Chapter 5
100
rather large measurement uncertainties which may occult small systematic devia-
tions of sulfate extractability. These measurement uncertainties resulted as uncer-
tainties from the determination of the total sulfate content – necessary for the calcu-
lation of the sample weight – sum to the uncertainties stemming from all steps of the
sulfate extractability and measurement procedure.
5.4.6 Porosity measurements
Porosity measurements were conducted for samples I1 (only background electrolyte
was added) and I15 (highest sulfate load) from the pH 4.0 and pH 4.5 isotherms and
for pristine allophane (Table 5.3). All samples were run in triplicate. Compared to
pristine allophane, mesoporosity did not change significantly for samples I1 from the
pH 4.5 experiment while all other samples had lower mesopore volumes (p < 0.05).
Especially small mesopores with diameters of 2 – 5 nm were affected while no
differences could be observed for pores with diameters of 5 – 10 nm and 10 –
50 nm. Microporosity was reduced for samples I15 from both the pH 4.0 and pH 4.5
experiments while micropore volumes of samples I1 were not significantly different
from the micropore volume of pristine allophane. Average micropore diameters were
smaller for I15 samples while I1 samples had similar average micropore diameters
as pristine allophane. Thus, a reduction of mesopore volume could be observed for
samples to which sulfate had been added as well for the pH 4.0 experiment samples
to which only background electrolyte had been added. Contrastingly, significant
changes in microporosity could only be observed for samples to which sulfate had
been added. No differences were detected between micropore volumes of pH 4.0
and pH 4.5 samples.
Decreases in micro- and mesopore volumes may be attributed to (i) aggregation
processes, (ii) allophane dissolution, and / or (iii) pore clogging due to precipitate
formation. By SEM imaging no hints on striking differences in aggregation could be
observed between samples with the highest sulfate load and pristine allophane.
Possibly, small changes in the aggregation degree may have occurred without being
visible at first glance. Nevertheless, decreases in micro- and mesoporosity are
rather large reaching values up to 22%. It is unlikely that such a rather large de-
crease may be caused by only small changes in the aggregation degree.
Sulfate retention isotherms
101
Table 5.3 Mesopore volumes (MePV), micropore volumes (MiPV) and average micropore
diameters (AMiPD) of pristine allophane and samples I1 and I15 from the pH 4.5 and pH 4.0
experiment. Values are given as means
±
standard error.
Sample MePV
(mm
3
g
-1
) MiPV
(mm
3
g
-1
) AMiPD
(nm)
total 2-5
nm 5-10
nm 10-50
nm
allophane 209±3
a
159±3
a
28±1
a
23±1
a
178±1
a
2.01±0.01
a
pH 4.5 I1 200±2
ab
151±2
ab
28±1
a
22±1
a
169±5
ab
2.01±0.01
ab
pH 4.5 I15
187±5
b
140±3
c
25±1
a
22±1
a
153±7
b
1.95±0.01
c
pH 4.0 I1 184±7
bc
137±5
bcd
26±1
a
22±1
a
168±5
ab
1.98±0.01
abc
pH 4.0 I15
170±1
c
124±2
d
26±1
a
21±1
a
152±7
b
1.91±0.01
d
Means were compared by the unpaired t-test. Values in the same column that are followed
by the same letter are not statistically different with p<0.05.
Allophane dissolution may contribute partially to the reduction in porosity. Neverthe-
less, the fact that no differences in microporosity could be observed between pH 4.0
and pH 4.5 samples - although allophane dissolution is larger at pH 4.0 than pH 4.5
- makes it unlikely that allophane dissolution caused the decrease in microporosity.
Only the decreases in mesopore volumes may be attributed partially to allophane
dissolution. Thus, the decrease in micropore volume and at least partially the de-
crease in mesopore volume may be attributed to pore clogging due to precipitate
formation.
5.4.7 Influence of sulfate concentration and pH on sulfate retention processes
The biphasic nature of the sorption isotherms and the trends of Al release and ba-
saluminite saturation indices suggest the formation of BAS. The results of STA-MS
and XANES analysis are in line with the assumption of BAS formation. However,
both STA-MS and XANES are “fingerprint” methods and potentially misestimate the
contribution of BAS to the total sulfate retention. In STA-MS analysis matrix effects
can distort SO and H
2
O signals (chapter 3). However, this effect should be negligi-
ble for samples of the present study as the allophane microstructure is expected not
to change substantially due to the short-term character of the conducted experi-
ments. This assumption was confirmed by IR spectra of selected samples which
showed additional bands attributable to sulfate retention but no visible change in the
characteristic allophane bands (data not shown).
Chapter 5
102
One source of differences between STA-MS and XANES may be that in XANES
analysis different spectra may be observed for inner sphere and outer sphere com-
plexes with spectra of inner sphere complexes being intermediate between those of
outer sphere complexes and those of precipitated phases. In XANES spectra the
white line at the absorption edge results from electron transitions between orbitals
while the post edge oscillations are caused by backscattering of ejected photoelec-
trons by neighboring atoms (Fendorf & Sparks, 1996). Prietzel et al. (2008) argued
that the height ratio of the white line to the post edge feature is smaller for precipi-
tated phases compared to adsorbed sulfate because the 3-dimensional system of
precipitated phases allows for a more intense backscattering of photoelectrons and
consequently an increase in the post-edge feature. According to Khare et al. (2005),
who differentiated between adsorbed and precipitated phosphate by XANES analy-
sis, the white lines of adsorbed phases are higher and sharper than those of precipi-
tated phases. The authors attributed this to the fact that in the two-dimensional
system of adsorbed phases atomic orbitals overlap to form discrete molecular orbi-
tals. Contrastingly, in three-dimensional precipitate phases, especially in amorphous
phases, atomic orbitals broaden because of a nearly continuous distribution of
energy differences between orbitals. The argumentations of Khare et al. (2005) and
Prietzel et al. (2008) make it likely that differences between adsorbed and precipi-
tated phases and not differences between inner and outer sphere complexes are
probed by XANES analysis. Nevertheless, Arai et al. (2001) could show that XANES
spectra of outer-sphere and inner-sphere As complexes on the aluminum oxide
interface differ from each other. Thus, it cannot be ruled out that similar differences
affect the quantification of adsorbed and precipitated sulfate when applying XANES.
For the present study, this would lead to an overestimation of BAS phases by
XANES analysis. However, XANES results are not systematically larger than STA-
MS results suggesting that potential differences in XANES spectra between inner
and outer sphere complexes do not distort substantially the results obtained by
XANES analysis for precipitate formation. A third potential source for differences
between STA-MS and XANES results may be the fact that both methods probe
different sample aliquots: while in STA-MS analysis a sample aliquot of 100 mg is
analyzed, the X-ray beam in XANES analysis probes a much smaller aliquot with
fluorescence signals stemming only from a depth of few micrometers and a sample
area of 0.13 mm
2
. Thus, it is probable that XANES results of few measurement
points per sample may not be representative for the whole sample, especially in
Sulfate retention isotherms
103
case that precipitates are not evenly distributed. This assumption is confirmed by
the observation that XANES results for each sample vary substantially from one to
another analysis point as indicated by minimum and maximum values of precipitated
sulfate in Table 5.2. Contrastingly, preliminary replicate measurements of samples
by STA-MS had given reproducible results. Based on the exposed I conclude that
STA-MS gives a fairly good estimate of BAS precipitates in samples of the present
study. Although XANES results may not be representative for the bulk sample the
overall agreement between STA-MS and XANES results suggests that each XANES
measurement gives a fairly good estimate of BAS precipitates of the selected meas-
urement point.
The formation of BAS phases during sulfate retention by allophane is in accordance
with Delfosse et al. (2005b, 2005c, 2006), Prietzel et al. (2008) and with my results
of chapter 4 in which BAS phases were detected on synthetic allophane which had
been exposed in soils close to SO
2
releasing active volcanoes. BAS precipitation
increased with decreasing pH. This may be attributed most probably to an increase
in aluminum concentrations resulting from increased allophane dissolution with
decreasing pH. This finding is in contradiction with Bigham and Nordstrom (2000)
who postulated that BAS only form under conditions under which aluminum hydroxo
complexes exist, i.e. when the pH is nearly 5.0 or higher. Nevertheless, the pres-
ence of alunitic phases in extremely acid hydrothermal systems of active volcanoes
(Hochstein & Browne, 2000) invalidates the postulation of Bigham and Nordstrom
(2000). This is also confirmed by my observation that the synthesis of basaluminite
and K-alunite resulted in a pH between 4.0 and 4.5 in the supernatant solution.
Biphasic retention isotherms may be interpreted by sulfate adsorption preceding
BAS precipitation at the allophane-water interface (Delfosse et al., 2006; Ishiguro et
al., 2006). However, my results show that adsorption and precipitation occur simul-
taneously, even at rather low sulfate loads. Sulfate adsorption is the dominant proc-
ess for sulfate retention for all investigated sulfate concentrations and pH conditions
and the relative contribution of precipitates to the total sulfate retention increases
only slightly with increasing sulfate load. Consequently, the overall isotherm shape
has to be explained primarily by changes in factors influencing sulfate adsorption.
Sulfate may compete more effectively with adsorbed silica with increasing sulfate
solution concentrations. Based on increasing silica solution concentrations, such an
effect has been suggested by Pardo & Guadalix (1990) and by Arai et al., (2005) for
phosphate and arsenate retention by allophane, respectively. Nevertheless, in my
Chapter 5
104
study a marked increase in sulfate adsorption at high sulfate loads is not corre-
sponded by a marked increase in silica release thus making this explanation little
probable. Large sulfate solution concentrations may destabilize allophane leading to
an increase in the number or size of defect sites and thus generating new sites for
sulfate adsorption. A similar mechanism has been postulated by Imai et al. (1981)
for phosphate retention by allophane. Additionally, increasing sulfate retention at the
allophane-water interface at high sulfate loads may be caused by an increase in
ionic strength of the surrounding solution as it has been described by Celi et al.
(2000) for phosphate retention on iron (hydr)oxide. In the present study and also in
sulfate sorption isotherms described by Delfosse et al. (2006) 0.1 M background
electrolyte was used to maintain the ionic strength constant. Nevertheless, as the
added sulfate solutions in both studies span a rather wide range of concentrations,
the ionic strength is not maintained strictly constant but increases from 0.1 M to
0.14 M from samples with the lowest to the highest sulfate load, respectively. Celi et
al. (2000) showed that a gradual increase in ionic strength can result in non-gradual
steps in ion retention and thus may be the cause of the observed biphasic isotherms
of the present study.
Based on results gained on phosphate retention by iron (hydr)oxides (Anderson et
al., 1985; Barrow et al., 1993; Liu et al., 1995; Strauss et al., 1997), aggregation
processes and concomitant ion burial or transformation of bidentate into monoden-
tate adsorption complexes may explain the observed. Additionally, increased sulfate
diffusion into the interdomain spaces of allophane may result in a retention increase
at high sulfate concentrations. Whether the results gained on P retention by sesqui-
oxides may be transferred to anion retention by allophane deserves further investi-
gation.
In the present study no BAS precipitates could be visualized by SEM-EDX although
STA-MS and XANES indicate the presence of such phases. Delfosse et al. (2005c)
observed structures with compositions similar to BAS having diameters of approxi-
mately 100 nm in Andosol samples from Masaya Volcano by transmission electron
microscopy. In case such precipitates would be present in my samples they should
be visible at least under the larger magnifications I applied. However, STA-MS
indicated that BAS contribute with maximal 14% to the total sulfur content of 1.25%
at the largest sulfate load of allophane samples at pH 4.0. Thus, the total amount of
BAS in allophane samples is rather low and such “needles in a haystack” may not
be visible by SEM-EDX although present in the samples. Besides, BAS precipitates
Sulfate retention isotherms
105
may have formed smaller particles or more or less continuous thin networks on
allophane surfaces with dimensions of a few atomic layers not detectable by SEM-
EDX. Additionally, a preferential BAS precipitation may have occurred in micro- and
mesopore voids of allophane aggregates as ion exchange between the void spaces
and the freely draining soil solution may be diffusion limited allowing for BAS satura-
tion in void spaces while the freely draining solution may be undersaturated.
5.4.8 Ecological implications
The increase in precipitate formation with increasing sulfate load and with decreas-
ing pH was rather small and was not reflected by a change in the amount of easily
exchangeable sulfate. Thus, although precipitate formation increases the total sul-
fate retention capacity it is of only minor relevance when considering sulfate avail-
ability. However, experiments of the present study were short-term experiments and
precipitate formation and the stability of precipitates may increase with increasing
reaction time.
Isotherm experiments yielded decreases in meso- and microporosity. The decrease
in micropore volume and at least partially the decrease in mesopore volume may be
attributed to pore clogging due to precipitate formation. Thereby, it is unclear
whether precipitates form on allophane aggregates with consequent clogging of
underlying pores or if precipitates are preferentially formed in micro- and small
mesopores caused by an increase in saturation indices due to diffusion limitations
as mentioned before. Micropore and mesopore clogging is of ecological relevance
as it may limit or impede ion diffusion into allophane aggregates thus influencing the
mobility and plant availability of nutrients or contaminants. Such limitations have
been described e.g. for molybdate and phosphate diffusion into iron (hydr)oxides by
pore clogging due to organic coatings (Lang and Kaupenjohann, 2003; Mikutta et
al., 2006b).
The laboratory systems which I used in this study are rather simple and pure. Obvi-
ously, volcanic soils are much more complex and the presence of other solid phases
in soils may influence sulfate retention by allophane. As an example, aluminum for
BAS precipitation may not stem necessarily from allophane dissolution but may be
provided by the dissolution of volcanic glass. Which processes dominate, will de-
pend ultimately on small scale heterogeneities of the solid and solution composi-
tions. In this view, my experiments represent one of a number of possible scenarios
occurring in real-word samples.
Chapter 5
106
5.5 CONCLUSIONS
The processes of sulfate adsorption and BAS precipitation at the allophane-water
interface do not occur successively but simultaneously, even at low sulfate loads.
Increasing sulfate addition and decreasing pH lead to an increase in BAS formation
although the effects are small. Sulfate adsorption is the dominant process for sulfate
retention for all analyzed sulfate concentrations and pH conditions. Consequently,
the biphasic isotherm shape has to be explained by processes which increase the
sulfate adsorption capacity. This may be (i) an increase in defect sites of allophane
spherules, (ii) an increase in ionic strength at high sulfate loads, (iii) the transforma-
tion of binuclear into mononuclear complexes or (iv) increased diffusion of sulfate
into the interdomain spaces of allophane at high sulfate concentrations. Additionally,
aggregation processes at high sulfate loads may lead to ion burial. Research is
needed to elucidate which of the mentioned processes govern ion adsorption at the
allophane-water interface.
Precipitation of BAS phases increases the total sulfate retention capacity of allo-
phane. Nevertheless, sulfate availability seems to be only little affected as precipi-
tate formation did not have a measurable influence on sulfate extractability in the
presented short-term experiments. The formation of BAS phases led to a decrease
in micro- and mesoporosity of allophane aggregates suggesting that BAS precipita-
tion may occur on surfaces or in the pore space of allophane aggregates which
reduces the volume of small pores. The decrease in meso- and microporosity is of
ecological relevance as it may limit or impede ion diffusion into or out of allophane
aggregates and thus affect the reactivity of allophane.
107
6 SULFATE RETENTION BY ALLOPHANE – ADSORPTION OR
PRECIPITATION? PART 2. KINETICS
6.1 SUMMARY
Short-term sulfate retention by allophane is dominated by adsorption processes and
retained sulfate can be easily exchanged by phosphate. I investigated whether
longer reaction times may increase the relevance of basic aluminum sulfate phases
(BAS) for the total sulfate retention and/or may alter the BAS composition and
thereby affect the sulfate extractability. I conducted pH
stat
sulfate retention experi-
ments (4 mM K
2
SO
4
) with synthetic allophane at pH 4.5 and pH 5.0 up to 6 months.
Solution concentrations of sulfate, aluminum and silica as well as acid or base con-
sumption were recorded. The solid experimental products were characterized by a
combination of S K-edge X-ray absorption near edge spectroscopy (XANES), simul-
taneous thermal analysis coupled with mass spectrometry (STA-MS), scanning
electron microscopy with energy dispersive X-ray analysis (SEM-EDX), infrared
spectroscopy (IR) and wet chemical extraction procedures. At pH 5.0, sulfate reten-
tion reached an apparent equilibrium after two weeks while it increased at pH 4.5 up
to six months. The precipitation of BAS contributed maximal by 8% to the total long-
term sulfate retention which can thus dominantly be explained by adsorption proc-
esses. Applying IR and STA-MS I could attribute the slow sulfate adsorption to
(i) the exposition of existing adsorption sites due to silica depolymerization and (ii) to
the generation of new sites caused by an increase in the number of defect sites of
the allophane spherules. These results stress the importance of analyzing the stabil-
ity of metastable sorbents like allophane when conducting long-term experiments.
Chapter 6
108
6.2 INTRODUCTION
Sulfate retention in SO
2
-impacted volcanic soils is governed by sulfate adsorption to
variable charge minerals like allophane and ferrihydrite (Shoji et al., 1993; Padilla et
al., 2002) and by the formation of basic aluminum sulfate phases (BAS, Delfosse et
al., 2005b, 2005c; Prietzel et al., 2008). The knowledge of the relevance of both
processes is important for the prediction of sulfate dynamics as (i) adsorption is
limited by the number of available adsorption sites while precipitation can be an
infinite sulfate sink and as (ii) the products of adsorption and precipitation processes
are expected to differ in their formation and dissolution kinetics. Transmission elec-
tron microscopy observations of Delfosse et al. (2005c) revealed that BAS can be
found in close vicinity to allophane particles. Short-term experiments confirmed that
BAS form during sulfate retention by allophane and that the relevance of BAS for-
mation increases with increasing sulfate concentration and decreasing pH (chapter
5). The formation of BAS in these experiments did not have an influence on the
sulfate extractability. Nevertheless, this may change with increasing time as (i) the
relevance of precipitation processes may increase with time and / or (ii) formed
precipitates may undergo aging processes with a concomitant decrease in sulfate
extractability.
Studies on the kinetics of sulfate retention by allophane are very scarce and they
present contradicting results on the question whether sulfate retention shows a slow
increase over longer time spans. While Rajan (1979) reported that sulfate retention
on synthetic and natural allophane at pH 5 did not increase from 3 hours up to 52
hours, Jara et al. (2006) observed an increase in sulfate retention by synthetic allo-
phane and an Andosol sample up to the end of their experiment after 2 weeks. No
explanation is given on the processes responsible for an increase in sulfate reten-
tion with time, however. Increasing sulfate incorporation into BAS may be one pos-
sible explanation. Similarly, Imai et al. (1981) and Parfitt (1989) attributed part of the
slow P retention by allophane at high P loads to the precipitation of aluminum phos-
phate phases. Imai et al. (1981) based their conclusions on macroscopic wet chemi-
cal data and microcalorimetric measurements. By applying these methods the same
authors reported that apart from precipitation processes the slow increase in phos-
phate retention is caused by the progressive adsorption of phosphate. They postu-
lated that the exposure of new high-energetic surfaces for anion retention caused by
the breakdown of the allophane microstructure is responsible for the observed. Imai
Sulfate retention kinetics
109
et al. (1981) attributed the breakdown to two potential processes, i.e. (i) the fracture
of Al-O-Al bonds and (ii) the substitution of structural silicate groups by phosphate.
Accordingly, a slow sulfate retention by allophane may be also caused by the expo-
sure of newly forming sites for sulfate adsorption. Nevertheless, until now no direct
evidence exists of changes in the allophane structure caused by anion retention.
Additionally, slowly increasing ion sorption may be attributed to (i) particle aggrega-
tion with concomitant ion burial and (ii) ion diffusion to adsorption sites i.e. micropore
diffusion. The latter two processes have been described for phosphate retention by
iron or aluminum (hydr)oxides (Anderson et al., 1985; Strauss et al., 1997) but their
relevance for ion retention by allophane has not been investigated so far. Parfitt
(1989) postulated that ion diffusion to adsorption sites is of no relevance in the case
of allophane as all AlOH sites, which are responsible for anion retention, are at the
surface of the hollow spherules. Nevertheless, allophane has a strong tendency to
form microaggregates. Thus, micropore diffusion of anions into aggregates may
affect the slow sulfate retention by allophane. Summarizing the before mentioned, it
is not clear whether sulfate retention by allophane shows a slow increase with time
and whether BAS precipitate formation may be responsible for that.
Besides, it is unclear whether rapidly formed BAS precipitates may transform with
time and if this process has an influence on the kinetics of sulfate release from these
phases. Precipitate aging effects have been described e.g. for Ni-Al layer double
hydroxide phases by Ford et al. (1999) who showed that Ni-Al LDH slowly trans-
formed to a Ni-Al phyllosilicate by the incorporation of silica leading to an increase of
the thermal stability of the formed precipitates. In the case of BAS precipitates it is
known that alunitic phases like K-alunite or Na-alunite are thermodynamically more
stable than e.g. basaluminite or aluminite (Bigham and Nordstrom, 2000). However,
natural precipitates mostly have a composition similar to basaluminite or hydroba-
saluminite and the formation of theses phases is kinetically favored compared with
alunitic phases (Nordstrom, 1982). This is in agreement with the results presented in
chapter 5, in which STA-MS analysis suggested the presence of a basaluminite-like
phase in short-term experiments on sulfate retention by allopane. The question
whether basaluminitic phases transform into alunitic phases with increasing time is
unclear. While in experiments with pure solutions containing aluminum, potassium
and sulfate such a transformation was only observed at elevated temperatures, the
transformation occurred already at ambient temperatures in solutions seeded with
bentonite (Adams and Rawajfih, 1977).
Chapter 6
110
In this paper I test the hypothesis that the contribution of BAS precipitates to the
total sulfate retention increases with increasing reaction time. Besides, I investigate
if BAS phases undergo transformation processes. The influence of both the potential
increase in BAS phases and BAS transformation on the sulfate extractability is
analyzed. Additionally, I evaluate if further processes as (i) the breakdown of the
allophane microstructure, (ii) aggregate formation with ion burial and (iii) micropore
diffusion may have an influence on the slow increase in sulfate retention by allo-
phane.
6.3 MATERIALS AND METHODS
6.3.1 Experimental approach
I conducted laboratory sulfate retention experiments with synthetic allophane under
pH
stat
conditions at pH 4.5 and pH 5.0. Experiments were stopped after 6 months
and 3 weeks, respectively. Sulfate concentrations of the added solutions were ap-
proximately 4 mM. The pH conditions and sulfate concentrations are representative
for conditions observed for equilibrium solutions of volcanic soil samples impacted
by large SO
2
input (Herre et al., 2007). During the experiments, the solution concen-
trations of sulfate, Al and Si and the amounts of consumed acid were determined.
Samples for the solid phase analysis were selected based on these macroscopic
results. Solid samples were then analyzed by (i) simultaneous thermal analysis
(STA = differential scanning calorimetry and thermogravimetry, DSC/TG) coupled
with mass spectrometry (MS), (ii) sulfur K-edge X-ray absorption near edge fine
structure spectroscopy (S K-edge XANES) and (iii) scanning electron microscopy
with energy-dispersive X-ray analysis (SEM-EDX). The combination of these meth-
ods was applied with the aim to distinguish between adsorbed and precipitated
sulfate. Additionally, STA-MS analysis allows the identification of alunitic phases.
Synthetic sulfate-containing standard compounds were characterized by all of the
mentioned methods and the results were compared with those of the solid experi-
mental products.
Selected solid samples from the kinetic experiment were analyzed for their sulfate
extractability. Besides, the potential breakdown of the allophane structure was in-
vestigated by the interpretation of the DSC/TG signals of the STA-MS analysis, by
attenuated total reflectance Fourier-transformed infrared spectroscopy (ATR-FTIR)
Sulfate retention kinetics
111
and by oxalate extraction. By SEM I looked for potential changes in the allophane
aggregation degree. Based on the sulfate extractability the potential role of ion burial
in newly forming aggregates or of ion diffusion into aggregate micropores is dis-
cussed.
6.3.2 Synthesis of allophane
Allophane was synthesized according to Ohashi et al. (2002) with minor modifica-
tions. Briefly, 225 g of 100 mM Na
4
SiO
4
and 300 g of 100 mM AlCl
3
* 6H
2
O were
mixed rapidly, stirred for one hour and centrifuged for two hours at 7600 g. After
decanting the clear supernatant, 200 mL of doubly deionized H
2
O were added, the
allophane precursors were resuspended and the suspension was boiled for 48
hours under a reflux condenser. Remaining salts were eliminated by dialysis (Spec-
tra/Por
Dialysis membrane, MWCO: 12-14,000, Spectrum Laboratories, Inc., Ran-
cho Dominguez, USA) until the electrical conductivity of the suspension was lower
than 10 µS cm
-1
. The suspension was centrifuged for four hours at 7600 g and the
clear supernatant was decanted. By the described procedure an allophane gel was
obtained which was re-suspended in H
2
O and freeze-dried after shock-freezing the
suspension by dropwise injection into liquid N
2
(-196 °C). The allophane was char-
acterized by X-ray diffraction (Philips PW 3710, Cu K
α
-radiation), differential scan-
ning calorimetry - thermogravimetry coupled with mass spectrometry (Netzsch 409
PC – Pfeiffer Thermostar MS), mid and far FT infrared spectroscopy (Thermo
Nicolet Nexus), oxalate extraction according to Schwertmann (1964), and transmis-
sion electron microscopy (Philips CM 12 TEM, FEI Company, Oregon, USA). Micro-
and mesoporosity of the allophane were measured by N
2
adsorption (Autosorb-1,
Quantachrome, Syosset, NY), micropore (<2 nm) volumes were calculated by the
method of Dubinin-Radushkevic (DR method; Gregg and Sing, 1982), mesopore (2
– 50 nm) volumes were obtained based on desorption data using the Barrett-Joyner-
Hallenda method (BJH, Barrett et al., 1951). By the described synthesis procedure
allophane with the formula 1.25 SiO
2
* Al
2
O
3
* 3.2 H
2
O and a Al:Si ratio of 1.6 was
obtained. Micro- and mesopore volumes were 178±1 and 209±3 mm
3
g
-1
, respec-
tively (mean values ± standard error). All other mentioned analyses gave results
characteristic for allophane.
Chapter 6
112
6.3.3 Synthesis of sulfate-containing reference compounds
I synthesized one standard reference compound for sulfate adsorbed to allophane
(further named Ref
ads
) and two standard reference compounds for precipitated
sulfate of the aluminum hydroxy sulfate group: (i) amorphous basaluminite and (iii)
K-alunite.
As reference compound for adsorbed sulfate I used a sample which I obtained
applying the same procedure as described in chapter 5 for samples of the sorption
isotherms adding a 6 mM K
2
SO
4
solution. As the only difference to the isotherm
experiments I did not manipulate the pH before and during sulfate retention. By this,
the pH of the sample was 5.2 at the beginning and 5.7 at the end of the experiment.
The Al solution concentration in this sample was 26 µg L
-1
at the end of the experi-
ment and BAS precipitates are not expected to form in measurable quantities. Nev-
ertheless, precipitation of Al hydroxy sulfate minerals cannot be ruled out com-
pletely. STA-MS analysis of Ref
ads
however showed that the contained sulfate can
be attributed dominantly (98%) to adsorbed sulfate and that precipitated phases only
contribute to a minor extent (2%) (chapter 4).
The reference mineral compounds of the aluminum hydroxy sulfate group were
synthesized, aged at 50 °C (K-alunite) and dried according to Prietzel and Hirsch
(1998). The synthesis products were characterized by X-ray diffraction (Philips PW
3710) and the Al and K contents were determined by atomic absorption spectrome-
try (Perkin Elmer 1100 B, Waltham, USA) after wet acid digestion (HNO
3
conc.), the
S content was quantified by a CNS analyzer (Elementar vario EL III). Elemental
compositions and XRD reflexes of the synthesized reference mineral compounds
gave characteristic results which were comparable to the results reported by Prietzel
and Hirsch (1998) for their synthetic BAS phases.
6.3.4 Retention experiments
During retention experiments samples were maintained at a preset pH under con-
stant stirring using automatic titrators (DL50 Graphix, Mettler Toledo) equipped with
combined glass electrodes with ceramic diaphragms (DG111-SC, Mettler Toledo)
and a Teflon blade stirrer. The titrators recorded acid consumption during the dura-
tion of the experiment. They were calibrated before each run using buffer solutions
of pH 4 and pH 7 (WTW, Weilheim, Germany, traceability to NIST/PTB) and the
calibration was controlled after each run. All samples were run in triplicate.
Sulfate retention kinetics
113
For the sorption experiment at pH 4.5 sample amounts of 1.25 g freeze-dried allo-
phane were weighed into 500 mL reaction vessels (PP) and 125 mL of 0.1 M KCl
were added; for the pH 5.0 experiment 0.4 g freeze-dried allophane were weighed
into 120 mL reaction vessels (PP) and 40 mL of 0.1 M KCl were added. Suspen-
sions were stirred for 10 minutes prior to the start of the experiment. Subsequently,
suspensions were titrated to the preset pH and the pH was maintained constant by
addition of 0.1 M HCl (Titrisol
) for 4.5 hours. The mean acid addition during this
period was 5.1 mL and 1.6 mL for the pH 4.5 and pH 5.0 experiments, respectively.
For the pH 4.5 experiment 125 mL 8 mM K
2
SO
4
(in 0.1 M KCl) were added and
sample aliquots were taken after 0.5, 1, 2, 4, 8, 24, 48, 168, 336, 672, 1344 and
4032 hours. For the pH 5.0 experiment 40 mL 8 mM K
2
SO
4
(in 0.1 M KCl) were
added and sample aliquots were taken after 0.5, 1, 2, 4, 8, 24, 48, 192, 336 and 504
hours. Corresponding to sample numbering from the first to the last sampling mo-
ment, the samples further will be named Ka1 – Ka12 for the pH 4.5 experiment and
Kb1 – Kb10 for the pH 5.0 experiment. For samples Ka1 – Ka5 and Ka7 – Ka9 an
aliquot of 10 mL was taken while for samples Ka6 and Ka10 – Ka12 an aliquot of
40 mL was taken. For samples Kb1 – Kb10 an aliquot of 6 mL was taken. From the
beginning of the experiment up to 24 hours the pH was maintained constant by
automatic titrators. After 24 hours, the pH was manually adjusted to the preset pH ±
0.1 by addition of 0.1M HCl (Titrisol
) or 0.1M KOH (Titrisol
). Aliquots were filtered
through membrane filters (cellulose nitrate, 0.45 µm, Sartorius AG, Göttingen, Ger-
many). Filters with solid products were frozen at –20° C, freeze-dried and the dry
solid products stored in a desiccator. Filtrates and aliquots of the K
2
SO
4
solutions
were analyzed for sulfate by anion chromatography (DX-120, Dionex, Sunnyvale,
USA, eluent: 3.5 mM NaCO
3
/ 1 mM NaHCO
3
), and for Al and Si by flame atomic
absorption spectroscopy (Perkin Elmer 1100 B, Waltham, USA). The amount of
retained sulfate was calculated as the difference of the amount of sulfate added and
the amount of sulfate remaining after centrifugation. On a selection of samples the
concentration of sulfur in the experimental products was also analyzed by a CNS
analyzer (vario EL III, Elementar, Hanau, Germany) and the obtained sulfur concen-
trations corresponded to 96 - 119% of the sulfur concentrations as determined by
the calculation method with deviations showing no systematic behavior. Saturation
indices for the minerals basaluminite, Al(OH)
3
and amorphous SiO
2
were calculated
with the geochemical equilibrium model Visual Minteq ver. 2.50 (Allison et al., 1991).
Chapter 6
114
6.3.5 Simultaneous thermal analysis coupled with mass spectrometry (STA-
MS)
In STA-MS analysis samples are heated at a given rate and under a controlled
atmosphere up to a predetermined temperature. Heat fluxes from or to the sample
(DSC signal) and mass losses (TG signal) are recorded by a thermobalance, re-
leased gases are identified by a coupled mass spectrometer as a function of sample
temperature. Mass spectrometry (MS) signals are more sensitive and allow the
detection of small quantities of different S-pools which cannot be seen in the DSC or
TG traces. Therefore, the MS signals are used to distinguish between different S-
pools. Experiments conducted in chapter 3 showed that STA-MS signals of sulfate
adsorbed to allophane, sulfate included in basaluminite or sulfate included in K-
alunite differ. Adsorbed sulfate and basaluminite can be distinguished from each
other by differences in the temperature dependence of SO release while K-alunite
can be identified by a characteristic combination of SO- and H
2
O-release. In case
that only sulfate adsorbed to allophane and amorphous basaluminite are present in
the system, peak fitting allows the quantification of the contribution of both species
to the total sulfate retention. This holds true if the matrix composition of the samples
remains unchanged. Changes in the matrix composition may influence the sulfur
release from samples substantially thus complicating signal interpretation (chapter
3).
All samples were weighed in Pt/Rh crucibles, tapped by a Pt/Rh lid with a small
orifice for gas release and analyzed with a STA 449C Jupiter (Netzsch, Selb, Ger-
many) equipped with a TG/DSC sample holder. Released gases were detected by a
coupled quadrupole mass spectrometer (QMS 403C Aëolos, Netzsch, Selb, Ger-
many) which was connected by a heated capillary to the STA 449C to prevent con-
densation during gas transport to the MS. Signals of masses 18 and 48 were re-
corded corresponding to H
2
O and SO, respectively. After a 10 minutes isothermal
segment at 35 °C, samples were heated under an atmosphere of air / N
2
(gas fluxes
80 / 20 mL min
-1
, respectively) up to 1400 °C with a heating rate of 10 °C min
-1
.
Measurement signals were analyzed with the software package Proteus Version
4.8.1 (Netzsch, Selb, Germany). Mass spectrometry signals for SO were smoothed
by the Golay Savitzky algorithm. In case that SO signals indicated that adsorbed
sulfate and basaluminite were the only S-containing phases, two Gaussian peaks
were fitted to the temperature region 700 – 1000 °C of the SO signals with the pro-
gram PeakFit
TM
Version 4 (Systat Software Inc., San Jose, USA). Peak centers and
Sulfate retention kinetics
115
peak widths were allowed to float without restrictions as the exact peak position and
peak width in STA-MS analysis depend slightly on the contents of the analyzed
phases (Smykatz-Kloss, 1974).
6.3.6 Sulfur K-edge X-ray absorption near edge spectroscopy (S K-edge
XANES)
By sulfur K-edge XANES spectroscopy I determined the percentage of adsorbed
and precipitated sulfate as described by Prietzel et al. (2008): XANES spectra of
adsorbed and precipitated sulfate differ in their ratio of the white line (WL) height at
2482.5 eV to the height of a post-edge feature (PEF) at 2499 eV (WL/[PEF-1] ratio).
The contributions of adsorbed and precipitated sulfate can be estimated by calculat-
ing the WL/[PEF-1] ratios of sample spectra or by linear combination fitting (LCF) of
sample spectra using reference compounds. For the present study I analyzed spec-
tra by LCF.
Sulfur K-edge XANES spectra were acquired with the scanning transmission X-ray
microscope at the ID-21 beamline at the European Synchrotron Radiation Facility
(ESRF, Grenoble, France) described by Barrett et al. (2000). The method and the
instrumental setup have been described in detail by Prietzel et al. (2007). The
monochromaticity of the beam and the energy scan were ensured by a fixed exit
double crystal Si <111> monochromator located upstream the microscope, which
offers an energy resolution of 0.5 eV necessary to resolve the XANES structures.
The X-ray transmission and fluorescence signals were recorded simultaneously with
a Si photodiode mounted downstream the sample and an energy-dispersive silicon
drift diode (XFLASH 2001, Röntec, Berlin, Germany), respectively.
I obtained XANES spectra by irradiating samples that had been sandwiched be-
tween two Ultralene® foils (thickness 4 µm, Spex-Certiprep Comp., Metuchen, USA)
with X-rays in the energy range between 2450 and 2530 eV with stepwise energy
increments of 0.2 eV and a dwell time of 1 s. For each sample at least 3 spectra
were recorded at different sample positions. Fluorescence spectra of all samples
were evaluated in my study. Former comparisons between the transmission and
fluorescence spectra of selected samples and reference compounds assured that
self-absorption did not constitute a problem during XANES measurements (Prietzel
et al., 2008). For each spectrum, the signals of 10 scans were compiled. The energy
calibration was done with pure CaSO
4
(white line: 2482.5 eV). I analyzed all refer-
ence compounds and samples by recording XANES spectra in a non-focused mode
Chapter 6
116
(with the zone plate removed), the size of the beam being determined by a 200 µm
pinhole. In this mode, the integrating signal of an area of 0.13 mm² of the sample
was analyzed.
Spectra were analyzed using the software package WinXAS 3.1 (Ressler, 1998). All
spectra were baseline-corrected and normalized to the edge jump. The contributions
of adsorbed and precipitated sulfate to the total sulfate content were estimated
applying linear combination fitting (LCF). As predictor component for precipitated
sulfate the spectra of amorphous basaluminite, aged basaluminite or K-alunite were
taken conducting a LCF run for each reference compound separately and not as
mixtures. The XANES spectrum of aged basaluminite was supplied by Prof. Dr. Jörg
Prietzel (TU München) and had been obtained by the same measurement proce-
dure as described above. As predictor component for sulfate adsorbed on allophane
I used sample Ref
ads
. The positions of white lines in LCF were restricted to shift not
more than 0.5 eV.
6.3.7 Scanning electron microscopy with energy dispersive X-ray analysis
(SEM-EDX)
Scanning electron microscopy with energy dispersive X-ray analysis was applied
with the aim to visualize BAS precipitates and to analyze the samples for changes in
allophane morphology as e.g. changes in the aggregation degree. Therefore, sam-
ples were mounted on a double-adhesive carbon tape and coated with a thin film of
gold or carbon. Samples were observed by high resolution SEM (Hitachi S-4000,
Maidenhead, UK) with magnifications up to 54 000-fold applying an accelerating
voltage of 20 kV. For selected samples, additionally S mapping was done with a
conventional scanning electron microsope coupled with EDX (Hitachi S-2700, Maid-
enhead, UK) to check whether sulfur is distributed homogeneously in samples or
localized in S hot-spots. Due to the instrumental setup, S-mappings with magnifica-
tions up to 1500-fold can be conducted with this instrument, potential hot-spots can
be observed with larger magnifications.
6.3.8 Attenuated total reflectance Fourier-transformed infrared spectroscopy
(ATR-FTIR)
With the aim to detect potential allophane transformations during the kinetic experi-
ment Fourier-transformed IR spectra were recorded by a Bruker IFS66/S (Bruker
Sulfate retention kinetics
117
Optics, Ettlingen, Germany) equipped with a “Golden Gate” attenuated total reflec-
tance micro-unit (Specac, London, UK) with a single reflection diamond crystal.
Spectra were recorded from 4000 – 400 cm
-1
with a resolution of 4 cm
-1
and 64
scans per sample. Therefore, samples were deposited on the diamond crystal and a
defined pressure was applied on the samples by tightening the sample screw with a
constant torque of 90 cNm with a dynamometric key. All spectra were background
and baseline corrected. As background signal the IR spectrum of the empty ATR-
diamond in air was used.
The ATR-FTIR technique was chosen instead of the more commonly applied KBr
technique, as the latter may have substantial impact on sample signals: High pres-
sure effects on allophane during the preparation of KBr pellets, ion exchange of KBr
with allophane functional groups or the influence of KBr on spectra due to differ-
ences in the refractive index of KBr and allophane (“Christiansen effect”) may all
distort IR spectra. The fact that the ATR technique is a more surface sensitive tech-
nique was not expected to influence allophane spectra substantially as the probing
depths of the ATR technique of a few µm is rather large compared to allophane
hollow spherule dimensions of 3.5 to 5 nm and is in the range of allophane aggre-
gate dimensions. In order to be able to compare IR spectra of the present study with
literature data which predominantly were obtained by the KBr technique, an aliquot
of pure allophane was analyzed additionally by the KBr technique. All bands associ-
ated with vibrations due to structural bonds in allophane were shifted by 30-40 cm
-1
to larger wavenumbers for the KBr technique compared to the ATR-FTIR technique
while the signal attributable to adsorbed water appeared for both techniques at
1635 cm
-1
. This suggests that the KBr technique has a substantial effect on struc-
tural allophane IR signals. Thus, for a comparison of ATR-FTIR signals of structural
vibrations of my samples with literature data my data should be shifted by 30-
40 cm
-1
to larger wavenumbers.
6.3.9 Wet chemical extraction methods
6.3.9.1 Sulfate extractability
I analyzed the sulfate extractability of samples from the kinetic experiment by ex-
change with phosphate to evaluate whether BAS precipitate formation – as deter-
mined by STA-MS and XANES – has an influence on the fraction of easily ex-
Chapter 6
118
changeable sulfate. Therefore, sample aliquots corresponding to 40 µg S were
weighed into 25 mL PE bottles and 5 mL of 0.016 M KH
2
PO
4
were added. Suspen-
sions were shaken for 1 hour on a rotary shaker, filtered through 0.2 µm membrane
filters (Minisart RC 15, Sartorius, Göttingen, Germany) and the filtrates were ana-
lyzed for sulfate by capillary electrophoresis (HP 3D CE, Agilent) using a buffer
solution contained in the Agilent anion analysis kit (product number: 5063-6511).
6.3.9.2 Oxalate extraction
Oxalate extractable Al and Si were determined by shaking sample aliquots for 2
hours with ammonium oxalate / oxalic acid solution according to Schwertmann
(1964) and a solid:solution ratio of 1:1000. After 2 hours, samples were filtrated and
the concentrations of Al and Si were determined by flame atomic absorption spec-
troscopy (Perkin Elmer 1100 B). For quality control, in each analysis batch aliquots
of synthetic allophane were used as in-house reference material.
6.4 RESULTS AND DISCUSSION
6.4.1 Analysis of sulfate, aluminum and silica concentrations and of acid con-
sumption
6.4.1.1 Sulfate retention kinetics at pH 4.5
Sulfate retention by allophane did not reach a plateau within the experimental period
of 6 months (Figure 6.1). Aluminum concentrations increased up to 24 hours after
which the concentrations decreased exponentially. Silica concentrations increased
up to 48 hours, remained then constant up to 1 week, decreased up to 1 month and
increased again until the end of the experiment. Acid consumption increased up to
48 hours and then decreased slightly up to 6 months. Saturation indices for basalu-
minite and Al(OH)
3
increased from Ka1 – Ka6 and decreased from Ka6 – Ka12.
Saturation indices of basaluminite were positive for samples Ka1 – Ka11 and nega-
tive for samples Ka12. For Al(OH)
3
different equilibrium constants exist in the litera-
ture, Visual Minteq includes two of them (log K
1
= 10.8, Nordstrom et al., 1990 and
log K
2
= 8.29, Gustafsson et al., 1998) and the saturation index for the samples of
this study vary substantially with the chosen equilibrium constant. When choosing
Sulfate retention kinetics
119
log K
1
the saturation indices were negative for all samples and decrease from sam-
ples Ka1 – Ka12 (Figure 6.1). When choosing log K
2
saturation indices for Al(OH)
3
were positive for all samples (data not graphed). Thus, a precipitation of Al(OH)
3
may not be ruled out based on the macroscopic data.
time (h)
0 1000 2000 3000 4000 5000
SO
4sorb
(mmol g
-1
)
0.10
0.15
0.20
0.25
0.30
0.35
0.40
time (h)
0 1000 2000 3000 4000 5000
Al (mmol g
-1
)
0.00
0.05
0.10
0.15
0.20
time (h)
0 1000 2000 3000 4000 5000
Si (mmol g
-1
)
0.04
0.08
0.12
0.16
0.20
time (h)
0 1000 2000 3000 4000 5000
H
+
(mmol g
-1
)
0.1
0.2
0.3
0.4
0.5
0 50 100 150 200
0.00
0.05
0.10
0.15
0.20
time (h)
0 1000 2000 3000 4000 5000
SI basaluminite
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
time (h)
0 1000 2000 3000 4000 5000
SI Al(OH)
3
-2.6
-2.5
-2.4
-2.3
-2.2
-2.1
-2.0
-1.9
-1.8
Figure 6.1 Sulfate retention, Al and Si net release and acid consumption during the kinetic
experiment at pH 4.5. Graphs show mean values
±
standard errors. In the case of Al a
second and smaller graph zooms into Al release up to 200 hours.
Chapter 6
120
The complex behavior of Si may be explained by the combination of allophane
dissolution and the exchange of adsorbed Si by sulfate at the beginning of the ex-
periment, the inclusion of Si in precipitating phases in the middle and the destabili-
zation of allophane at the end of the experiment.
Increasing sulfate retention, the decrease in aluminum concentrations from Ka6 –
Ka12, the release of H
+
from Ka9 – Ka12 and positive, decreasing saturation indices
for basaluminite suggest that the relevance of this phase may increase during the
kinetic experiment. Nevertheless, the interpretation of the macroscopic data is
speculative (Sposito, 1986; Ford et al., 2001) and an integrative discussion on the
relevance of different processes for sulfate retention will be given in section 6.4.9
based on the results of all applied methods.
6.4.1.2 Sulfate retention kinetics at pH 5.0
Sulfate retention at pH 5.0 increased up to two weeks but no further change was
observed between two and three weeks (Figure 6.2). Aluminum concentrations were
constant from Kb1 – Kb5 and decreased from Kb5 – Kb10 while silica concentra-
tions increased from Kb1 – Kb7 and decreased slightly up to Kb10. Acid consump-
tion increased up to Kb6 and did not change further up to Kb10. Saturation indices
for basaluminite were positive for samples Kb1 – Kb7 and negative for samples Kb7
– Kb10. They were constant for Kb1 – Kb5 and decreased from Kb5 to Kb10. Satu-
ration indices for Al(OH)
3
depended on the chosen equilibrium constant as de-
scribed for pH 4.5. When choosing log K
1
the saturation indices were negative for all
samples and decrease from samples Kb5 to Kb10. When choosing log K
2
saturation
indices for Al(OH)
3
were positive for samples Kb1 – Kb8 (data not graphed). Thus, a
precipitation of Al(OH)
3
may not be ruled out based on the macroscopic data.
The decrease in Al concentrations for samples Kb5 – Kb10 suggests that precipita-
tion processes occur. Nevertheless, saturation indices for basaluminite got negative
during the experiment, thus making it unlikely that this phase controlled the slow
sulfate retention. The formation of other BAS phases cannot be ruled out but de-
creasing Al concentrations may be equally explained by the precipitation of Al(OH)
3
or an alumino-silicate phase.
The slow increase in sulfate retention is more pronounced for samples of the pH 4.5
experiment than for samples of the pH 5.0 experiment. Thus, for solid phase analy-
sis only samples of the pH 4.5 experiment have been selected.
Sulfate retention kinetics
121
time (h)
0 100 200 300 400 500 600
SO
4sorb
(mmol g
-1
)
0.18
0.20
0.22
0.24
0.26
0.28
time (h)
0 100 200 300 400 500 600
Al (mmol g
-1
)
0.000
0.002
0.004
0.006
0.008
0.010
0.012
0.014
0.016
0 10 20 30 40 50
0.000
0.004
0.008
0.012
0.016
time (h)
0 100 200 300 400 500 600
Si (mmol g
-1
)
0.04
0.05
0.06
0.07
0.08
0.09
0.10
time (h)
0 100 200 300 400 500 600
H
+
(mmol g
-1
)
0.10
0.12
0.14
0.16
0.18
0.20
time (h)
0 100 200 300 400 500 600
SI basaluminite
-3
-2
-1
0
1
2
3
time (h)
0 100 200 300 400 500 600
SI Al(OH)
3
-3.0
-2.8
-2.6
-2.4
-2.2
-2.0
-1.8
-1.6
-1.4
Figure 6.2 Sulfate retention, Al and Si net release and acid consumption during the kinetic
experiment at pH 5.0. Graphs show mean values
±
standard errors. In the case of Al a
second and smaller graph zooms into Al release up to 50 hours.
Chapter 6
122
6.4.2 ATR-FTIR
Samples of the pH 4.5 experiment taken after 0.5, 24 and 48 hours, 1 week, 2
weeks, 1 month, 2 months and 6 months were analyzed in triplicate by ATR-FTIR.
For the sake of clarity, only spectra of sample aliquots of one replicate taken after 48
hours, 2 weeks, 2 months and 6 months are presented in Figure 6.3 together with
the spectrum of pure allophane. Spectra taken after 0.5 and 24 hours, 1 week and 1
month as well as spectra of the other replicates show the trends as described in the
following. Up to 48 hours, the signal intensity in the region 950 – 1200 cm
-1
in-
creases, if normalized to the maximum signal between 930 and 940 cm
-1
. As all
main sulfate bands both of adsorbed and precipitated sulfate species are located in
this region (e.g. Wijnja and Schulthess, 2000 and own measurements of reference
phases) the increase in signal intensity most probably reflects the increase in sulfate
retention. After 48 hours and up to 6 months the signal intensity between 950 and
1200 cm
-1
decreases. This decrease cannot be attributed to a change in sulfate
bands as the amount of retained sulfate increased up to 6 months. Thus, the de-
crease indicates changes in the IR bands of allophane. Infrared signals of allophane
in the region 800 - 1200 cm
-1
are attributed to Si-O vibrations with signals of
wavenumbers >1000 cm
-1
stemming from polymerized silica of the allophane struc-
ture (Harsh et al., 2002). As the main reduction in signal intensity can be observed
for wavenumbers around 1020 cm
-1
(corresponding to ~1060 cm
-1
in KBr-based
literature data taking into account the band shift as described in section 6.3.8), I
conclude that the fraction of polymeric silica decreased during the kinetic experi-
ment. This can be caused by the transformation of polymeric Si into monomeric Si
and / or a preferential Si release from the pool of polymeric silica compared with
orthosilicate groups of the allophane structure. Abidin et al. (2007) observed a simi-
lar shift of the Si-O band to lower wavenumbers as a consequence of dilute alkali
treatment of natural allophane. However, Abidin et al. (2007) observed the shift only
for Si-rich allophane but not for rather Al-rich allophane as I used it in my experi-
ments.
Sulfate retention kinetics
123
Figure 6.3 ATR-FTIR spectra of pure allophane and solid sample aliquots taken from the
kinetic experiment after 48 hours, 2 weeks, 2 months and 6 months.
6.4.3 Oxalate extraction
For oxalate extraction samples taken after 24 hours, 1 month, 2 months and 6
months have been selected. Silica and aluminum extractability decreased during the
experiment. The decrease is larger for Si than for Al leading to an increase of the
Al:Si ratio of the oxalate extracts (Table 6.1).
Table 6.1 Oxalate extractable aluminum and silica and molar Al:Si ratios from solid sample
aliquots taken after 24 hours, 1 month, 2 months and 6 months of the kinetic experiment.
Si Al Al:Si
------------------- g kg
-1
------------------- mol mol
-1
24 hours 131±2 a 211±2 a 1.67±0.02 a
1 month 131±1 a 207±2 ab 1.64±0.01 a
2 months 125±2 a 208±4 ab 1.72±0.01 b
6 months 118±1 b 199±3 b 1.75±0.01 b
Means were compared by the unpaired t-test. Values in the same column that are followed
by the same letter are not statistically different with p<0.05.
Chapter 6
124
6.4.4 STA-MS
6.4.4.1 Differential Scanning Calorimetry
Sample aliquots, which were taken after 24 hours, 2 weeks, 1 month, 2 months and
6 months, were analyzed by DSC. I observed a broad endothermic peak between
30 and 350 °C, which corresponds to water loss from allophane and which did not
change from samples Ka6 – Ka12. In addition, a very sharp exothermic peak around
995 °C could be observed for pure allophane and samples taken after 24 hours, 2
weeks, 1 month and 2 months. This peak is typical for allophane and reflects the
exothermic crystallization of the dehydroxylated phase into γ-alumina and/or spinel
like e.g. mullite (Smykatz-Kloss, 1974; Wada, 1989). For the sample aliquot taken
after 6 months, this exothermic peak broadens and shifts from 995 °C to 880 °C.
Smykatz-Kloss (1974) reported that the position of the exothermic peak of kaolinite
transformation into mullite shifts from approximately 1000 °C to 940 °C with increas-
ing kaolinite disorder. Accordingly, the shift of the exothermic peak for the sample
aliquot taken after 6 months may be interpreted as an increase in allophane disor-
der. This can be caused by an increase in defect sites of the hollow spherule parti-
cles. Such an increase in defect sites has been reported for alkali-treated allophane
by Abidin et al. (2007).
6.4.4.2 Mass spectrometry signals
For samples of the kinetic experiment no water vapour release could be observed
between 500 and 550 °C (data not shown). Thus, the formation of alunitic phases
can be ruled out. Mass spectrometry signals of SO for the before mentioned se-
lected samples are shown in Figure 6.4 together with signals of basaluminite mixed
with allophane and Ref
ads
. The signal of the sample aliquot taken after 24 hours
looks similar to the signal of Ref
ads
. Additionally to the main SO signal between 800
and 1000 °C further peaks appear between 500 and 70 0 °C and between 1000 and
1400 °C. With increasing time from 24 hours up to f our weeks, the additional peaks
between 500 – 700 °C and 1000 – 1400 °C increase. A fter two months the main SO
signal shifts towards higher temperatures with the additional peaks increasing fur-
ther and after 6 months the SO signal consists of one main signal between 900 and
1200 °C with a number of other signals at lower and higher temperatures.
Sulfate retention kinetics
125
Temperature (°C)
400 600 800 1000 1200
1400
Ion current (m/z=48, SO) / a.u.
24 hours
2 weeks
1 month
2 months
6 months
Ref
ads
basaluminite
Figure 6.4 Mass spectrometry signals of m/z=48 (SO) during STA-MS analysis of amorphous
basaluminite, Ref
ads
and samples taken after 24 hours, 2 weeks, 1 month, 2 months and 6
months from the kinetic experiment.
A comparison of sample SO signals with those of reference compounds reveals that
just the sample taken after 24 hours can be largely explained by the combination of
reference compounds, namely Ref
ads
and basaluminite. Fitting two Gaussian peaks
to the main SO signal by the procedure described in section 6.3.5 results in a por-
tion of 9% precipitated sulfate. For all other samples at least part of their SO signals
cannot be attributed to the analyzed reference compounds. This may be caused by
(i) the presence of sulfate phases different from the chosen reference compounds or
(ii) a matrix-induced change in the release pattern of SO from samples or (iii) a
combination of both effects. STA-MS analyses of reference compounds alone and in
physical mixture with allophane had shown that the matrix can have a substantial
effect on the release pattern of SO (chapter 3). The decreases in aluminum and
silica concentrations during the experiment as well as the results of the oxalate
extraction indicate that the samples undergo changes in their matrix composition
during the experiment. Thus, a potential matrix effect on the SO release pattern
caused by newly formed phases cannot be ruled out thus not allowing to distinguish
Chapter 6
126
unambiguously between adsorbed and precipitated S-containing phases in samples
of the kinetic experiment by STA-MS analysis.
6.4.5 S K-edge XANES
For XANES analysis samples taken after 2 hours, 8 hours, 2 weeks and 2 months
were selected. Samples taken after the experimental period of 6 months could not
be included in XANES measurements as samples were taken after the XANES
measurement campaign. In Figure 6.5 representative S K-edge XANES spectra of
samples and of reference compounds are shown. As STA-MS indicated that no
alunites formed during the experiment, LCFs were only run with amorphous and
aged basaluminite as predictor components for precipitated sulfate and Ref
ads
for
adsorbed sulfate. XANES spectra of amorphous and aged basaluminite are very
similar. Consequently, the results obtained by LCF for precipitated sulfate in sam-
ples are very similar irrespective if amorphous or aged basaluminite were used as
predictor component. When applying amorphous basaluminite, linear combination
fitting of spectra resulted in rather low portions of precipitated sulfate of 3 ± 2, 6 ± 2,
7 ± 1 and 8 ±1% for samples taken after 2 hours, 8 hours, 2 weeks and 2 months,
respectively. The contribution of precipitated phases tends to increase with increas-
ing time but the increase is not statistically significant (Kruskal-Wallis, p<0.1). Al-
though up to 6 measurements per sample were made at different sample locations,
XANES results are not necessarily representative for the whole sample as explained
in chapter 5. Nevertheless, the comparison of results obtained by XANES and STA-
MS for the isotherm experiments revealed that XANES measurements give an
estimate of the magnitude of BAS formation. Thus, the XANES results of the kinetic
experiment indicate that the increase in BAS formation with increasing time is rather
low. I calculated the contribution of precipitated phases to the overall increase in
sulfate retention between two hours and two months: while the amount of totally
retained sulfate increased from 2 hours up to 2 months by 0.14 mmol g
-1
, the
amount of precipitated sulfate increased by 0.02 mmol g
-1
thus representing only
14% of the increase in totally retained sulfate.
Sulfate retention kinetics
127
Energy / keV
2.47 2.48 2.49 2.50 2.51 2.52 2.53
2 months
2 weeks
8 hours
2 hours
Ref
ads
basaluminite
fluorescence / a.u.
Figure 6.5 S K-edge XANES spectra of amorphous basaluminite, Ref
ads
and sample aliquots
taken after 2 hours, 8 hours, 2 weeks and 2 months of the kinetic experiment.
6.4.6 SEM-EDX
Exemplarily, pristine allophane and samples taken after 2 and 6 months were ob-
served by scanning electron microscopy with energy dispersive X-ray analysis.
Surface structures of samples from the kinetic experiment were comparable to those
of pristine allophane (Figure 6.6) and repeated EDX measurements gave spectra of
constant ratios of Al, Si and S. Sulfur was homogeneously distributed in the samples
when mapped with a magnification of 1500-fold. These results suggest that (i) BAS
precipitates are rather small with maximum dimensions up to few hundreds of na-
nometers thus being not distinguishable by S-mapping with magnifications of 1500-
fold and / or that (ii) potentially larger precipitates may be present but possibly not
detected when observed at larger magnifications without preceding localization of S
hot-spots.
Chapter 6
128
Figure 6.6 Scanning electron microscopy images of (a) pristine allophane and (b) a sample
aliquot taken after 2 months of the kinetic experiment
6.4.7 Sulfate extractability
The sulfate extractability was determined for samples taken after 24 hours, 1 week,
2 weeks, 1 month, 2 months and 6 months. The percentage of sulfate extractable
within one hour by KH
2
PO
4
ranged between 82 and 103% showing no systematic
behavior with increasing time. The large percentage of extractable sulfate points
towards a dominance of adsorbed sulfate. Nevertheless, the presence of basalu-
minite also can contribute to extractable sulfate with its sulfate extractability depend-
ing on the size of precipitates and the precipitate solution ratio. No effect of BAS
formation on sulfate extractability could be detected. Partially, this may be due to
rather large measurement uncertainties of the sulfate extractability. These uncertain-
ties resulted as uncertainties from the determination of the total sulfate content –
necessary for the calculation of the sample weight – sum to the uncertainties stem-
ming from all steps of the sulfate extractability and measurement procedure. Thus,
small potential systematic trends of sulfate extractability with increasing reaction
time may not be detected.
6.4.8 Allophane dissolution / transformation during the pH 4.5 experiment
Aluminum and silica solution concentrations during the experiment, infrared spec-
troscopy, differential scanning calorimetry and oxalate extraction indicate that allo-
phane is partly dissolved and that the allophane structure undergoes transforma-
tions during the pH 4.5 experiment. A comparison of the reduction in silica
extractability (13 g kg
-1
) with the increase in released silica between 24 hours and 6
b
a
Sulfate retention kinetics
129
months during the kinetic experiment (0.55 g kg
-1
) indicates that the increase in
silica release during the kinetic experiment represents just 4% of the reduction of Si
extractable by oxalate. Thus, the decrease in oxalate extractable Si has to be attrib-
uted to changes of the solid phase. On one side adsorbed sulfate or BAS precipi-
tates on allophane aggregate surfaces may slow down the allophane dissolution
kinetics in oxalate solution. However, basaluminitic phases dissolve rapidly and
completely in oxalate (results not shown) and no hints on alunitic phases, which
have a lower oxalate extractability, could be found. Thus, a decrease in Si extracta-
bility most probably can be attributed to an allophane transformation resulting in a
Si-containing phase of lower oxalate extractability. Based on negative saturation
indices for amorphous SiO
2
it seems unlikely that this phase was precipitated. Nev-
ertheless, saturation conditions in micro- and mesopores of allophane aggregates
may differ substantially from conditions in the surrounding solution and precipitation
of amorphous SiO
2
thus cannot be ruled out. Additionally, solid solutions containing
Si may have formed. The formation of such mixed phases can decrease the solubil-
ity of an ion substantially in comparison with the formation of a pure compound
(McBride, 2000).
Infrared spectroscopy indicated that the release of polymeric silica is dominantly
responsible for the decrease in silica extractability. In the allophane structure the
amount of polymeric silica increases with increasing Si:Al ratio (Harsh et al., 2002).
According to Wada (1989) with increasing Si:Al ratio the amount of silica bound to
aluminum octahedra on the outer side of allophane spherules increases. Combining
both findings leads to the statement that during my experiment polymeric silica from
the outer side of the allophane spherules depolymerizes and may consecutively be
removed while orthosilicate groups, which are located on the inner side of the hollow
spherules, are less affected.
The results of differential scanning calorimetry indicated an allophane destabilization
for samples taken after 6 months. It is unlikely that such a destabilization may be
caused by the removal of polymeric silica as in the literature no changes in the
position of the exothermic DSC signal has been reported for allophanes of varying
Si:Al ratios, i.e. with varying relevance of polymeric silica. Accordingly, another
process must be responsible for allophane destabilization as it may be the genera-
tion of defect sites, i.e. an increase in the number or size of pores in the allophane
structure as is has been described for alkaline-treated allophane (Abidin et al.,
2007). The latter authors observed that alkaline treatment affected polymerized
Chapter 6
130
silica and the pore regions of allophane sperules but that the effect on the pore
regions was much smaller when polymerized silica was present. Thus I may hy-
pothesize that in my samples the slow removal of polymerized silica delayed the
substantial destabilization of allophane, which I only could observe for samples
taken after 6 months. Overall, it is unclear whether the observed changes in the
allophane microstructure during the kinetic experiment are caused by the influence
of sulfate or H
+
or both.
6.4.9 Kinetics of sulfate retention processes and consequences for the sulfate
extractability
The decrease of Al concentrations during the pH 4.5 experiment suggests that an
aluminum containing phase is precipitated. This may be a BAS phase. STA-MS
analysis of samples taken after 24 hours suggest the presence of a basaluminite-
like phase, interpretation of STA-MS for other samples is ambiguous. Hints on the
formation of alunitic phases could not be found in any of the analyzed samples.
XANES analyses showed that the percentage of precipitated sulfate generally is
rather low and although it tends to increase with time, this increase contributes only
to a minor extent (at most 14% of the increase from 2 hours up to 2 months) to the
overall increase in sulfate retention. In accordance with the low relevance of precipi-
tation processes and the absence of more stable alunitic phases, the sulfate extrac-
tability does not change during the total duration of the experiment at pH 4.5.
Whether precipitation processes are also of comparatively little relevance at pH 5.0
cannot be answered unambiguously without the analysis of the solid experimental
products. Nevertheless, silica and aluminum concentrations as well as saturation
indices show similar trends in the pH 5.0 experiment as in the pH 4.5 experiment
suggesting that the same processes proceed. In addition, I found that the relevance
of rapid precipitate formation is lower at pH 5.0 compared with pH 4.5 (chapter 5).
This suggests that precipitate formation is of even less importance at pH 5.0 com-
pared with pH 4.5.
The low relevance of precipitation processes indicates that adsorption processes
dominate the slow increase in sulfate retention with increasing time. Generally,
adsorption processes are rather fast. As explained before, the apparent contradic-
tion that sulfate adsorption increases only slowly during the experiment may be
explained by (i) the slow formation and/or exposure of new sites for sulfate adsorp-
tion, (ii) the slow diffusion of sulfate into intra- and interparticle pores (e.g. Barrow et
Sulfate retention kinetics
131
al., 1993; Fuller et al., 1993; Strauss et al., 1997; Lang and Kaupenjohann, 2003) or
(iii) an increasing aggregation during sulfate retention with concomitant anion burial
in newly forming intra-aggregate pores (Anderson et al., 1985). The analysis of
changes in the allophane structure showed that the percentage of polymeric silica
which is located on the outer side of the allophane spherules decreases and that
additionally the number or size of defect sites increases. Depolymerization of Si with
the formation of monomeric Si on the outer side of allophane spherules most proba-
bly increases the exchangeability of silicate by sulfate. Thus, Si depolymerization
makes existing aluminol sites accessible for sulfate adsorption. Additionally, an
increase in the number of defect sites leads to the formation of new sites for anion
adsorption. Defect sites have been proposed to be preferential sites for ion retention
(Parfitt and Henmi, 1980; Abidin et al., 2007). Thus, the exposure of existing sites
and the formation of new sites for anion adsorption are at least partly responsible for
the slow increase in sulfate retention with time. Consequently, my analyses
strengthen the assumptions made by Imai et al. (1981) who postulated that a desta-
bilization of allophane and the consecutive exposure of reactive surfaces are re-
sponsible for the slow increase in phosphate adsorption on allophane. These find-
ings are also in accordance with assumptions made by Arai et al. (2005) who
studied As retention by allophane by EXAFS analysis. The authors observed an
increase in As retention up to 720 hours and EXAFS spectra suggested that only
adsorbed arsenic species were present in the solid experimental products. To ex-
plain the apparent contradiction between the dominance of fast adsorption proc-
esses and the slow increase in As retention, the authors assumed that new sites for
As adsorption are only slowly exposed due to the desorption of Si.
Diffusion may additionally contribute to the slow sulfate retention as sulfate may
diffuse through interparticle pores into allophane aggregates or through defect sites
into allophane spherules. Nitrogen BET measurements and the calculation of micro-
porosity by the Dubinin-Radushkevic method for the synthetic allophane resulted in
rather large micropore volumes of 190 mm
3
g
-1
. Nevertheless, the large percentage
of sulfate exchangeable by phosphate within one hour from all samples makes both
slow diffusion of sulfate through micropores and the burial of sulfate in newly form-
ing aggregates unlikely. Hints on the formation of larger aggregates could neither be
found by SEM imaging of samples.
Chapter 6
132
6.5 CONCLUSIONS
My results show that the precipitation of BAS phases is of no or comparatively low
relevance for the slow sulfate retention by allophane under the applied experimental
conditions. Additionally, no transformation of rapidly forming basaluminite-like
phases into alunitic phases could be observed within the experimental period of 6
months. Thus, adsorption processes govern the slow sulfate retention by allophane.
A slow increase in sulfate adsorption is caused by two processes: (i) the exposure of
existing aluminol sites due to a slow removal of polymerized silica from the outer
surface of the allophane hollow spherules and (ii) the generation of new sites for
anion retention due to an increase in defect sites. In accordance with the dominance
of adsorption processes, the fraction of sulfate easily extractable by phosphate was
large and did not decrease during the experiment. The latter additionally indicates
that sulfate diffusion into allophane aggregates or sulfate burial in newly forming
aggregates is of no or only negligible relevance for the slow sulfate retention by
allophane. Overall, my results illustrate the importance of evaluating the stability of a
sorbent, especially when long-term retention experiments are conducted.
133
7. EXTENDED SUMMARY, GENERAL CONCLUSIONS AND OUTLOOK
7.1 EXTENDED SUMMARY
Sulfur dioxide deposition close to active volcanoes leads to soil acidification and to
an increase in soil sulfur contents. Both acidification and sulfur retention in soils can
affect microorganisms, plants and animals by influencing the concentration of toxic
aluminum or of available sulfur and other nutrients. For a reliable prediction of po-
tentially adverse effects, the knowledge of acid buffering and sulfur retention proc-
esses in soils is essential.
Our knowledge on acid buffering and S retention processes has mainly been gained
by studying soils of the temperate regions in Europe and North America. Carbonate,
silicate and sesquioxide dissolution as well as the protonation of variable charge and
the release of OH
-
during specific anion adsorption were identified as processes
responsible for acid buffering in soils. Sulfate retention is governed by S incorpora-
tion into organic matter, by sulfate adsorption to variable charge minerals (both
inner-sphere and outer-sphere complexes) and by the precipitation of basic alumi-
num sulfate phases (BAS). The question on the relevance of BAS phases for S
retention has been a matter of extensive and ongoing scientific debate.
Soils and climatic conditions close to active volcanoes, which are mainly located in
the (sub)tropics, differ substantially from those of formerly studied sites of the tem-
perate regions. Hence, existing knowledge on acid buffering and S retention proc-
esses is not necessarily transferable to acid impacted sites surrounding active vol-
canoes. Thus, the main objectives of my thesis were (i) to estimate the
contributions of different processes to the effective acid neutralization capac-
ity and (ii) to determine the relevance of adsorption processes and BAS pre-
cipitation for the total sulfate retention in SO
2
-impacted Andosols. The main
objectives of my thesis were split into six specific objectives which were addressed
by five experimental blocks. The results are summarized in the following:
Chapter 7
134
7.1.1 Which processes are relevant for the effective acid neutralization in SO
2
-
impacted Andosols and which consequences do these processes have for
secondary mineral formation (chapter 2)?
To identify buffering processes, short-term pH
stat
experiments were conducted with
soil samples taken from two sites close to Masaya Volcano and with soil samples
from five sites of a transect of decreasing acid input at Poás Volcano. The occur-
rence under field conditions of the thereby identified buffering processes was tested
by IR and SEM-EDX analysis of the soil samples of the Poás transect. For this site,
plagioclase and glass dissolution as well as protonation of functional groups of
organic matter were identified as dominant processes of acid buffering in the short-
term pH
stat
experiments. The contribution of protonation of organic matter to the
short-term acid neutralization capacity was smaller for samples from heavily acid
affected sites than for samples from less affected sites. This could be attributed to
differences in the quantity and quality of organic matter caused by different acid
deposition rates. Under field conditions only plagioclase and glass dissolution are
relevant for the long-term buffering of acid inputs at Poás. For Masaya samples,
primary mineral, glass and possibly allophane dissolution dominate acid buffering
both during laboratory experiments and under field conditions.
Acid buffering in Andosols affects secondary mineral formation: soil saturation ex-
tracts suggest that released aluminum is precipitated as basaluminite at Masaya,
while it seems to be susceptible to leaching at Poás. At both sites, acid buffering
resulted in the precipitation of amorphous silica.
7.1.2 Do soil hydrological conditions influence acid buffering processes
(chapter 2)?
Different leaching conditions were simulated by two types of pH
stat
experiments in
the laboratory: an open system in which leaching prevails was simulated by the
addition of protonated cation exchange resin as an active and “infinite” sink for
released cations while a closed system with no leaching was simulated by pH
stat
titration under cation accumulation. In the closed system, basic cation and aluminum
accumulation in solution due to rapid ion release resulted in a decrease in mineral
dissolution and thus, in a decrease of the effective acid neutralization capacity of the
soils. In line with literature, these results indicate that hydrologic conditions of soils
influence acid buffering processes.
Extended Summary, Conclusions and Outlook
135
7.1.3 Evaluation of the suitability of STA-MS for the differentiation between
BAS phases, organically bound estersulfates and adsorbed sulfate (chapter 3)
Reference materials of basaluminite, K-alunite, Na-alunite, Na-Dodecylsulfate and
“sulfate adsorbed to allophane” were analyzed by STA-MS in pure form and in
physical mixtures with simple matrix components (allophane, SiO
2
, Al
2
O
3
). The
reference compounds could be distinguished by their characteristic patterns of SO
and H
2
O release during STA-MS analysis when analyzed as pure substances.
However, the addition of simple matrix components led to overlapping STA-MS
signals of different reference substances thus limiting the applicability of STA-MS for
BAS identification to selected and very simple samples of known matrix composi-
tion.
7.1.4 Are BAS phases preferentially formed in situ at the allophane-water
interface (chapter 4)?
Both synthetic allophane and synthetic glass were exposed in soils close to the SO
2
-
emitting volcanoes Poás and Masaya in Central America. After 18 months the mate-
rials were characterized by a combination of wet-chemical methods, S K-edge
XANES, STA-MS, SEM-EDX and IR spectroscopy. In situ formed BAS phases could
only be detected on allophane but not on glass samples thus indicating a preferen-
tial BAS formation at the allophane-water interface. BAS phases could not be visual-
ized by SEM-EDX which may indicate that these phases are rather small and/or
may be formed in intra-aggregate pore spaces.
My experiment additionally showed that C-coatings on allophane surfaces cause a
decrease in the oxalate dissolution kinetics of allophane and in the extractability of
adsorbed sulfate. These results demonstrate that wet-chemical methods may lead
to largely erroneous results when applied for phase quantification in soil samples.
7.1.5 Effects of sulfate concentration and pH on the relative contribution of
adsorption and BAS precipitation to the total sulfate retention by allophane
(chapter 5)
The effects of increasing sulfate concentrations on S retention by synthetic allo-
phane were studied at pH 4.0, pH 4.5 and pH 5.0 by short-term pH
stat
experiments.
Solid phases were characterized by S K-edge XANES, STA-MS, SEM-EDX and by
porosity measurements with gas adsorption (N
2
). Sulfate precipitation and adsorp-
Chapter 7
136
tion are simultaneous processes. Precipitation occurs even at low sulfate loads and
despite solution undersaturation with respect to amorphous basaluminite. Increasing
sulfate concentrations and decreasing pH tend to increase BAS precipitation but the
effects are small and overall, sulfate adsorption dominates sulfate retention. The
formation of BAS does not affect the sulfate extractability but decreases the allo-
phane micro- and mesoporosity.
7.1.6 How do the relative contributions of adsorption and BAS precipitation to
the total sulfate retention by allophane change with time (chapter 6)?
The kinetics of sulfate retention by synthetic allophane was studied at pH 4.5 and
pH 5.0 by pH
stat
experiments lasting up to 6 months. Solid phases were character-
ized by a combination of wet-chemical methods, S K-edge XANES, STA-MS, SEM-
EDX and IR spectroscopy. While sulfate retention at pH 5.0 reached an apparent
equilibrium after two weeks, sulfate retention at pH 4.5 increased up to the end of
the experimental period of 6 months. Overall, precipitation of BAS phases contrib-
uted at most by 8% to the total sulfate retention and adsorption processes domi-
nated the slow increase in sulfate retention. The slow increase in sulfate adsorption
can be attributed to (i) depolymerization of silica on the outer surface of allophane
spherules thus increasing silica desorbability and making existing aluminol sites
available for sulfate adsorption and to (ii) an increase in defect sites of allophane
spherules generating new sites for sulfate adsorption.
7.2 GENERAL DISCUSSION AND CONCLUSIONS
7.2.1 Acid buffering in SO
2
-impacted Andosols of the (sub-)tropics – a com-
parison with soils of heavily acid-affected sites in the temperate regions
In the frame of research on courses of forest dieback in the temperate regions,
mostly such sites have been studied which have a rather low acid neutralization
capacity. For these sites, protonation of variable charge of organic matter and the
dissolution of aluminum hydroxides have been identified as the dominant acid buff-
ering processes (e.g. Matzner, 1989; Nätscher and Schwertmann, 1991; Süsser and
Schwertmann, 1991). Vitric Andosols of the (sub-)tropics are comparatively younger
soils and contain large amounts of weatherable primary minerals. Accordingly,
besides the protonation of variable charge of organic matter the dissolution of
Extended Summary, Conclusions and Outlook
137
plagioclase phenocrysts and glass and possibly the dissolution of allophane
are the main effective short-term acid buffering processes in vitric Andosols
(chapter 2).
For the C-rich vitric Andosols developing towards the alu-andic subtype at the Poás
site the protonation of functional groups of organic matter has an outstanding rele-
vance for the short-term acid neutralization. This can be attributed to large C con-
centrations of soil samples and to a comparatively large content of functional groups
which can be protonated (potential acidity) in the soil organic matter: Poás samples
show a maximum potential acidity of 23 mmol
c
g
-1
C (chapter 2), while the mean
potential acidity of humic acids in soils of the temperate regions is only 13 mmol
c
g
-1
C (Stevenson, 1994). The importance of protonation of organic matter for acid buff-
ering in alu-andic Andosols has also been suggested by Dahlgren and Saigusa
(1994) and Takahashi et al. (1995). However, large H
+
deposition rates at Poás
influence the quantity and quality of organic matter and lead to a decrease in
the total acid neutralization capacity attributable to the protonation of variable
charge (chapter 2). The opposite effect can be observed for acid impacted sites of
the temperate regions where acid deposition mostly led to an increase in the organic
matter content due to a decreased organic matter mineralization. To my knowledge,
no information exists on the effect of acid deposition in the temperate regions on the
potential acidity of soil organic matter.
Although the protonation of variable charge of organic matter is important for the
short-term acid buffering in vitric Andosols of the alu-andic subtype, it is of little or of
no importance for the long-term acid buffering under field conditions as cation
release due to mineral weathering reverses the protonation of functional
groups of organic matter. Consequently, mineral dissolution is the dominant
process for the acid neutralization of vitric Andosols under mean field condi-
tions. This is resembled by an increase in the number and size of weathering chan-
nels in plagioclase phenocrysts with increasing acid load at Poás sites (chapter 2).
For heavily acid-affected sites of the temperate regions the protonation of variable
charge of organic matter cannot be reversed as effectively by basic cation release
due to mineral dissolution as it is the case for vitric Andosols of the (sub-)tropics.
This can be attributed to the fact that heavily acid-affected sites in the temperate
regions are already depleted in the reserve of bases and/or that mineral weathering
is comparatively slow.
Chapter 7
138
No differences in pH could be observed for soil samples taken along the Poás tran-
sect of increasing acid input (chapter 2). Thus, plagioclase and glass dissolution are
effectively buffering the large acid input at these sites. Contrastingly, Delfosse et al.
(2005a) observed a marked decrease in pH along a transect of soil samples taken
close to Masaya Volcano although acid and sulfate deposition rates are smaller for
Masaya than for Poás sites (chapter 4, chapter 2). This difference between Poás
and Masaya samples in the effectiveness of mineral/glass dissolution for acid buffer-
ing may at least partially be explained by the influence of climatic conditions on acid
buffering processes as I could affirm in my lab experiments: Under the compara-
tively dry climatic conditions at Masaya elevated concentrations of basic
cations and/or aluminum in soil solution, which are caused by rapid protona-
tion of variable charge, hamper mineral weathering due to product limitation.
Consequently, the relevance of different acid buffering processes may change with
seasonal changes in hydrological conditions at Masaya while the relevance of acid
buffering processes may vary comparatively little under the more constant climatic
conditions at sites in the perhumid tropics or in the temperate regions. Overall, the
results of my study on acid buffering in SO
2
-impacted Andosols of the (sub-)tropics
affirm that the effectiveness of different buffering processes in these soils cannot be
inferred directly from existing knowledge of acid buffering in soils of the temperate
regions.
7.2.2 Suitability of different methods for BAS identification
When starting with the present study, thermodynamical equilibrium calculation,
selective extraction procedures and electron microscopy were the typically applied
methods for BAS identification in soils (e.g. Van Breemen, 1973; Nordstrom, 1982;
Prietzel, 1992; Delfosse et al., 2005b, 2005c). During my experiments I evaluated
STA-MS and contributed to the development of XANES (Prietzel et al., 2008) to
detect BAS phases. I applied a combination of all before mentioned techniques for
BAS identification in my field and laboratory experiments. The comparison of the
results obtained by the different methods allows an evaluation of both potentials and
limitations of the actually available methods for BAS identification and this evalua-
tion is presented in the following:
(i) Thermodynamical equilibrium calculations may give hints on potentially pre-
sent phases. Nevertheless, kinetic restrictions of phase formation, surface
precipitation processes, the formation of solid solutions and the influence of
Extended Summary, Conclusions and Outlook
139
the crystallinity degree on equilibrium constants limit the interpretability of
thermodynamical equilibrium calculations (chapter 5 and 6).
(ii) Selective extraction procedures represent a cheap and easy approach for
BAS identification but they are not selective sensus strictus (Prietzel, 1992). A
rather large sulfate extractability from BAS phases by a solution of KH
2
PO
4
,
which has often been applied for the determination of adsorbed sulfate, dem-
onstrates this (chapter 4, 5 and 6). Results of my experiments additionally
show that organic coatings can decrease the extractability of adsorbed sulfate
substantially thus leading to an overestimation of BAS phases (chapter 4).
Similarly, it may be hypothesized that organic coatings on BAS phases or – in
a more general sense - the physical accessibility of BAS phases affect BAS
extractability thus limiting the suitability of selective extraction procedures se-
verely.
(iii) Delfosse et al. (2005c) suggested electron microscopy as a potential tool for
BAS identification in soils. Nevertheless, sample preparation for electron mi-
croscopy, very small-sized BAS precipitates or the formation of BAS phases in
intra-aggregate void spaces limit the suitability of this method (chapters 4, 5
and 6).
(iv) STA-MS has some potential for BAS identification in simple systems of known
matrix composition but signal distortion by rather simple matrix components
makes this method unsuitable for BAS identification in complex samples like
soils (chapter 3).
(v) From all available methods for the differentiation between adsorbed sulfate
and sulfate precipitated in BAS, XANES is the most promising one. Neverthe-
less, the presence of estersulfates and spatial heterogeneity of samples limit
the applicability of XANES for BAS identification (Prietzel et al., 2008; chapters
4 and 5). Besides, no separate reference materials for inner-sphere and outer-
sphere adsorption complexes are available. Thus, based on XANES meas-
urements solely, it is unclear whether potential differences in spectra of inner-
sphere and outer-sphere complexes have to be taken into account when ap-
plying XANES for the (semi-)quantification of adsorption complexes and BAS
phases. However, the rather good match between STA-MS and XANES re-
sults shown in chapter 5 suggests that potential differences between XANES
spectra of inner-sphere and outer-sphere adsorption complexes do not influ-
ence the quantification of adsorbed and precipitated sulfate substantially.
Chapter 7
140
Overall, I summarize that XANES is the most promising method for BAS iden-
tification and (semi-)quantification in soils which is only limited by the pres-
ence of estersulfate compounds or by the spatial heterogeneity of samples
with respect to sulfate species. The latter demands a large number of replicate
measurements per sample to obtain results which are representative for the whole
soil. Unfortunately, the number of replicate measurements is largely limited by the
scarcity of available beamline measurement time.
7.2.3 Contribution of adsorption processes and BAS precipitation to the total
sulfate retention in SO
2
-impacted Andosols
In the experiments with synthetic allophane, BAS phases contributed 11% (Masaya
site, chapter 4), 19% (Poás site, chapter 4), 5 – 14% (chapter 5) and 2 – 8% (chap-
ter 6) to the total sulfur retention. For the Poás site, alunitic aerosol particles contrib-
ute to the total sulfur content of field exposed allophane, thus the contribution of in
situ formed BAS phases may be less than 19%. Non-adsorbing phases like glass do
not contribute in measurable quantities to the total sulfur retention (chapter 4). The
role of iron and aluminum sesquioxides for BAS precipitation has not been investi-
gated so far. Nevertheless, as sesquioxide contents in sil-andic Andosols are gen-
erally smaller than allophane contents, the results gained on allophane allow a
rough estimation of the role of BAS precipitation in these soils.
Delfosse et al. (2005c) estimated the contribution of BAS phases to the total S
retention in vitric and eutric Andosols of the Masaya region to be 10 – 50%. The
authors based their estimation on selective extraction data (Delfosse et al., 2005a,
Table 3; Delfosse et al., 2005b, Table 1). These data indicate that the larger per-
centages of BAS phases can be especially found in eutric Andosols. However,
Delfosse et al. (2005c) could not identify BAS in eutric Andosols by TEM-EDX.
Further, XANES-measurements of Prietzel et al. (2008) resulted in only low values
of BAS phases in these soils (2 – 7%). Thus, to my opinion, the large percentages of
BAS phases reported by Delfosse et al. (2005c) can at least partly be attributed to
methodological limitations of the selective extraction procedures.
According to XANES measurements of Prietzel et al. (2008), the contribution of BAS
phases to the total S retention in vitric Andosols was 5, 11 and 60% for three differ-
ent measurement points of one soil sample taken close to Masaya Volcano. Two of
the three XANES results are similar to the percentages of precipitated sulfate as
Extended Summary, Conclusions and Outlook
141
determined in the present study. The mean contribution of BAS phases to the total
sulfate retention in Andosols cannot be calculated based on the three XANES
measurements as they indicate a large spatial variability. Based on my measure-
ments, I summarize that in situ BAS precipitation contributes approximately
with 10 – 20% to the total inorganic sulfur retention in SO
2
-impacted vitric
Andosols of the sil-andic subtype. The contribution of BAS phases to the total
sulfate retention seems to be mainly triggered by the solution pH and to a minor
extent by the sulfate concentration or the reaction time (chapters 5 and 6).
In vitric Andosols of the alu-andic subtype, traces of allophane may contribute to
inorganic sulfate retention by adsorption and precipitation processes. The same may
hold true for aluminum and iron sesquioxides and the influence of these phases on
sulfate adsorption and BAS precipitation deserves further investigation. However,
according to studies of C-rich horizons of SO
2
-impacted soils of the temperate re-
gions (Erkenberg et al., 1996; Prietzel et al., 2004), sulfur incorporation into or-
ganic matter is very probably the most relevant process for the total S reten-
tion in vitric Andosols of the alu-andic subtype.
7.2.4 Ecological implications of BAS formation in SO
2
-impacted Andosols
As observed in my field and laboratory experiments, the precipitation of BAS phases
can increase the sulfate retention capacity of sil-andic Andosols by 10 – 20% thus
potentially decreasing the sulfate and concomitant cation leaching to groundwater. A
partially low sulfate extractability from synthetic BAS reference materials additionally
suggests that BAS precipitation may influence the sulfate availability in soils (chap-
ter 4) as it has been discussed for soils of the temperate regions (Courchesne and
Hendershot, 1990). Such an effect of BAS precipitation on the sulfate extractability
could not be detected in my laboratory experiments on sulfate retention by allo-
phane (chapter 5 and 6). However, BAS phases which formed under natural condi-
tions may differ in size, composition and crystallinity from those precipitated in labo-
ratory experiments and the mentioned characteristics can influence the sulfate
extractability from BAS phases severely. In the experiment I described in chapter 4 I
could observe that sulfate could not completely be extracted from field-exposed
allophane and this may indicate that BAS phases may limit the sulfate extractability
in these samples. The sulfate extractability of field-exposed allophane samples
ranged between 17 and 82% although BAS phases contributed with maximal 19% to
the total sulfate retention. This clearly indicates that other factors like e.g. carbon-
Chapter 7
142
rich coatings on allophane surfaces contribute at least partially to the low sulfate
extractability. As the effects of BAS precipitates and C-coatings on the sulfate ex-
tractability cannot be separated from each other by the present study, a final conclu-
sion on the effect of in situ formed BAS precipitation on the sulfate availability is not
possible.
Generally, the precipitation of BAS phases may decrease aluminum concentrations
in soil solution thus potentially lowering the risk of aluminum toxicity. However,
aluminum concentrations in saturation extracts of soil samples taken close to Ma-
saya Volcano reached values up to 38 mg L
-1
(chapter 2), which can be toxic to
plant roots. This indicates that BAS precipitation cannot impede aluminum toxicity
effectively. Contrasting to the rather low ecological impact of BAS phases consider-
ing aluminum toxicity, the precipitation of such phases can affect the micro- and
mesoporosity of allophane thus potentially influencing the availability of plant nutri-
ents which are retained in the porous minerals (chapter 5).
7.3. OUTLOOK
For vitric Andosols I could show that protonation of variable charge and mineral /
glass dissolution are the main short-term and long-term acid buffering mechanisms,
respectively. In other Andosol subtypes as e.g. eutric Andosols other processes may
play an important role for acid buffering and the contribution of each process to the
overall acid buffering deserves further investigation. This is especially important in
the light of increasing anthropogenic SO
2
and NO
x
emissions in rapidly developing
countries like the Phillipines and Indonesia (Rodhe, 1989; Galloway, 1989) where
volcanic soils of varying subtypes are omnipresent. Similar experiments and the
analysis of transects of increasing acid input as described in my thesis may be
conducted with samples of other Andosol subtypes.
Sulfur K-edge XANES is the most promising method for BAS identification in soils.
Nevertheless, a method for an unambiguous differentiation between BAS phases
and estersulfate compounds is still lacking. Nuclear magnetic resonance spectros-
copy (NMR) may potentially allow such a differentiation. To test the method for its
suitability, model compounds containing
33
S should be analyzed by
33
S-NMR.
Thereby, also the detection limit of the method should be identified to test whether
this method could be valuable for the analysis of natural samples or if it may be
limited to labeled systems which are enriched with
33
S.
Extended Summary, Conclusions and Outlook
143
In my thesis the role of allophane and glass for sulfate retention in Andosols has
been investigated, the effect of aluminum and iron sesquioxides on adsorption and
precipitation processes still needs further investigation. Therefore, similar field and
laboratory experiments as presented here may be conducted with ferrihydrite or
goethite. For allophane I could show that S retention is dominantly governed by
adsorption processes and the mechanisms underlying the slow sulfate adsorption
could be identified. Contrastingly, the mechanism governing BAS formation remains
unclear. Further studies should investigate whether sulfate adsorption is a prerequi-
site for BAS formation at the allophane-water interface or if these phases form due
to solution supersaturation in meso- and microporous void spaces. Products of the
two processes may be distinguished from each other applying methods which give
information on the atomic neighborhood of S atoms in samples. Nevertheless, to my
knowledge no such method is available. A second possibility to probe whether the
presence of adsorption sites or the presence of pore spaces decides over BAS
formation could be conducting sulfate retention experiments with microporous but
non-adsorbing materials. Such materials could be e.g. glasses with controlled pore
spaces as they can be purchased from various suppliers of laboratory chemicals.
The knowledge of the process underlying BAS formation in Andosols is important as
it may allow predicting the effects of other factors than pH, sulfate concentration and
time on BAS formation without the necessity of conducting a large number of ex-
periments. In this context, the influence of fluoride, phosphate and dissolved organic
carbon as well as a combination of various of these factors on BAS precipitation
should be investigated as these ions compete with sulfate both for adsorption sites
and for dissolved aluminum and thus, potentially influence the quantity and the
composition of precipitates. Besides, the aggregation degree of allophane and thus,
the amount of readily available sites for sulfate adsorption may influence BAS for-
mation and should be studied in further detail as allophane tends to form strong
aggregates under natural conditions. Finally, the question whether BAS precipitation
decreases the sulfate availability in Andosols or if sulfate is readily released from in
situ formed BAS phases has to be answered for an overall evaluation of the ecologi-
cal implications of BAS precipitation in volcanic soils.
144
145
8 REFERENCES
Abidin Z., Matsue N. & Henmi T. 2007. Nanometer-scale chemical modifications of nano-ball
allophane. Clays and Clay Minerals 55, 443-449.
Adams F. & Rawajfih Z. 1977. Basaluminite and alunite: a possible cause of sulfate retention
by acid soils. Soil Science Society of America Journal 41, 686-692.
Adriano D.C. & Doner H.E. 1982. Bromine, chlorine and fluorine. In: Methods of Soil Analysis
Part 2 - Chemical and Microbiological Properties (eds. A.L Page, R.H. Miller & D.R.
Keeney), pp. 449-483. American Society of Agronomy & Soil Science Society of
America, Madison, WI.
Alewell C., Manderscheid B., Lükewille A., Koeppe P. & Prenzel J. 1995. Describing soil
SO
42-
dynamics in the Solling roof project with two different modelling approaches.
Water, Air and Soil Pollution 85, 1801-1806.
Allison J.D., Brown D.S. & Novo-Gradac K.J. 1991. MINTEQA2/PRODEFA2, A geochemical
assessment model for environmental systems. Version 3.0. User´s Manual. United
States Environmental Protection Agency, Office of Research and Development,
Washington, D.C..
Alvarado Induni G.E., 2005. Costa Rica: Land of Volcanoes. Editorial Universidad Estatal a
Distancia, San José, Costa Rica.
Amrhein C. & Suarez D.L. 1992. Some factors affecting the dissolution kinetics of anorthite
at 25°C. Geochimica et Cosmochimica Acta 56, 1815-1826.
Anderson M.A., Tejedor-Tejedor M.I. & Stanforth R.R. 1985. Influence of aggregation on the
uptake kinetics of phosphate by goethite. Environmental Science and Technology
19, 632-637.
Andres R.J., Barquero J. & Rose W.L. 1992. New measurements of SO
2
flux at Poás Vol-
cano, Costa Rica. Journal of Volcanology and Geothermal Research 49, 175-177.
Arai Y., Elzinga E.J. & Sparks D.L. 2001. X-ray absorption spectroscopic investigation of
arsenite and arsenate adsorption at the aluminum oxide-water interface. Journal of
Colloid and Interface Science 235, 80-88.
Arai Y., Sparks D.L. & Davis J.A. 2005. Arsenate adsorption mechanisms at the allophane-
water interface. Environmental Science and Technology 39, 2537-2544.
Baba M. & Okazaki M. 2000. Changes in aluminum pools of Andisols due to soil acidifica-
tion. Soil Science and Plant Nutrition 46, 797-805.
Chapter 8
146
Baba M., Okazaki M. & Hashitani T. 1995. Effect of acid deposition on forested Andisols on
the Tama hill region of Japan. Environmental Pollution 89, 97-106.
Barrett E.P., Joyner L.G. & Halenda P.P. 1951. The determination of pore volume and area
distribution in porous substances. I. Computations from nitrogen isotherms. Journal
of the American Chemical Society 73, 373-380.
Barrett R., Kaulich B., Salomé M. & Susini J. 2000. Current status of the scanning X-ray
microscope at the ESRF. In X-ray Microscopy: 6th International Conference, August
2-6, 1999. Berkely, CA (eds. W.E. Meyer-Ilse, T. Warwick and D. Attwood), AIP Con-
ference Reports 507, pp. 458-463. American Institute of Physics, Melville, NY.
Barrow N.J., Brümmer G.W. & Strauss R. 1993. Effects of surface heterogeneity on ion
adsorption by metal oxides and by soils. Langmuir 9, 2606-2611.
Bartoli F., Poulenard A.J. & Schouller B.E. 2007. Influence of allophane and organic matter
contents on surface properties of Andosols. European Journal of Soil Science 58,
450-464.
Beck W. & Engelhardt H. 1992. Capillary electrophoresis of organic and inorganic cations
with indirect UV detection. Chromatographia 33, 313-316.
Berggren D., Mulder J. & Westerhof R. 1998. Prolonged leaching of mineral forest soils with
dilute HCl solutions: the solubility of Al and soil organic matter. European Journal of
Soil Science 49, 305-316.
Bigham J.M. & Nordstrom D.K. 2000. Iron and aluminum hydroxysulfates from acid sulfate
waters. In: Sulfate Minerals – Crystallography, geochemistry and environmental sig-
nificance (eds. Alpers C.N., Jambor J.L. and Nordstrom D.K.). Reviews in Mineral-
ogy and Geochemistry 40. Mineralogical Society of America.
Bischoff W.-A. 2007. Entwicklung und Anwendung der Selbst-Integrierenden Akkumulatoren:
Eine Methode zur Erfassung der Sickerfrachten umweltrelevanter Stoffe. PhD thesis.
TU Berlin, Germany. 172pp.
Bruggenwert M.G.M., Hiemstra T. & Bolt G.H. 1991. Proton sinks in soil controlling soil
acidification. In: Soil Acidity (eds. B. Ulrich & M.E. Sumner). Springer Verlag, Berlin
Heidelberg.
Casadevall T.J., Rose Jr. W.I., Fuller W.H., Hunt W.H., Hart M.A., Moyers J.L., Woods D.C.,
Chuan R.L. & Friend J.P. 1984. Sulfur dioxide and particles in quiescent volcanic
plumes from Poás, Arenal and Colima Volcanoes, Costa Rica and Mexico. Journal
of Geophysical Research 89, 9633-9641.
Celi L., Barberis E. & Marson F.A. 2000. Sorption of phosphate on goethite at high concen-
trations. Soil Science 165, 657-664.
References
147
Cipriani P., Marruzzo G., Piga L. & Pochetti F. 1997. Thermal behaviour of mixtures of an
alunite ore with K
2
CO
3
, CaCO
3
and Ca(OH)
2
. Thermochimica Acta 294, 139-146.
Courchesne F. & Hendershot W.H. 1990. The role of basic aluminum sulfate minerals in
controlling sulfate retention in the mineral horizons of two Spodosols. Soil Science
150, 571-577.
Courchesne F., Turmel M.-C. & Beauchemin P. 1996. Magnesium and potassium release by
weathering in Spodosols: grain surface coating effects. Soil Science Society of
America Journal 60, 1188-1196.
Dahlgren R.A. 1994. Quantification of allophane and imogolite. In: Quantitative Methods in
Soil Mineralogy (eds. J.E. Amonette & L.W. Zelazny), pp. 430-451. Soil Science So-
ciety of America, Madison, WI.
Dahlgren R.A. & Saigusa M. 1994. Aluminum release rates from allophanic and nonallo-
phanic Andosols. Soil Science and Plant Nutrition 40, 125-136.
Delfosse T., Delmelle P., Iserentant A. & Delvaux B. 2005a. Contribution of SO
3
to the acid
neutralizing capacity of Andosols exposed to strong volcanogenic acid and SO
2
deposition. European Journal of Soil Science 56, 113-125.
Delfosse T., Delmelle P. Givron C. & Delvaux B. 2005b. Inorganic sulphate extraction from
SO
2
-impacted Andosols. European Journal of Soil Science 56, 127-133.
Delfosse T., Elsass F. & Delvaux B. 2005c. Direct evidence of basic aluminium sulphate
minerals in an S-impacted Andosol. European Journal of Soil Science 56, 281-286.
Delfosse T., Delmelle P. & Delvaux B. 2006. Sulphate sorption at high equilibrium concentra-
tion in Andosols. Geoderma 136, 716-722.
Delmelle P., Stix J., Bourque C.P.A., Baxter P.J., García-Alvarez J. & Barquero J. 2001. Dry
deposition and heavy acid loading in the vicinity of Masaya Volcano, a major sulfur
and chlorine source in Nicaragua. Environmental Science and Technology 35, 1289-
1293.
Delmelle P., Stix J., Baxter P.J., García-Alvarez J. & Barquero J. 2002. Atmospheric disper-
sion, environmental effects and potential health hazard associated with the low-
altitude gas plume of Masaya Volcano, Nicaragua. Bulletin of Volcanology 64, 423-
434.
Delmelle P., Delfosse T. & Delvaux B. 2003. Sulfate, chloride and fluoride retention in An-
dosols exposed to volcanic acid emissions. Environmental Pollution 126, 445-457.
Dorodnikov M., Fangmeier A. & Kuzyakov Y. 2007. Thermal stability of soil organic matter
pools and their δ
13
C values after C
3
-C
4
vegetation change. Soil Biology and Bio-
chemistry 39, 1173-1180.
Chapter 8
148
Emmerich K. & Smykatz-Kloss W. 2002. Exothermic effects in soils during thermal analysis.
Clay Minerals 37, 575-582.
Erkenberg A., Prietzel J. & Rehfuess K.E. 1996. Sulfur status of selected European forest
soils as dependent on the atmospheric S-deposition. Zeitschrift für Pflanzen-
ernährung und Bodenkunde 159, 101-109.
Eusterhues K., Wagner F.E., Häusler W., Hanzlik M., Knicker H., Totsche K.U., Kögel-
Knabner I. & Schwertmann U. 2008. Characterization of ferrihydrite-organic matter
coprecipitates by X-ray diffraction and Moessbauer Spectroscopy. Environmental
Science and Technology 42, 7891-7897.
Evans A. Jr. 1991. The interaction of aliphatic acids with basic aluminum sulfates in a for-
ested Ultisol. Soil Science 152, 53-60.
Falkengren-Grerup U. 1989. Soil acidification and its impact on ground vegetation. Ambio
18, 179-183.
Farley K.J., Dzombak D.A. & Morel F.M.M. 1985. A surface precipitation model for the sorp-
tion of cations on metal oxides. Journal of Colloid and Interface Science 106, 226-
242.
Farmer V.C. 1990. The solubility characteristics of organic Al precipitates in allophanic Bs
horizons could mimic those of Al hydroxides, and indicate the solubility of associated
allophanes. European Journal of Soil Science 50, 713-715.
Fendorf S.E. & Sparks D.L. 1996. X-ray absorption fine structure spectroscopy. In: Methods
of soil analysis Part 3: Chemical Methods (ed. D.L. Sparks). SSSA Book Series 5,
1390pp.
Fendorf S.C., Sparks D.L., Lamble G.M. & Kelley M.J. 1994. Applications of X-ray absorption
fine structure spectroscopy to soils. Soil Science Society of America Journal 58,
1583-1595.
Ford R.G., Scheinost A.C., Scheckel K.G. & Sparks D.L. 1999. The link between clay min-
eral weathering and the stabilization of Ni surface precipitates. Environmental Sci-
ence and Technology 33, 3140-3144.
Ford R.G., Scheinost A. & Sparks D.L. 2001. Frontiers in metal sorption/precipitation
mechanisms on soil mineral surfaces. Advances in Agronomy 74, 41-62.
Friedrich A., Grunewald K., Klinnert S. & Bechmann W. 1996. Thermogravimetric and differ-
ential thermal analytical investigations on sewage farm soils. Journal of Thermal
Analysis 46, 1589-1597.
References
149
Frost R.L., Wain D.L., Wills R.-A., Musemeci A. & Martens W. 2006. A thermogravimetric
study of alunites of sodium, potassium and ammonium. Thermochimica Acta 443,
56-61.
Fuller C.C., Davis J.A. & Waychunas G.A. 1993. Surface chemistry of ferrihydrite: Part 2.
Kinetics of arsenate adsorption and coprecipitation. Geochimica et Cosmochimica
Acta 57, 2270-2282.
Galloway J.N. 1989. Atmospheric Acidification: projections for the future. Ambio 18, 161-166.
Gebhardt H. & Coleman N.T. 1974. Anion adsorption by allophanic tropical soils: II. Sulfate
adsorption. Soil Science Society of America Proceedings 38, 259-262.
Gislason S.R. & Eugster H.P. 1987. Meteoric water-basalt interactions. I. A laboratory study.
Geochimica et Cosmochimica Acta 51, 2827-2840.
Gregg S.J. & Sing K.S.W. 1982. Adsorption, surface area and porosity, second ed. Aca-
demic Press, New York.
Gustafsson J.P., Lumsdon D.G. & Simonsson M. 1998. Aluminium solubility characteristics
of spodic B horizons containing imogolite-type materials. Clay Minerals 33, 77-86.
Hamilton J.P., Pantano C.G. & Brantley S.L. 2000. Dissolution of albite glass and crystal.
Geochimica et Cosmochimica Acta 64, 2603-2615.
Harsh J., Chorover J. & Nizeyimana E. 2002. Allophane and Imogolite. In: Soil Mineralogy
with environmental applications (eds. J.B. Dixon and D.G. Schulze). Soil Science
Society of America Book Series 7, Madison. pp. 291-322.
Hayashi K. & Okazaki M. 2002. Effect of volcanic fumes from Mt. Oyama, Miyakejima Island,
on atmospheric deposition, soil solution and soil properties in Kumagaya, Central
Japan. I. Chemical properties of atmospheric desosition and soil solution. Soil Sci-
ence and Plant Nutrition 48, 401-411.
Hayashi K. & Okazaki M. 2003. Effect of volcanic fumes from Mt. Oyama, Miyakejima Island,
on atmospheric deposition, soil solution and soil properties in Kumagaya, Central
Japan. II. Soil properties. Soil Science and Plant Nutrition 49, 503-511.
Herre A., Lang F., Siebe Ch., Dohrmann R. & Kaupenjohann M. 2007. Mechanisms of acid
buffering and formation of secondary minerals in vitric Andosols. European Journal
of Soil Science 58, 431-444.
Hesterberg D., Zhou W., Hutchison J., Beauchemin S. & Sayers D.E. 1999. XAFS study of
adsorbed and mineral forms of phosphate. Journal of Synchrotron Radiation 6, 636-
638.
Chapter 8
150
Hochstein M.P. & Browne P.R.L. 2000. Surface manifestations of geothermal systems with
volcanic heat sources. In: Encyclopedia of Volcanoes (eds. H. Sigurdsson, B.F.
Houghton, S.R. McNutt, H. Rymer & J. Stix), pp. 835-855. Academic Press, San
Diego, CA.
Imai H., Goulding K.W.T. & Talibudeen O. 1981. Phosphate adsorption in allophanic soils.
Journal of Soil Science 32, 555-570.
Ishiguro M., Makino T. & Hattori Y. 2006. Sulfate adsorption and surface precipitation on a
volcanic ash soil (allophanic andisol). Journal of Colloid and Interface Science 300,
504-510.
Jara A.A., Violante A., Pigna M. & Mora M. 2006. Mutual interactions of sulfate, oxalate,
citrate and phophate on synthetic and natural allophanes. Soil Science Society of
America Journal 70, 337-346.
Johnson D.W. & Cole D.W. 1977. Sulfate mobility in an outwash soil in Western Washington.
Water, Air and Soil Pollution 7, 489-495.
Johnson D.W. & Cole D.W. 1980. Anion mobility in soils: relevance to transports from forest
ecosystems. Environment International 3, 79-90.
Johnson N. & Parnell Jr. R.A. 1986. Composition, distribution and neutralization of “acid rain”
derived from Masaya volcano, Nicaragua. Tellus 38B, 106-117.
Johnson D.W., Henderson G.S., Huff D.D., Lindberg S.E., Richter D.D., Shriner D.S., Todd
D.E. & Turner J. 1982. Cycling of organic and inorganic sulphur in a chestnut oak
forest. Oecologia 54, 141-148.
Jongmans A.G., Mulder J., Groenesteijn K. & Buurman P. 1996. Soil surface coatings at
Costa Rican recently active volcanoes. Soil Science Society of America Journal 60,
1871-1880.
Kaiser K. & Guggenberger G. 2003. Mineral surfaces and soil organic matter. European
Journal of Soil Science 54, 219-236.
Kaiser K. & Kaupenjohann M. 1998. Influence of the soil solution composition on retention
and release of sulfate in acid forest soils. Water, Air and Soil Pollution 101, 363-376.
Kaiser K., Kaupenjohann M. & Zech W. 2001. Sorption of dissolved organic carbon in soils:
effects of soil sample storage, soil to solution ratio and temperature. Geoderma 99,
317-328.
Kaupenjohann M. & Wilcke W. 1995. Heavy metal release from a serpentine soil using a pH-
stat technique. Soil Science Society of America Journal 59, 1027-1031.
References
151
Khanna P.K., Prenzel J., Meiwes K.J., Ulrich B. & Matzner E. 1987. Dynamics of sulfate
retention by acid forest soils in an acidic deposition environment. Soil Science Soci-
ety of America Journal 51, 446-452.
Khare N., Hesterberg D. & Martin J.D. 2005. XANES investigation of phosphate sorption in
single and binary systems of iron and aluminum oxide minerals. Environmental Sci-
ence and Technology 39, 2152-2160.
Kopacek J., Vesely J. & Stuchlik E. 2001. Sulphur and nitrogen fluxes and budgets in the
Bohemian Forest and Tatra Mountains during the Industrial Revolution (1850-2000).
Hydrology and Earth System Sciences 5, 391-405.
Lang F. & Kaupenjohann M. 2003. Immobilisation of molybdate by iron oxides: effects of
organic coatings. Geoderma 113, 31-46.
Langier-Kuzniarowa A. 2002. Thermal Analysis of Organo-Clay Complexes. In. Organo-Clay
Complexes and Interactions (eds. S. Yariv and H. Cross). Marcel Dekker, New York.
688pp.
Leinweber P. & Schulten H.-R. 1992. Differential thermal analysis, thermogravimetry and in-
source pyrolysis-mass spectrometry studies on the formation of soil organic matter.
Thermochimica Acta 200, 151-167.
Liu F., He J., Li X., Xu F. & He H. 1995. Chemical states of phosphorus adsorbed on goe-
thite surfaces at various phosphate concentrations. Chinese Science Bulletin 40,
506-511.
Lopez-Capel E., Sohi S.P., Gaunt J.L. & Manning D.A. 2005. Use of thermogravimetry -
differential scanning calorimetry to characterize modelable soil organic matter frac-
tions. Soil Science Society of America Journal 69, 136-140.
Lükewille A., Malessa V. & Alewell C. 1995. Measured and modelled retention of inorganic
sulfur in soils and subsoils (Harz Mountains, Germany). Water, Air and Soil Pollution
85, 683-688.
Martínez M., Fernández E., Valdés J., Barboza V., Van der Laat R., Duarte E., Malavassi E.,
Sandoval L., Barquero J. & Marino T. 2000. Chemical evolution and volcanic activity
of the active crater lake of Poás Volcano, Costa Rica, 1993-97. Journal of Volcanol-
ogy and Geothermal Research 97, 127-141.
Matzner E. 1989. Acidic precipitation: case study Solling. In: Acidic precipitation Volume 1:
Case studies (eds.: D.C. Adriano and M. Havas). Springer Verlag, New York, USA.
Mayer B., Prietzel J. & Krouse H.R. 2001. The influence of sulfur deposition rates on sulfate
retention patterns and mechanisms in aerated forest soils. Applied Geochemistry 16,
1003-1019.
Chapter 8
152
McBride M.B. 2000. Chemisorption and precipitation reactions. In Handbook of Soil Science
(ed. M.E. Sumner), CRC Press, Boca Raton, USA. Pp. B-265-B302.
Meijer E.L., Buurman P., Fraser A. & García-Rodeja E. 2007. Extractability and FTIR charac-
teristics of poorly-ordered minerals in a collection of volcanic ash soils. In. Soils of
Volcanic Regions in Europe (eds. Arnalds O., Bartoli F., Buurman P., Oskarsson H.,
Stoops G. and García-Rodeja E.). Springer Verlag, Berlin, Heidelberg. 644pp.
Mikutta C., Lang F. & Neumann G. 2006a. Phosphate desorption from goethite in the pres-
ence of galacturonate, polygalacturonate and maize mucigel (Zea mais L.). Soil Sci-
ence Society of America Journal 70, 1731-1740.
Mikutta C., Lang F. & Kaupenjohann M. 2006b. Citrate impairs the micropore diffusion of
phosphate into pure and C-coated goethite. Geochimica et Cosmochimica Acta 70,
595-607.
Mikutta C., Mikutta R., Bonneville S., Wagner F., Vögelin A., Christl I. & Kretzschmar R.
2008. Synthetic coprecipitates of exopolysaccharides and ferrihydrite. Part I: Charac-
terization. Geochimica et Cosmochimica Acta 72, 1111-1127.
Muir I.J. & Nesbitt H.W. 1997. Reactions of aqueous anions and cations at the labradorite-
water interface: coupled effects of surface processes and diffusion. Geochimica et
Cosmochimica Acta 61, 265-274.
Mulder J., van Breemen N. & Eijck H.C. 1989. Depletion of soil aluminium by acid deposition
and implications for acid neutralization. Nature 337, 247-249.
Nätscher L. & Schwertmann U. 1991. Proton buffering in organic horizons of acid forest
soils. Geoderma 48, 93-106.
Nordstrom D.K. 1982. The effect of sulfate on aluminum concentrations in natural waters:
some stability relations in the system Al
2
O
3
-SO
3
-H
2
O at 298 K. Geochimica et Cos-
mochimica Acta 46, 681-692.
Nordstrom D.K., Plummer L.N., Langmuir D., Busenberg E., May H.M., Jones B.F. & Park-
hurst D.L. 1990. Revised chemical-equilbrium data for major water-mineral reactions
and their limitations. ACS Symposium Series 416, 398-413.
Ochs M. 1996. Influence of humified and non-humified natural organic compounds on min-
eral dissolution. Chemical Geology 132, 119-124.
Oelkers E.H., Schott J. & Devidal J.-L. 1994. The effect of aluminium, pH, and chemical
affinity on the rates of aluminosilicate dissolution reactions. Geochimica et Cosmo-
chimica Acta 58, 2011-2024.
References
153
Ohashi F., Wada S.-I., Suzuki M., Maeda M. & Tomura S. 2002. Synthetic allophane from
high-concentration solutions: nanoengineering of the porous solid. Clay Minerals 37,
451-456.
Paces T. 1985. Sources of acidification in Central Europe estimated from elemental budgets
in small basins. Nature 315, 31-36.
Padilla G.N., Matsue N. & Henmi T. 2002. Adsorption of sulfate and nitrate on nano-ball
allophane. Clay Science 11, 575-584.
Pardo M.T. & Guadalix M.E. 1990. Phosphate sorption in allophanic soils and release of
sulphate, silicate and hydroxyl. Journal of Soil Science 41, 607-612.
Parfitt R.L. 1989. Phosphate reactions with natural allophane, ferrihydrite and goethite.
Journal of Soil Science 40, 359-369.
Parfitt R.L. & Henmi T. 1980. Structure of some allophanes from New Zealand. Clays and
Clay Minerals 28, 285-294.
Parnell R.A. Jr. 1986. Processes of soil acidification in tropical Durandepts, Nicaragua. Soil
Science 142, 43-55.
Parnell R.A. Jr. & Burke K.J. 1990. Impacts of acid emissions from Nevado del Ruiz volcano,
Colombia, on selected terrestrial and aquatic ecosystems. Journal of Volcanology
and Geothermal Research 42, 69-88.
Paterson E. 1977. Specific surface area and pore structure of allophanic soil clays. Clay
Minerals 12, 1-9.
Peak D., Ford R.G. & Sparks D.L. 1999. An in situ ATR-FTIR investigation of sulfate bonding
mechanisms on goethite. Journal of Colloid and Interface Science 218, 289-299.
Pigna M. & Violante A. 2003. Adsorption of sulfate and phosphate on Andisols. Communica-
tions in Soil Science and Plant Analysis 34, 2099-2113.
Plante A.F., Pernes M. & Chenu C. 2005. Changes in clay-associated organic matter quality
in a C depletion sequence as measured by differential thermal analysis. Geoderma
129, 186-199.
Poetsch T. 2004. Forms and dynamics of silica gel in a tuff-dominated soil complex: Results
of micromorphological studies in the Central Highlands of Mexico. Revista Mexicana
de Ciencias Geológicas 21, 195-201.
Prietzel J. 1992. Auswirkungen definierter Schwefel-Belastung auf die chemischen Eigen-
schaften von Waldböden. Dissertation. Ludwig-Maximilians-Universität München.
Prietzel J. & Feger K.-H. 1991. Al-Spezies im Sickerwasser saurer Waldböden – Einfluß von
Wasserbewegung und Löslichkeitsgleichgewichten. Zeitschrift für Pflanzen-
ernährung und Bodenkunde 154, 271-281.
Chapter 8
154
Prietzel J. & Hirsch C. 1998. Extractability and dissolution kinetics of pure and soil-added
synthesized aluminium hydroxy sulphate minerals. European Journal of Soil Science
49, 669-681.
Prietzel J., Weick C., Korintenberg J., Seybold G., Thumerer T. & Treml B. 2001. Effects of
repeated (NH
4
)
2
SO
4
application on sulfur pools in soil, soil microbial biomass, and
ground vegetation of two watersheds in the Black Forest/Germany. Plant and Soil
230, 287-305.
Prietzel J., Thieme J., Neuhäusler U., Susini J. & Kögel-Knabner I. 2003. Speciation of
sulphur in soils and soil particles by X-ray spectromicroscopy. European Journal of
Soil Science 54, 423-433.
Prietzel J., Mayer B. & Legge A.H. 2004. Cumulative impact of 40 years of industrial sulfur
emissions on a forest soil in west-central Alberta (Canada). Environmental Pollution
132, 129-144.
Prietzel J., Thieme J. Salomé M. & Knicker H. 2007. Sulfur K-edge XANES spectroscopy
reveals differences in sulfur speciation of bulk soils, humic acid, fulvic acid, and par-
ticle size separates. Soil Biology and Biochemistry 39, 877-890.
Prietzel J., Thieme J., Herre A., Salomé M & Eichert D. 2008. Differentiation between ad-
sorbed and precipitated sulphate in soils and at micro-sites of soil aggregates by
sulphur K-edge XANES. European Journal of Soil Science 59, 730-743.
Prosser J.T. & Carr M.J. 1987. Poás volcano, Costa Rica: geology of the summit region and
spatial and temporal variations among the most recent lavas. Journal of Volcanology
and Geothermal Research 33, 131-146.
Rajan S.S.S. 1979. Adsorption and desorption of sulfate and charge relationships in allo-
phanic clays. Soil Science Society of America Journal 43, 65-69.
Ressler T. 1998. WinXAS: a program for X-ray absorption spectroscopy data analysis under
MS-Windows. Journal of Synchrotron Radiation 5, 118-122.
Reuss J.O. & Johnson D.W. 1986. Acid deposition and the acidification of soils and waters.
Ecological Studies 59. Springer Verlag, New York, USA.
Reuss J.O. & Walthall P.M. 1990. Soil reaction and acidic deposition. In: Acidic precipitation
Volume 4: Soils, aquatic processes and lake acidification (eds. S.A. Norton, S.E.
Lindberg & A.L.Page). Advances in Environmental Sciences. Springer Verlag, New
York.
Rhoades J.D. 1982. Soluble salts. In: Methods of Soil Analysis Part 2 - Chemical and Micro-
biological Properties (eds. A.L Page, R.H. Miller & D.R. Keeney), pp. 167-197.
American Society of Agronomy & Soil Science Society of America, Madison, WI.
References
155
Rodhe H. 1989. Acidification in a global perspective. Ambio 18, 155-160.
Rogers W.B. & Rogers R.E. 1848. On the decomposition and partial solution of minerals,
rocks, etc. by pure water and water charged with carbonic acid. American Journal of
Science Series 2 5, 401-405.
Rowe G.L. Jr & Brantley S.L. 1993. Estimation of dissolution rates of andesitic glass, plagio-
clase and pyroxene in a flank aquifer of Poás Volcano, Costa Rica. Chemical Geol-
ogy 105, 71-87.
Rowe G.L. Jr, Brantley S.L., Fernandez M., Fernandez J.F., Borgia A. & Barquero J. 1992.
Fluid-volcano interaction in an active stratovolcano: the crater lake system of Poás
Volcano, Costa Rica. Journal of Volcanology and Geothermal Research 49, 23-51.
Scheel T., Haumaier L., Ellerbrock R.H., Rühlmann J. & Kalbitz K. 2008. Properties of or-
ganic matter precipitated from acidic forest soil solutions. Organic Geochemistry 39,
1439-1453.
Schulze D.G. & Bertsch P.M. 1995. Synchrotron X-ray techniques in soil, plant, and envi-
ronmental research. Advances in Agronomy 55, 1-66.
Schwarz A., Wilcke W., Styk J. & Zech W. 1999. Heavy metal release from soils by pH(stat)
experiments. Soil Science Society of America Journal 63, 290-296.
Schwertmann U. 1964. Differenzierung der Eisenoxide des Bodens durch Extraktion mit
saurer Ammoniumoxalat-Lösung. Zeitschrift für. Pflanzenernährung und
Bodenkunde 105, 194-202.
Sedov S. Solleiro-Rebolledo E., Morales-Puente P., Arias-Herreia A., Vallejo-Gómez E. &
Jasso-Castañeda C. 2003. Mineral and organic components of the buried paleosols
of the Nevado de Toluca, Central Mexico as indicators of paleoenvironments and
soil evolution. Quarternary International 106-107, 169-184.
Shoji S., Nanzyo M & Dahlgren R. 1993. Volcanic ash soils: genesis, properties and utiliza-
tion. Developments in soil science 21. Elsevier Science Publishers, Amsterdam.
Shotyk W. & Nesbitt H.W. 1992. Incongruent and congruent dissolution of plagioclase feld-
spar: effect of feldspar composition and ligand complexation. Geoderma 55, 55-78.
Siewert C. 2001. Investigation of the thermal and biological stability of soil organic matter.
Berichte aus der Umweltwissenschaft. Technische Universität Berlin. Shaker Verlag,
Aachen, Germany.
Skyllberg U. 1999. pH and solubility of aluminium in acidic forest soils: a consequence of
reactions between organic acidity and aluminium alkalinity. European Journal of Soil
Science 50, 95-106.
Chapter 8
156
Smeck N.E. & Novak J.M. 1994. Weathering of soil clays with dilute sulfuric acid as influ-
enced by sorbed humic substances. Geoderma 63, 63-76.
Smith R.A. 1852. On the air and rain of Manchester. Memoires of the Literary and Philoso-
phical Society Manchester Series 2 10, 207-217.
Smith R.A. 1872. Air and rain. Longmans, Green, London.
Smykatz-Kloss W. 1974. Differential thermal analysis – application and results in mineralogy.
Springer Verlag Berlin, Germany.
Sposito G. 1986. Distinguishing adsorption from surface precipitation. In Geochemical Proc-
esses at Mineral Surfaces (eds. J.A. Davis and K.F. Hayes). American Chemical So-
ciety, Washington D.C. pp. 217-228.
Stevenson F.J. 1994. Humus Chemistry - Genesis, Composition, Reactions. 2nd ed. Wiley.
New York.
Strauss R., Brümmer G.W. & Barrow N.J. 1997. Effects of crystallinity of goethite: II. Rates of
sorption and desorption of phosphate. European Journal of Soil Science 48, 101-
114.
Stumm W. 1997. Reactivity at the mineral-water interface: dissolution and inhibition. Colloids
and Surfaces A - Physicochemical and Engineering Aspects 120, 143-166.
Süsser P. & Schwertmann U. 1991. Proton buffering in mineral horizons of some acid forest
soils. Geoderma 49, 63-76.
Takahashi T., Fukuoka T. & Dahlgren R.A. 1995. Aluminum solubility and release rates from
soil horizons dominated by aluminum-humus complexes. Soil Science and Plant Nu-
trition 41, 119-131.
Takamatsu T., Boratynski J. & Satake K. 1992. Effects of volcanic acid deposition on soil
chemistry: I. status of exchangeable cations and sulfur. Soil Science 154, 435-449.
Thieme J., Prietzel J., Tyufekchieva N., Paterson D. & McNulty I. 2006. Speciation of sulfur
in oxic and anoxic soils using X-ray spectromicroscopy. Proceedings of the 8
th
Inter-
national Conference on X-ray Microscopy. 26 – 30 Juliy 2005, Egret Himeji, Japan
(eds. S. Aoki, Y. Kagoshima & Y. Suzuki), pp. 318-320. IPAP Conference Series 7.
Institute of Pure and Applied Physics, Tokyo, Japan.
Ulrich B. 1986. Natural and anthropogenic components of soil acidification. Zeitschrift für
Pflanzenernährung und Bodenkunde 149, 702-717.
Ulrich B. 1990. An ecosystem approach to soil acidification. In: Soil Acidity (eds. B. Ulrich &
M.E. Sumner). Springer Verlag, Berlin.
Ulrich B. & Sumner M.E. 1990. Soil acidity. Springer Verlag, Berlin Heidelberg, Germany.
References
157
Ulrich B., Mayer R. & Khanna R.K. 1980. Chemical changes due to acid precipitation in a
loess derived soil in Central Europe. Soil Science 130, 193-199.
UNA (Universidad Nacional Agraria) & ISRIC (International Soil Reference and Information
Centre) 1994. Soil reference profiles of Nicaragua. In: Field and Analytical Data,
Country Report 11 (profile NI005) (eds E. Acuña Espinales, J.H. Kauffmann & A.W.
Vogel), pp. 16-17. ISRIC Wageningen.
Usai M.-R. & Dalrymple J.B. 2003. Characteristics of silica-rich pedofeatures in a buried
paleosol. Catena 54, 557-551.
Van Breemen N. 1973. Dissolved aluminum in acid sulfate soils and in acid mine waters. Soil
Science Society of America Proceedings 37, 694-697.
Van Breemen N. 1991. Soil acidification and alkalinization. In: Soil Acidity (eds. B. Ulrich &
M.E. Sumner). Springer-Verlag, Heidelberg Berlin.
Van Breemen N., Burrough P.A., Velthorst E.J., van Dobben H.F., de Wit T., Ridder T.B. &
Reijnders H.F.R. 1982. Soil acidification from atmospheric ammonium sulphate in
forest canopy throughfall. Nature 299, 548-550.
Van Breemen N., Mulder J. & Driscoll C.T. 1983. Acidification and alkalinization of soils.
Plant and Soil 75, 283-308.
Van Breemen N., Driscoll C.T. & Mulder J. 1984. Acidic deposition and internal proton
sources in acidification of soils and waters. Nature 307, 599-604.
Vance G.F. & David M.B. 1991. Spodosol cation release and buffering of acidic inputs. Soil
Science 151, 362-368.
Van der Salm C., Westerveld J.W. & Verstraten J.M. 2000. Release rates of Al from inor-
ganic and organic compounds in a sandy podzol, during laboratory experiments.
Geoderma 96, 173-198.
Veerhoff M. & Brümmer G.W. 1993. Bildung schlechtkristalliner bis amorpher Verwitterung-
sprodukte in stark bis extrem stark versauerten Waldböden. Zeitschrift für Pflanzen-
ernährung und Bodenkunde 156, 11-17.
Wada K. 1989. Allophane and imogolite. In Minerals in Soil Environments, 2
nd
edition (eds.
J.B. Dixon and S.B. Weed). Soil Science Society of America Book Series no. 1,
Madison, USA. pp. 1051-1087.
Walker J.A., Williams S.N., Kalamarides R.I. & Feigenson M.D. 1993. Shallow open-system
evolution of basaltic magma beneath a subduction zone volcano: the Masaya Cal-
dera Complex, Nicaragua. Journal of Volcanology and Geothermal Research 56,
379-400.
Chapter 8
158
Wijnja H. & Schulthess C.P. 2000. Vibrational spectroscopy study of selenate and sulfate
adsorption mechanisms on Fe and Al (Hydr)oxide surfaces. Journal of Colloid and
Interface Science 229, 286-297.
Wolt J.D., Hue N.V. & Fox R.L. 1992. Solution sulfate chemistry in three sulfur-retentive
Hydrandepts. Soil Science Society of America Journal 56, 89-95.
WRB, 1998. World Reference Base for Soil Resources. World Soil Resources Reports, No.
84, FAO, Rome.
Zimmer M.M., Fischer T.P., Hilton D.R., Alvarado G.E., Sharp Z.D. & Walker J.A. 2004.
Nitrogen systematics and volatile fluxes of subduction zones: insights from Costa
Rica arc volatiles. Geochemistry, Geophysics, Geosystems 5, Q05J11, doi:
10.1029/2003GC000651.
159
160
ACKNOWLEDGEMENTS
I kindly thank
Prof. Dr. Martin Kaupenjohann for supervising this thesis, for giving me confi-
dence and freedom throughout my work, and for his guidance in science, during
which he always tried to bring out the best in me,
Dr. Friederike Lang for her ever-existing readiness for discussion, her confi-
dence and her overall support from first plans until the end of this project,
Prof. Dr. Jörg Prietzel for guidance in XANES-measurements, for supplying
aged basaluminite, for being co-examiner and for his beneficial disposition to
say things straight out,
Dr. Christina Siebe for encouraging me to start this journey and for her friend-
ship throughout all this time,
Dr. Guillermo Alvarado Induni, Carlos Ramírez Umaña, Raúl Mora Amador, Ma-
rita and Marta for their invaluable help during field work at Poás, for the unfor-
gettable excursions to Arenal Volcano and the crater lake of Poás and for giving
me shelter in San José,
Dr. Wilfried Strauch and Pedro Pérez from INETER for their invaluable help du-
ring field work and the possibility to look into the deep earth interior at Masaya,
Dr. Jürgen Thieme and Dr. Vincent de Andrade for introducing me to the fasci-
nating world of XANES instrumentation,
Dr. Katja Emmerich, Annett Steudel, Dr. Peter Weidler, Stefan Heißler and Do-
reen Rapp for receiving me cordially and for all the interesting discussions and
their overall help during and after my stay at the Forschungszentrum Karlsruhe,
Dr. Reiner Dohrmann, Dr. Stephan Kaufhold, Annette Kaufhold, Peter Rend-
schmidt and Detlev Klosa for their help with XRD-, IR-, STA-MS and ESEM-
EDX analysis at the Bundesanstalt für Geowissenschaften und Rohstoffe in
Hannover,
161
Ulrich Gernert, Jörg Nissen, Irene Preuss and Francois Galbert from ZELMI for
many hours at the petrographic and scanning electron microscopes as well as
the microprobe,
Dr. Christoph Böttcher for (Cryo-)TEM-imaging of synthetic allophane, Barbara
Büchtemann, Dr. Bernhard Durschang and Dr. Andreas Hösch for their help du-
ring glass synthesis, Petra Marsiske for her support with glass grinding, Dr.
Holger Müller and Mr. Stegemann for their help with particle size measurement,
Dr. Ruth Ellerbrock and Ralf Rath for support during IR measurements, Dr.
Wolf-Anno Bischoff for helpful discussions on the use of ion exchange resins in
SIA´s, Dr. Jörn Breuer for ICP-OES analysis and Dr. Elisabeth Irran for XRD
measurements,
Dr. Peter Buurman for helpful discussions on a first draft of the manuscript pre-
sented in Chapter 2,
Geerd Smidt and Juliane Hirte for their great and competent assistance during
laboratory work, Frédéric Hasché and Jördis Eisenblätter for help with sample
preparation,
the Deutsche Forschungsgemeinschaft (grant 1398/1) and the European Syn-
crotron Radiation Facility (grant EC151) for funding this work,
Dr. Christian Mikutta for his helpful criticism,
Claudia Kuntz, Sabine Rautenberg, Michael Facklam, Maike Mai, Karl Böttcher
and Christine Ehrlicher for support in the laboratory, Martin Kern for his prompt
help with any PC problem,
all other colleagues of the departments of “Soil Science” and “Site Evaluation
and Land Protection” for the nice working atmosphere, especially Jaane, Clau-
dia, Micha, Björn, Geerd, Eva and Heiner for the relaxing badminton dates,
Suse, Anne, Sabine, Tine, Gabi, my brother and my sister for giving me a
shoulder to lean on if things did not work as thought before and
my parents for opening me opportunities with their overall support.
162
163
CURRICULUM VITAE
1969 Born on 8 January in Laupheim (Germany)
1988 Graduation (Abitur) at the Carl-Lämmle-Gymnasium in
Laupheim (Germany)
1988-1990 Graduation (Landwirtschaftlich-technische Assistentin) at the
Landwirtschaftliche Lehranstalten Landsberg/Lech (Germany)
1990-1992 Study of Biology at the University of Konstanz (Germany)
1992-1999 Study of Agricultural Biology at the University of Hohenheim
(Germany)
1999 Diploma in Agricultural Biology; thesis at the Institute of Soil
Science, University of Hohenheim: “Auswirkungen einer
veränderten Qualität des Bewässerungswassers auf die C-
Mineralisierung und Schwermetallmobilisierung in langjährig
mit ungeklärten Abwässern bewässerten Böden Mexiko´s”
1999-2003 Head of laboratory of the “Laboratorio de edafología ambiental
(LEA)” at the Universidad Nacional Autónoma de México
(Mexico)
2004-2008 Staff member of the Institute of Ecology, Department of soil
science (Prof. Dr. Martin Kaupenjohann), TU Berlin
A-1
APPENDIX
Appendix
A-2
Chapter 2: Mechanisms of acid buffering and formation of secondary minerals
in vitric Andosols
App. 2.1 Results of short-term acidification experiments (method 1 with ion exchange resin)
Sample replicate DOC Si Al Ca Mg K Na
g kg
-1
mmol kg
-1
-------------------------- mmol c kg
-1
----------------------------
P1 a 0.28 15.0 90.2 22.4 0.98 0.66 1.46
b 0.36 19.4 102 22.6 1.12 0.58 1.67
c 0.29 16.1 94.7 21.3 1.02 0.50 1.45
P2 a 2.97 34.6 206 29.4 1.66 0.70 2.78
b 3.73 45.1 243 37.8 3.17 0.82 3.32
c 2.69 46.2 230 38.2 2.85 0.83 3.69
P3 a 2.83 59.4 314 47.8 6.57 0.91 4.46
b 3.13 55.3 331 47.8 6.37 0.89 4.79
c 3.02 55.3 324 49.8 6.84 0.86 4.90
P4 a 2.66 30.4 284 29.4 2.97 1.25 3.06
b 2.84 34.9 292 30.2 3.08 1.36 3.07
c 2.91 38.8 295 32.6 3.52 1.39 3.71
P5 a 7.10 66.8 600 53.8 5.91 1.10 5.66
b 7.11 70.8 643 57.8 6.31 1.13 5.53
c 7.20 66.6 600 59.5 6.29 1.11 5.38
M1 a 0.17 45.1 398 17.2 1.79 2.48 1.40
b 0.14 41.8 385 17.6 1.87 2.45 1.39
c 0.13 39.3 392 14.8 1.84 2.54 1.35
M2 a 1.57 132 755 90.5 13.7 6.41 6.07
b 1.45 127 734 92.9 15.4 6.41 5.06
c 1.47 127 763 90.6 16.0 6.77 5.37
A-3
App. 2.2 Results of short-term acidification experiments (method 2: pH
stat
titration)
Sample replicate DOC Si Al Ca Mg Na
g kg
-1
mmol kg
-1
------------------- mmol c kg
-1
---------------------
P1 a 0.16 4.8 61.9 19.1 0.96 1.5
b 0.16 4.4 56.7 20.7 1.00 1.5
c 0.12 3.7 46.6 19.5 0.87 1.4
P2 a 0.44 4.0 63.7 4.4 0.51 1.3
b 0.48 4.3 68.9 5.6 0.55 1.4
c 0.58 3.9 77.4 4.6 0.46 1.4
P3 a 0.84 5.1 89.6 4.6 0.56 0.9
b 0.95 6.9 105 5.4 0.75 1.6
c 0.73 5.1 89.0 4.5 0.64 1.8
P4 a 0.43 3.1 77.9 7.7 1.53 1.0
b 0.33 4.4 97.0 6.6 1.74 2.4
c 0.47 4.4 96.9 6.4 1.58 2.1
P5 a .1.60 7.6 224 5.8 1.08 1.8
b 1.44 7.3 195 5.9 1.36 3.7
c 1.12 5.4 173 5.0 1.11 3.3
M1 a 0.17 15.1 185 7.5 1.05 1.3
b 0.13 11.6 140 7.1 0.94 1.3
c 0.15 12.4 163 7.1 1.01 1.3
M2 a 0.15 8.4 98.4 60.1 4.38 1.3
b 0.17 9.9 105 58.6 4.81 1.4
c 0.18 8.8 115 59.2 4.55 1.2
Appendix
A-4
App. 2.3a Results of soil saturation extracts: pH, Si, Al, Fe, Ca, Mg and Na (ion concentrations in mg L
-1
)
Profile Horizon Depth pH Si Al Ca Mg Na
P1 Ah 0-6 4.57 23.4 0.9 24.2 4.4 16.3
2AC 6-9 4.77 38.0 <0.8 5.9 27.3 15.1
3Ah1 9-17 3.97 27.3 8.1 16.8 10.7 12.9
3Ah2 17-25 3.78 26.0 1.7 6.5 4.0 12.0
4AC 25-29 3.57 29.3 6.3 14.0 16.8 17.6
5AB1 29-51 4.00 22.0 6.8 9.4 9.2 7.0
? 51-54 4.03 22.9 10.1 11.9 11.0 11.5
mBs 54-54.4 4.07 31.7 7.3 12.2 7.5 15.2
? 54.4-61.4 4.12 28.9 4.4 14.9 9.8 8.3
9C 64.1-84.4 4.66 21.4 0.8 30.3 2.7 11.8
P2 L 19-9 na na na na na na
Ofh 9-0 na na na na na na
Ah 0-4 4.26 43.6 0.9 17.1 10.1 23.4
2AC 4-9 4.66 23.1 <0.8 13.4 3.1 11.9
3Ah 9-15 4.18 23.0 2.8 10.9 2.2 9.1
4AC 15-18 4.24 24.3 1.8 7.1 1.8 6.7
5Ah 18-48 4.14 24.1 1.3 5.0 1.7 7.7
6C 48-52 4.18 21.5 1.7 6.1 1.8 8.4
7AC 52-66/69 4.16 24.1 1.4 5.3 1.9 7.0
7mBs (1-2mm) na na na na na Na
7ABws 66/69-71 4.18 16.1 1.3 5.8 1.5 6.9
8AhC 71-83 4.15 23.0 1.6 5.8 1.6 10.6
8ACg 83-90 4.24 24.5 1.5 10.0 1.9 8.9
9C 90-109/118 4.7 18.1 <0.8 22.3 2.9 10.1
P3 L 12-8 na na na na na na
Ofh 8-0 3.6 34.7 2.2 8.6 3.3 7.5
Ah 0-4 3.89 56.9 1.7 8.9 7.7 8.6
2AC 4-11 4.69 23.1 <0.8 14.2 3.6 12.8
3Ah 11-25 4.27 22.1 1.6 5.6 2.1 7.5
4C 25-28 1.90 32.7 174 41.3 16.7 28.1
5Ah 28-54 4.01 16.8 1.7 5.3 1.4 5.6
6AC 54-57 4.12 18.9 1.5 10.4 1.8 8.4
7Ah 57-74 3.98 22.8 1.7 6.0 1.6 8.2
7ABg 74-82 4.04 19.7 1.7 5.0 1.4 6.6
8ACg 82-88 4.12 26.6 1.8 4.6 1.6 7.6
8mBs 88-88.4 4.35 23.4 1 7.1 1.7 11.5
9C 88.4-107 4.48 23.4 <0.8 24.9 3.7 14.9
na: not analyzed
A-5
App. 2.3b Results of soil saturation extracts: pH, Si, Al, Fe, Ca, Mg and Na (ion concentrations in mg L
-1
)
Profile Horizon Depth pH Si Al Ca Mg Na
P4 L 5-3 na na na na na na
Ofh 3-0 3.55 12.2 <0.8 10.0 3.1 5.5
2AC 0-7 4.62 31.8 <0.8 8.7 4.1 11.6
3Ah1 7-11 4.81 18.9 <0.8 1.3 0.9 5.7
3Ah2 11-15 4.28 20.8 1.2 7.3 1.5 6.0
4AC 15-18/21 4.41 24.0 <0.8 5.2 2.0 7.0
4mBs (1mm) na na na na na na
5AB 18/21-34 4.33 24.9 <0.8 1.5 0.6 4.2
5ABs 34-38 4.47 21.3 <0.8 2.0 1.1 7.1
6C 38-44 4.76 14.1 <0.8 20.0 1.2 6.1
7BC 44-53 4.73 14.3 <0.8 8.8 1.1 4.0
8Bw 53-80 4.52 19.1 <0.8 10.3 1.9 5.5
P5 L 33-29 na na na na na na
Ofh 29-0 4.05 41.7 1.7 9.0 3.7 12.9
2AC 0-3 4.79 15.2 <0.8 3.1 2.2 20.3
3Ah 3-9 4.70 14.7 <0.8 2.4 1.9 11.8
4C1 9-20 4.02 7.2 5.1 7.9 2.2 9.1
4mCs (2mm) 4.52 6.4 na 4.4 1.1 11.4
4C2 20-27 4.12 7.3 2.1 12.6 2.3 16.5
5Ah 27-53 4.35 16.9 0.9 5.0 1.2 9.7
5ABws1 53-60 5.11 12.9 <0.8 0.5 0.4 8.2
5ABws2 60-83 4.57 9.7 <0.8 2.5 0.9 8.7
6AC 83-87 5.10 7.1 <0.8 2.1 0.6 6.3
7AC 87-101 5.06 7 <0.8 1.6 0.8 7.1
M1 A 0-9 4.14 26.2 21.6 98.0 26.8 33.3
2BC 9-14 4.41 29.4 9.1 49.2 10.1 17.4
3A 14-24 4.49 29.4 6.6 28.1 8.9 15.4
4C 24-52 5.53 23.7 2.4 35.1 6.2 8.2
5C 52-58 5.16 25.9 5.7 77.5 10.4 8.5
6BC 58->80 5.67 20.0 2.4 63.4 11.5 9.1
M2 A 0-6 4.48 33.1 11.8 983 148 21.3
2BC 6-15 4.81 24.5 4.8 156 24.0 12.6
3A 15-21 4.93 27.6 1.4 163 26.6 11.7
4C 21-44 6.33 16.6 <0.8 23.8 5.0 4.1
5BC 44-50 6.12 19.0 <0.8 43.0 7.6 3.9
5C 50-54 6.40 16.9 <0.8 25.4 4.9 12.6
6BC 54-63 5.91 21.8 <0.8 67.6 11.8 8.1
na: not analyzed
Appendix
A-6
App. 2.4a Results of soil saturation extracts: DOC, sulfate, chloride, fluoride and nitrate
(DOC and ion concentrations in mg L
-1
)
Profile Horizon Depth DOC SO
4
Cl F NO
3
P1 Ah 0-6 5.4 105 4.8 <0.8 2.2
2AC 6-9 12.6 155 21.9 <0.8 <1.1
3Ah1 9-17 8.0 168 1.4 0.8 2.3
3Ah2 17-25 2.5 83 2.0 <0.8 <1.1
4AC 25-29 12.3 182 11.5 1.5 <1.1
5AB1 29-51 6.4 121 na 1.0 <1.1
? 51-54 8.6 174 3.4 <0.8 5.4
mBs 54-54.4 4.0 119 4.4 1.3 <1.1
? 54.4-61.4 1.6 116 3.3 0.8 <1.1
9C 64.1-84.4 2.8 91.3 10.8 <0.8 1.1
P2 L 19-9 na na na na na
Ofh 9-0 na na na na na
Ah 0-4 24.9 116 11.9 <0.8 <1.1
2AC 4-9 8.1 77.2 11.6 <0.8 <1.1
3Ah 9-15 10.1 86.3 9.5 <0.8 <1.1
4AC 15-18 6.4 83.4 4.7 <0.8 <1.1
5Ah 18-48 6.9 49.3 7.8 <0.8 <1.1
6C 48-52 na 43.7 9.2 <0.8 <1.1
7AC 52-66/69 8.1 46.6 5.7 0.9 <1.1
7mBs (1-2mm) na na na na na
7ABws 66/69-71 5.7 40.5 5.1 0.8 <1.1
8AhC 71-83 8.1 54.9 8.1 1.0 <1.1
8Acg 83-90 5.2 75.8 4.8 1.0 <1.1
9C 90-109/118 4.0 90.6 2.9 0.8 <1.1
P3 L 12-8 na na na na na
Ofh 8-0 73.4 236 21.0 0.8 <1.1
Ah 0-4 30.0 150 16.5 <0.8 <1.1
2AC 4-11 7.3 72.3 8.4 <0.8 25.5
3Ah 11-25 6.3 75.1 5.4 <0.8 <1.1
4C 25-28 340 7834 na 1.6 <1.1
5Ah 28-54 7.2 47.7 6.1 <0.8 <1.1
6AC 54-57 9.6 50.4 10.4 0.9 <1.1
7Ah 57-74 4.9 61.1 6.3 <0.8 <1.1
7Abg 74-82 2.0 42.4 4.2 <0.8 <1.1
8Acg 82-88 3.2 55.0 3.8 0.9 <1.1
8mBs 88-88.4 na 52.3 6.1 0.9 <1.1
9C 88.4-107 6.5 97.6 11.4 <0.8 <1.1
na: not analyzed
A-7
App. 2.4b Results of soil saturation extracts: DOC, sulfate, chloride, fluoride and nitrate
(DOC and ion concentrations in mg L
-1
)
Profile Horizon Depth DOC SO
4
Cl F NO
3
P4 L 5-3 na na na na na
Ofh 3-0 62.9 75.3 8.0 <0.8 <1.1
2AC 0-7 16.4 85.7 2.8 <0.8 <1.1
3Ah1 7-11 7.2 71.0 5.1 <0.8 10.4
3Ah2 11-15 8.8 38.9 8.2 <0.8 46.9
4AC 15-18/21 14.0 62.6 5.6 <0.8 36.2
4mBs (1mm) na na na na na
5AB 18/21-34 10.5 18.7 5.6 <0.8 <1.1
5ABs 34-38 na 23.6 6.6 <0.8 <1.1
6C 38-44 2.1 65.8 3.5 <0.8 1.9
7BC 44-53 0.8 33.4 2.1 <0.8 2.3
8Bw 53-80 1.1 46.1 2.5 <0.8 <1.1
P5 L 33-29 na na na na na
Ofh 29-0 90.3 48.6 42.0 <0.8 2.4
2AC 0-3 10.4 60.0 33.7 0.8 2.8
3Ah 3-9 7.1 50.0 12.2 <0.8 <1.1
4C1 9-20 6.0 27.5 9.3 0.9 49.9
4mCs (2mm) na 27.1 12.8 1.0 6.6
4C2 20-27 na 34.9 4.7 1.0 61.0
5Ah 27-53 2.3 16.3 6.1 0.8 46.6
5ABws1 53-60 7.7 20.2 9.0 <0.8 <1.1
5ABws2 60-83 na 8.9 4.7 <0.8 22.2
6AC 83-87 1.5 24.3 3.6 <0.8 <1.1
7AC 87-101 na 25.0 3.4 <0.8 <1.1
M1 A 0-9 76.5 159 252 8.8 82.9
2BC 9-14 41.5 77.1 108 4.6 20.9
3A 14-24 44.2 70.7 71.0 3.9 22.4
4C 24-52 31.7 58.1 49.0 4.3 13.3
5C 52-58 31.8 97.2 110 9.2 24.0
6BC 58->80 27.1 62.8 110 4.6 19.4
M2 A 0-6 415 706 642 10.9 4139
2BC 6-15 41.4 107 125 6.6 409
3A 15-21 50.9 85.2 238 2.2 273
4C 21-44 23.7 23.0 41.4 1.5 33.4
5BC 44-50 31.6 46.3 68.7 1.4 33.9
5C 50-54 28.7 20.9 43.2 1.4 22.1
6BC 54-63 48.1 55.6 118 0.9 58.5
na: not analyzed
Appendix
A-8
Chapter 4: In situ precipitation of basic aluminum sulfate phases at the
allophane-water interface
App. 4.1 Sulfur- and aluminum fluxes over the experimental period of 18 months in resin boxes exposed in soils
close to Poás and Masaya Volcano
Poás Masaya
box # sublayer # S (kg ha
-1
) Al (kg ha
-1
) S (kg ha
-1
) Al (kg ha
-1
)
1 1 292 29 142 94
2 nd 2 nd 2
3 nd 1 nd 2
4 nd 2 20 9
2 1 453 28 110 50
2 nd 2 nd 2
3 nd 1 nd 4
4 62 10 11 3
3 1 187 27 58 29
2 nd 1 nd 2
3 nd 1 nd 2
4 22 4 18 5
4 1 372 20 62 53
2 nd 2 nd 2
3 nd 1 nd 2
4 1 2 19 6
5 1 997 37 63 47
2 395 4 nd 2
3 nd 2 nd 1
4 86 10 10 2
6 1 712 33 36 32
2 128 3 nd 1
3 nd 2 nd 1
4 156 19 15 4
7 1 457 13 32 34
2 nd 1 nd 2
3 nd 1 nd 1
4 45 4 12 2
8 1 129 17 6 9
2 nd nd nd 2
3 nd nd nd 1
4 294 45 18 2
nd: not detectable
A-9
App. 4.2 Carbon and sulfur concentrations of glass and allophane samples (pristine and after field exposure at Poás
and Masaya)
glass allophane
sample replicate %C %S %C %S
pristine 1 0.06 0.0015 0.16 <0.001
2 0.06 0.0019 0.16 <0.001
3 0.03 0.0016 0.16 <0.001
4 0.03 <0.0010 na na
Poás 1 0.07 0.0023 1.39 1.13
2 0.13 0.0027 0.93 1.07
3 0.08 0.0020 1.87 1.18
4 0.10 0.0034 1.73 1.23
5 0.06 0.0029 1.07 1.23
Masaya 1 0.06 0.0012 0.32 0.65
2 0.06 0.0033 0.21 0.75
3 0.05 0.0012 0.22 0.75
4 0.04 0.0018 0.25 0.44
5 na na 0.40 0.41
na: not analyzed
Appendix
A-10
App. 4.3 Sulfate extractable by phosphate from field exposed allophane and S containing reference compounds (in
% of total S)
sample replicate Extraction time
1 hour 8 hours 24 hours 48 hours 168 hours
basaluminite 1 19.2 29.9 36.9 40.6 48.1
2 19.8 31.8 35.1 37.7 47.2
3 19.0 31.8 36.1 41.3 44.8
K-alunite 1.8 1.5 2.1 3.1 6.0
nd 1.4 2.6 3.1 6.4
1.1 1.2 2.4 3.0 6.6
Na-alunite 8.7 8.9 9.0 9.4 9.5
7.4 8.4 9.7 10.2 9.1
6.5 8.6 9.7 9.2 9.6
Na-Dodecylsulfate nd nd nd nd nd
nd nd nd nd nd
nd nd nd nd nd
Ref
ads
89.0 96.8 94.1 94.8 95.0
92.5 92.5 95.1 89.4 96.0
94.4 92.6 96.1 92.0 94.1
Poás 9.9 30.3 34.3 38.0 34.6
7.1 17.3 27.7 23.8 23.4
3.1 13.6 19.9 21.6 19.0
3.6 12.6 15.4 14.7 16.8
4.6 14.4 17.8 16.2 21.5
Masaya 15.6 30.7 44.8 45.4 45.4
61.3 86.8 83.6 80.7 80.9
67.6 85.2 83.5 81.8 82.0
62.1 83.2 80.9 73.6 75.0
63.1 83.1 84.0 79.3 79.7
nd: not detectable
A-11
App. 4.4 Oxalate extractable Si, Al and Fe (in g kg
-1
) of field exposed allophane with varying extraction times (2, 4
and 168 hours)
Extraction time
------------ 2 hours ------------ ------------ 4 hours ------------ ----------- 168 hours ----------
site replicate Si Al Fe Si Al Fe Si Al Fe
Poás 1 54 88 0.34 87 131 0.47 129 209 0.51
2 59 92 <0.28 101 148 0.48 132 212 0.42
3 42 66 0.41 73 115 0.44 125 206 0.53
4 32 52 0.38 65 99 0.52 127 204 0.64
5 44 66 0.28 77 117 0.42 132 210 0.52
Masaya 1 127 189 <0.28 142 207 <0.28 143 208 <0.28
2 145 216 <0.28 135 214 <0.28 145 217 <0.28
3 142 215 <0.28 144 210 <0.28 142 216 <0.28
4 140 213 <0.28 148 216 <0.28 144 212 <0.28
5 129 196 <0.28 136 200 <0.28 137 205 <0.28
Appendix
A-12
Chapter 5: Sulfate retention by allophane – adsorption of precipitation? Part 1.
Isotherms
App. 5.1 Retained sulfate, aluminum and silica net release and acid consumption (in mmol g
-1
) during the short-term
sulfate retention experiment at pH 5.0
sample replicate sulfate Si Al H
+
c. sample replicate sulfate Si Al H
+
c.
I1 1 0 0.035 0.012 0.04 I9 1 0.235 0.058 0.006 0.21
2 0 0.036 0.011 0.03 2 0.221 0.055 0.007 0.21
3 0 0.030 0.008 0.04 3 0.225 0.054 0.007 0.19
I2 1 0.011 0.037 0.017 0.06 I10 1 0.226 0.053 0.007 0.22
2 0.011 0.034 0.008 0.04 2 0.223 0.052 0.006 0.20
3 0.011 0.044 0.044 0.17 3 0.226 0.050 0.006 0.20
I3 1 0.043 0.036 0.009 0.06 I11 1 0.236 0.056 0.005 0.20
2 0.044 0.037 0.008 0.06 2 0.245 0.063 0.007 0.24
3 0.044 0.034 0.009 0.06 3 0.246 0.053 0.008 0.21
I4 1 0.079 0.040 0.008 0.08 I12 1 0.262 0.062 0.007 0.24
2 0.078 0.036 0.008 0.08 2 0.254 0.053 0.006 0.23
3 0.077 0.034 0.007 0.10 3 0.256 0.054 0.006 0.21
I5 1 0.139 0.052 0.011 0.45 I13 1 0.282 0.066 0.009 0.27
2 0.134 0.044 0.006 0.12 2 .0278 0.067 0.010 0.25
3 0.133 0.045 0.011 0.12 3 0.269 0.056 0.006 0.25
I6 1 0.169 0.051 0.006 0.15 I14 1 0.295 0.072 0.009 0.26
2 0.168 0.052 0.008 0.17 2 0.287 0.058 0.006 0.23
3 0.161 0.046 0.006 0.16 3 0.290 0.058 0.006 0.25
I7 1 0.215 0.059 0.009 0.20 I15 1 0.297 0.074 0.009 0.29
2 0.203 0.056 0.009 0.19 2 0.295 0.059 0.007 0.25
3 0.197 0.052 0.008 0.19 3 0.291 0.059 0.006 0.24
I8 1 0.234 0.059 0.007 0.22
2 0.232 0.058 0.010 0.21
3 0.221 0.050 0.008 0.18
A-13
App. 5.2 Retained sulfate, aluminum and silica net release and acid consumption (in mmol g
-1
) during the short-term
sulfate retention experiment at pH 4.5
sample replicate sulfate Si Al H
+
c. sample replicate sulfate Si Al H
+
c.
I1 1 0 0.087 0.233 0.43 I9 1 0.282 0.117 0.187 0.53
2 0 0.067 0.164 0.27 2 0.276 0.095 0.125 0.37
3 0 0.073 0.179 0.43 3 0.270 0.097 0.111 .038
I2 1 0.010 0.069 0.157 0.29 I10 1 0.272 0.099 0.122 0.39
2 0.010 0.068 0.146 0.29 2 0.279 0.094 0.097 0.29
3 0.010 0.071 0.183 0.37 3 0.279 0.095 0.099 0.36
I3 1 0.045 0.071 0.153 0.28 I11 1 0.290 0.109 0.131 0.46
2 0.046 0.075 0.161 0.35 2 0.286 0.108 0.118 .038
3 0.046 0.069 0.155 0.29 3 0.294 0.101 0.112 0.40
I4 1 0.086 0.083 0.176 0.38 I12 1 0.309 0.111 0.113 0.38
2 0.086 0.075 0.150 0.31 2 0.317 0.103 0.109 0.41
3 0.085 0.077 0.150 0.36 3 0.302 0.103 0.118 0.42
I5 1 0.147 0.088 0.154 0.37 I13 1 0.323 0.108 0.115 0.41
2 0.147 0.081 0.144 0.31 2 0.330 0.104 0.109 0.41
3 0.144 0.082 0.142 0.37 3 0.333 0.100 0.099 0.39
I6 1 0.192 0.098 0.175 0.38 I14 1 0.340 0.113 0.122 0.43
2 0.189 0.090 0.137 0.31 2 0.365 0.109 0.096 0.42
3 0.177 0.081 0.115 0.28 3 0.363 0.110 0.114 0.43
I7 1 0.232 0.095 0.131 0.36 I15 1 0.402 0.112 0.101 0.47
2 .0229 0.092 0.127 0.34 2 0.410 0.115 0.117 0.47
3 0.225 0.088 0.128 0.33 3 0.386 0.110 0.110 0.47
I8 1 0.261 0.094 0.116 0.33
2 0.258 0.096 0.117 0.35
3 0.255 0.089 0.116 0.32
Appendix
A-14
App. 5.3 Retained sulfate, aluminum and silica net release and acid consumption (in mmol g
-1
) during the short-term
sulfate retention experiment at pH 4.0
sample replicate sulfate Si Al H
+
c. sample replicate sulfate Si Al H
+
c.
I1 1 0 0.17 0.63 0.81 I9 1 0.275 0.24 0.64 1.09
2 0 0.16 0.60 0.79 2 0.287 0.22 0.65 1.18
3 0 0.16 0.58 0.85 3 0.285 0.22 0.61 1.13
I2 1 0.011 0.18 0.64 0.83 I10 1 0.303 0.24 0.66 1.12
2 0.010 0.17 0.61 0.84 2 0.308 0.22 0.63 1.14
3 0.010 0.17 0.60 0.90 3 0.307 0.23 0.63 1.20
I3 1 0.050 0.18 0.63 0.84 I11 1 0.301 0.24 0.64 1.13
2 0.047 0.17 0.61 0.89 2 0.314 0.26 0.68 1.26
3 0.047 0.17 0.57 0.84 3 0.308 0.23 0.63 1.20
I4 1 0.087 0.20 0.65 0.88 I12 1 0.314 0.25 0.67 1.18
2 0.088 0.19 0.63 0.89 2 0.322 0.25 0.66 1.21
3 0.088 0.19 0.63 1.03 3 0.323 0.24 0.68 1.38
I5 1 0.153 0.21 0.64 0.95 I13 1 0.323 0.26 0.68 1.19
2 0.153 0.19 0.62 0.95 2 0.335 0.25 0.65 1.20
3 0.154 0.20 0.61 1.01 3 0.341 0.24 0.65 1.28
I6 1 0.196 0.22 0.63 0.99 I14 1 0.349 0.27 0.66 1.22
2 0.199 0.21 0.61 1.04 2 0.367 0.27 0.71 1.35
3 0.197 0.20 0.61 1.02 3 0.378 .024 0.63 1.31
I7 1 0.231 0.22 .065 1.02 I15 1 0.419 0.27 0.70 1.31
2 0.234 0.20 0.59 1.02 2 0.392 .028 0.71 1.40
3 0.233 0.20 0.59 1.06 3 0.398 .026 0.68 1.51
I8 1 0.276 0.23 0.61 1.06
2 0.277 0.22 0.68 1.33
3 0.273 0.21 0.63 1.14
A-15
App. 5.4 Mesopore volume (MePV), micropore volume (MiPV) and average micropore diamter (AMiPD) of selected
samples from the short-term sulfate retention experiments
sample replicate MePV MiPV AMiPD
--------------------- mm
3
g
-1
--------------------- mm
3
g
-1
nm
2-5 nm 5-10 nm 10-50 nm
pristine 1 162 27 24 180 2.00
allophane 2 154 26 23 175 2.03
3 161 29 22 178 2.00
pH 4.5, I1 1 152 26 23 171 2.02
2 148 28 21 159 2.02
3 153 29 21 176 2.00
pH 4.5, I15 1 133 23 21 139 1.97
2 144 27 23 158 1.95
3 143 26 23 162 1.94
pH 4.0, I1 1 138 26 23 168 1.99
2 127 25 20 159 2.00
3 146 26 23 177 1.96
pH 4.0, I15 1 122 27 21 139 1.90
2 123 24 21 159 1.91
3 127 27 19 159 1.92
Appendix
A-16
Chapter 6: Sulfate retention by allophane – adsorption of precipitation? Part 2.
Kinetics
App. 6.1 Sulfate retention, aluminum and silica net release and acid consumption during the kinetic experiment at
pH 4.5
sample replicate Retained sulfate Si Al acid consumption
------------------------------------ mmol g
-1
------------------------------------
Ka1 1 0.160 0.068 0.123 0.14
2 0.162 0.065 0.106 0.16
3 0.160 0.063 0.100 0.14
Ka2 1 0.169 0.075 0.128 0.17
2 0.168 0.074 0.111 0.19
3 0.162 0.071 0.107 0.17
Ka3 1 0.176 0.081 0.136 0.22
2 0.175 0.078 0.116 0.23
3 0.171 0.077 0.115 0.21
Ka4 1 0.180 0.093 0.156 0.28
2 0.176 0.093 0.136 0.28
3 0.173 0.089 0.135 0.27
Ka5 1 0.195 0.117 0.176 0.34
2 0.195 0.115 0.150 0.33
3 0.187 0.109 0.151 0.33
Ka6 1 0.218 0.145 0.198 0.43
2 0.217 0.144 0.158 0.40
3 0.209 0.140 0.173 0.43
Ka7 1 0.243 0.162 0.168 0.43
2 0.248 0.157 0.139 0.41
3 0.235 0.158 0.149 0.44
Ka8 1 0.273 0.161 0.142 0.43
2 0.278 0.158 0.117 0.42
3 0.273 0.160 0.124 0.44
Ka9 1 0.297 0.160 0.124 0.43
2 0.290 0.152 0.102 0.42
3 0.285 0.153 0.105 0.44
Ka10 1 0.305 0.151 0.093 0.39
2 0.304 0.150 0.083 0.41
3 0.297 0.148 0.086 0.41
Ka11 1 0.322 0.154 0.064 0.37
2 0.319 0.151 0.062 0.40
3 0.312 0.153 0.060 0.40
Ka12 1 0.345 0.166 0.047 0.38
2 0.332 0.162 0.045 0.41
3 0.327 0.160 0.045 0.41
A-17
App. 6.2 Sulfate retention, aluminum and silica net release and acid consumption during the kinetic experiment at
pH 5.0
sample replicate Retained sulfate Si Al acid consumption
------------------------------------ mmol g
-1
------------------------------------
Kb1 1 0.190 0.050 0.014 0.11
2 0.199 0.049 0.012 0.11
3 0.190 0.050 0.012 0.11
Kb2 1 0.200 0.054 0.015 0.12
2 0.209 0.053 0.013 0.12
3 0.200 0.052 0.013 0.13
Kb3 1 0.207 0.058 0.013 0.14
2 0.216 0.059 0.010 0.14
3 0.198 0.060 0.012 0.14
Kb4 1 0.216 0.068 0.013 0.15
2 0.222 0.067 0.013 0.15
3 0.208 0.068 0.013 0.15
Kb5 1 0.223 0.075 0.015 0.17
2 0.234 0.076 0.013 0.17
3 0.227 0.076 0.013 0.17
Kb6 1 0.240 0.083 0.013 0.17
2 0.212 0.083 0.012 0.18
3 0.235 0.086 0.012 0.18
Kb7 1 0.243 0.092 0.006 0.17
2 0.250 0.090 0.005 0.18
3 0.237 0.093 0.005 0.18
Kb8 1 0.259 0.089 0.003 0.17
2 0.259 0.093 0.002 0.18
3 0.250 0.089 0.003 0.18
Kb9 1 0.263 0.088 0.002 0.17
2 0.259 0.087 0.002 0.18
3 0.251 0.089 0.001 0.18
Kb10 1 0.267 0.088 0.001 0.17
2 0.259 0.089 0.002 0.18
3 0.249 0.089 0.001 0.18
Appendix
A-18
App. 6.3 Oxalate extractable Si and Al (in g kg
-1
) of allophane samples of the experiment on sulfate retention
kinetics at pH 4.5
sample replicate Si Al
------------------ g kg
-1
------------------
Ka6 1 128 211
2 130 208
3 135 214
Ka10 1 133 210
2 129 203
3 130 209
Ka11 1 128 215
2 125 206
3 121 202
Ka12 1 120 205
2 117 195
3 118 198
App. 6.4 Sulfate extractable by phosphate from allophane samples of the experiment on sulfate retention kinetics at
pH 4.5 (in % of total S)
sample replicate % of S
tot
sample replicate % of S
tot
Ka6 1 98 Ka10 1 99
2 99 2 100
3 100 3 96
Ka8 1 99 Ka11 1 91
2 90 2 87
3 99 3 82
Ka9 1 90 Ka12 1 90
2 93 2 93
3 92 3 99
