Surfactant Self-assembly in a
Magnetic Room Temperature
Ionic Liquid
vorgelegt von
Diplom-Chemiker Andreas Klee
geb. in Geilenkirchen
von der Fakultät II - Mathematik und Naturwissenschaften
der Technischen Universität Berlin
zur Erlangung des akademischen Grades
Doktor der Naturwissenschaften
Dr.rer.nat.
genehmigte Dissertation
Promotionsausschuss:
Vorsitzender: Prof. Dr. Peter Strasser
1. Gutachter: Prof. Dr. Michael Gradzielski (Technische Universität Berlin)
2. Gutachter: Prof. Dr. Werner Kunz (Universität Regensburg)
Tag der wissenschaftlichen Aussprache: 22. Mai 2015
Berlin 2015
Danksagung
An erster Stelle möchte ich mich ganz herzlich bei Prof. Dr. Michael
Gradzielski bedanken, in dessen Arbeitskreis ich die letzten Jahre die Mög-
lichkeit hatte, meine Dissertation anzufertigen. Ich bedanke mich einer-
seits für das groÿe Vertrauen, mir so viele Freiheiten zur selbständigen Ar-
beit und Ideenentwicklung zu überlassen, andererseits bestand jederzeit die
Möglichkeit zur wissenschaftlichen Diskussion, welche während der gesamten
Zeit immer wieder sehr hilfreich war. Auch bedanken möchte ich mich für
die Ermöglichung von so vielen Konferenzbesuchen, welche einen erheblichen
Beitrag dazu lieferten, regelmäÿig über den Tellerrand schauen zu können.
Die gegebenen Experimentiermöglichkeiten durch die Ausstattung im In-
stitut vor Ort wurden massiv erweitert durch Messreisen zu verschiedenen
europäischen Neutronenquellen, wofür ich mich beim HZB Wannsee, PSI
Villigen, ILL Grenoble und LLB Saclay für die bewilligten Messzeiten be-
danken möchte. Ebenso bei Klaus Kiefer vom LaMMB (HZB Wannsee) für
die Ermöglichung der magnetischen SQUID-Messungen.
Groÿer Dank geht ebenfalls an alle Mitarbeiter des Stranski-Labors, für
eine sehr angenehme Arbeitsatmosphäre, in der mir viel Hilfs- und Disskus-
sionsbereitschaft entgegengebracht wurde. Ganz besonders danke ich Sylvain
Prevost, der einen erheblichen Anteil meines Verständnisses für Kleinwinkel-
streuung zu verantworten hat und Lisa Reile und Miriam Simon, die mich
im Rahmen ihrer Bachelorarbeit bzw. hilfswissenschaftlichen Tätigkeit un-
terstützt haben. Nicht zu vergessen sind auch die Mitarbeiter der Werkstatt
und der Glasbläserei, mit deren Hilfe verschiedenste Versuchsaufbauten über-
haupt erst realisiert werden konnten.
Neben aller fachlichen Unterstützung möchte ich mich jedoch ebenso bei
all den Menschen bedanken, mit denen ich in den letzen Jahren neben der
Doktorarbeit Zeit verbracht habe. Dies gab mir immer wieder die wichtige
Möglichkeit, Abstand zu nehmen von der spezialisierten Tätigkeit des Grund-
lagenforschers, was immer wieder Kraft, Gelassenheit und einen geschärften
i
Blick fürs wesentliche und groÿe Ganze gefördert hat. Neben all den neuen
und alten Freunden bedanke ich mich hierfür vor allem bei den Laubenmu-
sikern, meiner Familie und natürlich der phantastischen Katharina.
ii
Parts of this thesis has already been published as listed in the following:
•
Paper I:
Magnetic Microemulsions based on Magnetic Ionic Liq-
uid A. Klee, S. Prevost, W. Kunz, R. Schweins, K. Kiefer and M.
Gradzielski.
Phys. Chem. Chem. Phys.
,
2012
,
14
, 1535515360.
http://pubs.rsc.org/en/content/articlepdf/2012/cp/c2cp43048g Re-
produced by permission of the PCCP Owner Societies.
•
Paper II:
Self-Assembly of Imidazolium-Based Surfactants in Mag-
netic Room-Temperature Ionic Liquids: Binary Mixture A. Klee, S.
Prevost and M. Gradzielski.
ChemPhysChem
,
2014
,
15
, 40324041.
http://onlinelibrary.wiley.com/doi/10.1002/cphc.201402548/pdf Re-
printed with permission. Copyright 2014 Wiley-VCH.
•
Paper III:
Understanding and Optimising Microemulsions with Mag-
netic Room Temperature Ionic Liquids (MRTIL) A. Klee, S. Prevost,
U. Gasser and M. Gradzielski.
J. Phys. Chem. B
,
2015
,
119
, 4133
4142. http://pubs.acs.org/doi/pdf/10.1021/jp512545c Reprinted with
permission. Copyright 2015 American Chemical Society.
iii
iv
Contents
Abstract viii
Zusammenfassung x
1 Introduction 1
1.1 Self-assembly ........................... 1
1.2 Room Temperature Ionic liquids . . . . . . . . . . . . . . . . . 4
1.3 Self-assembly in non-aqueous systems . . . . . . . . . . . . . . 5
1.4 Aimofthiswork ......................... 8
2 Methods and Materials 11
2.1 UsedCompounds......................... 11
2.2 Recording phase diagrams by visual observation . . . . . . . . 13
2.2.1 Microemulsion systems . . . . . . . . . . . . . . . . . . 13
2.2.2 Binary MRTIL/alcohol mixtures . . . . . . . . . . . . . 14
2.3 Dierential scanning calorimetry (DSC) . . . . . . . . . . . . . 14
2.4 Conductivity ........................... 15
2.5 Magnetic susceptibility . . . . . . . . . . . . . . . . . . . . . . 15
2.6 Viscosity.............................. 15
2.7 Small angle scattering . . . . . . . . . . . . . . . . . . . . . . 15
2.7.1 SANS ........................... 17
2.7.2 SAXS ........................... 20
2.8 Surfacetension .......................... 20
2.9 Density .............................. 21
v
2.10 Polarized microscopy . . . . . . . . . . . . . . . . . . . . . . . 21
2.11 Emulsion Stability . . . . . . . . . . . . . . . . . . . . . . . . 22
3 Binary Mixtures: MRTIL/Surfactant 23
3.1 Results............................... 24
3.1.1 Temperature dependent binary phase diagrams . . . . 24
3.1.2 Surface tension measurements . . . . . . . . . . . . . . 28
3.1.3 Small angle scattering . . . . . . . . . . . . . . . . . . 32
3.2 Discussion............................. 40
3.2.1 Low and mid surfactant concentrations - critical aggre-
gation conditions . . . . . . . . . . . . . . . . . . . . . 40
3.2.2 Quantitative results on solvent quality for self-assembly 42
3.3 Conclusion............................. 43
4 Other Binary Mixtures 45
4.1 Binary Mixtures MRTIL/alkanol . . . . . . . . . . . . . . . . 46
4.2 Binary mixtures cyclohexane/decanol . . . . . . . . . . . . . . 48
5 Microemulsions 57
5.1 Microemulsions based on C
4
mimFeCl
4
............. 58
5.1.1 Micellization with decanol . . . . . . . . . . . . . . . . 58
5.1.2 Microemulsions . . . . . . . . . . . . . . . . . . . . . . 60
5.1.3 Conclusion......................... 77
5.2 Microemulsions containing dierent MRTIL . . . . . . . . . . 78
5.2.1 Macroscopic observations . . . . . . . . . . . . . . . . . 78
5.2.2 Mesoscopic structure . . . . . . . . . . . . . . . . . . . 80
5.3 Magnetic behaviour . . . . . . . . . . . . . . . . . . . . . . . . 86
5.3.1 Field Gradient . . . . . . . . . . . . . . . . . . . . . . 86
5.3.2 Homogeneous eld . . . . . . . . . . . . . . . . . . . . 88
6 Conclusion 99
References 102
vi
A Appendix Binary Systems I
A.1 SANSmodeltting........................ I
A.1.1 Spherical model as used in section 3 . . . . . . . . . . . I
A.1.2 Alternative spherical model . . . . . . . . . . . . . . . VIII
A.1.3 Ellipsoidal model . . . . . . . . . . . . . . . . . . . . . XI
A.2 Scattering invariant . . . . . . . . . . . . . . . . . . . . . . . . XII
A.3 Density ..............................XII
A.4 Dierential scanning calorimetry (DSC) . . . . . . . . . . . . . XIV
A.5 Surfacetension ..........................XXII
A.6 Polarized microscopy . . . . . . . . . . . . . . . . . . . . . . . XXII
A.7 SAXS ...............................XXIII
B Appendix Microemulsions XXV
B.1 Additional phase diagrams . . . . . . . . . . . . . . . . . . . . XXVI
B.2 The clipped random wave model . . . . . . . . . . . . . . . . . XXIX
B.3 The scattering invariant
Qinv
..................XXX
B.3.1 Calculation of theoretical invariant . . . . . . . . . . . XXXIII
B.4 SANS experiments recorded at D11 (ILL) . . . . . . . . . . . XXXIX
B.5 SANS curves measured at 36
◦
C.................XLII
B.6 SANS experiments recorded at SANSI (PSI) under magnetic
eld ................................XLIII
B.7 Surfacetension ..........................XLIV
B.8 Conductivity ...........................XLV
B.9 Viscosity..............................XLVI
B.10Cubemodel............................XLVI
B.11 Force Calculations in a Magnetic Field . . . . . . . . . . . . . XLVII
vii
Abstract
In this work the self-assembly in a nonaqueous system was investigated,
realized by using dierent magnetic room temperature ionic liquids (MRTIL,
alkylmethylimidazolium tetrachloroferrates, C
i
mimFeCl
4
with i = 2, 4, 6) as
solvent and imidazolium based surfactants (C
j
mimCl with j=12, 14, 16, 18)
as amphiphile.
In a systematic fashion the phase behavior was studied. For this pur-
pose we started with the simplest case of binary MRTIL/surfactant mixtures
where the alkyl chain length of surfactant and MRTIL was varied over a
broad temperature range and the complete range of compositions. In this
way it was possible to nd classical mesoscopic structures like micelles and
liquid crystalline structures. The complexity was extended by adding oil
and cosurfactant to the system which enabled us to formulate microemul-
sions. Again the inuence of surfactant and MRTIL alkyl chain lengths on
the phase behavior was investigated and additionally the investigation was
broadened by a versatile variation of the structure and amount of the cosur-
factant and oil. To ensure an as substantive and reliable picture as possible
it was made use of many complementary methods as calorimetry (DSC), po-
larized microscopy, neutron and X-ray scattering (SANS/SAXS), and surface
tension.
In general it was proven that it is possible to form typical self-assembled
structures in this MRTIL-based matrix like micelles, liquid crystals, emul-
sions and microemulsions as they are common for classical aqueous systems.
However in dierence to the latter ones it was shown that the ability to
self-assemble is weaker which is expressed e.g. by higher critical aggregation
concentrations leading to micelles with rather low aggregation numbers and
which are partly swollen by the solvent, or smaller tri-phasic regions for mi-
croemulsions, in which the mesoscopic domains show a less pronounced long
range ordering.
The weakness in self-assembly was quantied by the solvophobic eect of
the alkyl chain which is in the MRTIL only about a fth of that in water. It
viii
was distinguished between the eects of the solvophobic and -philic part of
the surfactant and as a result it was quantitatively shown that decits in the
ability to self-assemble are mainly present in the surfactant's solvophobic tail.
Two opposed trends for the amphiphilic stregth could be pointed out given
on the one hand by the length of surfactant alkyl chains which quanties the
solvophobicity of the amphiphile, and on the other hand by the length of
MRTIL alkyl chains, which quanties the solvent polarity.
As for this study ionic liquids with paramagnetic properties were cho-
sen, it was proven that this property is still present in the formulated mi-
croemulsion systems. As a second result it was possible to orient mesoscopic
structures in an external magnetic eld. However this was only possible for
certain locations in the phase diagrams and at rather high magnetic elds of
≥5.5
Tesla.
In summary, the here presented broad investigation yields quantitative
information on the composition-structure relationship and therefore gives
recipes to design magnetic self-assembled structures with optimised proper-
ties and structures, as it has not yet been done for such systems that can
be manipulated by a magnetic eld. These ndings are useful for designing
strategies for formulating microemulsions of a given structure with MRTILs
as polar component. This is important as such microemulsions could in the
future be employed as interesting reaction media which contain also a com-
ponent for separation via magnetic forces.
ix
Zusammenfassung
In der vorliegenden Arbeit wurde die Selbstorganisation von Tensiden in
einem wasserfreien System untersucht, und zwar durch Verwendung von ver-
schiedenen magnetischen Raumtemperatur ionischen Flüssigkeiten (MRTIL,
Alkylimidazoliumtetrachloroferraten, C
i
mimFeCl
4
mit i = 2, 4, 6) als Lö-
sungsmittel und Tenside mit einer Imidazolium-Kopfgruppe (C
j
mimCl mit
j=12, 14, 16, 18).
Das Phasenverhalten wurde zuerst für den einfachsten Fall von binären
MRTIL/Tensid-Mischungen untersucht, indem systematisch die Kettenlänge
von Tensid und MRTIL über einen breiten Temperaturbereich und in allen
Mischungsverhältnissen variiert wurde. Dabei wurden klassische mesoskopis-
che Strukturen gefunden wie z.B. Mizellen und Flüssigkristalle. Desweit-
eren wurde durch die Zugabe von Öl und Kotensid die Komplexität erhöht,
woduch es möglich war, Mikroemulsionen herzustellen. Auch hier wurde
der Einuss von MRTIL- und Tensidkettenlängen auf das Phasenverhal-
ten untersucht und zusätzlich durch eine vielfältige Variation von Menge
und Struktur des Öls und Kotensids das Beobachtungsspektrum erweitert.
Um ein möglichst fundiertes und vertrauenswürdiges Bild zu erhalten, wur-
den viele komplementäre Methoden wie Kalorimetrie (DSC), Polarisation-
smikroskopie, Neutronen- und Röntgenstreuung (SANS/SAXS) und Ober-
ächenspannung verwendet.
Generell konnten die für wässrige Systeme typischen Strukturen wie Mi-
zellen, Flüssigkristalle, Emulsionen und Mikroemulsionen in diesen ionischen
Flüssigkeiten dargestellt werden, jedoch mit einer schwächer ausgeprägten
Triebkraft, ausgedrückt z. B. durch höhere kritische Mizellisierungskonzen-
trationen, kleinere Aggregationszahlen für Mizellen, welche zusätzlich eine
teilweise Quellung mit Lösungsmittel aufwiesen, oder kleineren Dreiphasen-
gebieten für Mikroemulsionen, deren Domänen eine weniger ausgeprägte Fern-
ordnung zeigten.
Die schwächere Selbstorganisation wurde mit der Solvophobie der Alkyl-
ketten quantiziert, welche etwa nur ein fünftel in den MRTIL-Systemen ver-
x
glichen mit Wasser beträgt. Es konnte zwischen dem Eekt von solvophilem
und -phobem Tensidteil unterschieden werden mit dem Ergebnis, dass De-
zite in der Fähigkeit zur Selbstorganisation hauptsächlich auf den solvo-
phoben Tensidmolekülteil zurückzuführen sind. Zwei entgegengesetzte Trends
zur Beeinussung der amphiphilen Stärke des Systems konnten herausgestellt
werden: Einerseits quantiziert die Tensidkettenlänge die Solvophobizität
des Amphiphils und andererseits erhöht die MRTIL-Alkylkettenlänge die Lö-
sungsmittelpolarität.
Da in der vorliegenden Arbeit ionische Flüssigkeiten mit paramagneti-
schen Eigenschaften verwendet wurden, wurde veriziert, dass diese Eigen-
schaft in den untersuchten Mikroemulsionssystemen erhalten blieb. Deswei-
teren was es möglich, mesoskopische Strukturen in einem externen Magnet-
feld auszurichten, jedoch nur für ganz bestimmte Bereiche im Phasendia-
gramm und unter recht hohen Feldern von
≥5.5
Tesla.
Zusammenfassend liefert die hier vorgestellte Arbeit quantitative Infor-
mationen zur Struktur-Eigenschaftsbeziehung und gibt damit Anleitung zur
maÿgeschneiderten Formulierung von mesoskopischen Strukturen mit mag-
netischen Eigenschaften. Dies ermöglicht z. B. die gezielte Herstellung
von Mikroemulsionen mit bestimmten Strukturen, welche MRTIL als po-
lare Komponente enthalten. Dies ist nützlich, da solche Mikroemulsionen in
der Zukunft als interessante Reaktionsmedien mit der Option zur magnet-
feldinduzierten Separierung genutzt werden könnten.
xi
xii
1
Introduction
1.1 Self-assembly
Self-assembly in water is a phenomenon known for a long time and it is
known to be an important driving force leading to fundamental biological
mechanisms (e.g., protein folding, cell membranes) and is widely used in
applications (washing processes, solubilization, foams, gels, emulsions, food
industry). In general the basic building block essential for such structure for-
mation is an amphiphilic molecule (or particle) equipped with a hydrophilic
and a hydrophobic part. Models to describe this phenomenon have been de-
veloped like the hydrophobic eect or the concept of water structure, which
are based on the disruption of the water structure caused by the presence of
hydrophobic moieties.
1,2
An early approach to correlate the aggregation behaviour and the result-
ing structures with the molecular geometry of the amphiphile is the
HLB
(hydrophilic-lipophilic balance) value
3
which is related to the molar mass of
1
the hydrophilic (
Mh
) and lipophilic (
Ml
) part of the amphiphilic molecule:
HLB = 20 ·(1 −Ml
Ml+Mh
)
(1.1)
A disadvantage of this approach is that the amphiphile is characterized
disregarding its environment. Another famous model to explain dierent
structures accessible by amphiphilic self-assembly which overcomes this dis-
advantages is the packing parameter
p
. This widely used parameter charac-
terizes the amphiphile in a geometrical approach by the volume and shape
of its hydro- and lipophilic building blocks which gives the relation between
the area
as
occupied by the hydrophilic block and the length (
l
) and volume
(
v
) of the hydrophobic part given in eq. 1.2.
p=v
as·l
(1.2)
The energetically most favorable values for
as
,
l
and
v
dene the so-called
spontaneous packing parameter which is related to a preferred curvature
which then denes the expected possible structures as shown systematically
in Fig. 1.1. The beauty of this model relies on its simplicity and it is easy to
use as relationship between molecular structure of the building blocks and the
resulting mesoscopic hierarchical functions. Other widely used parameters
based on the same geometrical approach are the HLD (hydrophilic-lipophilic
deviation)
4
or the spontaneous curvature
5
which both can be related to
p
.
6
A disadvantage of this geometrical approach is that the values for
as
,
l
and
v
are dicult to generalize as they are strongly dependent on the
surrounding matrix. For example the headgroup spacing is strongly aected
by solvation, ionic strength (which screens neighboring ionic headgroups) or
co-surfactant (which can compensate frustrated volumes between surfactant
headgroup and tail).
7
This makes it obvious that the spontaneous packing
parameter cannot be seen as a characteristic of an isolated surfactant but
instead is dened for an amphiphile in its actual environment.
Furthermore beside the geometries dened by the spontaneous curvature,
more unfavorable headgroup spacings and with that packing parameters can
be formed depending on the energy cost for that whereby so-called frustrated
2
structures are accessible.
7,8
1.2 Room Temperature Ionic liquids
Figure 1.2
Overview of dierent ions commonly used to design room tempera-
ture ionic liquids (RTIL). Reprinted with permission from Olivier-Bourbigou, H.;
Magna, L.; Morvan, D.
Appl. Catal., A
2010
,
373
, 156. Copyright 2010 Elsevier.
Room temperature ionic liquids (RTIL) are molten salts with a melting
point under or near room temperature (often
≤100◦
C is used as an arbi-
trary denition). This is usually realized by bulky organic ions, which render
crystallization dicult. In Fig. 1.2 an overview is given for a few common
building blocks used to design RTILs. The possibility of a modular combi-
nation of these or others plenty known ions makes a great pool of properties
accessible. By varying the used ions or chemical groups within them to for-
mulate RTILs, physical properties can broadly be varied, i. e., values for the
viscosity, melting point, glass transition temperature or conductivity can be
changed by decades which is nicely summarized in a review by Handy.
9
By in-
troducing functional groups which are known to work as catalysts, the whole
RTIL can be transformed into a solvent with an intrinsic catalytic property.
This was sucessfully demonstrated for example for Lewis acid catalyst, Henry
reaction, or the use of transition metals
1012
Many common ions can give a luminescent behaviour to the RTIL
13,14
4
partially with remarkable uorescence quantum yields.
15
Furthermore ionic
liquids can be categorized in aprotic and protic ones where the latter ones
contain one ore more functional groups with an acidic proton.
9
Finally by using lanthanides or transition metals in the formulation of
RTIL, a paramagnetic behavior can be introduced which was for example
demonstrated for iron,
1618
cobalt
19
or dysprosium
14
in various dierent ion
architectures.
The here discussed fullness of possibilities to vary the solvent's physical
and chemical properties which of course can only be briey touched in this
context makes it clear that RTILs are interesting for various applications as
solvent, lubricant, additive or catalyst.
2022
1.3 Self-assembly in non-aqueous systems
Next to water systems, self-assembly was also found in non aqueous solvents,
like glycol, formamide or DMF, which are highly polar liquids and possess a
high surface tension (like water and typically higher than 45-50mN/m)).
2326
This has made it necessary to generalize the concepts which were originally
developed specically for water systems, from hydro- to solvophobic/-philic
amphiphiles. Old concepts had to be extended to this new eld, e. g. the
above discussed packing parameter can in general be used to describe non-
aqueous systems as well but as its quantities (e. g. the headgroup area) are
strongly dependent on the surfactant environment, ndings are very solvent-
specic and have therefore to be newly tabulated for the new environment.
As the concept of the hydrophobic eect relies on a disruption of the solvent
structure by solvophobic moieties as the driving force for self-assembly, to ap-
ply this concept to non aqueous systems it is necessary to be able to quantify
the cohesive energy of a solvent. An old denition for this it the Hildebrand
solubility parameter (
δh
) which relates the cohesive energy between solvent
molecules to its enthalpy of vaporisation (
∆Hvap
).
27
δh=√∆Hvap −RT
vNA
(1.3)
5
Here
v
is the molecular volume and
NA
the Avogadro constant. As many
ionic liquids have no measurable vapor pressure this quantity is impossible to
measure. For this reason the, Gordon parameter is a more suitable concept,
which quanties the solvent quality for self-assembly.
28,29
and is dened by
eq. 1.4.
G=γ
3
√vNA
(1.4)
Here the cohesive energy is alternatively expressed by the surface tension
(
γ
) which is easily empirically accessible.
One quite dierent class of these water free solvents are ionic liquids (IL)
which can again be divided into protic and aprotic ones as already introduced
in section 1.2. A very famous class of the latter type consist of imidazolium
based ILs.
3032
and already more than 30 years ago the formation of micelles
in ILs by normal surfactants has been observed.
33
Although plenty of stud-
ies are already published on self-assembly in IL,
3437
most of them focused
on the comparison of many, often very dierent systems with a lack in full
details. Generally it is by now accepted that surfactant solvation eects
play a key role for micelle formation in ILs.
38
Typically it is observed that
aggregate formation is less pronounced in ILs, longer chain surfactants are
needed for micelle formation and the micelles formed exhibit a higher cur-
vature than they would have in aqueous solution, as for instance seen for
classical nonionic alkylethoxy (C
i
E
j
) surfactants in ethylammonium nitrate
(EAN).
39
Such nonionic surfactants typically show much more pronounced
tendency for micellisation than equivalent ionic surfactants.
40
The formation
of micelles by nonionic surfactants can easily be followed by methods like
tensiometry, dynamic light scattering, or small-angle scattering and these
methods could also show that for C
4
mim based ionic liquids the variation of
the counterion can have a pronounced eect on the aggregation behaviour.
41
In a systematic variation of the chain length of C
i
E
j
surfactants and mea-
suring their
cmc
's in bmimBF
4
by
1
H NMR it has been found that the
cmc
decreases exponentially with the chain length. From the thermodynamic pa-
rameters derived it could be concluded that this process is entropy driven in
6
a similar fashion as for the hydrophobic eect in aqueous solutions but the
solvophobic eect in bmimBF
4
is much weaker then in water.
42
That solvo-
phobic eect does not require the formation of a hydrogen bonded network
and accordingly is similarly observed in protic and aprotic ILs.
34
It might also be noted that not only micelles can be formed in ILs but
also the formation of liquid crystals,
34,43
vesicles
4446
or emulsions
47,48
has
been reported in some cases. The formulation of microemulsions can also be
widely found in literature. This has mostly been done for RTIL replacing
water
43,4957
but it is also possible that the RTIL functions as the hydropho-
bic component of the microemulsion.
5860
However, in any case the range of
applicable surfactant is much more restricted than in the case of water. The
formation of RTIL containing microemulsions has mostly been investigated
for bmim based ILs and with nonionic surfactants, where in particular TX-
100 has been shown to be quite eective.
50,61
For these nonionic surfactants
one can observe as a function of temperature and surfactant concentration
in the phase behaviour the classical Kahlweit-Fisch,
62
as for instance it has
been demonstrated for the case of C
14
E
4
and various alkanes.
57
For the case
of ionic surfactants one typically has to resort to adding a cosurfactant in
order to raise the solubilisation capacity of the hydrocarbon in the IL mi-
croemulsion (in this respect microemulsion in polar ILs behave similar to
ones in water). For instance this has been successfully done for C
16
mimCl
in bmimBF
4
, where the solubilisation of dodecane could be facilitated by
the presence of decanol as cosurfactant.
63
Similarly for a system of CTAB
in 1-ethyl-3-methylimidazolium hexylsulfate (emim hexSO
4
) the solubilisa-
tion of toluene could be increased substantially by the use of pentanol as
cosurfactant and one observes the percolation phenomenon as in similar mi-
croemulsions in water.
64
Here in the droplet regime also an increase of the
droplet size with increasing content of IL was observed. Such microemulsions
can be quite robust with respect to temperature changes, as demonstrated
for the case of C
16
mimCl in bmimBF
4
with decanol as cosurfactant and dode-
cane as continuous oil phase, where stability up to 150
◦
C has been shown,
49
obviously much higher than it can be achieved with similar water based sys-
tems.
7
So, there exists already quite a bit of knowledge regarding the formulation
of structured solutions in ILs but the situation is in general more complex
than in water as it depends also subtly on the precise type of IL employed.
Nonetheless, up to now systematic information regarding the aggregation
process of surfactants in ILs is still far from being satisfactory and deduc-
ing general information is dicult (as depending on its building blocks the
properties of ILs scatter over a very broad range).
1.4 Aim of this work
The aim of this work is to get a deeper insight into the mechanisms and
driving forces of surfactant self-assembly. For that purpose self-assembly is
not studied in water but in a nonaqueous model system based on a magnetic
room temperature ionic liquid (MRTIL) as solvent combined with an ionic
surfactant. This gives the possibility by a variation of the molecular struc-
ture of both solvent and surfactant to systematically vary its properties (i.e.
attraction potentials, solvophobicity/-philicity) and by that extract informa-
tion on the correlation between these parameters and the driving force for
self-assembly. It will be made use of a broad palette of complementary meth-
ods to guarantee a well balanced and reliable view on the system, nonetheless
small angle neutron scattering will be a central method.
In a rst step a detailed picture of this model system and its phase be-
havior and with that ability to form structures as micelles, microemulsions
and liquid crystals will be investigated. By this and by comparing with
other (classical aqueous) systems, qualitative and quantitative information
on self-assembly will be derived.
By choosing an ionic liquid with paramagnetic properties as solvent, this
opens the possibility to also investigate the magnetic behavior of the result-
ing mesoscopic structures. This will include rst the proof of principle if a
paramagnetic behavior is still present in these structures and then an inves-
tigation on how and if an external magnetic eld can trigger the systems
stability, shape and orientation.
Next to the more theoretical insights in the origin and driving force of
8
self-assembly which are already of high value on its own, the work can as well
point to some potential future application. As already mentioned above there
are already plenty of known applications for ionic liquids (e.g as catalyst or
custom build solvent) just as it is known for mesoscopic systems (e.g. as
nano-sized containers, structured reaction media, shaping template). By a
combination of those both one opens up the possibility to an extended eld
of application. Examples are mesoscopic systems with a temperature range
beyond the limit of 0100
◦
C caused by water or microemulsions as anhydrous
reaction media. In particular the MRTILs used in this study oer by their
contain of the transition metal ion iron the possibility to function as media
for catalysis or as solvent which can be manipulated by an external magnetic
eld interesting e.g. for separation chemistry.
9
10
2
Methods and Materials
2.1 Used Compounds
In this study 1-alkyl-3-methylimidazoliumtetraferrates (C
i
mimFeCl
4
) with
dierent chain lengths (i=2, 4, 6) were used as solvents. This ionic liquids
have, due to its iron containing anion, a paramagnetic behavior. As listed
in table 2.1 the calculated Gordon parameters (see eq. 1.4) for all three
solvents with dierent chain lengths are far lower than highly structured
water (2.7J/m
3
) but still above 0.5J/m
3
, the border under which no self-
assembly was found till now.
34
Accordingly these solvents have potential to
function as matrix for self-assembly but a lower driving force compared to
water is expected and a graduation with respect to the chain length.
As surfactant 1-alkylimidazoliumchlorides (C
j
mimCl) with dierent chain
lengths (j=12, 14, 16, 18) were used. By reason of its very similar head group
with respect to the solvent cation, it is expected to be well soluble and with
that the imidazolium group function as the solvophilic part of the molecule.
11
Table 2.1
Summarized properties of the MRTIL used as solvent in this study.
v
is the molecular volume calculated from the density,
G=γ
3
√vNA
is the Gordon
parameter and
Tm
is the melting temperature.
C
i
mimFeCl
4
ρ(25 ◦
C
)
gcm
−1
γ(25 ◦)
mNm
−1
v
nm
3
G
Jm
−3
Tm
◦
C
2 1.44 52 0.368 0.86 16
4 1.36 47 0.411 0.75 -88
a
6 1.30 42 0.466 0.64 -86
a
a
Glass transition, taken from literature
65
The structures of solvents and surfactants are given in Fig. 2.1.
Figure 2.1
Structures of the used compounds.
Top:
1-alkyl-3-
methylimidazoliumtetraferrates (C
i
mimFeCl
4
with i=2,4,6) as solvent.
Bottom:
1-alkyl-3-methyimidazoliumchlorides (C
j
mimCl with j=12,14,16,18) as surfactant.
The ionic liquids were synthesized analogously as described for the butyl-
derivative in literature
16
C
14
mimCl, C
16
mimCl and C
18
mimCl were synthe-
sized as described in reference.
63
C
12
mimCl was a gift from Prof. Werner
Kunz (Universität Regensburg). Its synthesis is documented in literature.
66
The quality of the resultant materials was veried by NMR, ESI-MS and
DSC.
Further used chemicals (oils and alcohols) to formulate microemulsions
and binary mixtures were used as purchased. A summary is given in table
2.2.
12
Table 2.2
Suppliers and grades of oils and alcohols as used to formulate microemul-
sions and binary MRTIL/alcohol mixtures.
oil supplier grade/%
hexane Fluka 98.5
octane Fluka 99.5
decane Fluka 98
isooctane Fluka 99.8
cyclohexane Sigma-Aldrich 99
alcohol supplier grade/%
1-propanol Merck 99.8
1-butanol Merck 99
1-pentanol Fluka 99.5
1-hexanol Sigma-Aldrich 99
1-heptanol Merck 99
1-octanol Sigma-Aldrich 99
1-decanol Merck 99
1-dodecanol Fluka 98
3,7-dimethyloctanol Fluka 98
geraniol Sigma-Aldrich 98
cis
-nerolidol Fluka 98
2-butoxyethanol Sigma-Aldrich 99
2.2 Recording phase diagrams by visual obser-
vation
2.2.1 Microemulsion systems
The pseudo ternary phase diagrams were recorded at a constant temperature
of 24
±0.5◦
C. Mixtures of dierent ratios between cyclohexane and surfactan-
t/cosurfactant were titrated with C
4
mimFeCl
4
. In all cases the molar ratio
of C
j
mimCl/decanol was 1:2. The pseudo ternary phase diagrams to deter-
mine the cosurfactant and oil inuence were recorded by the same titration
method but starting with a constant mass ratio of 86.4% cyclohexane in all
samples. For the cosurfactant dependency the ratio between C
16
mimCl and
13
alcohol was then varied. The phase boundary was detected visually. The
added relative amount of C
4
mimFeCl
4
is calculated with respect to the nal
mixture.
Kahlweit sh diagrams
62
were recorded at 24
±0.1◦
C by titrating samples
with a volume ratio of C
i
mimFeCl
4
/oil=1:1 and dierent surfactant concen-
trations with decanol or by repeatedly adding of solid surfactant. The added
amounts were measured gravimetrically. For details see the appendix B. It
should be mentioned that for the system C
2
mimFeCl
4
/C
18
mimCl at 24
◦
C
a precipitation of solid surfactant was observed after some time. As it was
possible to prevent this by rising the temperature by 1-2
◦
C and the phase
diagram showed nearly no temperature dependency in this range (the shtail-
point was shifted by
≈2
wt% between 24 and 45
◦
C), this was not considered
furthermore.
2.2.2 Binary MRTIL/alcohol mixtures
Dierent C
4
mimFeCl
4
/alcohol mixtures, sealed in glass ampoules together
with a magnetic stirrer were heated up under stirring (0.5
◦
C-steps, including
an appropriate holding time of several minutes). The sample clearing was
detected visually and interpreted as the phase boundary.
2.3 Dierential scanning calorimetry (DSC)
Measurements of binary MRTIL/surfactant mixtures as presented in chapter
3 were performed on a multi-cell calorimeter (TA Instruments MC DSC)
at heating/cooling rates of 0.2
◦
C/min. More concentrated samples were
before homogenized by alternating mixing with a spatula and heating several
times. Results were extracted after observing at least two equivalent runs.
Phase transition temperatures (peak position) and transition heats were read
out from the heating cycles after background (unloaded cell) subtraction.
For binary MRTIL/alcohol mixtures as presented in chapter 4.1 the phase
boundary was read out from cooling cycles performed at rates of 0.5
◦
C/min.
14
2.4 Conductivity
Conductivity titration was done at a temperature of
24.0±0.1◦
C with
a home-build Pt-electrode connected to a Methrom 712 conductometer at
2.4kHz, starting with an oil rich microemulsion sample by stepwise addition
of an oil free sample with a syringe.
2.5 Magnetic susceptibility
Magnetic susceptibility measurements were done at 300K with a MPMS
(Quantum Design, located at the Laboratory for Magnetic Measurements,
HZB). Samples were placed in a home-made vacuum-sealed sample cham-
ber and scanned in a range from -5 to 5Tesla. The empty cell signal was
subtracted from each measurement.
2.6 Viscosity
Viscosity measurements were done at
(25.0±0.1) ◦
C with a micro-Ostwald
viscosimeter (Ic or IIc, SI Analytics, Mainz). The obtained kinematic viscos-
ity
ν
was multiplied with the density
ρ
of the same sample to calculate the
dynamic viscosity
η
.
η=ρ·K·t=ρ·ν
(2.1)
Here
t
is the retention time needed for a dened sample volume and
K
is the capillary constant. For each sample at least three measurements were
done. The deviation of the retention times gives the error for the viscosity
and was
≤0.3 %
.
2.7 Small angle scattering
As small angle scattering relies on a complex and quite voluminous theory,
for a better understanding it is referred to textbooks found in literature.
67,68
15
In the following only the basic principles are discussed. An ideal scattering
experiment is represented in Fig. 2.2 were an incoming wave, represented
by the vector
ki
, strikes a scattering center in the sample which produces a
scattered wave (
ks
) dependent on the angle
θ
. For purely inelastic scattering
(which will be the assumption throughout this work) the energy of incoming
and scattered wave is equal (
|
ki|=|
ks|=2π
λ
) which leads to the denition of
the magnitude of the scattering vector
q
given by eq. 2.2.
q=|q|=|
ki−
ks|=4π
λsin θ
2
(2.2)
Figure 2.2
Schematic representation of a scattering experiment: The explanations
are given in the main text.
For an assembly of scattering centers in a sample, the amplitude
A(q)
of
the elastic scattered wave is furthermore related to the Fourier transform of
the scattering length density function
β(r)
, which denes the relative position
of the scattering centers to each other.
A(q) = ∫ ∫ ∫ β(r)eiqr
d
3r
(2.3)
Eq. 2.3 gives the relation between the inverse length scale
q
and character-
istic distances
r= 2π/q
within the sample. As neutron and X-ray scattering
experiments give access to a
q
-range of typically 0.018nm
−1
, this denes
that structures of one to a few hundred nanometer sizes are visible by this
technique. The measurable scattering intensity
I(q)
is dened as
I(q) = ⟨A(q)A∗(q)
V⟩∝⏐⏐⏐⏐∫ ∫ ∫ β(r)eiqr
d
3r⏐⏐⏐⏐
2
(2.4)
16
which is the squared scattering amplitude per sample volume
V
, aver-
aged over time, all possible congurations and orientations. Its amplitude is
dened by the scattering invariant
Qinv
.
Qinv =∫∞
0
I(q)q2
d
q
(2.5)
To solve eq. 2.4, assumptions have to be made how the sample is struc-
tured which will dene
β(r)
and
Qinv
. Depending on the model it is oftenly
useful to split the intensity into two terms, called the form factor
P(q)
and
the structure factor
S(q)
where the rst accounts for the scattering intensity
originated by one single object and the latter one by the interaction of an
assembly of several of these objects.
I(q)∝P(q)·S(q)
(2.6)
In this work small angle scattering experiments were done using neutrons
(SANS) and X-rays (SAXS) as incoming beam.
2.7.1 SANS
For better contrast the microemulsion samples were prepared with D12-cyclo-
hexane (For consistency mass fractions in the phase diagrams were recalcu-
lated to H12-cyclohexane with the same volume.) and for the measurements
placed in cuvettes (Quarz, Hellma) of 1mm thickness. To extract the scat-
tering intensity purely originated by the sample, the intensity, measured at
each pixel (
i
) of the detector plate at a given sample to detector distance (
d
),
has to be corrected due to eq. 2.7.
Ii,d =
I(S)i,d
T(S)i,d −I(BG)i,d
T(BG)i,d
I(R)i,d
T(R)i,d −I(BG)i,d
T(BG)i,d
·scaling factor
(2.7)
Here
T
is the transmission,
S
accounts for the sample,
BG
for the back-
ground which is the empty cell and
R
for a reference. As reference H
2
O
placed in a similar cuvette as for the sample measurements was used. As
H
2
O is expected to have a constant scattering over all angles it corrects in-
17
equalities in the detector pixel sensitivity. Additionally it brings the data
to absolute scale when using the correct scattering cross section for water
as the scaling factor. For each sample to detector distance
Ii,d
was radial
averaged and the data for dierent distances was merged to result the nal
q
dependent 1D-spectra. When not stated dierently, this data reduction was
done with BerSANS
69
and tting of the resulting 1D-spectra with SASt.
70
For all measurements the samples were placed in a temperature controlled
sample holder. As the SANS data were recorded at dierent facilities, their
special characteristics are listed in the following.
V4 at HZB (Berlin)
An incoming beam of 4.6Å wavelength at two detector distances (1.3 and
6m with collimation at 2 and 8m, respectively) was used, resulting in a
q
-
range of 0.09nm
−1≤q≤6.5
nm
−1
. The scattering of water was used to
correct the detector eciency and to bring the data to absolute scale. Under
these conditions data were recorded for microemulsion contrast variation as
presented in Fig. 5.14 (except the system C
2
mimFeCl
4
/C
18
mimCl).
D11 at ILL (Grenoble)
An incoming beam of 6Å wavelength and a collimation length of 8m was
used. Scattered neutrons were recorded with a 2D-detector for sample-
to-detector distances of 8m and 1.2m resulting in an observed q-range of
0.09nm
−1≤q≤5.12
nm
−1
. Incoherent scattering of water was used to
correct the detector eciency and to bring the data to absolute scale. Data
reduction was done with LAMP.
71
Under these conditions data were recorded
for microemulsions as shown in the appendix B.4.
SANSII at PSI (Villigen)
An incoming beam of 6Å wavelength was used. Scattered neutrons were
recorded for sample-to-detector distances of 1.2m and 5m resulting in an
observed
q
-range of 0.09nm
−1≤q≤3.04
nm
−1
. Incoherent scattering of
water was used to correct the detector eciency. The data were brought to
18
absolute scale by comparing with the scattering intensity of glassy carbon
72
for a sample-to-detector distance of 5m. Under these conditions data were
recorded for binary MRTIL/surfactant mixtures as presented in chapter 3.
SANSI at PSI (Villigen) with cryo magnet
As sample environment a cryomagnet (MA11, Oxford Instruments), which
allows to generate a horizontal magnetic eld up to 8T perpendicular to the
incoming beam, was used. An incoming beam of 5Å wavelength was used.
Scattered neutrons were recorded for sample-to-detector distances of 2m,
8m and 18m resulting in an observed
q
-range of 0.03nm
−1≤q≤2.1
nm
−1
.
The
q
-range is shortened due to a partial screening of scattered neutrons by
the magnet coils at high angles. Incoherent scattering of water was used to
correct the detector eciency.
To extract information about the anisotropy of scattering data as pre-
sented in chapter 5.3, the program SASET,
73
version 7.01.30, was used.
From the 2D-detector image recorded at 2m distance, a radial segment of
1.25nm
−1≤q≤1.43
nm
−1
was analyzed to extract three dierent values for
the anisotropy:
1. The principal component analysis (PCA),
2. the alignment factor, dened as
Af=∫2π
0I(q, ϕ) cos(2ϕ)
d
ϕ
∫2π
0I(q, ϕ)
d
ϕ
(2.8)
with
ϕ
as the azimuthal angle,
3. and the order parameter
S
, dened as
S=⟨P2(cos ϕ)⟩I/⟨P2(cos ϕ)⟩Im
(2.9)
⟨P2(cos ϕ)⟩I=∫π
0I(q, ϕ)P2(cos ϕ) sin ϕ
d
ϕ
∫π
0I(q, ϕ) sin ϕ
d
ϕ
(2.10)
with
P2(·)
as the second-order Legendre polynomial. It should be men-
tioned that in Fig. 5.24
⟨P2(cos ϕ)⟩I
is plotted instead of the order pa-
19
rameter because the normalization value
⟨P2(cos ϕ)⟩Im
for a perfectly
aligned sample is not known.
For details it is recommended to use the given literature and the software
documentation.
73
Paxy at LLB (Saclay)
Scattered neutrons were recorded for sample-to-detector distances of 5m
and 1m with an incoming beam of 4Å resulting in an observed
q
-range
of 0.1nm
−1≤q≤6
nm
−1
. Additionally for the short sample-to-detector dis-
tance the detector tube was shifted by an angle of 13
◦
relative to the beam
pathway to extend the observation window up to
q= 8.35
nm
−1
. Incoherent
scattering of water was used to correct the detector eciency and to bring the
data to absolute scale. Under these conditions data were recorded for binary
cyclohexane/decanol mixtures as presented in chapter 4.2, contrast variation
for the system C
2
mimFeCl
4
/C
18
mimCl (Fig. 5.14) and curves presented in
Fig. 5.13.
2.7.2 SAXS
Small angle X-ray scattering (SAXS) experiments were performed on an An-
ton Paar SAXSess using a CCD-Kamera as detector. Samples were prepared
as described for DSC and measured in sealed quarz capillaries (Hilgenberg,
outer diameter 1.0mm, wall thickness 0.01mm).
2.8 Surface tension
Surface tension measurements were done by the pendant drop method. This
method bases on the idea that the shape of a pendant drop is dened by the
force equilibrium between the surface tension (
γ
), expressed by the Young-
Laplace law, and the weight, resulting in eq. 2.11.
γ=1
2[( 1
RA1
+1
RA2)A−(1
RB1
+1
RB2)B]ρgh
(2.11)
20
Here
g
is the acceleration due to gravity,
ρ
the density dierence between
drop and environment and
Ri
are the radii to dene the curvatures at points
A and B placed along the droplet surface with a dierence
h
in height. Mea-
surements were done on a Contact angle System OCA 15plus (Dataphysics)
by using a needle with an outer diameter of 1.83mm (NE45, Krüss) and a
homebuild temperature control cell. All samples were measured with dier-
ent droplet sizes and values were extrapolated to large droplet volumes to
cancel out eects arising from needle tilting. The droplet prole was deter-
mined by the instrument software and surface tension values were corrected
by the calculated sample density which is in good agreement with experi-
mental values measured with Anton Paar densiometer. For details see Fig.
S6 in the ESI
†
.
2.9 Density
The density was measured with an Anton Paar density-meter (DMA 4500).
Calibration was done with deionized water. For the density
ρmix
of an ideal
mixture (constant molecular volume) the following expressions are valid:
1
ρmix
=∑Vi
∑mi=∑φm,i
ρi
(2.12)
ρmix =∑mi
∑Vi=∑φiρi
(2.13)
With
mi
,
Vi
,
φm,i
,
φi
and
ρi
being the mass, volume, mass fraction volume
fraction and density of the pure component
i
, respectively.
2.10 Polarized microscopy
Polarized microscopy for identication of liquid crystal (LC) phases was done
on a Zeiss microscope (12.5x/0.25 planchromat pol objective) using the same
samples prepared for DSC measurements. Temperature scans (1
◦
C/min with
a linkam TMS91 hot stage) and photographs (1photo/min with a Canon EOS
21
5D camera) were done remote controlled with home-build software.
2.11 Emulsion Stability
The magnetic eld eect on the emulsion stability as presented in section 5.3
was done with a home-build temperature control cell at
(25.0±0.3) ◦
C. A
magnetic eld was applied by a Brucker magnet B-E15. The eld prole was
measured with a Brucker B-H11D Hall probe.
22
3
Binary Mixtures: MRTIL/Surfactant
As this study concentrates on a selected pool of molecular architectures, in
this chapter just a variation of the alkyl chain of the cationic surfactant is in-
vestigated, but giving for these a detailed view over a large temperature range
and the full concentration range using dierent complementary methods. For
that purpose results of self-assembly in 1-ethyl-3-methylimidazolium tetra-
chloroferrate (C
2
mimFeCl
4
) and 1-butyl-3-methylimidazolium tetrachlorofer-
rate (C
4
mimFeCl
4
) as solvents are presented. With the surface tension of the
pure solvents Gordon parameters of 0.86 and 0.75J/m
3
(for C
2
mimFeCl
4
and
C
4
mimFeCl
4
, respectively) can be calculated. This is far lower than highly
structured water (2.7J/m
3
) but still above 0.5J/m
3
, the border under which
no self-assembly was found till now.
34
In order to study the aggregation conditions in these MRTILs in a sys-
tematic fashion we choose cationic surfactants with imidazolium chloride as
head group, which has a high structural similarity to the solvent. Low inter-
action potentials and good solubility for this part of the molecule is expected.
23
As solvophobic part of the molecule, aliphatic hydrocarbon chains (with 14,
16 or 18 carbon atoms) are used because hydrocarbon oils with six and more
C-atoms were observed to be insoluble in C
4
mimFeCl
4
at the investigated
temperature. Accordingly alkylimidazolium chlorides (C
j
mimCl, with j=14,
16, 18) are expected to be suitable as amphiphiles in our solvent. It might
be noted that for such C
j
mimCl surfactants calorimetric investigations have
shown negative enthalpies of micellization in EAN, while they are positive
for micellization in water at not too high temperatures.
7476
3.1 Results
As the aim of our work was the systematic comparison of the aggregation
behaviour of alkylimidazolium chlorides of dierent alkyl chain length in
C
2
mimFeCl
4
and C
4
mimFeCl
4
as representative magnetic room temperature
ionic liquids (MRTIL), we rst made a determination of the temperature
dependent phase behaviour. Based on that the individual phases were as-
signed structurally by polarized microscopy and SAXS. The
cmc
(critical
micelle concentration) was determined by surface tension measurements and
the structural details of the aggregates were deduced from SANS experiments.
3.1.1 Temperature dependent binary phase diagrams
Fig. 3.1 shows binary phase diagrams for all three surfactants in both
C
2
mimFeCl
4
and C
4
mimFeCl
4
as solvent which were constructed from DSC
and polarized microscopy measurements. As a common observation for all
systems we found a phase boundary separating a multi-phase region at lower
temperatures from homogeneous phases formed at higher temperatures which
increases in size with the surfactant concentration. While at low concentra-
tions the homogeneous phase is an (optically) isotropic phase, a birefringent
liquid crystalline (LC) region was found at high surfactant concentration,
showing typical textures for lamellar structures (see inset of Fig. 3.8). The
LC-lamellar phase is directly connected with the pure surfactant isopleth
(and not separated by another isotropic one-phase region) as known for sur-
24
Figure 3.1
Temperature dependent phase diagrams for binary mixtures of
C
14
mimCl (left column), C
16
mimCl (middle column) and C
18
mimCl (right col-
umn) mixed with C
2
mimFeCl
4
(top row) or C
4
mimFeCl
4
(bottom row). Open
symbols are transitions observed by DSC, lled symbols are transitions observed
with polarized microscopy. Dierent symbol shapes are used to group similar tran-
sitions to enhance the clearness. Crosses give the position of samples analyzed by
SAXS (see Fig. 3.8). Lines give a guide to the eye for equilibrium (straight) and
metastable (broken) transitions. Multi-phase regions are displayed as gray area (it
might be noted that upon approaching 100wt% there will be single-phase regions
again, but these regions are experimentally dicult to access and were not in the
focus of our investigation).
25
factants having a thermotropic behavior.
77
A more detailed structural inves-
tigation of the LC-phase can be found in section 3.1.3. Beside the lamellar
phase no other mesophases like cubic or hexagonal phases were found in
equilibrium. This observation ts to earlier studies on cationic surfactants
in IL.
78
Only for the system C
14
mimCl/C
4
mimFeCl
4
a metastable hexagonal
phase was observed by polarized microscopy during cooling cycles (for details
see Fig. A.12 in the appendix).
Focussing on the multi-phase region a Krat discontinuity (at middle
concentrations) or Krat boundary (at lower concentrations) was observed
which is comparable to ndings for water systems.
77
A characteristic drop-
ping down while going to lower concentrations, known as the Krat knee
can be found which is much more pronounced and at higher concentrations
for the shorter chain surfactant systems which gives already a rough esti-
mate for the critical micellization concentration (
cmc
). Analyzing the sum
DSC-integral over all transition peaks for each sample gives enthalpy-values
of around 5-6kJ/mol per CH
2
unit of the surfactant chain (see table A.4,
Fig. 3.2), a value very common for melting alkyl chains.
79,80
With that the
multi-phase region can be interpreted as solid surfactant in equilibrium with
monomeric, micellar, or liquid crystal phases, separated by isothermal phase
boundaries.
The system C
14
mimCl/C
4
mimFeCl
4
shows by DSC an additional meta-
stable region at low surfactant concentrations which can be interpreted by as
due to a kinetically hindered transport of MRTIL molecules towards surfac-
tant crystals. The reason why this is only observed here could be a coinci-
dence of high solvent viscosity, low temperature and relatively high surfactant
concentrations compared to the other systems which favors a slow transport
kinetics. For C
2
mimFeCl
4
as solvent additional transitions are observed due
to the solvent's higher melting point of 18
◦
C showing a freezing-point de-
pression by adding surfactant. C
4
mimFeCl
4
has a much broader liquid range
down to a glass transition at -88
◦
C
65
and therewith has no such transitions
within our observation window. The solid surfactant phase could be dry sur-
factant or any kind of solvated crystals. Am indicator for the latter could be
the missing transitions at high surfactant concentrations/low temperatures
26
Figure 3.2
Integral over all peaks (excluding the low temperature peaks for the
C
2
mimFeCl
4
-system originated by the solvent melting point, see Fig. 3.1) extracted
from DSC-curves given in the appendix A.4 and normalized by the surfactant con-
centration for C
14
mimCl (squares), C
16
mimCl (circles) and C
18
mimCl (triangles).
The lines give the mean values from whose mean distance the heat per surfactant
CH
2
-unit can be deduced as dQ/d(CH
2
)
≈
5-6kJ/mol, a value common for alkyl
chain melting.
(but this could also be interpreted by diculties to reach equilibrium in a ki-
netically slow solid
→
solid transition). The pure surfactant is known to have
dierent solid structures (with the triclinic double bilayer as the most sta-
ble one) whose appearance and transformation strongly depends on thermal
history and the amount of additional solvent.
81
Nevertheless it is dicult to
elucidate this issue and this is not the subject of the actual study.
In both solvents the same trend can be observed, i.e., with longer sur-
factant chains the lamellar region is getting wider, shifted to lower concen-
trations, the Krat knee is shifted to lower concentrations which allows a
rst estimation in amphiphilic strength in the order C
18
mimCl
>
C
16
mimCl
>
C
14
mimCl as to be expected. An extraction of characteristic values is given
in table 3.1 and will be furthermore discussed in section 3.2. In Fig. 3.3
the position of the triple point at which lamellar LC, isotropic solution and
solid surfactant are in equilibrium is plotted and shows exemplary the men-
27
tioned dependency on surfactant chain length. The temperature increases
with chain length as the whole phase diagram is shifted to higher tempera-
tures while the mass fraction of surfactant (
φm,LC
) decreases as the LC region
is extended to lower surfactant concentrations.
Figure 3.3
Position by means of temperature (top) and composition (bottom) of
the triple point at which lamellar LC, isotropic solution and solid surfactant are in
equilibrium as a function of the surfactant chain length. Values are extracted from
Fig. 3.1
3.1.2 Surface tension measurements
As surface tension is a quantity that is intimately linked to the aggregation
behaviour of amphiphilic systems it was also studied for our systems. It can
be noted that the surface tensions of the pure ILs are 50 and 45mN/m for
C
2
mimFeCl
4
and C
4
mimFeCl
4
, respectively, substantially lower than that of
water and at the border where one may still expect the formation of aggre-
gates. A rst estimate of the
cmc
can be inferred from DSC measurements
as discussed in section 3.1.1. To investigate the herein presented system on
its aggregation behavior with common methods known from water systems
28
Figure 3.4
Surface tension measurements at 45
◦
C for binary mixtures of
C
14
mimCl (cubes), C
16
mimCl (circles) and C
18
mimCl (triangles) in C
2
mimFeCl
4
(top) or C
4
mimFeCl
4
(bottom) as solvent. Lines are ts performed with eq. 3.1a.
raises some practical problems: As UV and visible light is strongly absorbed
by the iron ions
82
a very low transmission prohibits a
cmc
determination by
light scattering, turbidity measurements or uorescence/ dye uptake meth-
ods. Its high conductivity (20 and 9mS/cm for C
2
- and C
4
mimFeCl
4
)
65
makes the conductivity method unfavorable. Expected
cmc
values at rela-
tively high concentrations give distractingly high diluting enthalpies for the
isothermal titration calorimetry (ITC) method.
Accordingly we turned to surface tension measurements and Fig. 3.4 shows
results as a function of surfactant concentration in C
2
mimFeCl
4
/C
j
mimCl
29
(top) and C
4
mimFeCl
4
/C
j
mimCl (bottom) binary mixtures. All measure-
ments were done at 45
◦
C to have monophasic samples for each system.
Adding surfactant decreases the surface tension and a (more or less pro-
nounced) discontinuity leading to a plateau (or less steep slope, depending on
amphiphilicity) in surface tension vs. concentration can be observed. Qual-
itatively two trends can be noticed: Shortening the solvent's alkyl chain or
lengthening the surfactant's one leads to a more pronounced knee which is
located at lower concentrations. This can be interpreted as an enhancement
in amphiphilic strength of the system by these molecular changes which is
congruent with observations in section 3.1.1.
Compared to strong amphiphiles in water, which usually give a very sharp
transition between the surface tension decrease and the plateau (which makes
the extraction of the
cmc
value at this knee straightforward) in our system
the nding is less clear. To extract quantitative information from the data
points nonetheless, surface tension curves were tted with a modication of
the Szyszkowski equation:
83,84
γ=γ0−RT ·Γ·ln [1 + K1·φm,b ·f]
(3.1a)
f= exp (b·φm,tot)
(3.1b)
φm,tot =K2·N·φN
m,b +φm,b
(3.1c)
with the monomer mass fraction in equilibrium (
φm,b
) calculated by a
simple mass action model (N[monomers]
[Micelle], leading to eq. 3.1c).
φm,tot
is the total surfactant mass fraction. To correct the nonideal behaviour
especially at high concentrations a coecient
f
was introduced expressed by
a simple exponential function (eq. 3.1b). This could be interpreted as an
activity coecient as it has a similar form as found for activity coecients in
mixed salt solutions.
85
Fit results are listed in table A.5 and in Fig. 3.5 the
concentration dependent evolution of
φm,b
and
φm,mic
vs.
φm,tot
is plotted.
From that the initial appearance of micelles was interpreted as
cmcγ
(for
details see appendix A). Apparently eq. 3.1a is describing the experimental
situation rather well and the extracted
cmc
values describe a quantitative
30
trend as a function of the surfactant chain length which conrm the trend
observed with the other methods. This model also yields that at concen-
trations above the
cmc
there is still a moderate increase of the molecularly
dissolved surfactant (Fig. 3.5).
Figure 3.5
Values for monomeric surfactant (open symbols) and surfactant in
micelles (lled symboles) derived from the modied Szyszkowski model. The cmc
γ
was determined by linear regression (broken line) of data points in the linear regime
at higher concentration and extrapolation to
φm,mic
=0.
Alternatively in order to quantify the eciency of dierent surfactants
in the same solvent the surfactant concentration needed to lower the surface
tension by 10mN/m from the value of the pure solvent (
c10
) can be extracted.
All parameters are summarized in table 3.1 and show that the eciency of
31
the surfactant is increasing by about a factor 1.5-2 per 2 CH
2
-groups, i.e., the
hydrophobic eect per CH
2
group is less than half the one in water,
2
which is
also reected in dierent CH
2
transfer energies for water and MRTIL which
are given in table 3.2 and will be discussed later in section 3.2.2.
3.1.3 Small angle scattering
Micellization - SANS:
Figure 3.6
Small angle neutron scattering curves for the system C
18
mimCl in
C
2
mimFeCl
4
. Open symbols are data measured at 45
◦
C. Concentration increases
gradually from blue to red. Lines are ts according to eq. 3.4a.
Due to the knowledge gained from DSC and surface tension measure-
ments, small angle neutron scattering (SANS) experiments were carried out
in the low and mid concentration range to prove the existence of aggregates
and characterize them structurally. SANS was necessary as the X-ray high
absorbance of iron renders SAXS not a useful technique here. Fig. 3.6 gives
the resulting curves exemplary for the system C
18
mimCl in C
2
mimFeCl
4
(for
the complete set of curves see Fig. A.2 in the appendix). With the enhanced
32
Table 3.1
Summary of the values extracted from surface tension (
cmcγ
and
c10
), SANS invariant (
cmcinv
), aggregation
number (
cmcagg
), DSC (
φm,LC
as the lowest concentration at which a LC phase occurs,
Tm,surf
and
TKrafft,disc
,
∆T=
Tm,surf −TKrafft,disc
) and SAXS measurements (
d
and
∆d
). For comparison values for comparable aqueous systems found
in literature are listed.
C
i
mimFeCl
4
C
j
mimCl
cmci(45 ◦
C
)
wt%
extracted from Fig. 3.1
d
nm
∆d
nm/CH
2
i=γ i =inv i =agg c10
wt%
φm,LC
wt%
TKrafft,disc
◦
C
Tm,surf
◦
C
∆T
◦
C
75
◦
C 100
◦
C 75
◦
C 100
◦
C
2 14 11.66 9.00 19 6.36 68 19 53 34 3.478 3.406
↑ ↑
2 16 6.75 6.50 13 3.51 65 36 64 28 3.746 3.657 0.141 0.131
2 18 4.90 4.00 8 1.91 61 43 70 27 4.043 3.932
↓ ↓
4 14 22.51 13.00 25 21.96 76 16 53 37 3.355 3.288
↑ ↑
4 16 15.76 9.60 18 15.44 69 29 64 35 3.597 3.521 0.129 0.122
4 18 8.09 6.20 12 9.60 66 36 70 34 3.870 3.775
↓ ↓
H
2
O 10 1.03
86
1.05
86
a
H
2
O 12 0.38,
86
0.35
66
0.39
86
a
H
2
O 14 0.09,
86
0.11
66
0.12
86
a
H
2
O 16 0.03
86,87
0,03
66,86
a
a
measured by conductivity method at 25
◦
C
33
scattering intensity in the observed
q
-range the existence of some kind of
aggregate structure or at least aggregate-like density uctuations is already
proven.
Qualitatively all data sets have a similar appearance. With higher sur-
factant concentrations the scattering intensity grows due to more scattering
aggregates in the sample. This as well enhances the interactions between
aggregates which leads to a correlation peak. The increasing scattering in-
tensity at low
q
-values could be due to attractive interactions but potentially
also to errors in background subtraction. As a rst approach to get quan-
titative information on the ratio of monomerically and micellar dissolved
surfactant with a model independent approximation the experimental scat-
tering invariant
Qex
inv
was compared with expectations calculated from the
sample composition. Another advantage of this method is its nearly inde-
pendence from very low
q
values which might bare some artifact problems as
mentioned above. Errors are mainly produced by the Porod extrapolation
and were estimated to be less than 10 % by choosing the data range for the
extrapolation carefully. The experimental invariant is dened as
Qex
inv =∫+∞
0
I(q)·q2
d
q
(3.2)
were the discrete data points were extrapolated to zero and
+∞
with
standard methods by Guinier and Porod approximations, respectively. A
theoretical two-level invariant
Qth
inv
was calculated by considering having a
bulk phase composed by the solvent with a quantity
cmcinv
of monomeric dis-
solved surfactant plus all the surfactant head groups (-mimCl) and a second
34
phase composed by the remaining surfactant hydrophobic tails as
Qth
inv = 2π2φbulk ·φmic ·∆SLD2
(3.3a)
φbulk = (mIL
ρIL
+msurf,b
ρsurf
+msurf −msurf,b
a·ρmimCl
)/Vtot
(3.3b)
φmic = 1 −φbulk
(3.3c)
cmcinv =msurf,b
msurf,b +mIL
(3.3d)
with
mi
as the masses of MRTIL (
i=IL
), total surfactant (
i=surf
) and
surfactant in the bulk (
i=surf, b
), and
ρi
as the corresponding densities.
Vtot
is the resulting total volume of the sample and
1/a
the mass ratio of mimCl
in C
j
mimCl. The volume ratios
φbulk
and
φmic
and the scattering length
density dierence (contrast)
∆SLD
were calculated assuming ideal mixing.
(For details see appendix A). As shown in Fig. 3.7 (top) it gives a quite
good agreement with the experimental data with the mass ratio of surfactant
dissolved in the bulk (
cmcinv
) being the only free parameter. Only at very
high concentration the experimental invariant has appreciable lower values.
This could be explained with a general failure of the simple assumption of
a xed
cmc
but an increasing monomeric surfactant concentration which
would lower the scattering contrast and with that
Qth
inv
. Such a scenario is
also indicated by the fact that the surface tension still decreases beyond the
cmc
(Fig. 3.4).
To get more detailed information on the aggregation behavior, model
tting was applied to the data with a spherical form factor
P(R)
. As most
of the samples had high concentrations, a structure factor was necessary to
simulate the correlation peak mentioned above. For this reason a hard sphere
structure factor
S(RHS, φHS)
88
was introduced and implemented in the local
35
Figure 3.7 Top:
Experimental (symbols) and theoretical (lines) invariants for
all systems.
Bottom:
Aggregation numbers derived from spherical model t.
In both diagrams binary mixtures of C
14
mimCl (cubes), C
16
mimCl (circles) and
C
18
mimCl (triangles) in C
2
mimFeCl
4
(open symbols, broken lines) or C
4
mimFeCl
4
(lled symbols, solid lines) as solvent are shown. Data can be found in table A.3.
36
monodisperse approach (eq. 3.4a).
I(q) = BG +∫LN(R0, σ)·P(R)·S(RHS, φHS)
d
R
(3.4a)
RHS =R+ ∆R
(3.4b)
φHS =φmicelle
(R+ ∆R)3
R3
(3.4c)
By calculating the micelle volume fraction
φmicelle
and scattering contrast
from the sample composition the nal set of t parameters was reduced to
the width (
σ
) and location (
R0
) of the size distribution (
LN(R0, σ)
), the
eective interaction radius of the hard spheres
∆R
, the volume ratio of
MRTIL in micelles (
α
), the ratio of surfactant in bulk phase (
y
) and the
background
BG
. It might be noted that here we imply that also some of the
solvent (MRTIL) can be contained within the micellar core, which is quite
a dierent situation compared to water based micelles. However, due to the
much weaker amphiphilicity of the surfactants here, this might be a realistic
scenario. A detailed description of the model tting is given in the appendix
A. From these the aggregation number (
Nagg
) can be extracted by comparing
the aggregates mean volume with the surfactant hydrocarbon chain volume
vHC
:
Nagg =4
3π⟨R3⟩1
vHC
(1 −α)
(3.5)
Here
⟨R3⟩
is the 3rd moment of the size distribution. Results are shown in
Fig. 3.7 (bottom). From the plot one can extract an aggregation concentra-
tion (
cmcagg
), where after an initial nearly constant single-digit aggregation
number at low concentration the aggregation number rises. Apparently the
aggregation here proceeds via a broad range of pre-aggregation and therefore
in not such a sharp fashion as typically observed in aqueous systems. Several
other SANS models have been examined and led to a qualitatively similar
result for the aggregation number (see appendix A.1, table A.3). In general,
the aggregation numbers increase with increasing length of the alkyl chain of
37
the surfactant and are higher in C
2
mimFeCl
4
compared to C
4
mimFeCl
4
as
solvent.
Figure 3.8
SAXS curve of 85wt% C
14
mimCl in C
2
mimFeCl
4
. Arrows indicate the
peak positions at
q=n·1.84
nm
−1
, characteristic for lamellar structures. The inset
shows a representative polarized microscopy image of the same system at 75
◦
C.
The lamellar Phase - SAXS:
The liquid crystal region found in all six systems shown in Fig. 3.1 were fur-
thermore characterized by polarized microscopy and small angle X-ray scat-
tering (SAXS). Due to low X-ray transmission of the iron containing solvents,
samples with solvent ratios above
≈
25wt% cannot be measured with the used
setup. Therefore samples at surfactant concentrations of 85wt% were inves-
tigated and are marked in Fig. 3.1 with crosses. Fig. 3.8 gives the SAXS
curve for the system C
2
mimFeCl
4
/C
14
mimCl (for others see appendix A).
From the periodicity of the correlation peaks appearing at
q=n·1.84
nm
−1
and the characteristic pattern (maltese crosses) in polarized microscopy the
liquid crystals can clearly be identied as lamellar structures. Due to limita-
tions in the temperature range oered by the SAXS machine it was dicult
38
to heat all measured samples to the isotropic state followed by a slow cool-
ing to the desired liquid crystal state before scattering experiment to ensure
a well ordered sample. Due to this limitation only the scattering curve of
the here shown C
2
mimFeCl
4
/C
14
mimCl system with a LC
→
isotropic transi-
tion at relatively low temperatures at 85wt% C
14
mimCl shows higher order
peaks. Nevertheless all six systems could be identied by the characteristic
polarized microscopy textures.
Figure 3.9
Lamellar spacing
d= 2·π/qmax
extracted from the peak position
qmax
from SAXS measurements. Solid lines are giving linear regressions.
From the rst correlation peak position, the periodic length
d= 2·π/qmax
can be calculated as given in table 3.1 and is plotted in Fig 3.9 for the dierent
surfactants at 75 and 100
◦
C. The change in
d
as a function of surfactant chain
length (as derived by linear regression, plotted in Fig. 3.9, listed in table
3.1) gives a value of 0.12-0.14nm per CH
2
group which is very close to the
projection of one C-C-bond on the chain axis suggesting an interpenetration
of the surfactant chains in a bilayer. This is in a good agreement with
similar structures found in IL systems.
81,8991
Changing the MRTIL from
C
2
mimFeCl
4
to C
4
mimFeCl
4
leads to a reduced value for the domain size
d
39
which is true for both temperatures. This can be due to a better solubility of
the hydrocarbon chains in the MRTIL with the longer alkyl chain. This then
would promote a deeper penetration of the solvent into the surfactant lamella
leading to a bigger headgroup area which would force the lamella to become
thinner (because the overall amount of surfactant chain stays constant). This
interpretation is supported by surface tension and SANS results which both
suggest a bigger headgroup area while using C
4
- instead of C
2
mimFeCl
4
as
solvent (see table A.5 and Fig. A.1, respectively). Increasing the temperature
as well reduces the size of
d
which can be interpreted analogously.
3.2 Discussion
3.2.1 Low and mid surfactant concentrations - critical
aggregation conditions
The
cmc
values derived from the three dierent methods are rather scattered
which indicates that the micellar transition is not so well-dened but instead
occurs over a rather broad concentration range. This phenomenon is also
known for water systems, especially for the case of rather low aggregation
numbers.
92,93
In our system these deviations are rather broad as the relatively
weak amphiphilicity leads to a correspondingly broad transition region of
aggregate formation. The here obtained surface tension curves show similar
critical aggregation concentrations compared to much shorter (C
6
or C
8
) sur-
factants in water. In contrast at very high concentrations these systems show
a surface tension plateau and not a smooth decrease after the
cmc
as observed
in the here presented MRTIL systems.
9395
For all systems the
cmcinv
derived
from the SANS invariant has always values lower than
cmcagg
(from the sur-
face tension) insofar as the scattering invariant is independent of the aggre-
gate size or shape and detects therefore already the very small pre-aggregates
formed by just a few molecules. Depending on the amphiphilic strength of
the system (which follows the order C
18
mimCl
>
C
16
mimCl
>
C
14
mimCl and
C
2
mimFeCl
4
>
C
4
mimFeCl
4
)
cmcγ
lies closer to
cmcinv
or
cmcagg
. With in-
formation from these complementary methods we can draw a picture of a
40
broad transition range with a concentration dependent aggregation, having
small pre-aggregates at lower concentration and continuously growing aggre-
gates with growing surfactant concentration. Additionally the SANS model
tting suggest that the micelles are partly (
∼
30-40vol%) swollen by the sol-
vent. The ratio of solvent in the micelle continuously declines with surfactant
concentration approaching zero at a concentration around the phase separa-
tion boundary at high concentration (see Fig. A.1). This observation can
be related to the continuously increasing average aggregation number (Fig.
3.7) and only for the high aggregation numbers achieved towards the phase
boundary really rather compact micellar aggregates can be formed that do
not contain some of the solvent.
Table 3.2
Free enthalpies of micellization (
∆Gmic
) and transfer energies for a
CH
2
-unit into the solvent (
∆GCH2
), calculated from
cmc
and
c10
values. Here
cmc
and
c10
values are used in units of mole fractions. The last column shows the mean
value of all four previous columns. For comparison values for comparable aqueous
systems found in literature are listed.
C
i
mimFeCl
4
C
j
mimCl
∆Gmic=RT ln cmci
RT
mean value
i=γ i =inv i =agg ln c10
∆GCH2
RT
2 14 -2.17 -2.43 -1.68 -2.77
↑
2 16 -2.79 -2.83 -2.13 -3.45 0.27
2 18 -3.19 -3.40 -2.70 -4.14
↓
4 14 -1.44 -1.98 -1.34 -1.46
↑
4 16 -1.86 -2.36 -1.73 -1.88 0.24
4 18 -2.60 -2.87 -2.21 -2.43
↓
H
2
O 10 -12.56
86
a
↑
H
2
O 12 -12.77
87
a
, -14.73
86
a
1.25
86
a
H
2
O 14 -15.24
87
a
, -17.31
86
a
1.22
87
a
H
2
O 16 -17.64
87
a
, -20.04
86
a
↓
a
calculated from data measured with conductivity
method at 25
◦
C considering a partial degree of dis-
sociation of micelle counter ions.
41
3.2.2 Quantitative results on solvent quality for self-
assembly
An observation which can be made by all used techniques is an amphiphilic
character much lower compared to similar water systems (compare table 3.1
and 3.2): Much higher concentrations are needed to form aggregates, and
a relatively small lamellar phase is found, which is the only representative
of LC-phases in these systems. For water the LC region is broader and
includes more diversity e.g. an additional hexagonal phase.
96
In general
the class of surfactants under investigation has a suppressed amphiphilic
behavior in the IL solvents investigated here compared to water. To nd the
reason it is useful to consider separately the solvophilic and -phobic character.
Qualitative information on the solvophilicity gives the
∆
T value which is the
dierence between surfactant melting point and Krat discontinuity. As in
both transitions the surfactant melts the value gives the compensation of
thermal energy by solvent attraction and can therefore directly be related
to the solvophilicity (deviations originating from the appearance of solvated
solids in the phase diagram are not considered here). Values for
∆
T extracted
from Fig. 3.1 are listed in Table 3.1. As expected the value is enhanced by
shorter surfactant chains and larger alkyl substituents in the solvent molecule.
The absolute values of around 3040
◦
C are comparable with similar water
systems
97
and for that certify a good solvophilicity.
On the other hand quantitative values on the solvophobicity can be ex-
tracted from the
cmc
values. The free enthalpy of micellization (
∆Gmic
) is
proportional to
ln cmc
and furthermore the variation of
∆Gmic
by the surfac-
tant's chain length gives the transfer energy for a CH
2
unit into the solvent
(
∆GCH2
) and therewith a quantity for its solvophobicity. Table 3.2 gives
the calculated values for both solvents. As expected C
2
mimFeCl
4
gives a
slightly higher value as its alkyl chain is shorter and therefore diers more
strongly from the surfactants than is the case for C
4
mimFeCl
4
. Compared to
water, where values of
1.2−1.4
RT are usual,
66,86,98
this is much smaller and
gives a quantitative value to explain the lower driving force to self-assembly.
Polypropylene oxide (PO) is an example for a hydrophobic unit having a
42
transfer energy of 0.15RT per PO unit into water
92
which is even less than
the here found value for the IL-systems. This ts perfectly to the observation
that PO units are normally used as long-chain polymer units and that the
aggregate core is swollen with water
99,100
in the same manner as here the
SANS model indicates micelles swollen with MRTIL.
For that reason it seems to be useful in future work to concentrate on
the solvophobic part of the molecule to enhance its amphiphilic character.
By introducing disturbances into the alkane chain like branching, or sub-
stituents like ester groups or unsaturated functions the crystalline ordering
of the hydrocarbon chain will be suppressed and the Krat discontinuity (as
well as the melting point of pure surfactant) is expected to be shifted to
lower temperatures. With that hidden mesoscopic phases (as observed in the
C
14
mimCl/C
4
mimFeCl
4
-system) could be discovered as thermodynamically
stable states. As aggregates in the isotropic phase region and the LC as well
contain uid chains they are expected to be not eected by these actions.
3.3 Conclusion
In this chapter a systematic study on imidazolium-based surfactant in param-
agnetic ionic liquids of the type C
i
mimFeCl
4
was done. By using a variety of
methods its phase and aggregation behavior was investigated. SANS exper-
iments clearly demonstrate the formation of aggregates at higher surfactant
concentrations but with rather low aggregation numbers. These micelles are
partly swollen by solvent and this process occurs over a rather broad con-
centration range. They become larger and contain then less MRTIL solvent
with increasing concentration. The tendency for micellization is higher the
longer the alkyl chain of the surfactant and the shorter the alkyl chain of
the solvent. At high concentration the only LC phase found is a lamellar
phase. The ability of long chain C
j
mimCl to self-assemble was proven and
evaluated compared to common water systems, where the solvophobic eect
of the alkyl chain in the MRTIL was determined to be only about a fth of
that in water.
From our analysis we were able to distinguish between the eects of the
43
solvophobic and -philic part of the surfactant. As a result it was quantita-
tively shown that decits in the ability to self-assemble are mainly present
in the surfactant's lipophobic tail which gives a good ngerpost on how to
enhance the tendency for self-assembly for these kind of systems in future
work.
44
4
Other Binary Mixtures
Compared to binary mixtures, microemulsions are self-assembled system with
a higher complexity, indicated by its higher number of components. To dis-
cuss the driving force of self-assembly in these kind of systems it is helpful
to know the location of all species in a mixture within the microstructure.
As a preliminary work one can consider isolated binary mixtures as they
give information about interactions between two species in an environment
easier to analyze compared to multi-component systems. In chapter 3 it was
already focused on the surfactant in binary mixtures and a result was a rea-
sonable amount of monomerically dissolved surfactant molecules within the
MRTIL-solvent dependent on surfactant and MRTIL chain length.
In chapter 5 four-component microemulsion systems including an alcohol
as cosurfactant will be discussed and as a preliminary work for this, this
chapter focuses on the interaction of the co-surfactant with both continuous
phases present in the microemulsions: the MRTIL and the oil.
45
Figure 4.1
Solubility curves for binary mixtures C
4
mimFeCl
4
/C
i
OH (colored sym-
bols) and C
2
mimFeCl
4
/C
10
OH (black symbols). Open symbols are data obtained
by visual observation, lled symbols are data obtained by DSC measurements, solid
lines are ts via
RT ln x=−∆Gmix =T∆S−∆H
, broken lines are theoretical
values obtained by extrapolating the measured results.
4.1 Binary Mixtures MRTIL/alkanol
Fig. 4.1 shows the solubility curves for dierent n-alkanols in C
4
mimFeCl
4
.
To extract quantitative values for the interaction between alcohol and ionic
liquid, the free enthalpy of mixing (
∆Gmix
) was calculated using the mole
fraction of alcohol (
x
). As already discussed in section 3.2.2, the dependency
of the free enthalpy with respect to the alcohol chain length gives directly
the transfer energy for one CH
2
-unit into the MRTIL given by eq. 4.1.
∆Gmix = ∆GCH2·i+ ∆Geg
(4.1)
Here
i
is the number of CH
2
-groups and
∆Geg
is the part of the free en-
thalpy of mixing which is originated by the endgroups OH and CH
3
in the
alcohol molecule. A calculated value for 45
◦
C is listed in table 4.1 and is com-
parable to the value measured for binary MRTIL/surfactant mixtures (see.
table 3.2). As expected the process of mixing is entropy driven as indicated
by positive
∆S
values and a cost in energy (positive
∆H
values). In Fig.
46
4.1 additionally the solubility curve for decanol in C
2
mimFeCl
4
is shown. As
expected the solubility is lower compared to C
4
mimFeCl
4
. An extrapolation
of the data for the shorter chain alcohols in C
4
mimFeCl
4
to the theoretical
curves for C
11
OH and C
12
OH shows that the solubility in C
2
mimFeCl
4
is
shifted by two CH
2
-units compared to C
4
mimFeCl
4
, an observation which
was as well also found for dierent quantities (e. g.
cmc
,
φLC
) in the binary
MRTIL/surfactant mixtures (see chapter 3).
As
n
-alcohols are regarded to function as cosurfactant in microemulsion
formulation it is important to see that there is only little solubility of de-
canol in the ionic liquid (around 1.7wt% and 2.8wt% for C
2
mimFeCl
4
and
C
4
mimFeCl
4
at 20
◦
C, respectively). Furthermore it is possible that decanol
is not dissolved monomerically but organizes in aggregates which would make
it even more likely to nd the decanol in the interface when one is oered, as
it is the case in a microemulsion. For these reasons the solubility of decanol
in the MRTIL was subsequently neglected.
Table 4.1
Free enthalpies of mixing (
∆Gmix
) and transfer energies for a CH
2
-unit
into the solvent (
∆GCH2
), calculated from the t parameters
∆S
and
∆H
.
C
i
mimFeCl
4
C
i
OH
t parameter mean value
∆S
J(Kmol)
−1
∆H
kJ(mol)
−1
∆Gmix(45 ◦
C
)
RT
∆GCH2(45 ◦
C
)
RT
4 4 28.4 10.3 0.46
↑
4 5 32.9 12.4 0.73
4 6 30.8 12.5 1.03 0.29
4 7 41.1 16.7 1.35
4 10 38.8 18.2 2.23
↓
2 10 39.5 19.9 2.78
47
4.2 Binary mixtures cyclohexane/decanol
With the oil (cyclohexane) decanol is miscible over a broad range of com-
position and therefore it has to be considered that maybe the decanol is
monomerically dissolved in the oil instead of a location in the interface which
would change a lot in the interpretation of microemulsion phase behaviour.
A starting point is the investigation of binary mixtures of decanol with cy-
clohexane.
Fig. 4.2 gives SANS spectra of dierent decanol concentrations in D12-
cyclohexane. As the incoherent scattering scales with the H/D-ratio in the
sample which leads to very dierent background levels for each concentration,
the background was subtracted to make the samples better comparable. For
this purpose the background (
BG
) for each curve was calculated by the sum
of the curve for pure D12-cyclohexane and 30wt% decanol/H12-cyclohexane,
weighted by the decanol volume fraction as given by eq. 4.2.
BG = (1 −φdecanol)·I(q)D12−cyclo +φdecanol ·I(q)H12−cyclo/decanol
(4.2)
To get a rst estimate for the size of the scattering objects, a Guinier plot
was done giving the radius of gyration
RG
and the zero angle intensity
I(0)
:
I(q) = I(0) exp −R2
Gq2
3
(4.3a)
I(0) = V·φ·∆SLD2
(4.3b)
The obtained results are listed in table 4.2. First of all it can be seen
that
RG
is similar for all four decanol concentrations. The sample at 20wt%
deviates from the Guinier-law at low
q
and this behaviour is even more
pronounced for the sample at 30wt%. This can be explained by interactions
between the scattering objects for which case the Guinier law is not dened.
The results from the Guinier ts can be compared with theoretical val-
48
Figure 4.2 left:
SANS curves for binary mixtures of decanol and D12-cyclohexane
with a subtracted background as calculated by eq. 4.2,
right:
Guinier-plot, t was
done in the range
7
nm
−2≤q2≤10
nm
−2
.
ues which are expected to be found for decanol molecules. For the case of
completely collapsed spheres the radius of gyration can be expressed by eq.
4.4a.
Rg=√3R2
5
(4.4a)
R=3
√3V
4π
(4.4b)
V=Nagg ·0.317
nm
3
(4.4c)
Here
0.317
nm
3
is the volume of one single decanol molecule and
Nagg
the number of decanol molecules per sphere. The resulting volume
V
of one
sphere can additionally be used to calculate
I(0)
via eq. 4.3b. The calculated
values are listed in table 4.2. Comparing the values obtained by Guinier ts
with the calculated ones it can be seen that
I(0)
suggests aggregates of
≈2
and
RG
of
≈12
decanol molecules. Although this lack in self-consistency
expose the used simple model as insucient, it is already demonstrated that
scattering can not be explained solely by single decanol monomers.
49
Table 4.2
Results for Guinier ts shown in Fig. 4.2 and theoretical values for
dierent decanol aggregation numbers, expressed by the sphere model.
Guinier t Sphere model
Nagg = 1 Nagg = 2 Nagg = 12
φdecanol I(0)
cm
−1
RG
nm
I(0)
cm
−1
RG
nm
I(0)
cm
−1
RG
nm
I(0)
cm
−1
RG
nm
0.30 0.772 0.691 0.466 0.328 0.932 0.413 5.592 0.750
0.20 0.647 0.731 0.311 0.328 0.621 0.413 3.728 0.750
0.10 0.355 0.750 0.155 0.328 0.311 0.413 1.864 0.750
0.05 0.167 0.764 0.078 0.328 0.155 0.413 0.932 0.750
To get a more detailed look on the distribution of molecules in the mixture
(and as a prework to get access to the background of contrast variated mi-
croemulsion studies, see section 5.2.2) for each decanol concentration shown
in Fig. 4.2 the scattering contrast was varied by using several dierent ratios
of deuterated and hydrogenated cyclohexane as oil. The resulting spectra
can be found in Fig. 4.3. Additionally shown in this gure are scattering
curves of dierent D/H-ratios of oil without decanol. In contrast to studies
based on H/D-water where the solvent background is at, it can be clearly
seen that this is not the case for the much bigger cyclohexane molecules who
give a reasonable scattering form factor by itself, visible at high
q
. This
leads to a superposition of the scattering of single cyclohexane molecules in
cyclohexane and decanol aggregates in cyclohexane.
The scattering proles of the binary mixtures of D12/H12-cyclohexane
as shown in Fig. 4.3 can be tted with a simple sphere form factor were the
radius is given by a lognormal size distribution. For a description of the model
see appendix A.1.1. The results are listed in table 4.3. The obtained average
radius is in a good agreement with the expected theoretical hard sphere radius
of a single cyclohexane molecule of around 0.27nm. The size distribution did
not improve the t results reasonably for which reason the width was set to
σ= 10−6
which makes the size distribution quasi monodisperse.
To substract the scattering of cyclohexane molecules from the contrast
50
Figure 4.3 a-d:
SANS curves for dierent decanol concentrations in cyclohexane
(symbols) and ts with eq. 4.5 (lines).
e:
Dierent D12/H12-cyclohexane mixtures
without decanol (symbols) and ts with a spherical form factor (lines, see table.
4.3). In all graphes the key indicates the volume fraction of deuterated cyclohexane
in the oil phase.
51
Table 4.3
Fit results for cyclohexane mixtures using a spherical form factor. Vol-
ume fractions and SLD were calculated from sample composition,
σ
and
R0
was
simultaneously tted for both samples,
3
√< R3>
was then calculated as the aver-
age radius.
φ
D12-cyclo
σ R0/
nm
∆SLD/10−4
nm
−2BG/
cm
−13
√< R3>/
nm
0.88
10−6
0.305 7.00 0.109 0.305
0.76 0.292 0.155 0.292
variation data shown in Fig. 4.3 a-d, the same spherical model was used.
Accordingly the SANS curves were tted as expressed by eq. 4.5 with a sum-
mation of a constant incoherent background (
BG
), a spherical form factor to
express the cyclohexane monomers and a summand to express the decanol
aggregates (
TS(q)dec
). The cross-terms between these two were assumed to
be negligible.
I(q) = TS(q)dec +∫LN(R0, σ)·P(q)sphere
d
R+BG
(4.5)
As summand to describe the scattering originated from the decanol, the
Teubner-Strey model (see eq. 5.2a) was used. This model was used because
it expresses the data by two length scales (the quasiperiodic repeat distance
Ds
and the correlation length
ξ
) and one parameter related to the scatter-
ing invariante (
⟨η2⟩
), but without making any restriction to the structural
shape. For each decanol concentration
ξ
and
D
were hold constant and t-
ted simultaneously as the structure of decanol in cyclohexane should not be
eected by the grade of oil deuteration. The t results are listed in table
4.4. The parameters which were obtained for the sphere summand are in
good agreement with the values obtained for pure oil samples (see table 4.3)
which supports that by the used t method the Teubner-Strey summand now
represents solely the scattering contribution of decanol in oil. Here it can be
seen that the domain size
D
is far too high compared to the dimensions for
single decanol moelcules which is in a good agreement with the results from
the Guinier analysis and speaks for a decanol aggregation.
ξ
shows relative
52
Table 4.4
Fit parameters due to eq. 4.5 for decanol/cyclohexane mixtures with
dierent scattering contrasts.
ξ
and
D
were tted simultaneously for each decanol
concentration,
σ
was xed at
10−6
to prevent overtting,
fp
and
∆SLD
were
calculated from the sample composition (
fp=φ
D
12
or
φ
H
12
).
composition TS summand sphere summand
φdec φ
D
12 φ
H
12 x
D12-cyclo
ξ
nm
D
nm
⟨η2⟩
cm
−1
nm
−3
BG
cm
−1fpσR0
nm
∆SLD
10−4
nm
−2
0.30 0.70 0.00 1.00 0.5269 3.4816 0.5481 0.2582 0.0000
0.30 0.61 0.09 0.87 0.4104 0.3173 0.0800
10−6
0.3400 7.00
0.30 0.53 0.17 0.75 0.3232 0.3823 0.1700 0.2971
0.30 0.35 0.35 0.50 0.1471 0.4800 0.3500 0.2724
0.30 0.18 0.52 0.25 0.0310 0.6408 0.1800 0.3120
0.30 0.00 0.70 0.00 0.0089 0.7899 0.0000
0.20 0.80 0.00 1.00 0.5694 3.9335 0.3795 0.1860 0.0000
0.20 0.69 0.11 0.87 0.2925 0.2535 0.1000
10−6
0.3408 7.00
0.20 0.60 0.20 0.75 0.2247 0.3043 0.2000 0.3165
0.20 0.40 0.40 0.50 0.0996 0.4281 0.4000 0.2852
0.20 0.20 0.60 0.25 0.0211 0.6178 0.2000 0.3139
0.10 0.90 0.00 1.00 0.5652 4.5316 0.2067 0.1115 0.0000
0.10 0.78 0.12 0.87 0.1555 0.1761 0.1100
10−6
0.3259 7.00
0.10 0.68 0.22 0.75 0.1192 0.2314 0.2200 0.3106
0.10 0.45 0.45 0.50 0.0547 0.3699 0.4500 0.2821
0.10 0.23 0.67 0.25 0.0094 0.5807 0.2300 0.3107
0.05 0.95 0.00 1.00 0.4892 5.2409 0.1123 0.0747 0.0000
0.05 0.82 0.13 0.87 0.0870 0.1449 0.1100
10−6
0.3165 7.00
0.05 0.71 0.24 0.75 0.0676 0.1963 0.2400 0.2999
0.05 0.47 0.47 0.50 0.0357 0.3567 0.4700 0.2784
0.05 0.24 0.71 0.25 0.0068 0.5692 0.2400 0.3176
53
low values for all decanol concentrations due to a very loosely structured
system without any long-range ordering.
Figure 4.4
Scattering invariant
Qinv = 2π2⟨η2⟩
calculated from the Teubner-
Strey t parameter listed in table 4.4 for dierent scattering contrasts and decanol
concentrations (symbols). Solid lines are giving linear ts (left graph) or calculated
values for a two-level model with a volume ratio of 0.30 cyclohexane in decanol and
0.02 decanol in cyclohexane domain (right graph). For comparison the values for
domains of pure cyclohexane and decanol without a mutual miscibility are given
in the right graph as broken lines.
Furthermore from the amplitude parameter (
⟨η2⟩
) the scattering invariant
can be calculated due to eq. 5.2d which is plotted in Fig. 4.4 for all D/H-
cyclohexane ratios. Compared to the MRTIL/surfactant binary systems,
where a kink in the concentration dependent development of
Qinv
indicates
a critical aggregation concentration (compare Fig. 3.7), here the trend is
strictly linear going approximately through the origin. When calculating (as
well analog to eq. 5.2d) a simple theoretical two-level invariant with sharply
separated decanol and oil domains, this gives much too high values. Instead
when considering an appropriate constant solubility of oil in the decanol
domain and decanol in the oil, a very good agreement with the experimental
results is obtained as shown in Fig. 4.4 (right). This model only gives
comparable values when just a little amount of decanol is dissolved in the oil
54
domain but a bigger amount of oil in the decanol which is quite intuitive as
the decanol has long enough alkyl chains to host the oil while the OH-groups
would nd a strictly hydrophobic environment in the cyclohexane domain
which is less favorable. In total this gives a picture of a solution were the two
components are not simply statistically mixed but where loosely associated
domains are present. In view of decanol as a cosurfactant in microemulsion
formulation this speaks, on the mesoscopic scale, for a driving force of decanol
molecules into an amphiphilic interface although macroscopic observations
show a broad miscibility of decanol and cyclohexane.
55
56
5
Microemulsions
The conditions for forming microemulsions based on non-aqueous solvents
are not really well understood and accordingly this point will be investigated
in detail in this chapter by using several dierent alcohols as cosurfactant, an
imidazolium chloride based surfactant with dierent alkyl chain lengths and
a systematic variation in composition with respect to the oil and cosurfactant
employed. In a second part the system is restricted to one representative oil
and cosurfactant (cyclohexane and 1-decanol, respectively) but instead the
MRTIL alkyl chain length will be varied giving information on the inuence
of the solvent polarity on microemulsion formation. In total this extends the
concept of structural control by variation of the chain length of the surfactant
and the solvent and gives a more complete view on the system providing
thereby the possibility to draw more general conclusions on self-assembly in
non-aqueous media. Hereby the main focus lies on elucidating the role of the
surfactant chain length in stabilizing microemulsions with MRTILs.
57
5.1 Microemulsions based on C
4
mimFeCl
4
5.1.1 Micellization with decanol
In a rst step we investigated the binary or pseudo-binary (containing cosur-
factant in addition) systems in order to elucidate their potential as a basis for
microemulsion formation. As shown earlier by recording temperature depen-
dent binary phase diagrams (see chapter 3), the Krat points of the pure sur-
factants in C
4
mimFeCl
4
are, dependent on the chain length, above or around
24
◦
C and with that no microemulsion formation is expected at ambient con-
ditions. Fig. 5.1 shows surface tension measurements at 45
◦
C as a func-
tion of surfactant+decanol concentration of solutions of C
j
mimCl/decanol
in C
4
mimFeCl
4
. While in the decanol free binary systems (Fig. 5.1 top)
the reduction of the surface tension depends on the surfactant chain length
(as already discussed in chapter 3), by adding decanol this eect is rst
damped (1mol decanol per 1 mole surfactant, Fig. 5.1 middle) and vanishes
completely by adding more cosurfactant (Fig. 5.1 bottom) what could be
explained by the rather high amount of decanol used, i. e., the decanol is
eectively determining the amphiphilic strength in these mixtures.
In the following microemulsion systems at ambient conditions (24
◦
C) will
be discussed, which works out despite the Krat point issue, as all formula-
tions studied contain rather large amounts of cosurfactant which reduce the
Krat temperature correspondingly. Although these surface tension measure-
ments shown here were recorded at higher temperatures (to avoid problems
with the Krat point) we may conclude from that data to the aggregation
behavior at room temperature as it does not change much in this tempera-
ture range. This was concluded from surface tension measurements at room
temperature (Fig. B.12, with systems including enough decanol to lower
the Krat point below room temperature) and having obtained comparable
SANS spectra at 24 and 36
◦
C (see Fig. B.11 in the appendix B). The surface
tension measurements at ambient conditions also show a signicant decrease
of the Krat point by adding alcohol which makes it possible to formulate
systems at room temperature.
58
Figure 5.1
Surface tension measurements at 45
◦
C for binary mixtures of
C
4
mimFeCl
4
/C
j
mimCl (top), C
4
mimFeCl
4
/C
j
mimCl+1mol decanol (middle) and
C
4
mimFeCl
4
/C
j
mimCl+2mol decanol (bottom) by using C
14
mimCl (cubes),
C
16
mimCl (circles) or C
18
mimCl (triangles) as surfactant.
59
5.1.2 Microemulsions
Adding an oil (which is insoluble in the MRTIL) to the system IL/C
j
mimCl
can lead to the formation of microemulsions if the amphiphilic strength of
the surfactant C
j
mimCl is high enough to stabilize the MRTIL/oil inter-
face. As discussed in section 5.1.1 surface tension measurements predict no
microemulsion formation with pure surfactant at room temperature as the
cmc
goes beyond the solubility of the surfactant. However, by adding the
cosurfactant decanol micelle formation is much facilitated and accordingly
they might be swollen by adding an oil, thereby leading to the formation of
microemulsions.
Macroscopic observations.
The main characteristics of a microemulsion system can already be noticed
by the simple observation of macroscopic phase separation. Fig. 5.2 shows
the ability to solubilize C
4
mimFeCl
4
in cyclohexane at 24
◦
C dependent on
the alcohol/C
16
mimCl ratio for dierent aliphatic alcohols as cosurfactant.
For our experiments we chose cyclohexane as oil as it demands a reduced
need of surfactant to form monophasic systems and shows a higher solubili-
sation capacity compared to other oils, such as isooctane and several linear
alkanes, see Fig. 5.3. As clearly seen by extrapolating to an alcohol free sys-
tem, the ability of MRTIL uptake is zero which means that no microemulsion
formation takes place. Adding alcohol supports an uptake of MRTIL what
can be interpreted as the formulation of microemulsions. Independent on
which alcohol was used qualitatively the uptake capacity is rst enhanced
by enhancing the alcohol ratio in the mixture and then passes through a
maximum which indicates an optimum cosurfactant/surfactant ratio. This
phenomenon could be explained either by entropy/synergism eects due the
preferred solubility of surfactant and alcohol mainly in the MRTIL and oil,
respectively, as proposed by Huibers et al. for mixed surfactant systems
101
or
by geometric considerations, as the surfactant/cosurfactant ratio inuences
the packing parameter and the maxima in MRTIL-uptake shown in Fig. 5.2
are located at the resulting optimal interfacial curvature for MRTIL uptake.
60
Figure 5.2
Maximum MRTIL uptake at a constant ratio of
cyclohexane/(cyclohexane+alcohol+C
16
mimCl) of 86.4wt%. for dierent
aliphatic alcohols as cosurfactant as a function of alcohol/C
16
mimCl mole ratio at
24
◦
C. Lines are guides to the eye. Details can be found in experimental section
2.2.
The shift of the maximum to higher alcohol content for shorter chain
alcohols can on one side be explained by a growing solubility in the MRTIL
and with that a growing part of alcohol which is not acting as a cosurfactant
but is solubilized monomerically (or at least as aggregates too small for a
solvent uptake) in the MRTIL. In addition, for conventional microemulsions
it has been observed before that the rigidity of the amphiphilic monolayer
becomes substantially reduced by the addition of shorter chain alcohols but
not so for longer alcohols like octanol or decanol.
102
As for the stability of the
microemulsions a certain rigidity of the amphiphilic lm is required which is
better achieved for the longer chain alcohols. Both eects (maximum shift
and lower eciency) are consistent with both proposed explanations for the
maximum: A growing solubility of alcohol in the MRTIL will as well reduce
the entropy/synergism eect mentioned before (leading as well to a lower
61
Figure 5.3 left:
Eect of dierent oils on the ability to form monophasic systems
by using decanol as cosurfactant.
right:
Eect of dierent alcohols on the ability to
form monophasic systems by using cyclohexane as oil. Both graphs were recorded
as described in the experimental section for pseudo binary phase diagrams. Start-
ing with homogeneous samples formulated with alcohol, C
16
mimCl and a starting
amount of 86.4wt% oil, C
4
mimFeCl
4
was added dropwise. In the left graph all
samples have a constant mol ratio of decanol/C
16
mimCl=2 (13.6wt%).
amplitude) and lowers the eective content in the interface (maximum shift
to higher concentrations). Similarly the packing parameter will be shifted
less by a shorter alcohol and therefore a larger amount would be needed to
achieve a balanced microemulsion (where one expects the peak of solubility).
In addition, with increasing alcohol chain length the amphiphilic system is
rendered more hydrophobic and having a stier monolayer, which apparently
favors oil solubilisation.
102
At higher alcohol/surfactant ratios the uptake
ability declines due to a lack of amphiphilicity as decanol itself is not a
feasible amphiphile in this system.
In addition to the linear aliphatic alcohols several other alcohols (3,7-
dimethyloctanol, geraniol,
cis
-nerolidol, 2-butoxyethanol) were tested but
gave no improvements in MRTIL solubilization (see Fig. 5.3), only the 3,7-
dimethyloctanol had a similar performance as the 1-octanol. Due to its good
ability to function as cosurfactant, decanol was used for further investiga-
62
Figure 5.4
Phase digrams observed by plotting surfactant/cosurfactant ratio
δ
(eq. 5.1) vs. wt% surfactant+cosurfactant at equal cyclohexane and MRTIL vol-
umes (Kahlweit sh) for C
14
mimCl (straight line), C
16
mimCl (dotted line) and
C
18
mimCl (broken line) at 24
◦
C. The red line gives the experimental path for
SANS experiments done at the C
14
mimCl system shown in Fig. 5.6.
tions. Although dodecanol enhances even more the MRTIL uptake it was
not taken into account as its melting point is above/around room temper-
ature and caused solubility problems which lead to solid precipitate at low
MRTIL concentrations (not shown in Fig. 5.2). Employing alkanes of dier-
ent chain length as oils was also investigated and showed that the extent of
the monophasic microemulsion region becomes smaller with increasing chain
length (Fig. 5.3).
Fig. 5.4 shows Kahlweit-sh diagrams for C
4
mimFeCl
4
/cyclohexane sys-
tems for all three surfactant chain lengths. As a control parameter to mod-
ulate the packing parameter not the temperature was used (as known for
nonionic surfactants in water) but the cosurfactant/surfactant molar ratio
δ
(as similarly done when investigating the eect of medium chain alcohols as
cosurfactants on the phase behavior of nonionic surfactants.
103
δ=n(
C
10
OH
)
n(
C
10
OH
) + n(
C
j
mimCl
)
(5.1)
As already shown in Fig. 5.2 for C
16
mimCl, one nds for all three surfac-
63
tant systems that the presence of alcohol to form microemulsions is crucial.
The eciency decreases from C
18
- over C
16
- to C
14
mimCl due to a rising
monomeric surfactant solubility in the MRTIL. Nevertheless the formation
of microemulsions instead of pure molecular solutions is denitely proven by
observing a three phase region. Characteristic parameters for the position
of the shtail (minimum amount of amphiphilic material and surfactant/-
cosurfactant ratio
δ
required for forming a single phase microemulsion) are
summarized in table 5.1. For the C
14
mimCl system no three phase system
was observed but SANS measurements (Fig. 5.6, along the path shown in
Fig. 5.4) proof the existence of mesoscopic structuring. Apparently the
C
14
mimCl is a much weaker structuring amphiphile and the 3-phase region
was either too small to be detected (hindered also by the rather high con-
centrations and corresponding slowness of the phase separation) or is simply
no longer appearing.
Table 5.1
Characteristic parameters extracted from Fig. 5.4 giving the shtail
position (lowest surfactant amount needed to form the monophasic region) and the
required cosurfactant content
δ
(eq. 5.1).
C
j
mimCl 14 16 18
C
10
OH+C
j
mimCl [wt%] 23 19 15
δ
0.68 0.64 0.61
Beside the phase behavior induced by surfactant/cosurfactant variation,
in addition compositions with dierent oil/MRTIL ratios were investigated.
Fig. 5.5 shows the pseudo ternary phase diagram of these systems with a
constant C
j
mimCl/decanol mole ratio of 1:2 (equals to
δ= 0.67
), this value
being chosen as here maximum solubilization occurs according to Fig. 5.2.
With a solid precipitate at low MRTIL-content, a multi-phase region in the
low surfactant region and a broad mono-phasic region above a certain sur-
factant concentration all three surfactants show similarities. An increasing
need of MRTIL to dissolve all surfactant in the oil rich region with longer
surfactant chain is due to a decreasing solubility in C
4
mimFeCl
4
. The re-
64
Figure 5.5
Ternary phase diagrams for C
14
mimCl (straight line, circles),
C
16
mimCl (dotted line, squares) and C
18
mimCl (broken line, triangles) at (
24.0±
0.1) ◦
C by weight. Thick red line/crosses gives the sample position/experimental
path for SANS experiments done with the C
14
mimCl system shown in Fig. 5.6,
thick dotted line for SANS experiments done at all three surfactant system shown
in Fig. 5.8
gion at high surfactant and MRTIL concentrations was not investigated in
detail but gives qualitatively a multi-phase region increasing in size with
longer surfactant chain. The only small dierences in the size of the multi
phase region at low surfactant concentrations could be misinterpreted as
only a weak enhancement of the ability to form microemulsions with longer
alkyl chains. However a comparison with Fig. 5.4 can explain this phe-
nomenon with the fact that a C
j
mimCl/decanol mole ratio of 1:2 (
δ= 0.67
) is
rather ideal for the C
14
mimCl but becomes increasingly less so for the longer
chain surfactants which illustrates the essential need of both MRTIL/oil- and
surfactant/cosurfactant-ratio variation to get a full picture of the surfactant
eciency.
65
Mesoscopic structure.
Low viscosities of samples located in the single-phasic region of the phase
diagram (see appendix B.9) already hints at the presence of a microemulsion
in this range and against formation of a liquid crystalline phase, which is fur-
ther conrmed by the fact of optical isotropy. Additionally by conductivity
measurements a percolation point can be observed (see appendix B.8), which
is typical of microemulsion systems.
51
To investigate the system on the meso-
scopic scale SANS measurements were done, one of the key methods to study
microemulsions.
104
The scattering curves in Fig. 5.6 show that the intensity
increases largely upon reducing the content of amphiphile in the system and
for the highest amphiphile content only very little coherent scattering is seen.
Also interesting to note is that only at intermediate amphiphile concentra-
tion a correlation peak is seen that vanishes again upon further dilution. In
order to deduce quantitative structural information from the SANS curves
we applied the phenomenological Teubner-Strey (TS) model,
105
in which the
scattering intensity is given by eq. 5.2a and is basically determined by the
quasiperiodic repeat distance
Ds
(eq. 5.2b) and the correlation length
ξ
(eq.
5.2c) of the structural units. Here
⟨η2⟩
is directly related to the scattering
invariant (
Qinv
) and accounts for the contrast
∆ρ
and volume fractions
Φ
of
the oil and MRTIL phase (eq. 5.2d).
I(q) = 8·π·c2·⟨η2⟩/ξ
a2+c1·q2+c2·q4+BG
(5.2a)
Ds
2π=[1
2√a2
c2−1
4
c1
c2]−1/2
(5.2b)
ξ=[1
2√a2
c2
+1
4
c1
c2]−1/2
(5.2c)
⟨η2⟩= ΦIL ·Φoil ·(∆ρ)2=Qinv/2π2
(5.2d)
Next to the background,
Ds
,
ξ
and
⟨η2⟩
were free parameters during tting
even though the latter value derives directly from the sample composition and
the distribution of the components in the two phases. As the distribution of
66
surfactant and decanol between polar, oily and interface domains is not clear
this approach is justied. More details on this are given in the last part of
this chapter. From the t parameters the amphiphilicity factor
fa
and the
renormalized mean bending modulus
κ
were calculated according to eqs. 5.3
and 5.4, respectively.
106,107
fa=c1/√4a2c2
(5.3)
κ
kBT=10√3π
64
ξ
Ds
(5.4)
The amphiphilicity factor quanties the amphiphilic strength of the sys-
tem whereby a value of
−1
corresponds typically to a highly structured lamel-
lar phases while higher values are due to a decreasing amphiphilicity. Well
structured (good) microemulsions are normally found to have negative val-
ues near
−1
the Lifshitz line is dened at
fa= 0
and the disorder line at
fa= +1
. Above this the triclinic point can be found.
108
Table 5.2
Teubner-Strey t parameter derived from SANS measurements shown
in Fig. 5.6 and calculated amphiphilicity factor
fa
and bending rigidity
κ
.
C
14
mimCl
+C
10
OH
ξ Ds⟨η2⟩BG faκ
[wt%] [nm] [nm]
[1
cm nm
3] [ 1
cm
]
[kT]
61 1.31 3.66 0.02 0.69 -0.67 0.31
53 1.34 3.86 0.04 0.66 -0.65 0.30
49 1.43 4.04 0.06 0.65 -0.67 0.30
42 1.50 4.63 0.11 0.64 -0.61 0.28
37 1.50 5.38 0.16 0.64 -0.51 0.24
29 1.55 8.64 0.26 0.61 -0.12 0.15
26 1.51 16.26 0.30 0.59 0.49 0.08
Fig. 5.6 shows SANS measurements at constant oil/MRTIL volume ratio
of 1:1 along the experimental path shown in Fig. 5.4 and 5.5. Decreasing
the amphiphile amount while keeping the oil/MRTIL ratio constant at 1:1
67
Figure 5.6
SANS data (symbols) for microemulsions formulated with C
14
mimCl
and a constant volume ratio oil/MRTIL=1:1. For sample positions see Fig. 5.4.
Straight lines are results from ts with eq. 5.2a.
Figure 5.7
Teubner-Strey t parameters
ξ
(open squares) and
Ds
(lled squares)
for C
14
mimCl derived from curves displayed in Fig. 5.6. Dashed line displays the
cube model (eq. 5.6) with
Σ = 1.0
nm
2
,
cmon = 0.17
and
xc= 0.5
. For details see
the appendix B.10.
68
leads to an increase in scattering intensity due to an increase of the size
of the structures present. The data were tted with the TS model (eq.
5.2a) and the obtained parameters are summarized in Fig. 5.7 and Table
5.2. The domain size
Ds
increases with decreasing surfactant concentration
which can be explained such that less amphiphile per uid is available to form
an oil/MRTIL interface and with that the oil and MRTIL domains have to
grow to house the volume of the two solvents with less surfactant stabilized
interface available.
Ds
increases largely upon reaching the emulsication
failure (cf. Fig. 5.4 and 5.5) and appears to be diverging there. A simple
geometrical model to estimate the domain size
Ds
was proposed by Jouroy
et al.
109
describing the microemulsion by a lattice of cubes lled with either
polar (water) or apolar (oil) solvent and a separating surfactant layer between
dierently lled cubes leading to eq. 5.5
Ds=6ωφpφap
Σφs
(5.5)
with
φs
,
φp
,
φap
as the volume fractions of surfactant in the interface,
polar and apolar phase, respectively, the molecular volume of one surfactant
molecule (
ω
) and the surfactant headgroup area (
Σ
). As several assumptions
done here (i.e. vanishing interfacial surfactant layer thickness compared to
the cube length, location of all surfactant molecules in the interface, very
small surfactant concentration) are not valid for our actual system, eq. 5.5
was modied in such a way that the surfactant was partly allowed to be
monomerically dissolved in the MRTIL phase and the surfactant interface was
divided between the apolar and polar phase, whereby the volume fraction of
the surfactant head (
φmimCl
) and decanol head (
φOH
) was added to the polar
phase and the location of the alkyl chains (
φC14
and
φC10
) was dened by
the ratio
xc
which is added to the polar phase. This model can be expressed
by the following set of equations:
69
Ds=6ωC14φpφap
ΣφC14,i
(5.6)
φC14,i =φC14 −φILcmon
(5.7)
φp=φIL +φmimCl +φOH + (φC10 +φC14,i)xc+φILcmon
(5.8)
φap =φoil + (φC10 +φC14,i)·(1 −xc)
(5.9)
cmon[
wt
%] = 100 ·[1 + ρIL
ρC14
MC14
MC14 +MmimCl
1
cmon ]−1
(5.10)
Here
ωC14 = 0.4051
nm
3
is the volume of one C
14
-chain,
φIL
and
φoil
are
the volume fractions of MRTIL and oil, respectively.
cmon
gives the amount
of surfactant chains not present in the interface but dissolved monomerically
in the MRTIL. With eq. 5.10 it can be expressed in a form which makes it
comparable to the denitions of the
cmc
which were made earlier (see Fig.
3.5 and eq. 3.3d). Here
ρ
and
M
are the density and molar mass, respec-
tively.
Σ
gives the average area occupied by one surfactant+decanol unit.
The model with suitable parameters is plotted in Fig. 5.7. (For curves with
dierent parameters see Fig. B.15 in the appendix B.10. Despite its simplic-
ity it suces to describe the experimental values quite well with values for
Σ
around 1nm
2
which is a reasonable value for the size of the surfactant head
group. For comparison the same surfactant was found to have a minimum
headgroup area of 0.7nm
2
at the water/air and EAN/air interface.
66
Simi-
larly the monomeric concentration of 14wt% surfactant in the MRTIL which
is needed to describe the data corresponds well to the
cmc
measurements
described before (cf. table 3.1). This ndings accord with the preliminary
made interpretations and yields a coherent model.
ξ
increases as well by low-
ering the amphiphile amount explained by more dened aggregates with a
lower polydispersity. Calculated values for the amphiphilicity factor and the
mean bending modulus as listed in table 5.2 are as well in a good agreement
with the here proposed trends.
κ
increases with increasing surfactant con-
centration as the structures are expected to become stier,
fa
is located well
under the Lifshitz-line
108
as expected for microemulsion structures at higher
surfactant concentrations. Only near the phase boundary at low surfactant
70
concentration a positive value points to a less structured system.
To compare structures formed with the dierent surfactants additional
SANS measurements were done along the experimental path shown in Fig.
5.5 (thick dotted line). For all samples the ratio between the molar quan-
tity of amphiphile and the solvent volume was held at a constant value of
n(CjmimCl)/(VMRTIL +Voil) = 0.68
mol/L. Then the only parameter varying
is the ratio between oil and MRTIL volume dened as
xMRTIL =VMRTIL
VMRTIL +Voil
(5.11)
In Fig. 5.8 the obtained SANS curves are shown and it is interesting
to note that the overall scattering intensity becomes lower with increasing
length of the surfactant. So apparently the structural units are the largest
for the shortest chain surfactant. In contrast, the correlation peak at inter-
mediate mixing ratio of MRTIL/oil becomes more prominent with increasing
surfactant chain length, which indicates that the degree of ordering increases
correspondingly.
Teubner-Strey ts as described above were carried out and the results
are summarized in table 5.3. In all three surfactant systems the amplitude
(quantitatively expressed by
⟨η2⟩
) is decreasing with increasing
xMRTIL
due
to a vanishing contrast by substituting deuterated cyclohexane with hydro-
genated MRTIL. Following the picture of an inversion of mean curvature
while going from the oil-rich to the MRTIL-rich side of the phase diagram a
maximum in domain sizes
Ds
is expected for intermediate
xMRTIL
and can
indeed be observed for all three surfactants as seen in Fig. 5.9.
Surprisingly the correlation length
ξ
is not following simultaneously the
same trend as known from water systems.
105,110
Instead after a similar rise
up to
xMRTIL ≈0.3
,
ξ
decreases again for higher MRTIL content. This can
be interpreted by a more exible and interpenetrating mesoscopic structure
in the MRTIL-dominated region which is supported by an increasing value
for
fa
. A second observation not being in accordance with common water
systems is the fact that the structures are getting bigger by shortening the
alkyl chains. Intuitively one would expect stier and bigger domains with
71
Table 5.3
Teubner-Strey t parameters (see eq. 5.2a) derived from SANS mea-
surements shown in Fig. 5.8 and calculated amphiphilicity factor
fa
and bending
rigidity
κ
calculated with eq. 5.3 and 5.4, respectively.
x
MRT IL ξ Ds⟨η2⟩
BG
faκ
[nm] [nm]
[1
cm nm
3] [ 1
cm
]
[kT]
C
14
mimCl
0.05 1.67 4.86 0.60 0.29 -0.65 0.29
0.12 1.74 5.67 0.57 0.38 -0.58 0.26
0.19 1.79 6.15 0.54 0.48 -0.54 0.25
0.20 1.82 7.19 0.51 0.54 -0.43 0.21
0.37 1.77 8.82 0.44 0.59 -0.23 0.17
0.42 1.69 9.88 0.41 0.63 -0.07 0.15
0.49 1.54 10.77 0.33 0.65 0.11 0.12
0.54 1.36 12.48 0.30 0.70 0.36 0.09
0.60 1.27 13.85 0.22 0.66 0.50 0.08
0.72 0.90 16.04 0.08 0.60 0.78 0.05
0.77 0.87 8.95 0.03 0.58 0.45 0.08
1.00 0.78 5.41 0.03 0.54 0.10 0.12
C
16
mimCl
0.05 1.69 4.76 0.56 0.29 -0.67 0.30
0.12 1.91 5.38 0.51 0.41 -0.66 0.30
0.19 2.05 5.91 0.46 0.47 -0.65 0.29
0.27 2.20 6.42 0.43 0.57 -0.65 0.29
0.35 2.15 6.90 0.36 0.58 -0.59 0.27
0.42 2.11 7.24 0.30 0.62 -0.54 0.25
0.50 1.95 7.54 0.24 0.63 -0.45 0.22
0.55 1.87 7.57 0.21 0.68 -0.41 0.21
0.61 1.71 7.45 0.15 0.66 -0.35 0.19
0.71 1.53 6.84 0.05 0.63 -0.32 0.19
0.79 1.56 5.96 0.01 0.57 -0.46 0.22
0.94 1.41 4.85 0.01 0.54 -0.54 0.25
1.00 1.07 4.80 0.03 0.55 -0.33 0.19
C
18
mimCl
0.12 2.09 5.21 0.44 0.44 -0.73 0.34
0.19 2.27 5.70 0.38 0.48 -0.72 0.34
0.26 2.37 6.11 0.38 0.57 -0.71 0.33
0.33 2.35 6.51 0.32 0.59 -0.67 0.31
0.40 2.36 6.81 0.29 0.68 -0.65 0.29
0.47 2.19 7.12 0.24 0.68 -0.58 0.26
0.54 2.03 7.07 0.19 0.68 -0.53 0.24
0.59 1.92 7.18 0.14 0.68 -0.48 0.23
0.65 1.77 6.89 0.09 0.67 -0.44 0.22
0.75 1.58 6.14 0.03 0.59 -0.45 0.22
0.79 1.57 5.75 0.01 0.55 -0.49 0.23
0.94 1.52 5.52 0.01 0.55 -0.50 0.23
73
longer chains. While a higher stiness is indeed conrmed by higher values for
ξ
and
κ
, the contrary behavior of the domain size can be explained by a better
molecular solubility of C
14
mimCl compared to the longer chain surfactants
in C
4
mimFeCl
4
leading to a smaller interface to volume ratio and a tendency
for building bigger structures.
Figure 5.9
Teubner-Strey t parameters
ξ
(correlation length, open symbols)
and
Ds
(quasiperiodic repeat distance, lled symbols) for C
14
mimCl (squares),
C
16
mimCl (circles) and C
18
mimCl (triangles) as a function of
xMRTIL
derived
from curves displayed in Fig. 5.8.
Again some limitations of comparability may arise from the fact that the
samples have dierent positions relatively to the shtail position (compare
Fig. 5.4), a complication not to be avoided as within a 4-component system
such compromises regarding the composition have to be done, in order to
have better comparability for other aspects. While the C
16
- and C
18
mimCl
containing systems are placed in a region relatively far from the shtail, the
C
14
mimCl containing system is located very close to its optimal
δ
-value to
produce bigger structures.
To get an even more detailed insight into the microemulsion structure,
74
Figure 5.10
Scattering invariant
Qinv
(open symbols) for C
14
mimCl (squares),
C
16
mimCl (circles) and C
18
mimCl (triangle) as a function of
xMRTIL
derived from
curves displayed in Fig. 5.8 with eq. 5.12. Small lled symbols are values calculated
from sample composition with eq. 5.2d for the two cases of all (broken line) and
only parts (straight lines) of the surfactant/alcohol CH
2
-units counted to the oil
phase. Inset cartoons illustrate the separation between oily and aqueous phase for
these two cases.
values of the theoretical invariant
Qinv
were calculated with eq. 5.2d. Al-
though the sample compositions are xed, the result is highly dependent
on the assumption of the partitioning of the surfactant/cosurfactant chains
between MRTIL and oil domain as this eects volume ratio and average scat-
tering length density of both domains. Dierent possible distributions were
calculated (for details see appendix B.3.1) that dier with respect to how
much of the alkyl chain of the surfactant is counted into the hydrophobic
part, and a comparison with experimental values is shown in Fig. 5.10. As
the invariant value obtained by the Teubner-Strey t is strongly eected by
the t quality, instead the experimental invariant was calculated by integra-
tion of the measured data by
75
Qinv =∫∞
0
I(q)q2
d
q
(5.12)
where extrapolation to zero and innity was done by the Guinier and
Porod approximation, respectively (see appendix B.3.1 for details). In the
oil rich region (small
xMRTIL
values) a very good agreement is obtained for
the case were the surfactant/alcohol alkyl chains are divided almost equally (6
and 7 C atoms of decanol and C
j
mimCl surfactant, respectively, are counted
into the polar phase; solid lines in Fig. 5.10). These results were conrmed
by data recorded in a second SANS experiment (see description in appendix
B.4) Going to the MRTIL richer region (
xMRTIL ≥0.4
) the experimental
values start to deviate and have smaller values as predicted by this model.
This can be explained by a less and less dened interface caused by a weaker
mesoscopic structuring which is in good agreement with ndings for
ξ
dis-
cussed above. For
xMRTIL
values near 1 the experimental values can be better
described by a model which counts all CH
2
-groups to the oil phase. This is
plausible as the portion of cyclohexane is getting more and more negligible
compared to the amount of surfactant/decanol.
Additionally to the Teubner-Strey model a clipped random wave (CRW)
model
111,112
was applied to the data. As the derived values for the lengthscale
parameters
Ds
and
ξ
are nearly identical with the ones obtained by TS it
is only described in detail in the appendix B.2. Nevertheless applying this
model to the scattering data gives additional information as it delivers a third
lengthscale (
c
) which accounts for the interfacial roughness. In the oil rich
region (low
xMRTIL
) this roughness parameter shows values comparable to
water systems
112114
(in our system they are slightly higher due to a higher
surfactant concentration) and the trend of a growing roughness value with
longer surfactant chains is as expected. Increasing the MRTIL content in
the system gives continuously bigger
c
-values for all three systems which ts
well to the general picture of a weakening of the mesoscopic structuring by
increasing the MRTIL ratio. Above
xMRTIL ≈0.4
the CRW-ts give random
high numbers for
c
. This is due to the fact that the roughness parameter is
not necessary anymore to simulate the SANS data (i. e. the t quality is
76
independent from
c
) as the TS-model itself gives already excellent t results.
Also remarkable is the fact that the simple two phase model to explain the
invariant (compare Fig. 5.10) as well fails above
xMRTIL ≈0.4
which gives
a hint to structural changes at this point.
5.1.3 Conclusion
The here presented study gives a detailed view on the phase behavior of
C
4
mimFeCl
4
containing microemulsions. As surfactants we employed dier-
ent 1-alkyl-3-methylimidazolium chlorides with C
14
, C
16
and C
18
chains, but
as alone its amphiphilic strength was not high enough to form microemul-
sions, it was employed in a 1:2 (C
i
mimCl/decanol) molar ratio with decanol
as cosurfactant. Studies of the phase behavior showed that alcohols become
increasingly eective as cosurfactants with increasing chain length, while
the range of having monophasic microemulsions becomes at the same time
smaller upon increasing the chain length of the oil. The variation of the
surfactant chain length shows on the one hand a classical behavior expressed
by an enhancement of solubilization strength or lm rigidity with increasing
chain length. On the other hand the eect is damped by the inuence of the
high amount of cosurfactant so that the surfactant chain length has nearly
no eect on surface tension.
The SANS data can be well described with the Teubner-Strey model and
show that microemulsion structures form most prominently in the region
of
xMRTIL = 0.2−0.6
. The degree of structuring increases with increasing
chain length of the surfactant and the size of the structural domains increases
largely upon approaching the emulsication failure. Values for
κ
,
c
(mean
bending modulus and roughness parameter from the CRW model) and
ξ
are
comparable with water systems in the oil rich region. With an increasing
content of MRTIL all this parameters point to less and less structured sys-
tems with interpenetrating phases leading to a rough and less sti interface
with less pronounced long range ordering.
The here presented broad investigation yields quantitative information on
the composition-structure relationship and therefore gives recipes to design
77
magnetic microemulsions with optimised properties and structures, as it has
not yet been done for such systems that can be manipulated by a magnetic
eld. These ndings are useful for designing strategies for formulating mi-
croemulsions of a given structure with MRTILs as polar component. This
is important as such microemulsions could in the future be employed as in-
teresting reaction media which contain also a component for separation via
magnetic forces.
5.2 Microemulsions containing dierent MR-
TIL
In the following chapter the inuence of MRTIL chain length on the mi-
croemulsion formation is investigated. To enhance clearity the number of
varied compounds is reduced to one oil (cyclohexane) and one cosurfactant
(1-decanol) as these two showed the best ability to form microemulsions in
the C
4
mimFeCl
4
system. Instead the chain length of the surfactant (n = 12,
14, 16 or 18) and MRTIL (n = 2, 4 or 6) were varied to have insight into the
dependency of microemulsion phase behavior on molecule chain length in a
more broad fashion.
5.2.1 Macroscopic observations
In Fig. 5.11 the resulting Kahlweit-sh phase diagrams are shown. The sys-
tem based on C
4
mimFeCl
4
was already discussed in detail in section 5.1.2
revealing that the ability to form microemulsions increases with longer sur-
factant alkyl chains expressed by a bigger three-phase region, a lower need
of amphiphile, and a lower ratio of cosurfactant needed. The same trend can
be observed in the C
2
mimFeCl
4
based microemulsions when lengthening the
surfactant chain from C
12
to C
18
. Additionally, when focussing on the eect
of the MRTIL alkyl chain one can see a reverse trend of these quantities:
For the microemulsion systems formulated with C
18
mimCl as surfactant, the
obtained three-phase region is getting bigger, less amphiphile is needed and
78
Figure 5.11
Kahlweit sh diagrams (volume ratio MRTIL/cyclohexane = 1:1)
of all investigated microemulsion systems with C
2
mimFeCl
4
(dashed lines),
C
4
mimFeCl
4
(straight lines) or C
6
mimFeCl
4
(dotted lines) as polar phase and
C
12
mimCl (black), C
14
mimCl (blue), C
16
mimCl (purple) or C
18
mimCl (red) as
surfactant. For details see Appendix B. The crosses (for C
2
mimFeCl
4
systems)
and stars (for C
4
mimFeCl
4
systems) show the sample positions for SANS investi-
gation.
the ratio of cosurfactant is lower to obtain single phasic systems when short-
ening the alkyl chain from C
6
over C
4
to C
2
(red curves in Fig. 5.11). The
same trend can be observed by using C
16
mimCl (purple curves) or C
14
mimCl
(blue curves).
As a working hypothesis the eects provoked by surfactant and MR-
TIL chain length can be summarized in a dependency of the ability to self-
assemble in the following form:
∝j−i
(5.13)
where
j
and
i
are the alkyl chain length of the surfactant and MRTIL,
respectively. Indeed this relation is valid for the shtail position and is shown
in Fig. 5.12 c. In general as well a few other values characterizing the
aphiphilicity of the binary systems discussed in chapter 3, namely the
cmc
,
the critical concentration to form liquid crystals (
φLC
) or the position of the
Krat discontinuity (
TKrafft,disc
) can be described by the same relation as
79
plotted in Fig. 5.12, too.
The linear dependency of the free energy of micellization (which is pro-
portionoal to
ln cmc
) on the surfactant chain length was already mentioned
in chapter 3.2.2 and was explained to have its origin in the transfer energy
for one CH
2
-group into the solvent. When this is true it is not surprising that
the amount of lipophilic moieties in the solvent, quantied by the number of
CH
2
groups in the solvent molecule, has the same eect but with opposite
sign. It is now straightforward to link also the other values plotted in Fig.
5.12 to this transfer energy. Consequently the solubility of aliphatic alcohols
in MRTIL as discussed in section 4.1 follows the same law (see table 4.1).
5.2.2 Mesoscopic structure
To investigate the mesoscopic structure, SANS measurements were done for
samples whose positions within the phase diagrams are marked in Fig. 5.11.
The positions were chosen to be in the monophasic regime close to the sh tail
position to have a comparable mean curvature and as less excess amphiphile
as possible for all systems. In Fig. 5.13 the resulting curves for all systems
are shown. The ndings here underline the trends which were concluded
from the phase diagrams. The system C
2
mimFeCl
4
/C
18
mimCl is expected
to show the strongest amphiphilicity and indeed this system shows a well
dened correlation peak. By shortening the surfactant alkyl chain, the peak
rst decreases and then completely vanished for C
14
mimCl. Lengthening
the MRTIL chain length gives a similar eect with respect to the surfactant
chain length but in general the peak feature is damped as it is expected
for systems with less long-range ordering. To extract quantitative values ts
were done with the CRW model (for the model description see appendix B.2)
and the resulting curves are additionally plotted in Fig. 5.13. The underlying
parameters are listed in table 5.4. Here one can see that the trend expressed
by eq. 5.13 is supported by the length parameter
ξ
which indicates the long
range correlation between the microemulsion domains and which is increasing
with surfactant and decreasing with MRTIL chain length. The trend for
the domain size
Ds
is counter intuitive as it was already observed for the
80
Figure 5.12
cmc
derived from SANS arregation number (agg), scattering invari-
ante (inv) and surface tension (
γ
) as summarized in table 3.1 (top), characteristic
values derived from DSC-measurements in binary mixtures (middle) and the sh-
tail position in microemulsions (bottom) as a function of the alkyl chain length
dierence between surfactant and MRTIL. The colors indicate the solvent being
C
2
mimFeCl
4
(red), C
4
mimFeCl
4
(black) or C
6
mimFeCl
4
(blue). The lines are lin-
ear ts.
81
Figure 5.13
Scattering curves for all microemulsion systems, formulated with
H12-cyclohexane as oil, The MRTIL is C
2
mimFeCl
4
(left) or C
4
mimFeCl
4
(right).
For sample positions see Fig. 5.11. Straight lines are ts with the CRW-model.
The underlying parameters are listed in table 5.4.
C
4
mimFleCl
4
system earlier (see chapter 5.1.2) showing bigger structures for
shorter surfactant chain length. For the C
4
mimFeCl
4
/C
14
mimCl systems this
leads even to sizes beyond the observation limit making it impossible to t
with the CRW-model.
Contrast variation
To evaluate the mesoscopic structure of the microemulsions in a more detailed
fashion, for the sample positions given in Fig. 5.11 a contrast variation study
was performed. For this reason for each position ve dierent levels of oil
deuteration were measured (pure H12-, 25% D12-, 50% D12-, 75% D12- and
pure D12-cyclohexane). The resulting spectra are plotted in Fig. 5.14.
With the present number of components and consequential high number
of degrees of freedom in the present systems, to get quantitative information
on the distribution of the dierent compunds (i.e. the surfactant and cosur-
factant) it would be deemed necessary to have a reliable value for
I(q→0)
and the scattering invariant for each system.
115,116
The rst requirement is
82
Table 5.4
Results from CRW-ts for curves shown in Fig. 5.13. Values for the
system C
4
mimFeCl
4
/C
14
mimCl are set in brackets due to its bad matching with
experimental data.
C
i
mimFeCl
4
C
j
mimCl
CRW-Model Porod/Guinier
ξ
/nm
Ds
/nm
Qinv
/cm
−1
nm
−3c
/nm
BG
/cm
−1Qinv
/cm
−1
nm
−3
4 14 (3.89) (300) (0.83) (5) 0.664 1.83
4 16 2.98 529 1.25 12 0.663 1.88
4 18 3.88 37 1.31
>25
0.658 1.55
2 14 3.24 44 1.64 15 0.626 1.51
2 16 5.50 21 1.67
>25
0.606 1.45
2 18 7.20 22 1.25
>25
0.551 0.85
at least dicult for the system C
4
mimFeCl
4
/C
14
mimCl as the measured in-
tensities do not converge to a plateau for low
q
-values. To extract the latter
one, additionally to this an extrapolation to high
q
-values is necessary. As
already discussed in chapter 4.2 and 5.1.2, this is as well dicult due to dis-
tracting scattering from cyclohexane on the one hand and eventually present
monomerically decanol dissolved in the oil phase. Due to these diculties the
results of the contrast variation experiments are analyzed only qualitatively
to avoid an over-interpretation of the data.
As one can see for all six systems, a variation of the cyclohexane scatter-
ing length density from
6.68 ·10−4
nm
−2
(100% D12-cyclohexane) to
−0.28 ·
10−4
nm
−2
(100% H12-cyclohexane) lets pass the scattering intensity through
a minimum as to be expected around the value for the MRTIL (
1.83·10−4
nm
−2
and
1.55 ·10−4
nm
−2
for C
2
mimFeCl
4
and C
4
mimFeCl
4
, respectively). For
the case of a well dened interface of hydrocarbon chains between the oil and
MRTIL domains, dierent parts of the microstructure would be visible as a
function of the oil SLD. As illustrated schematically in Fig. 5.15 for three
dierent cases, using pure H12-cyclohexane nearly matches the surfactant
interface (left), adjusting MRTIL and oil to the same density would make
only the interface visible (middle) and the use of pure D12-cyclohexane gives
84
Figure 5.15
Scattering length density prole of the interfacial lm for three cases:
The surfactant is contrast matched by the pure H12-cyclohexane (left), oil and
MRTIL are contrast matched (middle) and dierent levels for oil, MRTIL and
interface when using pure D12-cyclohexane (right). The proles are given for the
assumption of sharply separated domains (straight lines, illustrated by the cartoons
above the graphs) and a partly interpenetration (broken lines, illustrated by the
cartoons under the graphs)
a 3-level prole of oil, MRTIL and interface domain (right). Consequently
this results in completely dierently appearing SANS spectra as a function
of the oil SLD variation, which is indeed commonly observed for microemul-
sions.
117119
The data presented in Fig. 5.14 do not show any variation beside
the scattering originating by the cyclohexane molecules at high
q
and a scal-
ing factor due to the overall contrast. This suggest a lack of well dened
microphases. In Fig. 5.15 the SLD-prole is additionally shown assuming a
partial monomerical solubility of the surfactant in the MRTIL (which lowers
the SLD of the polar domain) and a partial interpenetration of the interface
by MRTIL and oil. This smearing out of the interface results in a SLD prole
which would then accord with the measured data. This ndings underline the
picture drawn in the earlier sections which characterizes the microemulsions
by oil and MRTIL domains separated by a rough and poorly sharp interface.
85
5.3 Magnetic behaviour
5.3.1 Field Gradient
Due to its paramagnetic nature, the MRTIL is per denition attracted by a
magnetic eld, a behavior which was already discussed in the introduction
and this is demonstrated by the ability to move a macroscopic MRTIL phase
with a magnet. Going from a macroscopic portion to micron-sized MRTIL
droplets as it is given in an emulsion, this driving force is still valid, although
the forces which stabilize an emulsion are now appearing as opponents.
Figure 5.16 Left:
Sedimentation measurements with (blue) and without (red)
a magnetic eld gradient to investigate the emulsion stability. The emulsion is
composed of cyclohexane (17.4wt%) , C
4
mimFeCl
4
(60.4wt%), decanol (11.2wt%)
and C
18
mimCl (11.0wt%).
Right:
Image of the emulsion directly after dispersion
with a vortex mixer (top) and after complete demixing (bottom).
Fig. 5.16 shows an emulsion formed by immiscible cyclohexane and
C
4
mimFeCl
4
, stabilized by C
18
mimCl. This emulsion is only weakly kinet-
ically stable as indicated by a relatively fast phase separation into a lower
MRTIL phase and an upper microemulsion phase. Under ambient conditions
at zero magnetic eld this process is completed after around 30min. The
86
phase separation velocity (
vsep
) was measured quantitatively by measuring
the height of the precipitate phase as a function of time.
vsep =
d
V
d
t=A
d
x
d
t
(5.14)
with
V
and
x
being the volume and relative height of the lower MRTIL
phase, respectively, and
A
being the cross-section of the sample vial. This
sedimentation is in a simple approximation related to Stokes law which de-
scribes the sedimentation velocity
vs
as a function of droplet radius
R
, solvent
viscosity
η
and the density dierence between them (
ρdrop −ρsolv
):
vs=2R2(ρdrop −ρsolv)g
9η
(5.15)
The underlying force can be expressed as:
Fgrav =4
3πR3(ρdrop −ρsolv)g
(5.16)
Here
g
is the acceleration due to gravity. By exposing to a eld gradient
the separation kinetic is enhanced by a factor of around 2. This can be
expressed by a second force term (
Fmag
) added to
Fgrav
which is dependent
on the dierence in volume magnetic suszeptibility
χ
and the magnetic eld
gradient (
∇B
):
Fmag. =4
3πR3(χdrop −χsolv)B(∇B)
µ0
(5.17)
The mentioned enhancement of the separation kinetics by a factor of
2 consequently is due to the condition
Fgrav =Fmag
which indeed can be
calculated for the given system as demonstrated in the appendix B.11.
A MRTIL containing microemulsion still keeps a paramagnetic behaviour
which can be seen by the fact that the meniscus can be manipulated by a
magnetic eld gradient as shown in Fig. 5.17 and this applies to the whole
microemulsion range that contains at least a few percent of MRTIL. This
is due to the paramagnetic behaviour of the complete macroscopic sample
and can be explained analogously to the pure MRTIL. The magnitude of
87
the meniscus deformation scales with the average susceptibility of the whole
sample and makes no statement about the microstructure. Nevertheless the
stabilizing forces of the microstructure are far higher than in an unstable
emulsion (see eq. 5.17 where the sedimentation force induced by a eld
gradient scales with
R3
) and at the given elds the microemulsion can not
be broken.
Figure 5.17
Response of the MRTIL containing microemulsion to the eld gradi-
ent of an electromagnet. The sample shown consists of 31.2wt% D12-cyclohexane,
46.1wt% C
4
mimFeCl
4
, 11.8wt% C
16
mimCl and 10.9wt% decanol. The magnetic
eld is oriented parallel to the liquid surface.
5.3.2 Homogeneous eld
A quantitative insight in the magnetic behaviour of the microemulsions has
been gained by measurements of the magnetic susceptibility, which were done
along the dashed line given in Fig. 5.18 covering the full range from the pure
oil to pure MRTIL as solvent. All samples show a paramagnetic behavior
indicated by a linear eld dependence of the magnetization plotted in Fig.
5.18. The magnetic susceptibility shown in Fig. 5.19 increases in almost per-
fect linear fashion with the increasing weight content of the MRTIL, which
means that the magnetic properties of the microemulsion are not aected
by its mesoscopic structure which conrms the ndings under a eld gradi-
ent discussed in section 5.3.1. Extrapolation to pure C
4
mimFeCl
4
gives a
magnetic susceptibility of
40·10−6
emu/g. For comparison the initial suscep-
tibility of ferrouids can be found to be of the order of
104
times higher.
120
In fact the magnetic susceptibility is dimensionless but can be dened by
88
Figure 5.18 left:
Example SQUID measurements of the microemulsions and the
emty cell at 300K. The labels indicate the wt% of MRTIL in the sample.
right:
Location within the phase diagram (dots) of the samples given the results showed
on the left.
dierent systems of units and related to mole, mass or volume, which is of-
ten not clearly labeled in literature, and this can lead to some confusion.
For this reason in table 5.5 the magnetic suszeptibility is given for dierent
denitions.
Table 5.5
Magnetic suszeptibility calculated due to the denitions given by the SI
and EMU system of units, with respect to the sample mass (
χg
), volume (
χv
) or
molar amount (
χm
) with the relation
χv=χgρ=χmρ/M
, whereby
ρ
and
M
are
the density and molar mass, respectively. Values for cyclohexane are taken from
literature.
121
χgχmχv
SI EMU SI EMU SI EMU
C
4
mimFeCl
4
5.0·10−7
4.0
·10−5
1.7
·10−7
1.4
·10−2
6.8
·10−4
5.4
·10−5
cyclohexane
−1.0·10−8−8.1·10−7−8.6·10−10 −6.8·10−5−7.9·10−6−6.3·10−7
Considering the orientation of microscopic domains in a homogeneous
magnetic eld this can be described by two dierent mechanisms:
122
1. With the existence of an
intrinsic anisotropy
of the magnetic suscep-
89
Figure 5.19
Mass magnetic susceptibility as a function of wt% MRTIL. Measure-
ments were done along the experimental path shown in Fig. 5.18.
tibility of the material. (For that
χ
has to dier at least for two geomet-
rical axes.) the orientation of the highest susceptibility parallel to the
magnetic eld minimizes the magnetic energy by
E= 0.5V∆χB2
µ0cos2θ
,
with
θ
as the angle between the axis of the lowest susceptibility and
the applied magnetic eld.
2. When having objects with a
shape anisotropy
, orientation of the
longest axis parallel to the magnetic eld minimizes the magnetic en-
ergy due to the so-called demagnetization eect.
To nd out if the mesoscopic structure of the microemulsion systems can
be manipulated by a magnetic eld, neutron scattering was done under mag-
netic eld. In case of an inuence of the magnetic eld on the microemulsion
domain size, spontaneous curvature or lm rigidity this should be seen in a
change in the (isotropic) spectra shape. A complete destabilization (passing
the phase boundary to a multi-phase region) would not imperatively be seen
in the scattering pattern. Instead of that due to the slow remixing of a phase
separated system in a cuvette of 1mm thickness, this case is believed to be at
least recognized after a measurement cycle during sample changing. In case
of an orientation parallel or perpendicular to the eld the two-dimensional
detector image would show an anisotropic pattern. In Fig. 5.20 the location
of the samples in the phase diagram are shown. These samples were scanned
90
Figure 5.20
Kahlweit sh diagrams of all investigated microemulsion systems
with C
2
mimFeCl
4
(dashed lines) or C
4
mimFeCl
4
(straight lines) as polar phase and
C
14
mimCl (blue), C
16
mimCl (purple) or C
18
mimCl (red) as surfactant. The crosses
(for C
2
mimFeCl
4
systems) and stars (for C
4
mimFeCl
4
systems) show the sample
positions for samples investigated with SANS under magnetic eld.
δ
accounts for
the ratio between cosurfactant and surfactant and is dened by eq. 5.1.
at elds of 0, 2, 4 and 8T. The location within the phase diagram were mainly
chosen to be near the sh tail to have as little excess surfactant as possible
and to have as big structures as possible. Additionally several samples in
the oil rich region were measured. All samples measured under magnetic
eld can be found in the appendix B.6. Because these samples and all the
microemulsion samples showed no eect by the applied magnetic elds, the
spectra are not discussed furthermore here. As shown in Fig. 5.20, for the
system C
2
mimFeCl
4
/C
16
mimCl a series at constant surfactant/decanol-ratio
and with dierent surfactant ratios was done. In the following this series is
described in more detail.
At a starting point the mesoscopic structure along this experimental path
at zero magnetic eld was investigated and a rst insight into that is al-
ready given by a simple observation with crossed polarizers which is shown
in Fig. 5.21. While the samples at lower surfactant concentrations are opti-
91
Figure 5.21
Photographs of samples for the system C
16
mimCl/C
2
mimFeCl
4
be-
tween (top) and without (bottom) cross-polarizers. for position in the phase di-
agramm see Fig. 5.20. The Labels indicate the wt% of surfactant+decanol in
the sample. It should be mentioned that the sample at 64.8wt% shows a macro-
scopic phase separation. The sample at 36.3wt% is only partly birefringent due to
an ongoing slow transition after shaking. The picture on the right is a polarized
microscopy image of the sample at 42.0wt%.
cal isotropic and easy-owing which points to microemulsion structures, the
higher concentrated samples are gels and optically birefringent. Observation
of these samples in thin layers under polarized microscopy identies these
samples as lamellar liquid crystals. The sample at 36.3wt% shows birefrin-
gence which vanished by shaking the sample and the structure is reversibly
build up again after some minutes of resting. This is a hint to a thixotropic
behaviour, or a shear induced phase transition. It should be mentioned that
usually surfactant systems are known to show a shear induced transition from
an isotropic to a lamellar phase and the here reported behaviour shows the
reverse direction.
123,124
The observed SANS spectra of these samples at zero magnetic eld are
shown in Fig. 5.22. All samples show a correlation peak which becomes
sharper and is shifted to higher
q
values with increasing surfactant concen-
tration. Including the macroscopically observed behaviour (uidity, birefrin-
gence) this can be interpreted as a smooth transition from a low viscous,
optically isotropic microemulsion at low surfactant concentration to a high
92
Figure 5.22 left:
SANS curves for microemulsion systems
C
16
mimCl/C
2
mimFeCl
4
, for the location in the phase diagram see Fig. 5.20
right:
peak position vs. surfactant concentration. (The symbol for 64.8wt% expresses
q≥2.1
nm
−1
because no peak was observed within the observation range ending
at this
q
-value.)
viscous (gelly), birefringent (lamellar) liquid crystal phase at high concen-
tration. From the peak position
qmax
a length scale
R= 2π/qmax
for these
structures can be calculated and is as well shown in Fig. 5.22. Of high inter-
est is the transition area between these two well dened structures because
next to the sample with a shear dependent structuring (observed by polar-
ized microscopy) at 36.3wt%, the sample at 42.0wt% is the only mixture
responding to the exposed magnetic eld. The 2D detector image at 2m of
this sample under a magnetic eld of 8Tesla is shown in Fig. 5.23 and shows
a clear orientation in the magnetic eld.
To get a quantitative value for the anisotropy of the scattering pattern
(which is a measure of the degree of orientation in the magnetic eld), the
detector image was averaged slice-wise as shown in Fig. 5.23. The shown
resulting 1D-spectra for an isotropic sample (at 0Tesla) are identically inde-
pendent on the azimuthal angle, an orientation (8Tesla) leads to spectra with
dierent peak amplitudes dependent on the azimuthal angle while the peak
position is not aected and stays constant for dierent azimuthal angles. In
93
Figure 5.23 upper row:
2D-detector images (2m sample to detector distance)
of the sample with 42.0wt% surfactant+decanol (for the position in the phase dia-
gram see Fig. 5.20) at dierent times during the eld-ramp experiment. Triangles
indicate the pixels which were averaged.
bottom row:
Resulting
q
dependent
1D-spectra by performing a slice-wise averaging.
94
Fig. 5.24 (inset) the intensity of the scattering peak (at
q= 1.3
nm
−1
) is
plotted for dierent magnetic elds and accordingly shows no variation with
respect to the azimuthal angle at a eld of 0T. Two maxima appear with
increasing magnetic eld at an angle of 90
◦
and 270
◦
due to the orientation
in the magnetic eld. Additionally several alternative values to estimate the
relative anisotropy were calculated with the help of the program SASET
73
(for details see chapter 2.7.1).
The freshly homogenized sample (by heat) was then investigated while
applying dierent magnetic elds. Results are shown in Fig. 5.24. At an ini-
tal eld free measurement the sample shows spectra whose peak amplitudes
are independent from the analyzed detector angle which indicates a statisti-
cal random distribution. Rising the magnetic eld in a rate of 0.2T/min rst
doesn't change this behavior which indicates that the force on the mesoscopic
domaines induced by the magnetic eld is to weak. Above a critcal eld of
around 5.5T (here, of course, a small temporal retardation might be present)
the sample starts to become more and more anisotropic and this eect is pro-
portional to the rising eld. Holding then the eld constant at a value of 8T
the anisotropy still increases but on a much smaller timescale. When then
applying a eld ramp back to 0T, a slow relaxation of the anisotropy is visi-
ble but even after hours the sample still shows an anisotropy. This behaviour
can be interpreted by three dierent processes: Starting above the critical
eld strength, the resulting magnetic force is strong enough to overcome the
(sterical) hindrance to move domains, this process (I) is fast and propor-
tional to the magnetic eld. At a constant eld above the critical eld (in
our experiment 8T) a second process (II) of orientation is visible which has a
much slower kinetic rate. Finally the relaxation below the critical eld value
back to the isotropic state (III) should be due to thermal uctuations and
is therefore much slower. For process I it is dicult to extract quantitative
kinetic values as the magnetic eld is not constant, for II and III, by applying
a simple single-exponential law as done with eq. 5.18 and 5.19, respectively,
one can estimate the range of characteristic times for these processes.
95
Figure 5.24
Parameters to represent a change in sample anisotropy vs. time while
applying a magnetic eld prole.
top:
Peak-maxima from 1D-spectra extracted by
a slicewise analysis of the 2D-detector images as described in Fig. 5.23.
bottom:
Dierent parameters quantifying the anisotropy as determined with the program
SASET. Details are described in chapter 2.7.1. Lines are ts with eq. 5.18 (broken
line) and 5.19 (dotted line) with parameters listed in table 5.6.
inset:
Averaged
intensity for the radial segment of 1.25nm
−1≤q≤1.43
nm
−1
as a function of the
azimuthal angle for selected magnetic elds.
96
II:
A=A∞−(A∞−A0)·exp [t0−t
τ2]
(5.18)
III:
A=A0·exp [t0−t
τ3]
(5.19)
Here it is assumed that the alignment factor
Af
is proportional to the
amount of aligned domains,
A0
and
A∞
are the alignment factor at time
t0
and innity, respectively. Fit results are given in table 5.6. Although
the underlying data is quite incomplete to make substantiated statements, it
gives an estimate for the maximum alignment at 8T (
A∞
) and a dierence
in the rate of process II and III of about a factor of 10.
Table 5.6
Fit parameters derived when applying eq. 5.18 and 5.19 to the alignment
factor (
Af
) shown in Fig. 5.24.
t range/min
t0
/min
A0A∞τ
/min
II
42 −77
42 0.14 0.18 12.6
III
77 −100
77 0.18 150.2
Process III can be interpreted as a rotational diusion of the aligned
domains back to the random isotropic state. The corresponding rotational
diusion coecient (
D⊥
r
) can be calculated via eq. 5.20. Assuming a simple
cylindrical or disk-like shape of the rotating domains, characterized by a
width length (
L
) and a diameter (
d
), the order of the size of the domains can
be estimated via eq. 5.21.
125
D⊥
r=1
6τ3
= 1.85 ·10−5
s
−1
(5.20)
= 3kT(ln p−0.662 + 0.917/p −0.050/p2)
πη0L3
(5.21)
Here
p=L/d
is the aspect ratio and
η0
is the viscosity of the solvent. In
Fig. 5.25, values for
L
and
d
are plotted for dierent aspect ratios, which
97
fulll eq. 5.21. It can be seen that from this model domain sizes in the range
of micrometers are predicted.
Figure 5.25
Values for
L
and
d
as a function of the aspect ratio
p
, which are
fullling eq. 5.21. As solvent viscosity the value for cyclohexane
η0= 0.89
mPas
was used.
126
The fact that only the sample at 42.0wt% was able to orient in the eld
could be explained due to a coincide of structure domains which are big and
sti enough to produce a reasonable eld induced force which can compete
against Brownian motion with a not to viscous sample (due to not too big
domains) to make the domains still able to move and orient in the magnetic
eld in a reasonable time. Naturally although these conclusions are based
on a weak empirical foundation and the experiments should be veried by
a more detailed study, the results can be seen as a starting point and a
working hypothesis on which ongoing research can build up in the eld of
systems based on MRTILs, where the orientation of a LC phase and thereby
the optical properties are controlled by the magnetic eld.
98
6
Conclusion
In this work the surfactant self-assembly in a nonaqueous system was investi-
gated, realized by using dierent alkylmethylimidazolium tetrachloroferrates
(C
i
mimFeCl
4
with i = 2, 4, 6) as solvent and imidazolium based surfactants
(C
j
mimCl with j=12, 14, 16, 18) as amphiphile.
In a systematic fashion the phase behavior was studied. For this purpose
we started with the simplest case of binary IL/surfactant mixtures where the
alkyl chain length of surfactant and IL was varied over a broad temperature
range and the complete range of compositions. In this way it was possible to
nd classical mesoscopic structures like micelles and liquid crystalline struc-
tures. The complexity was extended by adding oil and cosurfactant to the
system which enabled us to formulate microemulsions. Again the inuence
of surfactant and IL alkyl chain lengths on the phase behavior was investi-
gated and additionally the investigation was broadened by a wide variation
of the structure and amount of the cosurfactant and oil. To ensure an as
substantive and reliable picture as possible it was made use of many comple-
99
mentary methods as calorimetry (DSC), polarized microscopy, neutron and
X-ray scattering (SANS/SAXS), and surface tension.
In general it was proven that it is possible to form typical self-assembled
structures in this IL-based matrix like micelles, liquid crystals, emulsions and
microemulsions as they are common for classical aqueous systems. However
in dierence to the latter ones it was shown that the ability to self-assemble is
weaker which is expressed e.g. by higher critical aggregation concentrations
leading to micelles with rather low aggregation numbers and which are partly
swollen by the solvent, or smaller tri-phasic regions for microemulsions, in
which the mesoscopic domains show a less pronounced long range ordering.
For the formulated microemulsions it was found that they were only to be
formed when adding a cosurfactant. Here a broad range of dierent branched
and aliphatic alcohols was investigated whereby the eciency to form single-
phasic systems was found to scale with the alkyl chain length of the alco-
hol while the range of having monophasic microemulsions becomes at the
same time smaller upon increasing the chain length of the used aliphatic oil.
Furthermore cyclohexane showed outstanding quality for the microemulsion
formulation. The variation of the surfactant chain length shows a classical
behavior expressed by an enhancement of solubilization strength and lm
rigidity with increasing chain length. Values for
κ
,
c
(mean bending modu-
lus and roughness parameter from the CRW model, respectively) and
ξ
are
comparable with water systems in the oil rich region. With an increasing
content of MRTIL all this parameters point to less and less structured sys-
tems with interpenetrating phases leading to a rough and less sti interface
with less pronounced long range ordering. The local separation between the
hydrophilic and hydrophobic domains in the microemulsion was found to
be along a surface which divides the surfactant/alcohol alkyl chains almost
equally between them.
The weakness in self-assembly was quantied by the solvophobic eect of
the alkyl chain which is in the MRTIL only about a fth of that in water. It
was distinguished between the eects of the solvophobic and -philic part of
the surfactant and as a result it was quantitatively shown that decits in the
ability to self-assemble are mainly present in the surfactant's solvophobic
100
tail. Two opposed trends for the amphiphilic strength could be pointed
out given on the one hand by the length of surfactant alkyl chains which
quanties the solvophobicity of the amphiphile, and on the other hand by
the length of MRTIL alkyl chains, which quanties the solvent polarity. It
could be shown, that this can be expressed by the working hypothesis
∝
j−i
(with j and i as the number of methylene groups in the surfactant and
MRTIL chains, respectively) whose proportionality is followed by a lot of
values characterizing the amphiphilicity of the dierent systems, such as the
cmc
, the critical concentration to form liquid crystals or the sh tail position
in microemulsions.
As for this study ionic liquids with paramagnetic properties were cho-
sen, it was proven that this property is still present in the formulated mi-
croemulsion systems. As a second result it was possible to orient mesoscopic
structures in an external magnetic eld. However this was only possible for
certain locations in the phase diagrams where a lamellar phase was present
and at rather high magnetic elds of
≥5.5
Tesla.
In summary, the here presented broad investigation yields quantitative
information on the composition-structure relationship and therefore gives
recipes to design magnetic self-assembled structures with optimised proper-
ties and structures, as it has not yet been done for such systems that can
be manipulated by a magnetic eld. These ndings are useful for designing
strategies for formulating microemulsions of a given structure with MRTILs
as polar component. This is important as such microemulsions or emulsions
could in the future be employed as interesting reaction media which contain
also a component for separation via magnetic forces.
101
102
References
[1] Tanford, C.
Science
1978
,
200
, 10121018.
[2] Tanford, C.
The hydrophobic eect: formation of micelles and biological
membranes
, 2nd ed.; Wiley Interscience Publications; Wiley, 1980.
[3] Grin, W. C.
J. Soc. Cosmet. Chem.
1949
,
1
, 311326.
[4] Salager, J.-L.; Marquez, N.; Graciaa, A.; Lachaise, J.
Langmuir
2000
,
16
, 55345539.
[5] Helfrich, W.
Z. Naturforsch.
1973
,
28c
, 693703.
[6] Kunz, W.; Testard, F.; Zemb, T.
Langmuir
2008
,
25
, 112115.
[7] Kumar, P. In
Handbook of microemulsion science and technology
; Ku-
mar, P., Mittal, K. L., Eds.; Dekker, 1999; p 849.
[8] Duvail, M.; Dufreche, J.-F.; Arleth, L.; Zemb, T.
Phys. Chem. Chem.
Phys.
2013
,
15
, 71337141.
[9] Handy, S. T.
Curr. Org. Chem.
2005
,
9
, 959988.
[10] Chen, Y.; Zu, Y.; Fu, Y.; Zhang, X.; Yu, P.; Sun, G.; Eerth, T.
Molecules
2010
,
15
, 94869495.
[11] Duan, Z.; Gu, Y.; Deng, Y.
Catal. Commun.
2006
,
7
, 651656.
[12] Olivier-Bourbigou, H.; Magna, L.; Morvan, D.
Appl. Catal., A
2010
,
373
, 156.
103
[13] Yoshida, Y.; Saito, G.
Phys. Chem. Chem. Phys.
2010
,
12
, 167584.
[14] Mallick, B.; Balke, B.; Felser, C.; Mudring, A. V.
Angew. Chem., Int.
Ed.
2008
,
47
, 76357638.
[15] Chen, Z.; Zhang, S.; Qi, X.; Liu, S.; Zhang, Q.; Deng, Y.
J. Mater.
Chem.
2011
,
21
, 89798982.
[16] Hayashi, S.; Hamaguchi, H.
Chem. Lett.
2005
,
34
, 740740.
[17] Del Sesto, R. E.; McCleskey, T. M.; Burrell, A. K.; Baker, G. A.;
Thompson, J. D.; Scott, B. L.; Wilkes, J. S.; Williams, P.
Chem. Com-
mun.
2008
, 447449.
[18] Krieger, B. M.; Lee, H. Y.; Emge, T. J.; Wishart, J. F.; Cast-
ner, E. W., Jr.
Phys. Chem. Chem. Phys.
2010
,
12
, 89198925.
[19] Peppel, T.; Kockerling, M.; Geppert-Rybczynska, M.; Ralys, R. V.;
Lehmann, J. K.; Verevkin, S. P.; Heintz, A.
Angew. Chem., Int. Ed.
2010
,
49
, 71167119.
[20] Plechkova, N. V.; Seddon, K. R.
Chem. Soc. Rev.
2008
,
37
, 123150.
[21] Kokorin, A., Ed.
Ionic Liquids: Theory, Properties, New Approaches
;
InTech, 2011; p 748.
[22] Welton, T.
Coord. Chem. Rev.
2004
,
248
, 24592477.
[23] Ray, A.
J. Am. Chem. Soc.
1969
,
91
, 65116512.
[24] Ward, A. J.; du Reau, C. In
Surface and Colloid Science
; Matije-
vic, E., Ed.; Surface and Colloid Science; Plenum Press, New York,
1993; Vol. 15; Chapter 4.
[25] Singh, H. N.; Saleem, S. M.; Singh, R. P.; Birdi, K. S.
J. Phys. Chem.
1980
,
84
, 21912194.
[26] Seguin, C.; Eastoe, J.; Clapperton, R.; Heenan, R. K.; Grillo, I.
Colloids
Surf., A
2006
,
282283
, 134142.
104
[27] Hildebrand, J.
Solubility of Non-electrolytes
; ACS monograph; Rein-
hold Publishing Corporation, 1936.
[28] Gordon, J. E.
The organic chemistry of electrolyte solution
; Wiley, New
York, 1975; pp 158162.
[29] Evans, D. F.
Langmuir
1988
,
4
, 312.
[30] Welton, T.
Chem. Rev.
1999
,
99
, 20712083.
[31] Wasserscheid, P.; Keim, W.
Angew. Chem., Int. Ed.
2000
,
39
, 3772
3789.
[32] Huddleston, J. G.; Visser, A. E.; Reichert, W. M.; Willauer, H. D.;
Broker, G. A.; Rogers, R. D.
Green Chem.
2001
,
3
, 156164.
[33] Evans, D.; Yamauchi, A.; Roman, R.; Casassa, E. Z.
J. Colloid Inter-
face Sci.
1982
,
88
, 8996.
[34] Greaves, T. L.; Drummond, C. J.
Chem. Soc. Rev.
2008
,
37
, 1709
1726.
[35] Hao, J. C.; Zemb, T.
Curr. Opin. Colloid Interface Sci.
2007
,
12
, 129
137.
[36] Sharma, S. C.; Atkin, R.; Warr, G. G.
J. Phys. Chem. B
2013
,
117
,
1456814575.
[37] Chaudhary, G. R.; Bansal, S.; Mehta, S.; Ahluwalia, A.
J. Chem. Ther-
modyn.
2012
,
50
, 6370.
[38] Anderson, J. L.; Pino, V.; Hagberg, E. C.; Sheares, V. V.; Arm-
strong, D. W.
Chem. Commun.
2003
, 24442445.
[39] Araos, M. U.; Warr, G. G.
Langmuir
2008
,
24
, 93549360.
[40] Fletcher, K. A.; Pandey, S.
Langmuir
2003
,
20
, 3336.
105
[41] Patrascu, C.; Gaure, F.; Nallet, F.; Bordes, R.; Oberdisse, J.;
de Lauth-Viguerie, N.; Mingotaud, C.
ChemPhysChem
2006
,
7
, 99
101.
[42] Inoue, T.; Yamakawa, H.
J. Colloid Interface Sci.
2011
,
356
, 798802.
[43] Atkin, R.; Bobillier, S. M. C.; Warr, G. G.
J. Phys. Chem. B
2010
,
114
, 13501360.
[44] Kimizuka, N.; Nakashima, T.
Langmuir
2001
,
17
, 67596761.
[45] López-Barrón, C. R.; Li, D.; DeRita, L.; Basavaraj, M. G.; Wag-
ner, N. J.
J. Am. Chem. Soc.
2012
,
134
, 2072820732, PMID:
23030359.
[46] Rao, K. S.; So, S.; Kumar, A.
Chem. Commun.
2013
,
49
, 81118113.
[47] Li, J.; Zhang, J.; Zhao, Y.; Han, B.; Yang, G.
Chem. Commun.
2012
,
48
, 994996.
[48] Li, J.; Zhang, J.; Han, B.; Zhao, Y.; Yang, G.
J. Colloid Interface Sci.
2012
,
368
, 395399.
[49] Zech, O.; Thomaier, S.; Kolodziejski, A.; Touraud, D.; Grillo, I.;
Kunz, W.
J. Colloid Interface Sci.
2010
,
347
, 227232.
[50] Eastoe, J.; Gold, S.; Rogers, S. E.; Paul, A.; Welton, T.; Heenan, R. K.;
Grillo, I.
J. Am. Chem. Soc.
2005
,
127
, 73027303.
[51] Li, N.; Zhang, S. H.; Zheng, L. Q.; Gao, Y.; Yu, L.
Langmuir
2008
,
24
, 29732976.
[52] Mehta, S. K.; Kaur, K.
Anal. Chem. Indian Journal of Chemistry,
Section A: Inorganic, Bio-inorganic, Physical, Theoretical & Analytical
Chemistry
2010
,
49
, 662684.
[53] Zech, O.; Kunz, W.
Soft Matter
2011
,
7
, 55075513.
106
[54] Liu, L.; Bauduin, P.; Zemb, T.; Eastoe, J.; Hao, J.
Langmuir
2009
,
25
, 20552059.
[55] Rojas, O.; Koetz, J.
J. Surface Sci. Technol.
2010
,
26
, 173196.
[56] Gao, Y.; Hilfert, L.; Voigt, A.; Sundmachert, K.
J. Phys. Chem. B
2008
,
112
, 37113719.
[57] Atkin, R.; Warr, G. G.
J. Phys. Chem. B
2007
,
111
, 93099316.
[58] Seth, D.; Chakraborty, A.; Setua, P.; Sarkar, N.
Langmuir
2006
,
22
,
77687775.
[59] Gao, Y.; Li, N.; Zheng, L. Q.; Zhao, X. Y.; Zhang, S. H.; Han, B. X.;
Hou, W. G.; Li, G. Z.
Green Chem.
2006
,
8
, 4349.
[60] Behera, K.; Kumar, V.; Pandey, S.
ChemPhysChem
2010
,
11
, 1044
1052.
[61] Gao, H.; Li, J.; Han, B.; Chen, W.; Zhang, J.; Zhang, R.; Yan, D.
Phys. Chem. Chem. Phys.
2004
,
6
, 29142916.
[62] Kahlweit, M.; Strey, R.
Angew. Chem., Int. Ed.
1985
,
24
, 654668.
[63] Zech, O.; Thomaier, S.; Bauduin, P.; Ruck, T.; Touraud, D.; Kunz, W.
J. Phys. Chem. B
2009
,
113
, 465473.
[64] Rabe, C.; Koetz, J.
Colloids Surf., A
2010
,
354
, 261267.
[65] Yoshida, Y.; Saito, G.
J. Mater. Chem.
2006
,
16
, 12541262.
[66] Thomaier, S. Formulation and characterization of new innovative col-
loidal systems involving ionic liquids for the application at high tem-
peratures. Ph.D. thesis, University of Regensburg, 2009.
[67] Sivia, D.
Elementary Scattering Theory
; Oxford University Press, 2011.
[68] Lindner, P., Zemb, T., Eds.
Neutrons, X-rays and Light: Scattering
Methods Applied to Soft Condensed Matter
, 1st ed.; Elsevier Science
B.V., 2002.
107
[69] Keiderling, U.
Appl. Phys. A: Mater. Sci. Process.
2002
,
74
, s1455
s1457.
[70] Kohlbrecher, J.; Bressler, I. http://kur.web.psi.ch/sans1/SANSSoft/sast.html.
[71] Richard, D.; Ferrand, M.; Kearley, G. J.
J. Neutron Res.
1996
,
4
,
3339.
[72] Zhang, F.; Ilavsky, J.; Long, G. G.; Quintana, J. P. G.; Allen, A. J.;
Jemian, P. R.
Metall. Mater. Trans. A
2010
,
41
, 11511158.
[73] Muthig, M.; Prevost, S.; Orglmeister, R.; Gradzielski, M.
J. Appl. Crys-
tallogr.
2013
,
46
, 11871195.
[74] Heintz, A.; Lehmann, J. K.; Kozlova, S. A.; Balantseva, E. V.; Bazyl-
eva, A. B.; Ondo, D.
Fluid Phase Equilib.
2010
,
294
, 187196, Ionic
Liquids Special Issue.
[75] uczak, J.; Jungnickel, C.; Joskowska, M.; Thöming, J.; Hupka, J.
J.
Colloid Interface Sci.
2009
,
336
, 111116.
[76] Galgano, P. D.; Seoud, O. A. E.
J. Colloid Interface Sci.
2011
,
361
,
186194.
[77] Laughlin, R. G.
The Acqueous Phase Behaviour of Surfactants
; Colloid
Science; Academic Press Inc., 1996.
[78] Greaves, T. L.; Weerawardena, A.; Fong, C.; Drummond, C. J.
Lang-
muir
2007
,
23
, 402404.
[79] Dirand, M.; Bouroukba, M.; Briard, A.-J.; Chevallier, V.; Petitjean, D.;
Corriou, J.-P.
J. Chem. Thermodyn.
2002
,
34
, 12551277.
[80] Mosselman, C.; Mourik, J.; Dekker, H.
J. Chem. Thermodyn.
1974
,
6
,
477487.
[81] Li, L.; Groenewold, J.; Picken, S. J.
Chem. Mater.
2005
,
17
, 250257.
108
[82] Hayashi, S.; Saha, S.; Hamaguchi, H. O.
IEEE T. Magn.
2006
,
42
,
1214.
[83] Meissner, H. P.; Michaels, A. S.
Ind. Eng. Chem. Res.
1949
,
41
, 2782
2787.
[84] Szyszkowski, B.
Z. physik. Chemie
1908
,
64
, 385.
[85] Harned, H. S.; Owen, B. B.
The physical chemistry of electrolytic so-
lutions
, 3rd ed.; Reinhold Publishing corporation, New York, 1958;
Chapter 14, p 603.
[86] Seoud, O. A. E.; Pires, P. A. R.; Abdel-Moghny, T.; Bastos, E. L.
J.
Colloid Interface Sci.
2007
,
313
, 296304.
[87] Thomaier, S.; Kunz, W.
J. Mol. Liq.
2007
,
130
, 104107.
[88] Ashcroft, N. W.; Lekner, J.
Phys. Rev.
1966
,
145
, 8390.
[89] Bowlas, C. J.; Bruce, D. W.; Seddon, K. R.
Chem. Commun.
1996
,
16251626.
[90] Bradley, A. E.; Hardacre, C.; Holbrey, J. D.; Johnston, S.; McMath, S.
E. J.; Nieuwenhuyzen, M.
Chem. Mater.
2002
,
14
, 629635.
[91] Li, C.; He, J.; Chen, J.; Liu, J.; Zhang, Q.; Yu, Z.
J. Colloid Interface
Sci.
2011
,
359
, 474480.
[92] Watson, L. In
Encyclopedia of Surface and Colloid Science
, 2nd ed.;
Hubbart, A. T., Somasundaran, P., Eds.; Encyclopedia of Surface and
Colloid Science; Taylor & Francis, 2006; Vol. 2; pp 1014 1025.
[93] Vaghela, N. M.; Sastry, N. V.; Aswal, V. K.
Colloids Surf., A
2011
,
373
, 101 109.
[94] Sastry, N. V.; Vaghela, N. M.; Macwan, P. M.; Soni, S. S.; Aswal, V. K.;
Gibaud, A.
J. Colloid Interface Sci.
2012
,
371
, 52 61.
109
[95] Gómez-Díaz, D.; Navaza, J. M.; Sanjurjo, B.
J. Chem. Eng. Data
2007
,
52
, 889891.
[96] Wu, F.-G.; Wang, N.-N.; Zhang, Q.-G.; Sun, S.-F.; Yu, Z.-W.
J. Colloid
Interface Sci.
2012
,
374
, 197205.
[97] Laughlin, R. G. In
Cationic Surfactants: Physical Chemistry
, 2nd ed.;
Rubingh, D., Holland, P., Eds.; Surfactant Science; Taylor & Francis,
1990; Vol. 37; pp 140.
[98] Rosen, M. J.
Surfactants and Interfacial Phenomena
, 3rd ed.; John
Wiley & Sons, Inc., 2004; Chapter 3, pp 105177.
[99] Pedersen, J. S.; Gerstenberg, M. C.
Colloids Surf., A
2003
,
213
, 175
187.
[100] Yang, L.; Alexandridis, P.; Steytler, D. C.; Kositza, M. J.;
Holzwarth, J. F.
Langmuir
2000
,
16
, 85558561.
[101] Huibers, P. D. T.; Shah, D. O.
Langmuir
1997
,
13
, 57625765.
[102] Gradzielski, M.
Langmuir
1998
,
14
, 60376044.
[103] Penders, M. H. G. M.; Strey, R.
J. Phys. Chem.
1995
,
99
, 1031310318.
[104] Gradzielski, M.
Curr. Opin. Colloid Interface Sci.
2008
,
13
, 263269.
[105] Teubner, M.; Strey, R.
J. Chem. Phys.
1987
,
87
, 31953200.
[106] Gradzielski, M.; Langevin, D.; Sottmann, T.; Strey, R.
J. Chem. Phys.
1996
,
104
, 37823787.
[107] Gompper, G.; Endo, H.; Mihailescu, M.; Allgaier, J.; Monkenbusch, M.;
Richter, D.; Jakobs, B.; Sottmann, T.; Strey, R.
Europhys. Lett.
2001
,
56
, 683689.
[108] Schubert, K. V.; Strey, R.; Kline, S. R.; Kaler, E. W.
J. Chem. Phys.
1994
,
101
, 53435355.
110
[109] Jouroy, J.; Levinson, P.; de Gennes, P. G.
J. Physique
1982
,
43
,
12411248.
[110] Sottmann, T.; Strey, R.; Chen, S. H.
J. Chem. Phys.
1997
,
106
, 6483
6491.
[111] Chen, S.-H.; Lee, D.; Chang, S.-L.
J. Mol. Struct.
1993
,
296
, 259264.
[112] Chen, S.-H.; Choi, S.-M.
J. Appl. Crystallogr.
1997
,
30
, 755760.
[113] Choi, S.; Chen, S.; Sottmann, T.; Strey, R.
Physica A
2002
,
304
, 85
92, Scattering Studies of Mesoscopic Scale Structure and Dynamics in
Soft Matter.
[114] Chen, S.-H.; Choi, S.-M.
Physica A
1997
,
236
, 3851, Proceedings of
the Workshop on Current Problems in Complex Fluids.
[115] Kotlarchyk, M.
Physica B & C
1986
,
136
, 274280.
[116] Auvray, L.; Cotton, J. P.; Ober, R.; Taupin, C.
J. Phys. Chem.
1984
,
88
, 45864589.
[117] Arleth, L.; Pedersen, J. S.
Phys. Rev. E: Stat., Nonlinear, Soft Matter
Phys.
2001
,
63
, 061406.
[118] Klostermann, M.; Foster, T.; Schweins, R.; Lindner, P.; Glatter, O.;
Strey, R.; Sottmann, T.
Phys. Chem. Chem. Phys.
2011
,
13
, 20289
20301.
[119] Balogh, J.; Olsson, U.; Pedersen, J. S.
J. Phys. Chem. B
2007
,
111
,
682689.
[120] Odenbach, S.
J. Phys.: Condens. Matter
2004
,
16
, R1135R1150.
[121] Kumar, M.; Gupta, R. In
Diamagnetic Susceptibility of Organic Com-
pounds, Oils, Parans and Polyethylenes
; Gupta, R., Ed.; Landolt-
Börnstein - Group II Molecules and Radicals; Springer Berlin Heidel-
berg, 2008; Vol. 27B; pp 11901190.
111
[122] Rikken, R. S. M.; Nolte, R. J. M.; Maan, J. C.; van Hest, J. C. M.;
Wilson, D. A.; Christianen, P. C. M.
Soft Matter
2014
,
10
, 12951308.
[123] Porcar, L.; Hamilton, W. A.; Butler, P. D.; Warr, G. G.
Phys. Rev.
Lett.
2004
,
93
, 198301.
[124] Léon, A.; Bonn, D.; Meunier, J.; Al-Kahwaji, A.; Kellay, H.
Phys. Rev.
Lett.
2001
,
86
, 938941.
[125] Ortega, A.; Garciá de la Torre, J.
J. Chem. Phys.
2003
,
119
, 9914
9919.
[126] Wohlfarth, C. In
Supplement to IV/18
; Lechner, M., Ed.; Landolt-
Börnstein - Group IV Physical Chemistry; Springer Berlin Heidelberg,
2009; Vol. 25; pp 377380.
[127] Sears, V. F.
Neutron News
1992
,
3
, 2637.
[128] Lagourette, B.; Peyrelasse, J.; Boned, C.; Clausse, M.
Nature
1979
,
281
, 6062.
112
A
Appendix Binary Systems
A.1 SANS model tting
A.1.1 Spherical model as used in section 3
I(q) = BG +∫LN(R, R0, σ)·P(R)·S(RHS, φHS)
d
R
(A.1)
where
BG
is the background. The polydisperse form factor is given by a
spherical model:
P(q, R) = [4
3πR3∆SLD 3sin qR −qR cos qR
(qR)3]2
(A.2)
with
∆SLD
as the scattering contrast between micelles and bulk phase
and where
R
is expressed by the lognormal distribution:
LN(R, R0, σ) = N
σR√2πexp(−ln(R/R0)2
2σ2)
(A.3)
I
The
nth
moment
⟨Rn⟩
of the micelle radius can be calculated as
⟨Rn⟩=Rn
0exp 1
2σ2n2
(A.4)
S(q, RHS, φHS)
is given by the structure factor of a hard sphere
88
whereby
its number density is equal to the form factor which is ensured by its radius
and volume fraction dened as
RHS =R+ ∆R
(A.5a)
φHS =φmicelle
(R+ ∆R)3
R3
(A.5b)
Scattering contrast and volume fraction were calculated from the sample
composition whereby it was assumed that the micelles are partly swollen
by solvent (given by the ratio
x
of the total MRTIL volume located in the
micelles) and the surfactant is partly monomerically dissolved in the bulk
phase (given by the ratio
y
of the total surfactant chains located in the
micelles). The surfactant head groups are assumed to be counted to the bulk
phase. This is quantied by the following set of equations:
x=φMRTIL,m/φMRTIL
(A.6a)
y=φCn,b/φCn
(A.6b)
w=φCn/φsurfactant
(A.6c)
φbulk =φMRT IL(1 −x) + φsurfactant(1 −w) + φsurfactantwy
(A.6d)
φmicelle = 1 −φbulk
(A.6e)
SLDbulk = [sldMRTILφMRTIL(1 −x) + sldmimClφsurfactant(1 −w)
+sldCnφsurfactantwy]/φbulk
(A.6f)
SLDmicelle = [sldMRT ILφMRTILx+sldCnφsurfactantw(1 −y)]/φmicelle
(A.6g)
α=xφMRTIL
φmicelle
(A.6h)
II
with
sldi
as the scattering length density for the pure molecule (part)
indicated by
i
(see table A.2).
φMRTIL
,
φsurfactant
and
φCn
are the total
volume fractions of MRTIL, surfactant and the surfactant alkyl chain in a
sample, respectively.
φMRTIL,m
is the total volume fraction of MRTIL located
in micelles and
φCn,b
is the total volume fraction of alkyl chains located in the
bulk (continuous solvent). The volume fraction of ionic liquid in the micelle
relative to the micelle volume is then given by
α
. Used densities are listed
in table A.1. The average aggregation number
Nagg
of surfactant molecules
per micelle is then
Nagg =4
3π⟨R3⟩1
vHC
(1 −α)
(A.7)
The average headgroup spacing
as
is then given by
as=4π⟨R2⟩
Nagg
(A.8)
Table A.1
Densities used for the SANS model tting.
compound
density
g cm−3
C
2
mimFeCl
4
1.440
C
4
mimFeCl
4
1.360
C
14
mimCl 0.970
C
16
mimCl 0.960
C
18
mimCl 0.952
C
14
0.809
C
16
0.812
C
18
0.820
mimCl 1.452
Fig. A.2 shows the measured SANS curves including the ts and table
A.2 gives the underlying t parameters. The calculated aggregation num-
bers, surfactant headgroup spacings and solvent ratio in the micelles show a
coherent picture of the surfactant concentration dependent aggregation be-
III
haviour (see Fig. 3.7 and A.1): At very low concentrations the aggregates
contain only a few surfactant molecules swollen by a great amount of solvent
leading to more loosely packed micelles with high headgroup spacings. With
increasing surfactant concentration the aggregation number increases which
favors more compact micelles with less solvent inside and smaller surfac-
tant headgroups. Finally at high surfactant concentrations (near the phase
boundary)
α
approaches zero and
as
reaches values comparable to the results
from surface tension measurements (see table A.5) and values known from
water systems.
Figure A.1
Volume ratio of MRTIL in micelles (
α
) and headgroup area (
as
) as
derived from SANS data described in section A.1.1 for systems with C
2
mimFeCl
4
(open symbols) and C
4
mimFeCl
4
(lled symbols) as solvent.
IV
Table A.2
Input parameters and t results for the model described in section A.1.1 for all SANS samples.
C
i
mimFeCl
4
C
j
mimCl
given by sample composition
sldi
10−4
nm
−2
given by literature
t parameter calculated
φm,surf
wt%
φIL φsurf. wi=MRT IL i =
C
ni=
mimCl
x y σ R0
nm
∆R
nm
BG
cm
−1αas
nm
2
2 14 50 0.40 0.60 0.751 1.64 -0.38 2.41 0.18 0.11 0.159 1.675 0.174 0.610 0.12 0.77
2 14 40 0.50 0.50 0.751 1.64 -0.38 2.41 0.29 0.00 0.201 1.566 0.047 0.546 0.22 0.90
2 14 31 0.60 0.40 0.751 1.64 -0.38 2.41 0.28 0.00 0.256 1.286 0.024 0.503 0.30 1.14
2 14 21 0.72 0.28 0.751 1.64 -0.38 2.41 0.20 0.10 0.323 0.845 0.000 0.460 0.36 1.73
2 14 18 0.75 0.25 0.751 1.64 -0.38 2.41 0.09 0.23 0.333 0.634 0.006 0.451 0.27 1.99
2 14 16 0.77 0.23 0.751 1.64 -0.38 2.41 0.11 0.24 0.345 0.598 0.002 0.442 0.34 2.28
2 14 14 0.80 0.20 0.751 1.64 -0.38 2.41 0.04 0.50 0.286 0.576 0.000 0.430 0.22 2.20
2 14 12 0.83 0.17 0.751 1.64 -0.38 2.41 0.04 0.62 0.135 0.800 0.000 0.423 0.32 2.15
2 14 10 0.86 0.14 0.751 1.64 -0.38 2.41 0.04 0.64 0.063 0.825 0.000 0.411 0.41 2.49
2 16 24 0.68 0.32 0.774 1.64 -0.37 2.41 0.20 0.00 0.254 1.442 0.039 0.478 0.30 1.17
2 16 18 0.76 0.24 0.774 1.64 -0.37 2.41 0.15 0.09 0.306 1.076 0.001 0.442 0.33 1.52
2 16 14 0.81 0.19 0.774 1.64 -0.37 2.41 0.08 0.27 0.332 0.828 0.003 0.422 0.32 1.87
2 16 11 0.84 0.16 0.774 1.64 -0.37 2.41 0.01 0.53 0.296 0.674 0.000 0.419 0.13 2.19
2 16 9 0.87 0.13 0.774 1.64 -0.37 2.41 0.01 0.63 0.318 0.595 0.021 0.398 0.18 2.68
2 16 7 0.90 0.10 0.774 1.64 -0.37 2.41 0.03 0.59 0.178 0.739 0.003 0.393 0.36 1.89
2 18 20 0.72 0.28 0.792 1.64 -0.37 2.41 0.14 0.01 0.223 1.810 0.125 0.465 0.27 1.03
2 18 15 0.79 0.21 0.792 1.64 -0.37 2.41 0.14 0.00 0.261 1.575 0.057 0.440 0.33 1.24
2 18 12 0.83 0.17 0.792 1.64 -0.37 2.41 0.12 0.02 0.313 1.285 0.000 0.417 0.38 1.52
2 18 10 0.86 0.14 0.792 1.64 -0.37 2.41 0.05 0.33 0.348 0.927 0.005 0.408 0.30 1.74
2 18 7 0.89 0.11 0.792 1.64 -0.37 2.41 0.01 0.60 0.287 0.931 0.000 0.393 0.23 1.78
2 18 6 0.92 0.08 0.792 1.64 -0.37 2.41 0.00 0.81 0.184 0.848 0.091 0.383 0.06 1.76
(Continues on next page.)
V
(Continued from previous page.)
C
i
mimFeCl
4
C
j
mimCl
given by sample composition
sldi
10−4
nm
−2
given by literature
t parameter calculated
φm,surf
wt%
φIL φsurf. wi=MRT IL i =
C
ni=
mimCl
x y σ R0
nm
∆R
nm
BG
cm
−1αas
nm
2
4 14 67 0.26 0.74 0.751 1.55 -0.38 2.41 0.00 0.22 0.130 1.581 0.262 0.683 0.00 0.75
4 14 55 0.37 0.63 0.751 1.55 -0.38 2.41 0.21 0.08 0.153 1.570 0.122 0.639 0.12 0.84
4 14 43 0.49 0.51 0.751 1.55 -0.38 2.41 0.31 0.00 0.200 1.411 0.097 0.597 0.23 1.02
4 14 31 0.61 0.39 0.751 1.55 -0.38 2.41 0.30 0.01 0.257 1.112 0.116 0.559 0.32 1.38
4 14 28 0.65 0.35 0.751 1.55 -0.38 2.41 0.27 0.01 0.288 0.915 0.088 0.555 0.33 1.63
4 14 25 0.68 0.32 0.751 1.55 -0.38 2.41 0.24 0.00 0.309 0.809 0.110 0.543 0.34 1.81
4 14 23 0.70 0.30 0.751 1.55 -0.38 2.41 0.25 0.00 0.312 0.781 0.110 0.530 0.38 1.98
4 14 21 0.73 0.27 0.751 1.55 -0.38 2.41 0.25 0.03 0.307 0.758 0.127 0.524 0.41 2.19
4 14 15 0.80 0.20 0.751 1.55 -0.38 2.41 0.27 0.00 0.321 0.658 0.200 0.498 0.51 2.96
4 16 50 0.42 0.58 0.774 1.55 -0.37 2.41 0.16 0.12 0.154 1.701 0.187 0.629 0.11 0.86
4 16 38 0.53 0.47 0.774 1.55 -0.37 2.41 0.27 0.00 0.189 1.597 0.085 0.588 0.24 1.03
4 16 28 0.64 0.36 0.774 1.55 -0.37 2.41 0.28 0.00 0.245 1.297 0.071 0.569 0.33 1.36
4 16 20 0.74 0.26 0.774 1.55 -0.37 2.41 0.16 0.21 0.279 0.929 0.155 0.519 0.37 1.95
4 16 18 0.76 0.24 0.774 1.55 -0.37 2.41 0.23 0.00 0.316 0.809 0.049 0.516 0.42 2.30
4 16 16 0.78 0.22 0.774 1.55 -0.37 2.41 0.14 0.33 0.267 0.817 0.174 0.510 0.44 2.50
4 16 14 0.81 0.19 0.774 1.55 -0.37 2.41 0.17 0.34 0.297 0.741 0.256 0.486 0.53 3.16
4 16 12 0.84 0.16 0.774 1.55 -0.37 2.41 0.20 0.31 0.309 0.715 0.294 0.480 0.59 3.73
4 16 10 0.86 0.14 0.774 1.55 -0.37 2.41 0.29 0.20 0.310 0.726 0.294 0.477 0.69 4.88
4 18 25 0.67 0.33 0.792 1.55 -0.37 2.41 0.25 0.00 0.230 1.542 0.079 0.538 0.34 1.32
4 18 19 0.75 0.25 0.792 1.55 -0.37 2.41 0.24 0.00 0.282 1.205 0.013 0.517 0.42 1.80
4 18 15 0.80 0.20 0.792 1.55 -0.37 2.41 0.20 0.09 0.290 1.000 0.000 0.496 0.47 2.36
4 18 12 0.84 0.16 0.792 1.55 -0.37 2.41 0.15 0.18 0.283 0.841 0.003 0.492 0.48 2.94
4 18 10 0.86 0.14 0.792 1.55 -0.37 2.41 0.07 0.55 0.201 0.903 0.149 0.476 0.48 5.08
4 18 8 0.89 0.11 0.792 1.55 -0.37 2.41 0.16 0.51 0.124 1.018 0.082 0.469 0.71 2.87
VI
A.1.2 Alternative spherical model
Alternatively the model described in section A.1.1 was used with
∆SLD
and
volume fractions calculated as described in section A.2. With that eq. A.7
simplies to
Nagg =4
3π⟨R3⟩1
vHC
(A.9)
Resulting t curves and extracted aggregation numbers are plotted in
Fig. A.4 and A.3, respectively.
Figure A.3
Aggregation numbers calculated from SANS model tting described
in section A.1.2 (left) and A.1.3 (right) and listed in table A.3 for surfactant mixed
with C
2
mimFeCl
4
(open symbols) and C
4
mimFeCl
4
(lled symbols).
VIII
Table A.3
Scattering invariant as calculated with eq. 3.2 and 3.3a. The aggregation numbers are as derived for the three
dierent SANS models described in section A.1.1 (I), A.1.2 (II) and A.1.3 (III). Plots are found in Fig. A.3 and 3.7.
C
j
mimCl
C
2
mimFeCl
4
C
4
mimFeCl
4
φm,surf
wt%
Qex
inv
10−7
nm
−4
Qth
inv
10−7
nm
−4Nagg
(I)
Nagg
(II)
Nagg
(III)
φm,surf
wt%
Qex
inv
10−7
nm
−4
Qth
inv
10−7
nm
−4Nagg
(I)
Nagg
(II)
Nagg
(III)
14 50 1.51 2.02 48.0 33.0 65.7 67 1.44 2.16 43.6 47.2 64.2
14 40 1.35 1.72 36.9 18.6 54.8 55 1.43 1.86 38.7 25.6 49.2
14 31 1.23 1.33 20.7 9.0 33.4 43 1.42 1.47 26.5 8.8 29.5
14 21 0.80 0.83 6.4 2.9 15.7 31 1.10 1.01 12.8 4.1 14.7
14 18 0.56 0.64 3.2 2.4 8.8 28 0.80 0.86 7.6 3.9 10.4
14 16 0.50 0.54 2.5 1.7 7.1 25 0.75 0.71 5.5 3.7 7.6
14 14 0.35 0.38 2.2 2.3 4.8 23 0.55 0.61 4.7 2.6 6.4
14 12 0.29 0.26 3.9 2.2 4.8 21 0.47 0.48 4.0 2.3 4.9
14 10 0.20 0.08 3.5 15 0.16 0.16 2.3 1.9 2.4
16 24 1.02 1.16 25.5 12.1 45.7 50 1.39 1.83 44.4 30.9 59.5
16 18 0.84 0.81 11.5 5.1 22.4 38 1.29 1.48 33.4 16.5 43.6
16 14 0.52 0.54 5.8 2.8 16.4 28 1.08 1.08 17.5 5.8 24.4
16 11 0.41 0.38 3.6 2.1 7.6 20 0.67 0.64 6.5 2.5 10.1
16 9 0.26 0.20 2.5 2.6 5.3 18 0.62 0.55 4.4 3.1 7.4
16 7 0.13 0.06 2.7 16 0.40 0.44 3.9 1.4 5.5
16 14 0.31 0.30 2.6 1.3 3.4
16 12 0.18 0.17 2.1 1.9 2.6
16 10 0.15 0.03 1.6
18 20 1.04 1.17 44.1 23.7 70.2 25 0.93 1.15 25.1 8.7 38.5
18 15 0.84 0.87 28.8 12.8 54.8 19 0.71 0.84 11.9 4.0 24.7
18 12 0.58 0.62 16.6 6.8 42.1 15 0.52 0.57 6.3 2.3 12.9
18 10 0.46 0.46 7.8 2.9 12.0 10 0.26 0.28 3.8 0.5 5.0
18 7 0.36 0.29 7.4 1.7 8.9 8 0.16 0.15 2.6 0.4 2.4
18 6 0.21 0.13 5.4 1.4 4.6 12 0.42 0.42 3.6 1.4
X
A.1.3 Ellipsoidal model
I(q) = BG +P(R, ν)·S(RHS, φHS)
d
R
(A.10)
The form factor was chosen to be monodisperse to avoid overtting and
is given by an ellipsoidal model characterized by two equal semi-axis
R
and
a semi-principal axes
νR
.
P(q, R, ν) = ∫π/2
0
K2(q, x) sin Θ
d
Θ
(A.11a)
x=R√ν2cos2Θ + sin2Θ
(A.11b)
K(q, R, ν) = 4
3πνR3∆SLD3sin qx −qx cos qx
(qx)3
(A.11c)
∆SLD
and volume fractions were calculated as described in section A.2.
The hard sphere radius is given by eq. A.12a and denes the hard sphere
volume fraction via eq. A.12b.
RHS =3
√ν(R+ ∆R)
(A.12a)
φHS =φmicelle
ν(RHS)3
R3
(A.12b)
With that the aggregation number can be calculated as
Nagg =4
3πνR31
vHC
(A.13)
Extracted aggregation numbers are plotted in Fig. A.3. The general
trend of the aggregation number found for the other methods/models can be
reproduced.
XI
A.2 Scattering invariant
The scattering length densities needed to calculate the theoretical invariant
described in chapter 3 were derived as followed:
sldsurf =w·sldCn+ (1 −w)·sldmimCl
(A.14a)
SLDbulk =sldMRTIL mIL
ρIL +sldsurf
msurf,b
ρsurf +sldmimCl
msurf −msurf,b
a·ρmimCl
Vtotφbulk
(A.14b)
∆SLD =SLDbulk −sldCn
(A.14c)
A.3 Density
As shown in Fig. A.5 (left) the density of C
4
mimFeCl
4
and a mixture of
C
4
mimFeCl
4
/C
14
mimCl show a linear dependency on temperature. Linear
regression gives an equation to correct density values from literature given
for room temperature as surface tension measurements were done at 45
◦
C.
The linearity of the reciprocal density of C
4
mimFeCl
4
/C
14
mimCl mixtures
shown in Fig. A.5 (right) validates the simple ideal mixing approximation
used to calculate the density of the mixtures.
XII
Figure A.5 left:
Density (
d
) as a function of temperature measured with Anton
Paar density-meter DMA4500 for pure C
4
mimFeCl
4
(blue squares, linear regression
gives:
d= [(−8.23±0.012) 10−4◦C−1·x+(1.37833±0.000048)] g/cm3
) and 16.3wt%
C
14
mimCl in C
4
mimFeCl
4
(red circles, linear regression gives:
d= [(−7.66 ±
0.034) 10−4◦C−1·x+ (1.29287 ±0.00013)] g/cm3
).
right:
Density at 35
◦
C as a
function of surfactant concentration with Anton Paar density-meter for a binary
solution of C
4
mimFeCl
4
/C
14
mimCl. (linear regression gives:
1/d = [(3.769 ±
0.031) 10−3·x+ (0.74063 ±0.00021)] cm3/g
)
XIII
A.4 Dierential scanning calorimetry (DSC)
Figure A.6
DSC curves (background subtracted) for the system
C
18
mimCl/C
4
mimFeCl
4
. Increasing surfactant ratio is in ascending order.
Curves are shifted relative to each other to enhance clarity.
XIV
Figure A.7
DSC curves (background subtracted) for the system
C
16
mimCl/C
4
mimFeCl
4
. Increasing surfactant ratio is in ascending order.
Curves are shifted relative to each other to enhance clarity. In case of samples
with a low reproducibility several curves are plotted in the chronological order
straight
→
broken
→
dotted line.
XV
Figure A.8
DSC curves (background subtracted) for the system
C
14
mimCl/C
4
mimFeCl
4
. Increasing surfactant ratio is in ascending order.
Curves are shifted relative to each other to enhance clarity.
XVI
Figure A.9
DSC curves (background subtracted) for the system
C
18
mimCl/C
4
mimFeCl
4
. Increasing surfactant ratio is in ascending order.
Curves are shifted relative to each other to enhance clarity.
XVII
Figure A.10
DSC curves (background subtracted) for the system
C
16
mimCl/C
2
mimFeCl
4
. Increasing surfactant ratio is in ascending order.
Curves are shifted relative to each other to enhance clarity.
XVIII
Figure A.11
DSC curves (background subtracted) for the system
C
14
mimCl/C
2
mimFeCl
4
. Increasing surfactant ratio is in ascending order.
Curves are shifted relative to each other to enhance clarity.
XIX
Table A.4
Heats determined by the DSC integrals over all peaks (excluding the low temperature peaks for the C
2
mimFeCl
4
-
system, see description above) normalized by the total sample mass
msample
, the surfactant mass
msurf
and the surfactant
amount
nsurf
.
Q/n
is plotted in Fig. 3.2.
C
14
mimCl/C
2
mimFeCl
4
C
16
mimCl/C
2
mimFeCl
4
C
18
mimCl/C
2
mimFeCl
4
φm,tot
wt%
Q/msample
J/g
Q/msurf
J/g
Q/nsurf
kJ/mol
φm,tot
wt%
Q/msample
J/g
Q/msurf
J/g
Q/nsurf
kJ/mol
φm,tot
wt%
Q/msample
J/g
Q/msurf
J/g
Q/nsurf
kJ/mol
26.2 35.2 134.6 42.41 11.1 15.6 140.1 48.07 8.8 11.7 132.7 49.24
28.1 37.4 133.1 41.91 16.6 23.9 144.2 49.46 14.2 21.8 153.5 56.94
31.2 40.4 129.4 40.77 20.4 29.5 144.6 49.61 20.8 34.6 166.2 61.66
41.5 59.6 143.6 45.22 23.7 35.2 148.1 50.81 26.4 42.9 162.5 60.27
46.4 64.8 139.5 43.95 36.4 57.8 158.8 54.45 29.5 46.7 158.3 58.75
49.4 66.0 133.5 42.06 39.9 70.0 175.5 60.18 33.5 56.7 169.0 62.71
55.3 78.6 142.2 44.81 44.3 71.4 160.9 55.20 39.5 64.5 163.3 60.59
61.1 86.5 141.4 44.55 53.2 83.8 157.5 54.01 46.7 78.3 167.6 62.20
58.0 80.1 138.1 43.51 59.8 93.9 157.0 53.84 49.4 83.3 168.5 62.51
68.4 94.6 138.3 43.56 65.3 102.0 156.1 53.55 54.4 93.2 171.4 63.59
78.0 109.5 140.5 44.25 69.1 112.5 162.7 55.80 61.3 102.8 167.6 62.16
77.2 112.7 146.0 45.99 74.0 121.9 164.6 56.46 63.9 112.5 176.0 65.31
80.8 112.8 139.5 43.95 79.5 128.8 162.0 55.57 68.9 122.4 177.6 65.87
84.2 120.9 143.6 45.24 83.2 134.6 161.8 55.51 77.5 135.7 175.1 64.97
95.1 132.6 139.4 43.91 90.2 158.1 175.2 60.10 86.3 155.6 180.2 66.86
92.9 159.1 171.2 58.72 88.8 157.7 177.5 65.87
95.1 156.8 164.8 61.15
(Continues on next page.)
XX
(Continued from previous page.)
C
14
mimCl/C
4
mimFeCl
4
C
16
mimCl/C
4
mimFeCl
4
C
18
mimCl/C
4
mimFeCl
4
φm,tot
wt%
Q/msample
J/g
Q/msurf
J/g
Q/nsurf
kJ/mol
φm,tot
wt%
Q/msample
J/g
Q/msurf
J/g
Q/nsurf
kJ/mol
φm,tot
wt%
Q/msample
J/g
Q/msurf
J/g
Q/nsurf
kJ/mol
4.5 4.5 99.9 31.46 5.5 8.5 155.7 53.41 33.7 54.3 161.2 59.79
8.7 9.4 108.1 34.05 8.8 13.3 151.2 51.87 50.0 82.8 165.7 61.46
12.0 14.2 118.0 37.18 16.1 24.7 153.4 52.61 51.0 80.1 157.1 58.27
17.5 22.1 126.5 39.83 21.4 32.4 151.3 51.90 56.2 93.0 165.4 61.37
20.1 26.7 133.2 41.97 25.5 37.4 146.5 50.26 58.7 99.7 169.8 62.98
24.9 35.0 140.6 44.28 31.4 53.1 169.2 58.02 60.3 104.6 173.5 64.36
29.8 39.7 133.5 42.06 37.1 60.6 163.2 55.96 65.2 109.9 168.5 62.51
36.8 43.8 119.0 37.48 41.3 67.2 162.8 55.83 67.0 113.5 169.4 62.85
42.8 55.4 129.5 40.79 47.4 71.0 149.9 51.40 70.8 112.6 159.0 58.98
44.8 58.2 129.9 40.90 51.5 80.4 156.2 53.58 73.2 138.3 189.0 70.12
48.3 56.9 117.9 37.14 57.1 81.2 142.1 48.73 84.9 139.5 164.4 60.97
55.7 65.6 117.9 37.12 62.8 95.3 151.8 52.06 90.8 160.1 176.3 65.41
59.7 74.2 124.3 39.16 69.9 98.6 141.0 48.37 96.0 150.5 156.8 58.15
64.9 70.1 108.0 34.03 73.4 104.4 142.2 48.77 100.0 166.1 166.1 61.64
68.4 81.6 119.3 37.58 79.7 121.6 152.6 52.35
73.4 83.8 114.1 35.95 84.5 116.3 137.6 47.19
78.1 95.5 122.2 38.49 88.9 133.4 150.1 51.47
81.2 102.6 126.3 39.79 94.9 135.4 142.6 48.91
86.2 116.8 135.5 42.67 100.0 146.4 146.4 50.22
90.8 127.5 140.4 44.23 100.0 143.4 143.4 49.19
94.5 145.5 154.0 48.51
100.0 146.9 146.9 46.28
XXI
A.5 Surface tension
Table A.5
Fit parameters for the modied Szyszkowski model using eqs. 3.1a.
The headgroup spacing is given by
as= 1/Γ
.
C
i
mimFeCl
4
C
j
mimCl
24
◦
C 45
◦
C t parameter calc.
c10
wt%
cmcγ
wt%
Γ
nm
−2K1K2b N γ0
mN/m
as
nm
2
2 14 6.36 11.66 1.58 0.409 8.21E-31 0.023 27.42 49.65 0.63
2 16 3.51 6.75 1.71 0.891 4.98E-38 0.020 43.37 51.05 0.58
2 18 1.91 4.90 2.07 1.186 6.96E-30 0.000 40.61 50.77 0.48
4 14 18.7 21.96 22.51 0.99 0.193 6.45E-58 0.035 41.69 44.92 1.01
4 16 12.3 15.44 15.76 1.01 0.373 2.53E-52 0.028 42.29 44.99 0.99
4 18 9.60 8.09 2.75 0.128 2.35E-38 0.020 40.20 44.70 0.36
A.6 Polarized microscopy
Figure A.12
Example for a polarized microscopy image at 31.1
◦
C of the
metastable hexagonal phase in the system C
4
mimFeCl
4
/C
14
mimCl exhibiting a
typically fan-shaped texture.
XXII
A.7 SAXS
Figure A.13
SAXS curves for surfactant in C
2
mimFeCl
4
(straight lines) and
C
4
mimFeCl
4
(broken lines) at 100
◦
C (top) and 75
◦
C (bottom). All samples have
a content of 85wt% surfactant.
XXIII
XXIV
B
Appendix Microemulsions
XXV
B.1 Additional phase diagrams
Figure B.1
Detailed view of all data points recorded for determining the sh-
diagrams for C
4
mimFeCl
4
-based systems with C
18
mimCl (top left), C
16
mimCl (top
right) and C
14
mimCl (bottom). By visual observation the samples were classied
as monophasic (blue squares), triphasic (black diamonds) or biphasic (triangles)
with a bigger upper (yellow) or bottom (green) phase. From their position phase
boundaries were interpolated (black line).
XXVI
Figure B.2
Detailed view of all data points recorded for determining the sh-
diagrams for C
2
mimFeCl
4
-based systems with C
18
mimCl (top left), C
16
mimCl (top
right) and C
14
mimCl (bottom). By visual observation the samples were classied
as monophasic (blue squares), triphasic (black diamonds) or biphasic (triangles)
with a bigger upper (yellow) or bottom (green) phase. From their position phase
boundaries were interpolated (black line).
XXVII
Figure B.3
Detailed view of all data points recorded for determining the
sh-diagrams for the C
6
mimFeCl
4
-based system with C
18
mimCl (top) and the
C
2
mimFeCl
4
-based system with C
12
mimCl as surfactant (bottom). By visual ob-
servation the samples were classied as monophasic (blue squares), triphasic (black
diamonds) or biphasic (triangles) with a bigger upper (yellow) or bottom (green)
phase. From their position phase boundaries were interpolated (black line).
XXVIII
B.2 The clipped random wave model
111,112
The scattering intensity is described by an inverse eighth-order polynomial,
which is an extension of the Teubner-Strey model. It contains three length-
scale parameters with the inter-domain distance
Ds
, the coherence length
ξ
of the local domain order, and the surface roughness parameter
c
. While the
rst two parameters are similar to the analogous values in the Teubner-Strey
model, the roughness parameter gives additional information and improves
the model mainly in the high
q
range:
I(q) = 4Qinv b(a2+ (b+c)2)/π
(q2+c2)(q4−2(a2−b2)q2+ (a2+b2)2)+BG
(B.1)
a= 2π/Ds
(B.2)
b= 1/ξ
(B.3)
The amplitude scaling parameter
Qinv
gives directly the scattering invari-
ant. The t parameters are listed in table B.1 and plotted in Fig. B.6. It can
be seen that
Ds
and
ξ
are in a very good agreement with the Teubner-Strey
model. In the oil rich region (low
xMRTIL
) the surface roughness parameter
c
shows values comparable to water systems
112114
(in our system they are
slightly higher due to a higher surfactant concentration) and the trend of
a growing roughness value with longer surfactant chains (means an increas-
ing roughness) is as expected. Increasing the MRTIL content in the system
gives continuously bigger
c
-values for all three systems which ts well to the
general picture of a weakening of the mesoscopic structuring by increasing
the MRTIL ratio. Above
xMRTIL ≈0.4
the CRW-ts give high and erratic
numbers for
c
. This is due to the fact that the roughness parameter is not
necessary anymore to describe the SANS data (i. e. the t quality is inde-
pendent of
c
) as the TS itself gives already good t results. Also remarkable
is the fact that the simple two level model to explain the invariant (compare
Fig. B.5) as well fails above
xMRTIL ≈0.4
which gives a hint to structural
changes at this point.
XXIX
B.3 The scattering invariant
Qinv
The scattering invariant is dened as
Qinv =∫∞
0
I(q)·q2
d
q
(B.4)
During data analysis the invariant was determined by three dierent
methods:
1. With the help of the program SASt
70
the measured SANS data were
integrated numerically whereby extrapolation to
I(0)
and innity was
done by Guinier and Porod approximation, respectively. The back-
ground determination by Porod's law is shown in Fig. B.4 and all
results are listed in Table B.1.
2. The CRW-model gives directly the invariant as a t parameter.
3. The contrast t parameter in the Teubner-Strey model (see eq. 2d) is
directly related to the invariant by
Qinv = 2π2⟨η2⟩
.
All t parameters are listed in Table B.1. The invariants determined by
the three methods are plotted in Fig. B.5. The calculated invariants are
strongly dependent on the chosen
BG
which explains the deviations between
the three methods especially for low
xMRTIL
(where the TS-ts have an
obviously to low
BG
). For this reason the rst method is the most reliable
and was therefore chosen for further discussion in the main article.
XXX
Table B.1
Results for numerical SANS data analysis (including
I(0)
from Guinier approximation and Porod's Law
I(q) =
BG +C·q−4
) and t parameters for Teubner-Strey and CRW ts.
Teubner-Strey CRW Porod/Guinier
x
MRTIL ξ DsBG Qinv ξ Dsc BG Qinv I(0) C BG Qinv
[nm] [nm]
[1
cm
] [ 1
cm nm
3]
[nm] [nm] [
1
nm
] [nm]
[1
cmnm
3] [ 1
cm
] [ 1
cmnm
3] [ 1
cmnm
3] [ 1
cmnm
3]
C
14
mimCl
0.05 1.67 4.86 0.29 11.87 2.12 4.24 1.20 0.48 7.24 2.48 5.57 0.48 8.83
0.12 1.74 5.67 0.38 11.22 1.92 4.92 1.35 0.54 7.29 3.75 5.84 0.51 9.15
0.19 1.79 6.15 0.48 10.69 1.82 5.55 1.80 0.60 7.38 4.32 6.26 0.57 9.32
0.20 1.82 7.19 0.54 10.04 1.75 6.71 2.30 0.62 7.38 6.38 6.98 0.58 9.59
0.37 1.77 8.82 0.59 8.76 1.75 8.73 5.80 0.61 7.44 9.18 7.94 0.57 9.49
0.42 1.69 9.88 0.63 8.09 1.69 9.86 12.73 0.63 7.43 10.43 4.91 0.65 7.46
0.49 1.54 10.77 0.65 6.54 1.54 10.77
>25
0.65 6.54 9.73 6.22 0.63 7.02
0.54 1.36 12.48 0.70 5.86 1.36 12.48
>25
0.7 5.86 8.39 5.18 0.69 5.79
0.60 1.27 13.85 0.66 4.35 1.27 13.86
>25
0.66 4.35 6.44 3.75 0.66 4.18
0.72 0.90 16.04 0.60 1.64 0.90 16.04
>25
0.60 1.64 1.32 3.20 0.57 2.24
0.77 0.87 8.95 0.58 0.62 0.87 8.94
>25
0.58 0.57 0.23 1.94 0.55 1.17
1.00 0.78 5.41 0.54 0.59 0.78 5.42
>25
0.54 0.55 0.06 1.88 0.52 1.07
C
16
mimCl
0.05 1.69 4.76 0.29 11.12 2.22 4.24 1.32 0.48 6.69 2.00 5.76 0.47 8.47
0.12 1.91 5.38 0.41 9.98 2.19 4.82 1.44 0.55 6.60 2.61 5.58 0.52 8.46
0.19 2.05 5.91 0.47 9.01 2.15 5.41 1.69 0.58 6.36 3.17 5.88 0.52 8.50
0.27 2.20 6.42 0.57 8.45 2.18 6.13 2.48 0.64 6.52 3.86 5.11 0.61 8.13
0.35 2.15 6.90 0.58 7.11 2.15 6.84 6.09 0.60 6.21 3.90 5.59 0.58 7.56
0.42 2.11 7.24 0.62 5.96 2.10 7.25
>25
0.62 5.98 4.05 4.69 0.60 6.40
0.50 1.95 7.54 0.63 4.66 1.95 7.54
>25
0.63 4.66 3.27 3.59 0.62 4.86
(Continues on next page.)
XXXI
(Continued from previous page.)
Teubner-Strey CRW Porod/Guinier
x
MRTIL ξ DsBG Qinv ξ Dsc BG Qinv I(0) C BG Qinv
[nm] [nm]
[1
cm
] [ 1
cm nm
3]
[nm] [nm] [
1
nm
] [nm]
[1
cmnm
3] [ 1
cm
] [ 1
cmnm
3] [ 1
cmnm
3] [ 1
cmnm
3]
(C
16
mimCl)
0.55 1.87 7.57 0.68 4.14 1.86 7.58
>25
0.68 4.14 3.04 3.68 0.66 4.51
0.61 1.71 7.45 0.66 3.05 1.71 7.45
>25
0.66 3.05 2.01 3.12 0.64 3.43
0.71 1.53 6.84 0.63 1.09 1.53 6.84
>25
0.63 1.09 0.54 1.38 0.62 1.29
0.79 1.56 5.96 0.57 0.22 1.56 5.96
>25
0.57 0.22 0.06 1.59 0.53 0.95
0.94 1.41 4.85 0.54 0.20 1.41 4.85
>25
0.54 0.20 0.01 0.62 0.53 0.36
1.00 1.07 4.80 0.55 0.63 1.07 4.80
>25
0.55 0.63 0.10 0.67 0.55 0.53
C
18
mimCl
0.12 2.09 5.21 0.44 8.68 2.41 4.76 1.50 0.56 5.91 2.20 6.17 0.50 8.27
0.19 2.27 5.70 0.48 7.60 2.41 5.31 1.79 0.57 5.52 2.42 5.77 0.51 7.77
0.26 2.37 6.11 0.57 7.46 2.39 5.88 2.60 0.63 5.86 2.87 5.41 0.58 7.76
0.33 2.35 6.51 0.59 6.32 2.35 6.45 5.54 0.61 5.52 2.83 5.98 0.55 7.41
0.40 2.36 6.81 0.68 5.82 2.35 6.82
>25
0.68 5.83 2.78 4.91 0.64 6.61
0.47 2.19 7.12 0.68 4.70 2.19 7.12
>25
0.68 4.70 2.85 4.28 0.65 5.40
0.54 2.03 7.07 0.68 3.69 2.03 7.07
>25
0.68 3.66 2.03 4.12 0.64 4.54
0.59 1.92 7.18 0.68 2.74 1.91 7.18
>25
0.68 2.74 1.54 2.85 0.66 3.23
0.65 1.77 6.89 0.67 1.87 1.76 6.89
>25
0.67 1.87 1.05 2.14 0.65 2.25
0.75 1.58 6.14 0.59 0.51 1.58 6.14
>25
0.59 0.50 0.17 0.42 0.58 0.48
0.79 1.57 5.75 0.55 0.17 1.57 5.75
>25
0.55 0.17 0.01 0.23 0.55 0.18
0.94 1.52 5.52 0.55 0.24 1.52 5.52
>25
0.55 0.24 0.01 -0.08 0.55 0.04
XXXII
B.3.1 Calculation of theoretical invariant
Using eq. 5.2d,
⟨η2⟩
was calculated for two dierent cases: All hydrocarbon
chains of the surfactant/cosurfactant are belonging to the oil phase (case 2),
partitioning of the hydrocarbon chains (case 1, see table B.2). Scattering
length densities (SLD) were calculated as
SLD =∑n
i=1 bci
vm
(B.5)
where
bci
is the bound coherent scattering length of atom i (taken from
127
)
in the chemical group with molecular volume
vm
. Used densities and resulting
SLD are summarized in B.2. Resulting volume ratios and
ρ
for the two phases
in case 1-3 were calculated under the assumption of invariant densities of
chemical groups during mixing:
ΦIL =∑Φj
(B.6)
ρIL =∑(Φj·SLDj)
ΦIL
(B.7)
with the jth chemical group belonging to the IL phase.
Φoil
and
ρoil
where
calculated analogously.
XXXIII
Figure B.4
Determination of the background (
BG
) by Porod's Law (
I(q) = BG+
C·q−4
) Symbols are measured SANS data and lines are ts with Porod's law in
the range
2.45 nm−1≤q≤2.86 nm−1
(
36.4nm−4≤q4≤67.3nm−4
).
XXXIV
Figure B.5
Values for the scattering invariant as determined directly from the
experimental data and using the Porod/Guinier approximation for the range where
no experimental data is available. (a, open symbols), CRW t (b, open symbols)
and Teubner-Strey t (c, open symbols). Small symbols are calculated invariants
for two dierent distributions of surfactant/cosurfactant chains as described in the
main article and in section B.3.1.
XXXV
Figure B.7
SANS data (symbols) which were not plotted in the main article for
the systems with C
14
mimCl (top row), C
16
mimCl (middle row) and C
18
mimCl
(bottom row) as surfactant. Lines are ts with the CRW (left column) and TS
(right column) model. The corresponding parameters are listed in table B.1.
XXXVII
Table B.2
Characteristic parameters of dierent chemical groups used to calculate
values for
Qinv = 2π2⟨η2⟩
.
parameters location in IL/oil-phase
chem. group density/gcm
−3
SLD/10
−4
nm
−2
case 1 case 2
C
4
mimFeCl
4
1.36 1.55 1/0 1/0
D12-cyclohexane 0.89 6.68 0/1 0/1
-mimCl 1.45 2.41 1/0 1/0
decyl 0.79 -0.41 6/4 0/1
-OH 1.42 1.04 1/0 1/0
tetra- 0.81 -0.38 7/7 0/1
hexa- decyl 0.82 -0.37 7/9 0/1
octa- 0.82 -0.37 7/11 0/1
B.4 SANS experiments recorded at D11 (ILL)
The SANS data were analyzed by tting with the Teubner-Strey model sim-
ilar to the data measured at PSI (SANSII). Results are plotted in Fig. B.9
and listed in table B.4. As described in section B.3.1, the tted value for
⟨η2⟩
was compared with calculated values from the real sample compositions
for three possible divisions of the interface. Results are listed in table B.4
and plotted in Fig. B.10.
XXXIX
Figure B.9
SANS curves with D12-cyclohexane as the oil phase. The samples are
along the experimental path shown in Fig. 5.18. Ratios between MRTIL and oil
are
xMRTIL
=0.05-0.48 (left) and
xMRTIL
=0.53-1.00 (right) and are given in the
inset. Lines are ts with the TS model.
Figure B.10
⟨η2⟩
as deduced from the TS ts (lled squares) compared to values
calculated as described in section B.3.1 from the real sample compositions for three
possible divisions of the interface: The whole interface is counted into the IL phase
(dashed line), all hydrocarbon chains of the surfactant/cosurfactant belong to the
oil phase (dotted line), and partial partitioning of the hydrocarbon chains (full
line).
XL
Table B.3
Characteristic parameters of dierent chemical groups used to calculate
values for
⟨η2⟩
.
parameters location in IL/oil-phase
chem. group density/gcm
−3
SLD/10
−4
nm
−2
case 1 case 2 case 3
C
4
mimFeCl
4
1.36 1.55 1/0 1/0 1/0
D12-cyclohexane 0.89 6.68 0/1 0/1 0/1
-mimCl 1.45 2.41 1/0 1/0 1/0
decyl 0.79 -0.41 1/0 0.6/0.4 0/1
hexadecyl 0.82 -0.37 1/0 0.375/0.625 0/1
-OH 1.42 1.04 1/0 1/0 1/0
Table B.4
Parameter for the TS ts shown in B.9 and calculated values for
⟨η2⟩
.
t parameter SASt calculated
⟨η2⟩/
cm
−1
nm
−3
xMRTIL ξ
nm
Ds
nm
⟨η2⟩
cm
−1
nm
−3
BG
cm
−1
case 1 case 2 case 3
0.05 2.05 5.44 0.4356 0.3500 0.9161 0.4195 0.0709
0.11 1.90 6.53 0.4623 0.3669 0.9259 0.4388 0.1037
0.17 2.29 7.48 0.4349 0.4313 0.9297 0.4519 0.1317
0.25 2.28 8.59 0.4331 0.4851 0.9191 0.4536 0.1509
0.33 2.29 10.76 0.4184 0.5359 0.8936 0.4404 0.1570
0.40 2.30 13.39 0.3827 0.5657 0.8537 0.4149 0.1523
0.48 1.86 27.36 0.3614 0.5881 0.7975 0.3745 0.1361
0.53 1.96 60.00 0.3034 0.5981 0.7493 0.3398 0.1203
0.70 1.74 60.00 0.1043 0.6352 0.5466 0.1953 0.0511
0.77 1.14 60.00 0.0653 0.5951 0.4509 0.1331 0.0252
0.85 0.35 6.00 0.1128 0.5757 0.3059 0.0534 0.0019
1.00 0.65 5.64 0.0576 0.6000 0.0000 0.0299 0.0687
XLI
B.5 SANS curves measured at 36
◦
C
Figure B.11
SANS data measured at 24
◦
C (lled symbols) and 36
◦
C (open
symbols) for three dierent microemulsion systems.
XLII
B.6 SANS experiments recorded at SANSI (PSI)
under magnetic eld
Table B.5
Composition of all samples measured in SANS experiments under mag-
netic eld. It should be mentioned that in Fig. 5.20 and 5.22 for consistency the
mass fractions were recalculated to H12-cyclohexane with the same volume.
C
2
mimFeCl
4
wt% C
4
mimFeCl
4
wt% C
14
mmCl
wt% C
16
mimCl
wt% C
18
mimCl
wt%
decanol D12-cyclo-
hexane
45.7 12.2 12.7 29.5
45.1 13.8 11.5 29.5
49.1 10.4 8.3 32.2
47.8 12.0 8.9 31.3
51.5 9.4 5.9 33.2
50.1 10.8 6.0 33.1
33.0 8.3 5.2 20.4
32.7 9.7 4.9 20.1
34.4 7.4 2.6 21.2
33.8 9.1 2.7 20.7
10.9 8.9 3.0 77.2
21.6 48.5 15.1 14.9
32.9 34.3 10.6 22.2
35.6 31.0 9.7 23.7
38.9 26.9 8.2 26.1
35.2 6.7 1.2 21.7
35.1 7.1 1.1 21.6
10.0 9.2 1.6 79.2
XLIII
B.7 Surface tension
Figure B.12
Surface tension of C
j
mimCl without (open symbols) and with 2
molecules decanol per surfactant molecule (lled symbols) in C
4
mimFeCl
4
at 45
◦C
(top) and 24
◦C
(bottom).
XLIV
B.8 Conductivity
A rst insight into this mesoscopic microemulsion structure can be deduced
from the conductivity titration shown in Fig. B.13. With increasing MR-
TIL content the conductivity increases by one order of magnitude, which is
due to the formation of a structure continuous in MRTIL. In water and oil
containing microemulsions as well as in systems with an IL substituting the
water such a percolation behavior of the conductivity is well known and can
be explained by structural transitions between a droplet and a bicontinuous
structure.
51
The behavior of our MRTIL system is analogous and shows a
percolation point of conductivity around
xMRTIL ≈0.1
, determined by plot-
ting the specic conductivity with an exponent of
5/8
versus
xMRTIL
(Fig.
B.13).
63,128
Figure B.13
Specic electric conductivity as a function of the MRTIL volume ratio
(eq. 5.11) for the microemulsion system C
4
mimFeCl
4
/C
16
mimCl. The broken line
is a guide to the eyes. Measurements were done along the experimental path shown
in Fig. 5.18.
XLV
B.9 Viscosity
Figure B.14
Dynamic viscosity (squares) and density (diamonds) of microemul-
sions with a constant amphiphile mass ratio (C
16
mimCl+decanol) of 23wt%.
along the experimental path shown in Fig. 5.5. Viscosities are low and in
the range obtained for assuming a linear relation for the MRTIL/cyclohexane
mixture viscosities, i.e., they are determined simply by the liquid components
of the microemulsion.
B.10 Cube model
Figure B.15
Ds
as derived experimentally with Teubner-Strey ts (lled squares)
from SANS curves displayed in Fig. 5.7. Dashed lines display the cube model
dened with eq. 5.6 for various dierent parameters.
XLVI
B.11 Force Calculations in a Magnetic Field
Using the volume magnetic susceptibilities given in table 5.5, and densities of
1.36g/cm
3
(C
4
mimFeCl
4
) and 0.779g/cm
3
(cyclohexane), the eld gradient
to fulll the condition
Fgrav =Fmag
can be calculated to be
∇B= 4.4
Tm
−1
,
which eectively ts into the range of
3
Tm
−1≤ ∇B≤46
Tm
−1
which is
given by the eld prole along the sample estimated in Fig. B.16.
Figure B.16
Magnetic eld prole of the magnet used for the data shown in Fig.
5.16 and measured as described in section 2.11. The position of the edge of the
magnet plates where the sample was located is set to zero. The lines give linear
ts whose slope gives directly the eld gradient of
∇B= 46
Tm
−1
(broken line) or
3Tm
−1
(straight line). The double-arrow gives the approximate sample position.
XLVII
