Document [original]

Experimental and Numerical Studies of Laminar

Counter-flow Diffusion Flames Using Biomass-based

Gaseous Fuels

vorgelegt von

Dipl.-Ing.

Marie-Theres Scharl

an der Fakultät III - Prozesswissenschaften

der Technischen Universität Berlin

zur Erlangung des akademischen Grades

Doktorin der Ingenieurswissenschaften

– Dr.-Ing. –

genehmigte Dissertation

Promotionsausschuss:

Vorsitzender: Prof. Dr.-Ing. Dietmar Auhl

Gutachter: Prof. Dr. rer. nat. Frank Behrendt

Gutachter: Prof. Dr. rer. nat. Volker Sick

Tag der wissenschaftlichen Aussprache: 26. August 2021

Berlin 2022

Ich erkläre hiermit, dass ich die vorliegende Arbeit selbständig verfasst und keine

anderen als die angegebenen Quellen und Hilfsmittel verwendet habe.

Berlin, den 26. Dezember 2021

Acknowledgements

My gratitude and thankfulness go out to many people regarding the completion of

this thesis.

I wish to thank my supervisors and mentors, Prof. Dr. Frank Behrendt, Prof. Dr.

Douglas Greenhalgh, and Prof. Dr. Alba Dieguez Alonso, for their support and

challenges in the years at EVUR. Furthermore, a big thank you is directed at Prof.

Dr. Volker Sick for reviewing this work and Prof. Dr. Dietmar Auhl for taking the

time to serve as the chairman of the dissertation committee.

My work would not have been possible without the heart and soul of the EVUR

chair Uwe Röhr, Susanne Hoffmann, and Ines Preschel; a warm thank you also goes

to Birgit Packeiser and Michaela Riese.

Omid Elhami, Thomas Mouton, Jenny Rieck, Lina Taube, Hernán Almuiña Villar,

Maximilian Mehnert, and many other colleagues of EVUR have enriched the time

of my thesis with support, laughter, discussions, and memories for a lifetime, which

I will never forget.

Carsten Waechtler, Fabian Schmid, Arndt Kobusinski, Robin Reinking, Oliver

Löschke, Mohamad Hussein, and Madeleine Lange, you were all somehow „my stu-

dents“ during the summer schools, household tasks, and my thesis - thank you for

being part of this journey.

The support at Technische Universität Berlin goes beyond our faculty III when help

is needed. A big thanks goes to many colleagues and friends from all the years at

TUB, you know who you are.

Finally, I would like to thank my friends and, foremost, my family for the endless

love, encouragement, and support.

Mama, Papa, Curt, Max, Ferdi, Mone, Leo, Linchen - this thesis is dedicated to

you.

Ivo Schneider, you will never be forgotten.

Abstract

Thermochemical conversion processes are regarded as promising alternatives for de-

centralized energetic utilization of biomass. Pyrolysis and gasification of (woody)

biomasses present a possibility to produce a fuel gas to be used, among other ap-

plications, in internal combustion engines or turbines as part of combined heat and

power generation (CHP).

This thesis aims to gain a deeper understanding of the utilization of biomass-based

gaseous model fuels in combustion systems. This is investigated in the present

thesis, combining both experimental studies and numerical simulations of laminar

non-premixed flames in a counter-flow burner set-up.

To this end, a counter-flow burner system was designed and built as the central

part of a spectroscopic measuring system to validate an in-house time-dependent

implicit Fortran code for diffusion flames named DIFFLA. This was foremost

achieved by spectroscopic laser-induced fluorescence measurements of formaldehyde

and Rayleigh scattering experiments, deducing temperature fields in flames. Fur-

thermore, CH* chemiluminescence experiments were performed. These diagnostic

tools were all utilized to analyze flame behavior and flame structure by providing an

accurate understanding via temperature distributions, the tracking of intermediate

species, and by localizing the flame front.

This study focuses on synthetic gas mixtures consisting of multiple components,

resembling, to different extents, typical compositions of the product gas obtained in

biomass gasification or pyrolysis processes. Diluted methane as reference fuels and

further fuels composed of N2, H2, CO, CO2, and CH4were investigated, partially

adding O2to fuel or oxidizer. In all cases, the oxidizer was air, and a wide range

of air-fuel ratios was considered. Furthermore, the presented system comprised a

range of flames combusting at strain rates starting at around 60 s-1 leading up to

circa 250 s-1. Therefore, incorporating the boundary condition at very low fuel- and

oxidizer velocities for combustion, though not including the other limiting condition,

at strain out.

The influence of the product gas composition on the flame behavior and flame struc-

ture, with respect to the changes of the species profiles and peak temperatures with

changing flow velocities, was investigated. All in all, illuminating the combustion

mechanisms of these hydrocarbon mixtures from multiple perspectives and provid-

ing valuable information about the operations with different synthetic biomass-based

gaseous model fuels.

Zusammenfassung

Thermochemische Konversionsverfahren werden als vielversprechende Alternativen

zur dezentralen energetischen Nutzung von Biomasse gehandelt. Pyrolyse und Ver-

gasung von (holzigen) Biomassen stellen Möglichkeiten dar ein Brenngas zu erzeu-

gen, das, unter anderem, in Verbrennungsmotoren oder Turbinen im Rahmen der

Kraft-Wärme-Kopplung (KWK) eingesetzt werden kann.

Ziel dieser Arbeit ist es, ein tieferes Verständnis für die Verbrennung von gasförmigen

Brennstoffen, die aus thermochemischen Konversionsverfahren gewonnen wurden,

zu erlangen. Dies wird durch die Gegenüberstellung von experimentellen Unter-

suchungen zu numerischen Simulationen verschiedener laminarer, nicht vorgemis-

chter Flammen in einem Gegenstrombrenner erreicht.

Zu diesem Zweck wurde ein Gegenstrombrennersystem als zentraler Teil eines spek-

troskopischen Messsystems entworfen und gebaut, um den hauseigenen zeitabhängi-

gen und impliziten Fortran-Code für Diffusionsflammen namens DIFFLA zu vali-

dieren. Dies geschah in erster Linie durch spektroskopische laserinduzierte Fluo-

reszenzmessungen von Formaldehyd und durch Rayleigh Streuungsexperimente, die

auf Temperaturfelder in den Flammen schließen lassen. Des Weiteren wurden Mes-

sungen von CH*-Chemilumineszenz durchgeführt. Diese Diagnosewerkzeuge wurden

alle genutzt, um das Flammenverhalten und die Flammenstruktur zu analysieren,

indem sie einen genauen Zusammenhang über Temperaturverteilungen, die Verläufe

der Zwischenspezies und die Lokalisierung der Flammenfront herstellen.

Diese Studie befasst sich mit synthetischen Gasgemischen, welche, in unter-

schiedlichem Maße, den typischen Zusammensetzungen der Produktgase ähneln, die

bei Vergasungs- oder Pyrolyseprozessen von Biomasse entstehen. Untersucht wurde

verdünntes Methan als Referenz und Brennstoffe, welche zusammengesetzt waren

aus N2, H2, CO, CO2und CH4. Zusätzlich wurde dem Brennstoff oder Oxidator

teilweise O2zugegeben. In allen Fällen war der Oxidator Luft und es wurde ein

breiter Bereich von Luft-Kraftstoff-Verhältnissen berücksichtigt. In dieser Studie

wurden Flammen die bei der Dehnungsgeschwindigkeiten von ca. 60 s-1 bis hin zu

ca. 250 s-1 verbrannten in Betracht gezogen. Damit ist die Randbedingung sehr

niedriger Brennstoff- und Oxidationsströmungsgeschwindigkeiten beachtet worden,

weniger die mögliche andere Randbedingung des sogenannten „strain-out“.

Untersucht wurde der Einfluss der Produktgaszusammensetzung auf das Flammen-

verhalten und die Flammenstruktur im Hinblick auf die Änderungen der Speziespro-

file und Spitzentemperaturen bei wechselnden Strömungsgeschwindigkeiten.

Insgesamt werden die Verbrennungsmechanismen dieser Kohlenwasserstoffgemische

aus mehreren Blickwinkeln beleuchtet und liefern somit wertvolle Informationen für

die Nutzung verschiedener gasförmiger Brennstoffe auf Basis von holziger Biomasse.

Contents

1 Introduction 1

2 Biomass as a source of energy 5

2.1 Conversion processes of biomass . . . . . . . . . . . . . . . . . . . . . 7

2.1.1 Pyrolysis.............................. 9

2.1.2 Gasification ............................ 13

2.1.3 Combustion............................ 18

2.1.4 Liquefaction............................ 20

3 Combustion of gaseous fuels 21

3.1 Laminar and turbulent combustion . . . . . . . . . . . . . . . . . . . 21

3.1.1 Laminar premixed combustion . . . . . . . . . . . . . . . . . . 22

3.1.2 Laminar non-premixed combustion . . . . . . . . . . . . . . . 27

3.1.3 Turbulent combustion . . . . . . . . . . . . . . . . . . . . . . 30

3.2 Counter-flow combustion . . . . . . . . . . . . . . . . . . . . . . . . . 34

3.3 Laminar flamelets [in turbulent combustion] . . . . . . . . . . . . . . 37

3.4 Gaseousfuels ............................... 38

3.4.1 Methane and diluted Methane . . . . . . . . . . . . . . . . . . 38

3.4.2 Biomass-based fuels . . . . . . . . . . . . . . . . . . . . . . . . 39

3.5 Combustion properties . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3.5.1 Gasifyingagents ......................... 41

3.5.2 Enhancement with oxygen . . . . . . . . . . . . . . . . . . . . 42

4 Spectroscopic techniques 43

4.1 Laser-induced fluorescence . . . . . . . . . . . . . . . . . . . . . . . . 43

4.2 Rayleighscattering............................ 48

4.3 Chemiluminescence............................ 53

4.4 Techniques................................. 56

XII Contents

5 Experimental set-up and conditions 61

5.1 Experimental set-up . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

5.1.1 Counter-flow burner . . . . . . . . . . . . . . . . . . . . . . . 61

5.1.2 Nd:YAGlaser........................... 64

5.1.3 Spectroscopic set-up . . . . . . . . . . . . . . . . . . . . . . . 66

5.1.4 Analytical components . . . . . . . . . . . . . . . . . . . . . . 69

5.2 Experimental conditions . . . . . . . . . . . . . . . . . . . . . . . . . 75

5.2.1 Fuel mixtures - PG, GGL2, GGL3 . . . . . . . . . . . . . . . . 75

5.2.2 Strain rates - Velocities and straining out . . . . . . . . . . . . 79

5.3 Dataprocessing.............................. 79

6 Numerical investigations 85

6.1 DIFFLA.................................. 85

6.1.1 Introduction to DIFFLA and comparable models . . . . . . . 85

6.1.2 Governing equations . . . . . . . . . . . . . . . . . . . . . . . 88

6.1.3 Boundary layer assumptions, equations, and conditions . . . . 89

6.1.4 Transport and thermodynamic data, reaction mechanisms . . 92

6.2 CHFITandRAYFIT........................... 94

7 Results, discussion, and conclusions 97

7.1 Dilutedmethane ............................. 97

7.1.1 Rayleigh measurements of diluted methane fuels . . . . . . . . 98

7.1.2 Laser-induced fluorescence of diluted methane fuels . . . . . . 103

7.1.3 CH* chemiluminescence and CH of diluted methane fuels . . . 104

7.2 Biomass-based gaseous fuels - GGL2, GGL3, and PG . . . . . . . . . 107

7.2.1 Rayleigh measurements of biomass-based gaseous fuels . . . . 107

7.2.2 Laser-induced fluorescence of biomass-based gaseous fuels . . . 108

7.2.3 CH* chemiluminescence and CH of biomass-based fuels . . . . 111

7.3 Biomass-based gases with additional oxygen on the fuel side . . . . . 113

7.3.1 Temperature evolution with additional fuel side oxygen . . . . 114

7.3.2 LIF of biomass-based gases with additional fuel side oxygen . 115

7.3.3 Chemiluminescence with additional fuel side oxygen . . . . . . 118

7.4 Biomass-based gases with additional oxygen on the air side . . . . . . 119

Contents XIII

7.4.1 Temperature evolution with additional air side oxygen . . . . 119

7.4.2 Formaldehyde of biomass gases with additional air side oxygen 121

7.4.3 Chemiluminescence with additional air side oxygen . . . . . . 122

7.5 Combustion behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

7.5.1 Combustion behavior of diluted methane flames . . . . . . . . 123

7.5.2 Combustion behavior of biomass-based flames . . . . . . . . . 124

7.6 Criticalpoints...............................128

7.6.1 Gas cleaning and composition . . . . . . . . . . . . . . . . . . 128

7.6.2 LHV................................128

7.6.3 Low strain rates . . . . . . . . . . . . . . . . . . . . . . . . . . 129

8 Summary and outlook 131

8.1 Futurework................................134

A Appendix 135

A.1 Fuels....................................136

A.2 Experimentalset-up ...........................138

A.3 Equipmentlist ..............................139

A.3.1 Camera, energy monitor, and electrical equipment . . . . . . . 139

A.3.2 Gasmixing ............................139

A.3.3 Optical lenses, mirrors, filters, and equipment . . . . . . . . . 140

A.3.4 Gases ...............................140

A.3.5 Counter-flow burner . . . . . . . . . . . . . . . . . . . . . . . 141

A.3.6 Laser................................142

B Publications 145

B.1 Peer-reviewed journals . . . . . . . . . . . . . . . . . . . . . . . . . . 146

B.2 Presentations ...............................146

Bibliography 147

List of Figures

2.1 Schematic illustration of the conversion routes of biomass based on [1–4] 7

2.2 Model for cellulosic pyrolysis mechanism by Broid-Shafizadeh based

on[5].................................... 9

2.3 Model for lignocellulosic biomass pyrolysis, commonly known from [6],

basedon[5] ................................ 10

2.4 Steps of biomass gasification in a gasifier [7], reprinted with permission 13

2.5 Fixed-bed updraft (left) and downdraft (right) gasifiers [8], reprinted

withpermission.............................. 16

2.6 Fluidized-bed bubbling (left) and circulating (right) gasifiers [8], re-

printed with permission . . . . . . . . . . . . . . . . . . . . . . . . . 17

2.7 The four steps of biomass combustion in a boiler [7], reprinted with

permission................................. 19

3.1 Schematic overview of a bunsen burner [9] . . . . . . . . . . . . . . . 23

3.2 Schematic overview of a 1-D premixed planar combustion wave based

on[10]................................... 24

3.3 A non-premixed co-flow flame (left) and the scheme of a non-reacting

fuel jet (right) [11], reprinted with permission . . . . . . . . . . . . . 28

3.4 Over- and under-ventilated Burke-Schumann flame [11], reprinted

withpermission.............................. 29

3.5 Turbulent non-premixed combustion regime diagram based on [12] . . 32

3.6 Classification of counter-flow flames into Types I-IV from [13] com-

monly known from [14], reprinted with permission . . . . . . . . . . . 34

3.7 Type I counter-flow burner set-up with a steady strained 1-D diffusion

flame [11], reprinted with permission . . . . . . . . . . . . . . . . . . 35

4.1 Jablonski diagram [15], the basic version is commonly known

from [16], reprinted with permission . . . . . . . . . . . . . . . . . . . 44

4.2 Electromagnetic spectrum, defining the regions of electronic, vibra-

tional and rotational absorption after light is induced [17], reprinted

withpermission.............................. 45

List of Figures XV

4.3 Simultaneous electronic and vibrational change of energy levels with

semi-stationary nuclei during the absorption or emission of a photon,

which can be explained by the Franck-Condon principle [18], reprinted

withpermission.............................. 46

4.4 Principle of Rayleigh scattering, Raman scattering and fluore-

scence [19], reprinted with permission . . . . . . . . . . . . . . . . . . 49

4.5 Principle of polarization [15], reprinted with permission . . . . . . . . 52

4.6 CH*, OH* and C2* spectra of a methane-oxygen-hydrogen flame [20],

reprinted with permission . . . . . . . . . . . . . . . . . . . . . . . . 53

4.7 CH*, OH* with and without CO2* spectra in a hydrocarbon fla-

me [21], reprinted with permission . . . . . . . . . . . . . . . . . . . . 55

5.1 Schematic sectional view of a McKenna flat flame burner [23] . . . . . 62

5.2 Diffusion flames at L/D = 0.36 (left) and L/D= 0.83 (right) [23] in-

duced by two opposed McKenna flat flame burners . . . . . . . . . . 63

5.3 Scheme of (left) and actual constructed (right) counter-flow burner

system. (1): inner gas chamber top burner, (2): outer gas chamber

top burner, (3): alignment system for top and bottom burners, (4):

stage to move bottom burner, (5): gas input top burner, (6): cooling

waterinputtopburner........................... 64

5.4 Simplified four-level transition scheme of Nd:YAG energy levels based

on[24]................................... 65

5.5 Scheme of spectroscopic set-up. (1): Nd:YAG laser head, (2): perisco-

pe, (3): periscope, (4): beam-shaping lenses, (5): counter-flow burner,

(6): intensified charge-coupled device (ICCD) camera, (7): apertures,

(8): beam dump, (9): power unit Nd:YAG laser, (10): cooling water,

(11): mass flow controllers, (12): PTU [23] . . . . . . . . . . . . . . . 66

5.6 Schematic representation of the arrangement of the optical lenses used

for beam widening based on [23] . . . . . . . . . . . . . . . . . . . . . 67

5.7 Schematic representation of the technical operation of an ICCD ca-

mera[23].................................. 69

5.8 Top view scheme of spectroscopic set-up. (1): Nd:YAG laser head,

(2): periscope, (3a-c): beam-shaping lenses, (4): energy monitor, (5):

counter-flow burner, (6): beam dump, (7): aperture, (8): ICCD came-

ra, (9): interference / absorption / polarization filters . . . . . . . . . 72

5.9 Schematic representation of the burner set-up including gas inputs

and mixing chamber (MC), details of the gases used for the experi-

mental work can be found in Appendix A . . . . . . . . . . . . . . . . 73

XVI List of Figures

5.10 Diffusion flame (a) excited at 355 nm with fuel (PG-5-O2-F-100) co-

ming from the top and oxidizer (air) from the bottom nozzle at a

velocity of 30 cms−1, formaldehyde flat fielding image (b), flat fiel-

ded diffusion flame (c) with centerline from bottom to top nozzle

and formaldehyde signal (d) all at a burner separation distance of

18.9mm. An increase of intensity is represented by the color change

of dark blue, to blue, to green, to yellow, to red, to white. . . . . . . . 81

5.11 Polarized diffusion flame (a) excited at 532 nm with fuel (GGL3-2.5-

O2-F) coming from the top and oxidizer (air) from the bottom nozzle

at a velocity of 60 cms−1, depolarized diffusion flame (b) air flat fiel-

ding image (c), flat fielded diffusion flame (d) with centerline from

bottom to top nozzle and Rayleigh signal (e) all at a burner separati-

on distance of 18.9mm. An increase of intensity is represented by the

color change of dark blue, to blue, to green, to yellow, to red, to white. 82

5.12 The solid line represents the temperature distribution and normalized

formaldehyde profiles in a CH4–O2counter-flow and nitrogen diluted

diffusion flame. Solid squares represent the raw, uncorrected profiles

and correlate with profiles corrected with the assumption Q12 ∼T−1

(represented by the cross) and the assumption Q12 ∼T−0.5(repre-

sented by the open circle). The thickness of the formaldehyde zone is

characterized as FWHM (full width half maximum). Based on [26] . . 83

5.13 Normalized with respect to their value at T= 300 Kcorrection coef-

ficients were applied as a function of T, with the cross representing a

correction due to the temperature dependence of the partition func-

tion; the dashed line representing a quenching correction assuming

Q12 ∼T−0.5, the solid line representing a total correction assuming

Q12 ∼T−0.5, the open square representing a quenching correction as-

suming Q12 ∼T−1and the total correction assuming Q12 ∼T−1.

Basedon[26] ............................... 83

6.1 Schematic representation of the numerical and experimental data flows 95

7.1 Sooting methane flame (left) and non-sooting diluted methane flame

(right) with air as oxidizer input from the top and fuel from the

bottomnozzle............................... 97

7.2 Experimentally derived and modelled Rayleigh scattering and tempe-

rature curves with a fuel mixture of CH4-002-N2 at 60 cms−1..... 98

7.3 Experimentally derived and modelled Rayleigh scattering and tempe-

rature curves with a fuel mixture of CH4-002-CO2 at 60 cms−1. . . . 98

7.4 Minima of experimentally and numerically derived Rayleigh scatte-

ring curves with fuel mixtures including varying fractions of nitrogen

and carbon dioxide as diluents at 60 cms−1............... 99

7.5 Experimentally derived Rayleigh scattering and temperature curves

with fuel mixtures of CH-002-N2 in comparison to CH4-004-N2 and

CH4-007-N2 at 60 cms−1.........................100

List of Figures XVII

7.6 Experimentally derived Rayleigh scattering and temperature curves

with fuel mixtures of CH-001-N2 in comparison to CH4-002-CO2 and

CH4-004-CO2 at 60 cms−1........................100

7.7 Minimum Rayleigh scattering intensities of experimentally and nume-

rically derived Rayleigh scattering curves with fuel mixtures including

varying fractions of nitrogen and carbon dioxide as diluents at 60 cms−1101

7.8 Experimentally derived Rayleigh scattering and corresponding tem-

perature curves with fuel mixtures of CH-001-N2 and CH4-004-N2 in

comparison to CH-001-N2-O2-F and CH4-004-N2-O2-F at 60 cms−1. 101

7.9 Exp. derived Rayleigh scattering and corresponding temperature cur-

ves with fuel mixtures of CH-001-CO2 and CH4-004-CO2 in compa-

rison to CH-001-CO2-O2-F and CH4-004-CO2-O2-F at 60 cms−1. . . 101

7.10 Experimentally derived Rayleigh scattering and corresponding tem-

perature curves with fuel mixtures of CH-001-N2 and CH4-004-N2 in

comparison to CH-001-N2-O2-A and CH4-004-N2-O2-A at 60 cms−1. 102

7.11 Exp. derived Rayleigh scattering and corresponding temperature cur-

ves with fuel mixtures of CH-001-CO2 and CH4-004-CO2 in compa-

rison to CH-001-CO2-O2-A and CH4-004-CO2-O2-A at 60 cms−1. . 102

7.12 Numerically derived formaldehyde curves with fuel mixtures of CH-

002-N2 - CH4-007-N2 at 60 cms−1....................103

7.13 Experimentally derived laser-induced fluorescence curves at 355 nm

with fuel mixtures of CH-002-N2 - CH4-007-N2 at 60 cms−1. . . . . 103

7.14 Numerically derived formaldehyde curves with fuel mixtures of CH-

001-CO2 - CH4-004-CO2 at 60 cms−1..................104

7.15 Experimentally derived laser-induced fluorescence curves at 355 nm

with fuel mixtures of CH-001-CO2 - CH4-004-CO2 at 60 cms−1. . . . 104

7.16 Numerically derived CH and experimentally derived CH* chemilumi-

nescence curves with fuel mixtures of CH-002-N2 at 60 cms−1. . . . 105

7.17 Numerically derived CH and experimentally derived CH* chemilumi-

nescence curves with fuel mixtures of CH-002-CO2 at 60 cms−1. . . . 105

7.18 Numerically derived CH curves with fuel mixtures of CH-002-N2, CH-

004-N2, and CH-007-N2 at 60 cms−1..................105

7.19 Experimentally derived CH* chemiluminescence curves with CH-002-

N2, CH-004-N2, and CH-007-N2 fuels at 60 cms−1...........105

7.20 Numerically derived CH curves with fuel mixtures of CH-001-CO2,

CH-002-CO2, and CH-004-CO2 at 60 cms−1..............106

7.21 Exp. derived CH* chemiluminescence curves with CH-001-CO2, CH-

002-CO2, and CH-004-CO2 fuels at 60 cms−1..............106

7.22 Peak positions of experimentally and numerically derived CH* che-

miluminescence and CH curves with methane fuel mixtures including

varying fractions of nitrogen and carbon dioxide as diluents at 60 cms−1106

XVIII List of Figures

7.23 Experimentally derived Rayleigh scattering and temperature curves

with fuel mixtures of GGL2 at 30 cms−1,60 cms−1, and 100 cms−1. . 107

7.24 Experimentally derived Rayleigh scattering and temperature curves

with fuel mixtures of GGL3 at 30 cms−1,60 cms−1, and 100 cms−1. . 107

7.25 Experimentally derived Rayleigh scattering and temperature curves

with fuel mixtures of PG at 30 cms−1,60 cms−1, and 100 cms−1. . . 108

7.26 Numerically derived formaldehyde curves with fuel mixtures of GGL2

at 30 cms−1,60 cms−1, and 100 cms−1..................109

7.27 Experimentally derived laser-induced fluorescence curves at 355 nm

with fuel mixtures of GGL2 at 30 cms−1,60 cms−1, and 100 cms−1. . 109

7.28 Numerically derived formaldehyde curves with fuel mixtures of GGL3

at 30,60, and 100 cms−1.........................110

7.29 Experimentally derived laser-induced fluorescence (LIF) curves at

355 nm with fuel mixtures of GGL3 at 30,60, and 100 cms−1. . . . . 110

7.30 Numerically derived formaldehyde curves with fuel mixtures of PG at

30,60, and 100 cms−1..........................110

7.31 Experimentally derived LIF curves at 355 nm with fuel mixtures of

PG at 30,60, and 100 cms−1.......................110

7.32 Numerically derived formaldehyde curves with fuel mixtures of GGL2,

GGL3, and PG at 100 cms−1.......................111

7.33 Experimental LIF curves at 355 nm with fuel mixtures of GGL2,

GGL3, and PG at 100 cms−1.......................111

7.34 Numerically derived CH curves with fuel mixtures of GGL2 at velo-

cities of 30 cms−1,60 cms−1, and 100 cms−1..............111

7.35 Experimentally derived CH* chemiluminescence curves with fuel mix-

tures of GGL2 at velocities of 30,60, and 100 cms−1..........111

7.36 Numerically derived CH curves with fuel mixtures of PG at velocities

of 30 cms−1,60 cms−1, and 100 cms−1..................112

7.37 Experimentally derived CH* chemiluminescence curves with fuel mix-

tures of PG at velocities of 30,60, and 100 cms−1...........112

7.38 Numerically derived CH curves with fuel mixtures of GGL3 at velo-

cities of 30 cms−1,60 cms−1, and 100 cms−1..............112

7.39 Experimentally derived CH* chemiluminescence curves with fuel mix-

tures of GGL3 at velocities of 30,60, and 100 cms−1..........112

7.40 Peak positions of experimentally and numerically derived CH* che-

miluminescence and CH curves with fuel mixtures of GGL2, GGL3,

and PG at 30 cms−1,60 cms−1, and 100 cms−1.............113

7.41 Experimentally derived Rayleigh scattering and temperature curves

with fuel mixtures of GGL2 at 60 cms−1and additional fuel side oxy-

gen of 2.5,5, and 8.5vol. −%......................114

List of Figures XIX

7.42 Exp. derived Rayleigh scattering and temperature curves with fuel

mixtures of GGL3 at 60 cms−1and additional fuel side oxygen of

2.5vol. −%,5vol. −%, and 8.5vol. −%................114

7.43 Experimentally derived Rayleigh scattering and temperature curves

with fuel mixtures of PG at 60 cms−1and additional fuel side oxygen

of 2.5vol. −%,5vol. −%, and 8.5vol. −%...............114

7.44 Minima of experimentally and numerically derived Rayleigh scatte-

ring curves with fuel mixtures of GGL2, GGL3, and PG at 60 cms−1

and additional fuel side oxygen of 0vol. −%,2.5vol. −%,5vol. −%,

and 8.5vol. −%..............................115

7.45 Numerically derived formaldehyde curves with fuel mixtures of GGL2

at 60 cms−1with and without additional fuel side oxygen of 5vol. −%116

7.46 Experimentally derived LIF curves at 355 nm with fuel mixtures of

GGL2 at 60 cms−1with and without additional fuel side oxygen of

5vol. −%.................................116

7.47 Numerically derived formaldehyde curves with fuel mixtures of GGL2

at 30,60, and 100 cms−1with additional fuel side oxygen of 5vol. −%116

7.48 Experimentally derived LIF curves at 355 nm with fuel mixtures of

GGL2 at 30,60, and 100 cms−1with additional fuel side oxygen of

5vol. −%.................................116

7.49 Numerically derived formaldehyde curves with fuel mixtures of GGL3

at 30,60, and 100 cms−1with additional fuel side oxygen of 5vol. −%117

7.50 Experimentally derived LIF curves at 355 nm with fuel mixtures of

GGL3 at 30,60, and 100 cms−1with additional fuel side oxygen of

5vol. −%.................................117

7.51 Numerically derived formaldehyde curves with fuel mixtures of PG at

60 cms−1with additional fuel side oxygen of 2.5,5, and 8.5vol. −%. 117

7.52 Experimentally derived LIF curves at 355 nm with fuel mixtures

of PG at 60 cms−1with additional fuel side oxygen of 2.5,5, and

8.5vol. −%................................117

7.53 Numerically derived CH curves with fuel mixtures of GGL2, GGL3,

and PG at 60 cms−1with additional fuel side oxygen of 5vol. −%. . 118

7.54 Experimentally derived CH* chemiluminescence with fuel mixtures of

GGL2, GGL3, and PG at 60 cms−1with additional fuel side oxygen

of 5vol. −%................................118

7.55 Peak positions of experimentally and numerically derived CH* che-

miluminescence and CH curves with fuel mixtures of GGL2, GGL3,

and PG at 60 cms−1with additional fuel side oxygen of 0,2.5,5, and

8.5vol. −%................................119

7.56 Numerically derived Rayleigh scattering and temperature curves with

fuel mixtures of GGL2 at 60 cms−1with additional air side oxygen of

2.5,5, and 8.5vol. −%..........................120

XX List of Figures

7.57 Numerically derived Rayleigh scattering and temperature curves with

fuel mixtures of GGL3 at 60 cms−1with additional air side oxygen of

2.5,5, and 8.5vol. −%..........................120

7.58 Numerically derived Rayleigh scattering and temperature curves with

fuel mixtures of PG at 60 cms−1with additional air side oxygen of

2.5,5, and 8.5vol. −%..........................121

7.59 Numerically derived formaldehyde curves with fuel mixtures of GGL2,

GGL3, and PG at 60 cms−1with additional air side oxygen of 5vol.−%122

7.60 Numerically derived formaldehyde curves with fuel mixtures of GGL2,

GGL3, and PG at 60 cms−1with no additional air side oxygen . . . . 122

7.61 Numerically derived CH curves with fuel mixtures of GGL2, GGL3,

and PG at 60 cms−1with additional air side oxygen of 5vol. −%. . 122

7.62 Experimentally derived CH* chemiluminescence with fuel mixtures of

GGL2, GGL3, and PG at 60 cms−1with additional air side oxygen of

5vol. −%.................................122

7.63 Minimum Rayleigh scattering intensities of experimentally and nu-

merically derived Rayleigh scattering curves with fuel mixtures of

GGL2, GGL3, and PG at 60 cms−1with additional fuel side oxygen

of 0,2.5,5, and 8.5vol. −%.......................124

7.64 Numerically derived maximum temperatures with fuel mixtures of

GGL2, GGL3, and PG at 60 cms−1with additional fuel- and air side

oxygen of 0,2.5,5, and 8.5vol. −%...................125

A.1 Side view of experimental set-up, including laser, periscopes, and len-

ses[23]...................................138

A.2 Arrangement of the lenses as schematically represented in Figure 5.6 [23]138

A.3 Periscopes for spectroscopic measurements [23] . . . . . . . . . . . . . 138

A.4 Inside and outside view of in-house built counter-flow diffusion burner 141

A.5 Inside view of in-house built counter-flow diffusion burner with di-

mensions..................................141

A.6 Specifications of Spectra Physics pulsed Nd:YAG lasers [23] . . . . . . 142

A.7 Further specifications of Spectra Physics pulsed Nd:YAG lasers [23] . 143

List of Tables

2.1 Thermal stages of the pyrolysis process based on [5,28–30] . . . . . . 11

2.2 Basic chemical reactions during gasification based on [7] [31] . . . . . 14

3.1 Examples of stoichiometric, fuel-rich and fuel-lean laminar premixed

combustion[32].............................. 26

3.2 Categories in premixed combustion based on the air equivalence ratio

Φer and the fuel equivalence ratio λer [32]................ 27

3.3 Three-step methane reaction mechanism [33] . . . . . . . . . . . . . . 39

3.4 BDG fuel mixtures from thermochemical conversion processes, found

in[47,101–108] .............................. 40

4.1 Depolarization ratio and differential scattering cross sections of major

BDG gases and air [34] and own calculations . . . . . . . . . . . . . . 51

5.1 Technical parameters for experimental measurements with an ICCD

camera[23] ................................ 71

5.2 Defining technical parameters of experimental fluorescence, Rayleigh

scattering, and chemiluminescence measurements . . . . . . . . . . . 73

5.3 Investigated diluted methane mixtures; X representing N2or CO2. . . 76

5.4 Investigated diluted methane mixtures enriched with oxygen on the

fuel side; X representing N2or CO2................... 77

5.5 Investigated diluted methane mixtures enriched with oxygen on the

oxidizer side; X representing N2or CO2................. 77

5.6 Investigated basic fuel mixtures from thermochemical conversion pro-

cesses ................................... 78

5.7 Investigated fuel-oxidizer mixtures enriched with oxygen on the fuel

side..................................... 78

5.8 Investigated fuel-oxidizer mixtures enriched with oxygen on the oxi-

dizerside ................................. 79

6.1 Considered species in the combustion model DIFFLA [220,221] . . . 86

XXII List of Tables

7.1 Temperature evolution of GGL2 fuels with additional oxygen on fuel

orairside .................................126

7.2 Temperature evolution of GGL3 fuels with additional oxygen on fuel

orairside .................................126

7.3 Temperature evolution of PG fuels with additional oxygen on fuel or

airside...................................126

A.1 Investigated fuels with fuel- and oxidizer-velocities and strain rates

for biomass-based mixtures . . . . . . . . . . . . . . . . . . . . . . . . 136

A.2 Investigated fuels with fuel- and oxidizer-velocities and strain rates

for diluted methane mixtures . . . . . . . . . . . . . . . . . . . . . . 137

A.3 List of the main components for the spectroscopic electrical system . 139

A.4 List of the main components for the gas mixing system . . . . . . . . 139

A.5 List of the main components for the spectroscopic system . . . . . . . 140

A.6 List of the utilized gases . . . . . . . . . . . . . . . . . . . . . . . . . 140

A.7 List of the main components for the counter-flow burner system . . . 141

Abbreviations

ADC analog-to-digital converter

BDG biomass-derived gases

CCD charge-coupled device

CHP combined heat and power

Da Damköhler number

DME dimethyl ether

FHG fourth-harmonic generation

ICCD intensified charge-coupled device

L/D ratio of nozzle separation distance to nozzle diameter

Le Lewis number

LFA laminar flamelet assumption

LES large eddy simulations

LHV lower heating value

LIF laser-induced fluorescence

PLIF planar laser-induced fluorescence

Ka Karlovitz number

MC mixing chamber

MFC mass flow controller

MCP microchannel plate

NOxnitrogen oxides

PAH polycyclic aromatic hydrocarbons

PIV particle image velocimetry

PCI peripheral component interconnect

Pr Prandtl number

PTU programmable timing unit

RANS Reynolds-averaged Navier-Stokes equations

Re Reynolds number

RME rape methyl ester

Sc Schmidt number

SHG second-harmonic generation

STP standard temperature and pressure

THG third-harmonic generation

Symbols

Roman letters

Sign Description

ASurface area

aStrain rate

A21 Einstein coefficient

AfConical surface area

a(λ)Mean molecular volume polarizability

ANAvogadro constant

ArCollision factor

AtCross-sectional area of a tube

brTemperature exponent

fMass flux through a wave

CReaction progress variable

cpSpecific heat capacity

dDiameter

DimModified binary diffusion coefficient

DiTThermodiffusion

Di,j Mixture-averaged diffusion coefficient

Dmdc Mean diffusion coefficient

EEnergy of a photon

irepresents the vibrational levels

aActivation energy

Eelectronic Electronic energy of a photon

Erotational Rotational energy of a photon

Evibrational Vibrational energy of a photon

f1, f2Focal lengths

Fk(λ)King correction factor

GFunction representing a scalar field

G0Flame front position in G

r,f Free enthalpy

h=HFFlame height

hPPlanck’s constant

hi(T)Absolute enthalpies

hu,hbEnthalpy upstream and downstream of a wave

IIncident laser intensity

jiDiffusive mass flux

List of Tables XXV

Roman letters

Sign Description

KFlame stretch

Kem Constant for the detector efficiency and the solid collection angle

kiThermal diffusion ratio

KqStern-Volmer quenching constant

krReaction rate coefficients

rEquilibrium constant

kxRate constants

LBurner separation distance

ldistance Distance between lenses

lkKolmogorov turbulent length scale

ltTurbulent length scale

MiMolar mass

MjMean molar mass of the mixture

nAmount of substance

NNumber density

N∗Number density of excited species

n(λ)Refractive index

p,pu,pbPressure

QQuencher

RUniversal gas constant

rRadius

rdMass scale chemical rate of formation

riSpecific rate of formation

RLMolar refractivity

STotal Rayleigh scattering signal

S0Ground state

S1,S2Electronic states

Sem Chemiluminescence signal

si(T)Absolute entropies

SLPropagation speed

T,Tu,TbTemperature

T1Excited triplet state

uVelocity

u0Turbulent intensity

ukKolmogorov turbulent intensity

uu,ubVelocity upstream and downstream of a combustion wave

VVolume

VeAxial velocity

Vem Volume of a pixel

VfVelocity of fuel

VoVelocity of oxidizer

VRObservation volume

r,i,v00

r,i Stoichiometric coefficients

VtAverage flow velocity of a tube

xfFlame front location in G

yu,ybFuel concentration in an unburned and burned region

ZMixture fraction

Zstagnation Distance of the stagnation plane to the bottom nozzle

XXVI List of Tables

Greek letters

Sign Description Unit

δlThickness

αhThermal diffusivity

χair Air mole fraction

χfuel Fuel mole fraction

δDPreheat zone

δRFinite reaction zone

∆Ω Solid angle of the collection optics

(δσ/δσ)mix Differential scattering cross section

ηOptical collection efficiency

γ(λ)Mean molecular volume anisotropy

λWavelength

λbHeat conductivity

λer Fuel equivalence ratio

λtc Thermal conductivity

µDynamic viscosity

νKinematic viscosity

νiAbsorption frequency

νmf Number of moles for a complete reaction

Φer Air equivalence ratio

ΦfFluorescence quantum yield

ΨReactive scalars

ρDensity

ρ0(λ)Depolarization ratio

ρbMass density of a fluid downstream of a combustion wave

ρfDensity of fuel

ρh(λ)Horizontally polarized light

ρoDensity of oxidizer

ρuMass density of a fluid upstream of a combustion wave

ρv(λ)Vertically polarized light

τDetectors exposure time

τcChemical reaction time-scale

τfFlow time-scale

τkKolmogorov time-scale

τtTurbulent flow time-scale

σTotal Rayleigh cross section

Chapter 1

Introduction

One of the greatest challenges of our time is global warming and the ongoing climate

change. The use of limited fossil fuels needs to be re-thought, and the utilization of

renewable and sustainable energy systems finally needs to gain further importance.

In 2014, the leaders of the European Union agreed to achieve a reduction of green-

house gas emissions by more than 40 % by 2030, in comparison to the level in the

year 1990. At the beginning of 2015, an energy union strategy for the European

Union was published by the Juncker commission, basing the provisioning on secure,

sustainable, competitive, and affordable measures. These should be established with

the help of a plan split into five dimensions: ensuring energy security, improving en-

ergy efficiency, supporting research and innovation in clean technologies, integrating

an internal energy market, and decreasing the use of fossil fuels [35].

The energy system needs to be revolutionized holistically, not only focusing on the

electricity sector but also concentrating on other sectors such as agriculture, trans-

portation, industrial applications, and heating and cooling. From this point forward,

the main targets should be based on sector coupling, and furthermore, a carbon re-

duction plan for all situations and phases of our society.

Efforts are in progress to overcome the disadvantages of the utilization of renew-

able energy sources. Such may be, i.e., the specialization of countries in assorted

technologies. These can be chosen based on local and geographical circumstances,

therefore touching economic and environmental areas of expertise. Also, overcoming

cost detriments in comparison to traditional fuels like nuclear power, coal, lignite,

and/or gas, which have to be carried by the European Union and by its member

states [36]. Furthermore, spatial and temporal availability issues need to be over-

come, also having an influence on provisioning energy exclusively on the basis of

renewable sources by combining more than one source [4].

2 Chapter 1 Introduction

Biomass as a key renewable energy resource could be recognized as the answer to

many concerns and dependencies previously presented. To act as a source for any

of the possible conversion processes, some requirements need to be met regarding

up-scaling: an increased conversion efficiency, a dependable and stable source of

feedstock production, and a profitable but also sustained agricultural land manage-

ment [4].

Furthermore, for the ever-increasing request for energy, combustion has been and will

remain a simple key technology to utilize. The world’s energy demand has been met

by 90 % by the combustion of fossil and biofuels. Though this increasing demand has

been escalating the three major problems of: decreasing fossil fuel resources, steady

global warming, and increased air pollution [37]. To govern these major problems,

combustion technologies need to be substantially advanced.

Technical and industrial combustion applications, such as engines and gas turbines,

are dominated by turbulent, rather than laminar, combustion processes. Heat re-

leases and mixing processes accelerate flow instabilities and, therefore, turbulent

flow fields [38]. To advance a gain in knowledge in this field, initially, numerical

modelling of turbulent combustion is necessary. Multiple approaches are used to

numerically simulate such processes, such as laminar flamelet models. Here laminar,

one-dimensional, and time-dependent thin layers, namely flamelets, are covering and

can present a turbulent combustion regime [39]. Finally, these models need to be

validated by experimental solutions.

The needs for biomass as a renewable energy source and further improvements in

combustion technologies motivated this investigation of synthetic model fuels with

biomass-based compositions. Since combustion can be more efficient with syngas

from biomass gasification or pyrolysis processes, rather than combusting the feed-

stock directly. This is based on higher temperatures or the possibility of utilizing

other technical applications [40].

This work focuses on the experimental validation of a laminar model simulat-

ing one-dimensional counter-flow diffusion flames named DIFFLA using multiple

biomass-based synthetic fuels and, furthermore, the flames behavior during combus-

tion, with and without oxygen enhancement. Chapter 2 will present an overview of

biomass as a source of energy and the conversion processes leading to the investi-

gated fuels. This is followed by an outline of the varying combustion techniques of

gaseous fuels in Chapter 3. Chapter 4 presents the applied experimental spectro-

scopic techniques Rayleigh scattering, laser-induced fluorescence of formaldehyde,

and CH* chemiluminescence used for validating the model DIFFLA, which is fur-

thermore introduced in Chapter 6. To preliminarily verify the in-house designed

and built counter-flow burner set-up described in Chapter 5, diluted methane com-

bustion was conducted experimentally and confirmed numerically. The results and

conclusions are presented extensively in Chapters 7 and 8.

Chapter 2

Biomass as a source of energy

One of the major appeals of biomass as a source of energy or green chemicals is

the wide availability and its possibility to be quickly renewed. In contrast to fossil

fuels, biomass is formed from living species that are alive now or were briefly ago,

and it does not take several million years to be replenished. Atmospheric CO2is

metabolized by plants via sunlight using photosynthesis, instigating their growth [28,

29]. Commonly only about 1 % of the sunlight is used by photosynthesis. The origin

of the biomass can be treated as organic material that accumulates solar energy

in the form of chemical bonds. When these bonds are broken, for example, by

thermochemical conversion processes, the materials release their chemical energy.

As part of this process, the deduced carbon material is oxidized to again produce

CO2and water. Therefore it is regarded as a cycle when being available to create

new biomass [41].

In the sense of conversion, a wide range of biomass, like small grasses to immense

trees, can be useful and generally be classified in lignocellulosic material [42], sugar,

starch [28], oilseeds [41], and agricultural, forest, municipal or industrial waste [1].

Biofuels can also be further subclassified in first- (1G) and second-generation (2G)

biofuels. Grains, sugar cane, vegetable oils, i.e., are either food crops themselves or

mainly produced from food crops. 1G biofuels can therefore be linked to increasing

food prices. This promoted the development of 2G fuels from non-food biomass,

residues from agriculture or forestry, or co-products from production processes to

confine the „food vs. fuel“ conflicts [43,44].

Lignocellulosic biomass is dominantly composed of the aromatic polymer lignin and

the carbohydrate polymers cellulose and hemicellulose, which basically constitute

the cell walls in biomass. Humans can not digest lignocellulose; hence it is not

part of the human food chain and not ominous for food supplying processes [28,42].

6 Chapter 2 Biomass as a source of energy

The structure of lignin is a highly complex phenylpropenyl formation, which is ran-

domly branched and transforms based on the biomass’s origin. The phenypropenyl

macromolecule is, amongst others, made of syringols and guaiacols and joined by

carbon-oxygen or carbon-carbon bonds [42]. Cellulose is a crystalline structured

long-chain polymer made up of many glucose molecules. It can be represented by

the generic formula (C6H10O5)nand is highly polymerized, with a massive molecular

weight [28]. Hemicellulose is, unlike cellulose, of a weak, random, and amorphous

structure, which is not resistant to hydrolysis and can be presented by the generic

formula (C5H8O4)n[29]. It is made of a big group of carbohydrates, a typical hemi-

cellulose molecule being xylan, and is of a lower degree of polymerization. Further-

more, lignocellulosic biomass includes extractives and small amounts of inorganic

components such as iron, calcium, phosphorus, and/or sodium [28].

For lignocellulosic biomass, some pretreatment can be necessary due to the enclosing

of the lignin and hemicellulose around the cellulose. Pretreatment techniques there-

fore not only liberate the cellulose from other surrounding components of biomass

but also lower the cellulosic crystallinity [42]. The output depends on a criterion

named the „severity factor“, which is the joined effect of the duration, tempera-

ture, and acidity of the process [45]. Torrefaction can be used as a pretreatment

for thermochemical conversion processes by improving multiple properties of the

lignocellulosic biomass [46]. Pretreatment techniques should be utilized on low op-

erational and capital costs, preparations of the biomass before the pretreatment

should be kept to a minimum, and lastly, it should run on a low energy demand, or

it shall be possible for the energy to be reused, i.e., for secondary heating [45].

Biomass-based fuels, as presented in 3.4.2, based on thermochemical conversion pro-

cesses such as gasification and pyrolysis, represent carbon-neutral alternative fuels

for combustion processes. In comparison to fossil fuels, their energy density is lower;

therefore, their combustion behavior has to be investigated thoroughly to exploit

their full potential. Furthermore, due to the difference in energy efficiency, typi-

cal applications for fossil fuels would have to be modified geometrically, or co-fir-

ing would have to be applied. Finally, a systematic overview of biomass-derived

gases (BDG) has not been achieved yet, due to the extreme diversity of composi-

tions due to numerous dependencies on process and application parameters [47].

2.1 Conversion processes of biomass 7

2.1 Conversion processes of biomass

The conversion of biomass to energy incorporates various decisions, mainly which

type of energy is required for which purpose and also what type of biomass will be

utilized. Which conversion pathway will be pursued, what kind of infrastructure is

available or needed, and lastly, what application will be prosecuted [3].

Figure 2.1: Schematic illustration of the conversion routes of biomass based on [1–4]

The different conversion paths for a.) biological, b.) thermochemical, and c.) physi-

cal conversion processes are represented in Figure 2.1. The routes a.), and c.) will be

presented briefly, as the focus of this study was on product gas of the thermochemical

conversion processes gasification and pyrolysis.

Biological processes can be subdivided into the methods of anaerobic digestion by

bacteria or fermentation with the use of enzymatic conversion. Anaerobic digestion

can be applied to convert organic material by bacteria into biogas, which consists

mainly of methane and carbon dioxide (and other gases in small fractions like hydro-

gen sulphide). This biogas can be utilized directly in gas turbines and gas engines

and can furthermore be upgraded by removing carbon dioxide. Ethanol can be

produced from sugar and starch crops via fermentation [3]. The use of enzymes is

resorted to when converting macromolecular starch. This type of starch can not

be fermented by conventional technologies; therefore, it is ground up, mixed with

water, heated, and treated with enzyme preparations to produce ethanol or amino

acid [1].

Physical conversion techniques include distillation, briquetting of biomass, and me-

chanical or solvent extraction. To extract essential oils, the method of distillation

8 Chapter 2 Biomass as a source of energy

is frequently used. Biomass is steam distilled, the oils are vaporized, and the more

volatile fractions of a composition can be separated from the non-volatiles [1]. Es-

pecially biomass waste as agricultural or forest residues are frequently tricky to

handle due to their uneven and burdensome characteristics. By briquetting or pel-

letisation, this waste can be formed into a densified and homogeneous product for

reuse [48,49]. Extraction as a mechanical conversion process yields oil from seeds of

various biomass crops, to be processed in further steps to obtain bio-diesel or rape

methyl ester (RME). A residual solid or „cake“ emerges herewith as a byproduct

that can be disposed as animal food. Finally, primary and secondary metabolites

can be extracted from biomass using solvents [1,3].

Finally, the thermochemical conversion processes pyrolysis, gasification, combustion,

and liquefaction of biomass, especially lignocellulosic materials, will be presented.

These are key technologies for the production of industrial or domestic heat and

electricity or the production of fuel for a combined heat and power (CHP) cogen-

eration [50]. As biomass is heated at varying temperatures and under sufficient or

insufficient oxygen conditions, thermal degradation and chemical reformation lead

to gaseous, liquid, and/or solid outputs. Compared to other conversion techniques,

the advantage of thermochemical processing is the possibility of transforming all the

organic components of the biomass [1].

Pyrolysis is the thermochemical decomposition of biomass in the absence of exter-

nally provided oxygen, leading to liquid bio-oils, gases, and solid products like char-

coal. The produced char can be utilized as „green coal“, while the bio-oil can either

be upgraded to be possibly used as renewable diesel or jet fuel or applied directly in

power plants, furnaces, or boilers [51,52]. Lastly, a noncondensable primary gas mix-

ture made of carbon dioxide, carbon monoxide, water, methane, ethane, ethylene,

and other components is produced [28]. During gasification, biomass is converted

into a combustible gas mixture, typically by partial or complete oxidation, depend-

ing on the gasifying agent. The biomass feedstock is heated in a medium like air,

oxygen, or steam, commonly with an insufficient supply of oxygen. The produced

fuel can either be burned directly in gas engines or gas turbines or used as syngas for

the manufacturing of chemicals or liquid fuels [31,51]. Combustion or burning is the

direct exothermic chemical reaction between a fuel and an oxidizer, commonly air,

releasing carbon dioxide and water and leading to heat and therefore, i.e., mechanical

energy or electricity. Stoves, furnaces, boilers are based on this process [3].

2.1 Conversion processes of biomass 9

Finally, during the thermochemical process liquefaction, a liquid is obtained us-

ing low temperatures, high pressure, and a catalyst, also commonly with hydrogen

present [51].

2.1.1 Pyrolysis

Pyrolysis is the thermochemical degradation of materials in the absence of oxygen

up to around 500 ◦C, which not only results in the production of gases but also solid

(charcoal) and liquid (bio-oil water) outputs [3]. The general chemical equation is

shown as follows [28]:

(CnHmOp) + Heat −−→ Xliquid CaHbOc+Xgas CxHyOz+Xsolid C(2.1)

This is the fundamental chemical process, making it the precursor both of the gasi-

fication and the combustion processes. The products depend on the feedstock, the

heating rate, and the final temperature of the procedure. Furthermore, as the typi-

cal lignocellulosic feedstocks are poor heat conductors, the size of the particles needs

to be on a small-scale [42].

The reaction kinetics during pyrolysis of cellulose have been studied intensely, and

abundant mechanisms have been proposed. The model of Broid-Shafizadeh, as dis-

played in Figure 2.2, is broadly known.

Figure 2.2: Model for cellulosic pyrolysis mechanism by Broid-Shafizadeh based on [5]

Based on this model, the mechanism includes three steps: first, activation of the

cellulose, second, the transformation of this active cellulose into volatiles, and third

the conversion of this active cellulose into char and gas. The thermal degradation

of cellulose starts at 52 ◦C; at these low temperatures, the degree of polymerization

starts to decrease, and charring occurs. At higher temperatures, volatilization takes

place rapidly [53]. The determination of a mechanism for hemicellulose is complex

due to the varying and inconstant branched structures in the material when com-

paring woody biomass amongst each other. The most abundant polysaccharide is

10 Chapter 2 Biomass as a source of energy

xylan, this is also the part of the hemicellulose that is the least thermally stable.

The degradation process starts at 200 ◦C-260 ◦C. Lignins are also highly branched

and complex in the form of mononuclear aromatic polymers, often bound to border-

ing cellulose fibers. The pyrolysis of lignin has been investigated intensively, leading

to the assumption that the primary pyrolysis products are guaiacol and pyrogallol

dimethyl ether. The degradation reaction of lignin has maxima between 225 ◦C-

450 ◦C, depending on the gasifying agent [5]. The overall pyrolysis reactions for

woody lignocellulose are arranged schematically in Figure 2.3 [6].

Figure 2.3: Model for lignocellulosic biomass pyrolysis, commonly known from [6], based on [5]

This model insinuates that lignocellulosic biomass degrades in a competitive reac-

tion mechanism to volatiles, gases, and char in a primary pyrolysis. Furthermore,

the produced volatiles and gases react again with char to modified kinds of volatiles,

gases, and char in secondary interactions. [5]. Based on the heating rate, overall

temperature, feedstock residence time, and particle size of the solid feedstock, the

pyrolysis process is classically subdivided into slow pyrolysis, fast pyrolysis, and flash

pyrolysis [5,53].

The slowest type of pyrolysis process is carbonization. The biomass is heated slowly,

even for days, at very low heating rates, up to a temperature of 400 ◦C, with charcoal

production as a primary goal. The typical slow pyrolysis has a medium heating rate,

a residence time of minutes to hours, as opposed to days, and produces all types of

pyrolysis products previously mentioned at a maximum temperature of 600 ◦C[28].

The vapor residence time is considerably longer during slow pyrolysis, as in any

other form of this thermochemical conversion, hence giving the vapor additional

time to react while the char and oils are being formed. Otherwise, vapors can also

be removed constantly while being produced [54]. From a thermal perspective, the

process of pyrolysis can be separated in the stages listed in Table 2.1, lacking sharp

boundaries.

2.1 Conversion processes of biomass 11

By increasing heating rates, the yield of char decreases. In turn, a higher output

of liquid products can be achieved via the form of conversion named fast pyroly-

sis [51]. Bio-oils are formed by rapidly depolymerizing and fragmenting cellulose,

hemicellulose, and lignin, and finally quenching intermediate products that would

react further, with longer residence times, during high temperatures [54].

Table 2.1: Thermal stages of the pyrolysis process based on [5,28–30]

Temperature Processes Products

Drying (~100 ◦C) Free moisture and mildly bound

water evaporates, while this heat

is absorbed into the biomass.

Melting of lignic fraction at high

humidty.

Initial stage

(~100 ◦C-~300 ◦C)

Exothermic dehydration of the

feedstock, hence water

elimination. Depolymerization

and free radical formation.

Carbonyls, carboxyls, CO,

CO2, charred residue.

Intermediate stage

(>~200 ◦C)

Primary pyrolysis takes place

from ~200 ◦C-~600 ◦C, producing

most of the bio-oil.

Primary char, condensable

gases, noncondensable

gases.

Intermediate stage

(~300 ◦C-~450 ◦C)

Breaking of glycosidic bonds and

linkages of polysaccharides. The

secondary cracking of volatiles

into noncondesables and char,

and therefore the final stage

begins. Hydrocarbon evolution.

Exothermic peaks from lignin

decomposition appear at around

~280 ◦C-~390 and ~420 ◦C.

Tar in form of

levoglucosan, anhydrides,

and oligosaccharides

(important parts of

bio-oil).

Final stage

(>~450 ◦C)

Dehydration, displacement, and

decomposition of sugar units.

Forming of carbonyl

products such as glyoxal,

acrolein, and

acetaldehyde.

Final stage

(>~600 ◦C)

A higher production of H2is

favored, with a further increase

of the yield above ~900 ◦C.

H2and a mixture of all

the previously mentioned

compounds.

The produced bio-oil is a homogeneous liquid, exhibiting half the heating value

of conventional fuel oil. To obtain this output, the biomass feedstock should be

< 3 mm, have vapor residence times of < 2 s, at best < 10 % moisture in the biomass

feed and sustain rapid cooling of the vapors. Furthermore, the ash content of the

biomass and the char separation influence the catalytic effect on the vapor cracking

and hence the liquid yields [52].

12 Chapter 2 Biomass as a source of energy

Bio-oils need to be upgraded before most applications due to their low pH (leading to

corrosion), high viscosity (predicament to transport in pipes), alkali metals content

(damage or sedimentation), thermal instability (decomposition), and so forth. This

enhancement can be achieved by filtration, hydrogenation, or catalytic cracking of

the liquid product [7].

One of the biggest challenges during fast pyrolysis is the heat transfer to the vessel.

A considerable amount of thermal input is needed to heat the biomass feedstock

to a reaction temperature in this endothermic process. Possible solutions represent

multiple heat transfer surfaces, heating the fluidization gas, or heating the bed

material (if included) [52].

During flash pyrolysis, the feedstock is heated to a moderate temperature of 450 ◦C-

650 ◦C, but in an extremely short period of time of 30-1500 ms. During this process,

the liquid yield can be increased even further than during fast pyrolysis, up to

70-75 %, due to rapid quenching. High heating rates favor gaseous products at the

expense of bio-oil due to gas-phase cracking reactions [28,55]. Flash pyrolysis can

be achieved when involving a heat carrier solid. This ensures a high heat transfer

while the solid undergoes a rapidly fast mixing process with the biomass feedstock.

The peak liquid yields are reached at around 650 ◦C, and gaseous products can be

maximized close to 1000 ◦C[28].

Pyrolysis processes can occur in reactor types that will be presented thoroughly in

Section 2.1.2. The principal designs that are utilized are: fixed- or moving-bed,

bubbling fluidized-bed, circulating fluidized-bed (will be presented in the follow-

ing section), entrained flow, rotating cone, ultra-rapid, ablative, and vacuum reac-

tors [28,56]. The rotating-cone reactor works as a transport-bed reactor, with the

movement realized by centrifugal forces [52]. In an ultra-rapid pyrolyzer, both the

inerting agent and heat-carrier solids are heated externally and blasted on a biomass

feedstock stream. Such forces result in a heating rate of a few milliseconds. The

ablative technique is substantially different compared to the other pyrolysis pro-

cesses presented. A decrease in heat transfer limitation is established by pressing

biomass against the reactor’s wall at high pressure mechanically or via centrifugal

forces. The liquid rather melts out of the biomass, while the liquid film that devel-

ops on the walls vaporates. Reaction rates are not limited by the heating rates any

longer; larger sizes of feedstock can be utilized [28, 52]. Lastly, vacuum pyrolysis

with numerous stacked and heated circular plates at different temperatures, where

the biomass falls through with the help of knife scrapers, can be employed [28].

2.1 Conversion processes of biomass 13

2.1.2 Gasification

During the process of gasification, biomass is converted to gases, vapors, and

char/ash, achieving high yields of combustible gases like hydrogen, carbon monox-

ide and methane [56], and diluents like nitrogen and carbon dioxide, with the final

composition depending on the process conditions. Furthermore, tars and char are

byproducts of this process. To enhance the production of gaseous products, [57]

gasification is carried out at temperatures >800 ◦C[31], typically at partial oxida-

tion or partial combustion. [57] Coal gasification was used since the early 1800s for

illumination reasons (town gas) and utilized for multiple other applications until the

mid-1900s. The conversion from coal to gas was also regarded as an opportunity

to minimize the dependency on petroleum from uncertain sources. Biomass as a

feedstock for gasification offers an opportunity to use wastes and residues and si-

multaneously lower carbon emissions [7]. The produced gas is very versatile to be

utilized in gas turbines or power gas engines and can be further processed into liquid

fuels [31].

Figure 2.4: Steps of biomass gasification in a gasifier [7], reprinted with permission

The gasification of solid biomass also includes multiple steps: heating and drying,

pyrolysis, gasification, and oxidation (in case air or oxygen is used). All of these

reactions occur both in solid- and gas-phase, as pictured in detail schematically in

Figure 2.4.

The phase of heating and drying is essential to remove the moisture from the solid

biomass at around 100 ◦C. During this time, no chemical reaction takes place [7].

The time required for this step depends primarily on the moisture degree, which is

highly variable in biomass, and secondly on the temperature, the particle size, and

density of the feedstock [7,8].

14 Chapter 2 Biomass as a source of energy

After the drying process, the temperature of the biomass particle rises until the

chemical decomposition in the absence of oxygen, namely, pyrolysis, starts. First,

the hemicellulose breaks down around 225 ◦C-325 ◦C, then cellulose at around 300 ◦C-

400 ◦C. Lastly, the highly complex and robust lignin decomposes at temperatures up

to 500 ◦C, though the process can also begin at cooler temperatures. Lignin is the

component that produces the greatest quantity of solid products. This will be the

determining factor for the char production in the next steps of the mechanism. A

gaseous stream consisting of hydrogen, carbon monoxide, carbon dioxide, methane,

light hydrocarbons, and high molecular weight hydrocarbons are formed during the

pyrolysis step. The hydrocarbons with a high molecular weight are either combusted

in a later step or contribute to the tar fraction. Furthermore, the pyrolysis leads to

a stream of volatiles and a solid charcoal product, their yield depending on heating

rates, particle size, the interaction of temperatures, effects of catalysts, the nature

of the feedstock, and the existence of non-combustible components [7,56].

During the latter steps, the char reacts primarily with water and carbon dioxide

during gas-solid reactions. Furthermore, homogeneous gas-phase reactions take

place [7,58].

Table 2.2: Basic chemical reactions during gasification based on [7] [31]

Reaction

C + 1

2O2←−→ CO Partial oxidation

C + O2←−→ CO2Complete oxidation

C + 2 H2←−→ CH4Hydrogenation

C + CO2←−→ 2 CO Boudouard Reaction

C + H2O←−→ CO + H2Water gas reaction

CO + H2O←−→ CO2+ H2Water gas shift reaction

CO + 3 H2←−→ CH4+ H2O Methane formation

The first five reactions in Table 2.2 show heterogeneous gas-solid reactions, con-

verting solid carbon into carbon monoxide, carbon dioxide, methane, and hydrogen.

The heat release is largest from the complete oxidation, while partial oxidation only

generates 65 % ([31]) of this heat in comparison. The last two reactions show further

development of carbon monoxide and hydrogen and, furthermore, a higher yield of

methane and carbon dioxide in the produced fuel gas via gas-phase reactions. The

last two reactions depend very much on the conditions, shifting in any direction.

2.1 Conversion processes of biomass 15

Overall, the final gas composition will be determined by the residence time, temper-

ature, and pressure within the gasifier, by the fuel composition and amount, by the

heat distribution within the gasifier, and the type and amount of gasifying agent that

is utilized, all influencing the calorific value of the product gas. A low calorific gas is

produced when using air as a gasifying agent, which is used directly in combustion

or as an engine fuel. A product gas with a medium calorific value is produced when

utilizing oxygen or pure steam as a gasifying agent, which serves as a feedstock for

a conversion into chemicals like methane or methanol. Lastly, a product gas with

the highest calorific value is achieved when using hydrogen as a gasifying agent or

during hydrogenation [7,31,57]. At usual gasification conditions, though, hydrogen

is not an optimal gasifying agent due to more complex reaction mechanisms than,

i.e., with steam. During carbon conversion in chars and coals with hydrogen, the

reactivity can fluctuate greatly [59].

To achieve the conversion of biomass via gasification, multiple gasifier types subdi-

vide primarily into fixed-bed, fluidized-bed, and entrained flow gasifiers, utilized as

continuous processes, as exemplified in Figures 2.5 and 2.6 [8]. They differ in how

fuel and gas get in contact, the heating mode, and the flow patterns between the

biomass and gas [7]. The choice of gasifier depends, amongst others, on the quantity

of produced gas, the utilization [52], and primarily on the power-to-be-installed and

the available biomass.

For small and medium-scaled applications, the most traditional fixed-bed gasifier

systems are most suitable, generally leading to low calorific product gases and oper-

ating at around 1000 ◦C. Depending on the direction of flow of the gasifying agent,

these technologies can be classified as updraft (Figure 2.5, left), downdraft (Figure

2.5, right) or with a cross-flow [31,42].

In an updraft gasifier, the biomass feedstock is introduced gastight at the top and

the oxidizer at the bottom of the gasifying unit. This is considered to be the simplest

technique. The feedstock and oxidizer move in an opposing direction, the biomass

going through the stages of drying, devolatilization (pyrolysis), reduction (gasifica-

tion), and combustion. Some of the char settles on the grate in the combustion zone,

or „hearth zone“ state, ashes fall through the grate at the bottom. The gasifying

agent enters the combustion zone, reacting with the char at temperatures as high as

1200 ◦C, leading to carbon monoxide, carbon dioxide, and steam. This hot gas sup-

plies the energy for the other process steps and gets cooled while it moves upwards.

In the pyrolysis zone, considerable amounts of volatiles lead to large quantities of

tar in the product gas, making the exit pipes prone for plugging. Due to the up-

16 Chapter 2 Biomass as a source of energy

draft, the tar does not have the opportunity to pass through the hot temperature

zone for further reactions. On the other hand, a product gas with a low amount

of particulates is produced due to the filtering effect of the feed. The gas exits the

gasifier at around 80 ◦C-100 ◦C, leading to a high overall energy efficiency for this

process [8,31,42].

Figure 2.5: Fixed-bed updraft (left) and downdraft (right) gasifiers [8], reprinted with permission

In a downdraft gasifier (Figure 2.5, right), the gasifying agent is introduced in the

middle or at the top of the reactor in a throat region, as opposed to the bottom in

the updraft system. This leads to a coinciding flow of the solid biomass feedstock

and the produced gas, which now exits the vessel at the bottom. Furthermore, the

order of the hearth and reduction zones are reversed. This change in the configu-

ration of the gasification zones dramatically reduces the tar content in the syngas

when compared to the products from the updraft technique. The tar-rich volatiles

now pass through a high-temperatured (800 ◦C-1200 ◦C) combustion zone after the

pyrolysis. The tar can therefore efficiently be cracked, leading to high conversion

rates. To achieve this, the introduced biomass feedstock must have a moisture con-

tent of less than 20 % to reach sufficiently high temperatures [8, 42]. When using

the cross-flow technique, the feed of the biomass is descended, while the oxidizer is

moreover introduced from one side and the product gas is eliminated on the other

side. The combustion and reduction zones develop around the access point of the

gasifying agent, the pyrolysis and drying zones are positioned higher in the reac-

tor. A low efficiency due to product gas temperatures of up to 900 ◦Cand high tar

contents are achieved during this process [31].

2.1 Conversion processes of biomass 17

Fluidized-bed reactors can be subclassified in bubbling (Figure 2.6, left) and circu-

lating (Figure 2.6, right) gasifiers and are generally known for their homogeneous

mixing and temperature distributions. The biomass feedstock, in combination with

the gasifying agent and the material bed, is brought into a quasi-suspended „flu-

idized“ state.

Figure 2.6: Fluidized-bed bubbling (left) and circulating (right) gasifiers [8], reprinted with per-

mission

The bubbling gasifier was developed by Winkler in 1921 and is, therefore, the oldest

commercially used fluidized-bed reactor. Primarily utilized for the gasification of

coal then, it is now a popular choice for the gasification of biomass, especially due

to its insensitivity to the quality of the feedstock and applicability for medium-sized

units of <25 MWth [28]. While the gasifying agent flows upwards through the bed,

the material emerges into an emulsion of particles and gas bubbles, physically resem-

bling a boiling fluid. The bed commonly consists of sand, dolomite, or alumina and

can typically be fluidized by the agents air, oxygen, or steam. To maintain the flu-

idization in the bed, the volumetric flow has to be controlled, to finally be decreased

in the freeboard by enlargement of the cross-sectional area. This lowers the velocity

to return the particles to the material bed and increases the residence time of the

gas-phase, prompting gas-solid and gas-tar reactions [8]. Alongside the uniformity

of the temperature and mixing in the reactor, advantages of a bubbling bed include

a yield of homogeneous product gas and high heat transfer. Downsides include

large bubbles in the bed and hence some gas bypassing [7], a limit of the operating

temperature due to ash slagging, and a high tar and fines content in the produced

18 Chapter 2 Biomass as a source of energy

gas [31]. Circulating fluidized-bed systems operate on the principle of evading the

void of the material bed during the gasification process and are additionally useful

when coping with high capacity throughputs. As the velocity of the gas increases,

also the load of solids in the freeboard extends, resulting in a deficiency of particles

in the bed material. This is avoided by including cyclones in the reactor, returning

bed material and char via a downcomer to the reaction vessel, and to remove ash.

Advantages of this process are the possibility to operate with rapid reactions, and,

furthermore, achieve high heat transfer rates and high conversion rates. Disadvan-

tages appear to be temperature gradients and a heat exchange, which is limited in

comparison to the bubbling fluidized-bed gasification process [7,8,31].

For the sake of completeness, entrained-flow gasification is be presented shortly. This

technique is favorable and widely used for large-scale coal, coke, and refinery residue

processes, utilizing top-fed or side-fed gasifiers. Due to the need for finely ground

feedstock material <0.1-0.4mm, this process is not suitable for fibrous materials

like biomass. Such pulverization is hardly possible, and the very short residence

times in these reactors are therefore unfit for gasification of lignocellulosic material.

This limits the use of entrained-flow processes for biomass feedstock strongly on a

commercial scale [28,31].

2.1.3 Combustion

One of the oldest thermochemical conversion process known to humankind is the

combustion or burning of biomass. Until the early 1900s, a broad spectrum such as

heating, cooking, the generation of steam, mechanical and electrical power, and the

chemical and charcoal production was met by this technique. The basic stoichiomet-

ric formula for the burning of wood bases on the formula of cellulose (C6H10O5)n[29]:

(C6H10O5)n+ 6 nO2−−→ 6 nCO2+ 5 nH2O(2.2)

The understanding, the science, the knowledge of the chemical mechanisms, the

thermal efficiencies, and emission rates of woody combustion have significantly

developed since complex feedstocks and furthermore, waste feedstocks are being

utilized in combustion or co-combustion systems [29].

2.1 Conversion processes of biomass 19

Figure 2.7: The four steps of biomass combustion in a boiler [7], reprinted with permission

This technique is based on the three Ts: high temperatures for the ignition process,

enough turbulence for a complete mixture during the gas-phase combustion of the

fuel and oxidizer, and a sufficient amount of time for the oxidation to occur. This

is achieved during the four steps of heating and drying, pyrolysis, flame- and finally

char combustion, as can be seen in Figure 2.7.

The steps of heating and drying and pyrolysis have been thoroughly discussed in

section 2.1.2. Because these are both globally endothermic steps, they require an

external energy supply to take place. This is achieved by providing an external

source of heat or by the addition of air for partial oxidation [7,58].

To complete the thermochemical processes during combustion, the volatile gases

act around the solid fuel. Additionally, an adequate amount of oxygen needs to be

present for the final two steps. If this is not the case, incomplete combustion can

result in the formation of polycyclic aromatic hydrocarbons (PAH), soot, and other

organic compounds in the resulting flue gas. The charcoal, mainly consisting of

carbon, continues to react to carbon monoxide and carbon dioxide, being controlled

by the transfer of oxygen to the surface of the char. The reaction of the char takes

place on the surface or inside of its pores [7,56].

The scale of combustion plants ranges from primitive, small-scale traditional cook-

ing stoves to furnaces used for CHP or power generation applications, ranging from

kilowatt to multi-megawatt sizing [3,8]. Three classic types of combustion systems

are fixed-bed, fluidized-bed, and entrained flow reactors. The smallest fixed-bed

combustors are simply made of one combustion chamber and a grate, with pri-

mary and secondary air supply above and below this grate, to ensure the combus-

tion of volatiles and char, respectively [56]. Further applications are found using

manually-fed systems, spreader-stoker systems, underscrew systems, throughscrew

20 Chapter 2 Biomass as a source of energy

systems or static and inclined grates. In comparison to the fixed-bed technology,

fluidized systems seem more sufficient in large scale applications and exhibit greater

combustion efficiencies due to a homogeneous distribution of the temperature in

the reactor, operating at slightly lower temperatures from 700 ◦C-1000 ◦C. Silica

sand, limestone, dolomite, or other non-combustibles are utilized as the bed for the

biomass and act as the heat transfer medium by high-pressured air flowing from the

bottom up. Based on the velocity of the air, this technique can be subdivided into

bubbling or circulating fluidized-bed systems [42, 56]. Lastly, combustion can be

conducted in an entrained flow by transporting the feed into an externally heated

tube by a cooled injector. The feed consists of fuel particles and air, and it is ignited

by a burner at the access to the reactor [42].

2.1.4 Liquefaction

The final thermochemical conversion process to be presented is the biomass lique-

faction. This process was introduced at the beginning of the 20th century, peak-

ing in 1914 when Bergius introduced his procedure of coal conversion in the pres-

ence of hydrogen. [8] The feedstock is converted into oxygenated hydrocarbons in

high pressured hydrogen environment at low temperatures [3]. The lignocellulosic

macro-molecules are broken down with a catalyst present, leading to fragments of

light molecules. These unstable and reactive fragments continue to repolymerize

into oily liquids. The primary goal is liquids of higher quality than those attained

during the pyrolysis process, in the sense of lower oxygen content or higher heating

values, to avoid extensive upgrading due to hydrogen [5].

The main objective of the liquefaction process is the increase of the H/C ratio, with

a necessary decrease of the O/C ratio, from the biomass to the product. Hence,

adding hydrogen or carbon monoxide as reducing gas is essential. The removal

of oxygen is chemically feasible by producing water or carbon dioxide; the latter

process leads to a product with a higher H/C ratio and, therefore, a higher lower

heating value (LHV) [5]. The utilization of catalysts is especially critical in this

thermochemical conversion process to reduce the reaction temperature, increase the

yield of desired liquids, enhance reaction kinetics, and decrease the formation of

residues [56]. Typical catalysts are acids and alkalis. Furthermore, the consistency

of the feedstock slurry in the employed solvent (water or other) is crucial [5].

Chapter 3

Combustion of gaseous fuels

A deeper understanding of the diverse laminar and turbulent combustion processes

can be gained with this chapter. The basics of non-dimensional numbers or burning

effects for premixed and non-premixed techniques are investigated. Finally, the

focus is on counter-flow combustion, which is the main interest of these numerical

and experimental investigations. Furthermore, biomass-based and methane fuels are

presented in reference to properties like gasifying agents and oxygen enhancement.

3.1 Laminar and turbulent combustion

In this chapter, the fundamental differences between laminar and turbulent combus-

tion will be discussed. Combustion regimes will be defined, also by further charac-

terizing non-dimensional numbers. The concept of laminar flamelets as thin layers

embedded in turbulent combustion flow fields is also presented [39].

Combustion processes are based on the mixing and burning of fuel and oxidizer.

This procedure can be further categorized by mixing fuel and oxidizer and then

burning it subsequently (premixed) or by mixing and burning fuel and oxidizer

simultaneously (non-premixed). These processes can again be subdivided based on

laminar or turbulent flow types of fuel and oxidizer [32].

In basic terms, a laminar flow can be identified as a regular, steady, and smooth flow

of fluids when they are highly viscous or slowly moving. When these characteristics

change, by regarding fluids with a lower viscosity or by increasing the velocity, the

movement of the flow gradually becomes chaotic and irregular, therefore turbulent.

The instabilities caused by flow perturbations induce the change from a laminar to

a turbulent flow type, also passing a transitional state [60,61]. The non-dimensional

22 Chapter 3 Combustion of gaseous fuels

Reynolds number (Re) is used to categorize the effects of inertial and viscous forces

on flow fields. Furthermore, it helps to distinguish between the laminar and turbu-

lent flows of fluids, as well as the transitional period [61,62].

The Reynolds number can be defined as follows [62]:

Re =ρud

µ(3.1)

With ρas the density of the fluid, u as the velocity of the fluid, d as the diameter as

a characteristic dimension, and µas the dynamic viscosity of the fluid. Depending

on disturbances, a change in flow type is usually expected around Re ~2000; this is

also based on the type of reactor and further installations [62].

3.1.1 Laminar premixed combustion

The process of laminar premixed combustion can primarily be applied in industrial,

residential, or commercial processes like, i.e., heating appliances, ovens, or bunsen

burners. Basically, the comprehension of the laminar flame is a crucial requirement

to further understand the mechanisms in turbulent flames. Similar physical processes

underly both in laminar and turbulent combustion; therefore, laminar structures

often establish the basis of turbulent flow theories [63]. A central task for studies

of premixed and also non-premixed combustion processes is to establish the laminar

flame speed, next to concentrations and temperature distributions as further points

of interest. The flames burning velocity Slcan be defined as the speed at which the

flame front propagates relative to the unburnt gases of the combustion process [64].

The first laminar premixed flame created for laboratory use was the Bunsen burner,

developed by Bunsen in 1855 [64]. It is shown schematically in Figure 3.1.

The structure of a premixed flame includes four main parts: the preheat zone (where

the reactants are being heated), the inner-layer (where the radical formation takes

place), the oxidation-layer (where the remaining reactions occur), and the post-flame

zone (where hot products and NOxevolve). The oxidizer (air) is induced by numer-

ous ports on the side of the burner and introduced into the mixing chamber. The

fuel is supplied from the side (or bottom) and adjusted via an aperture („orifice“) at

the bottom, entering the mixing chamber, thus being premixed with the air. Upon

arrival at the top of the burner, the gases are considered well mixed and, therefore,

homogeneous. Via the orifice, the flow of the fuel can be adjusted, and the mixing

of the fuel and air can be optimized accordingly.

3.1 Laminar and turbulent combustion 23

When the fuel feed is regulated and constant, the flame is anchored near the top

of the burner, remaining stationary. This Bunsen flame propagates normal to itself

with Slinto the unburnt fuel mixture [9,64].

Figure 3.1: Schematic overview of a bunsen burner [9]

The laminar flame speed or laminar burning velocity for a premixed combustion

with a Bunsen burner Slcan be presented as follows [64]:

Sl=Vt

(3.2)

with Vtas the average flow velocity in the tube, Atas the cross-sectional area of

the tube, and Afas the conical surface area of the innermost cone of the flame [64].

Chemical and physical properties like temperature and also the type and composition

of the fuel type have an impact on the flame speed [11].

The combustion process of a premixed flame in a Bunsen burner is initiated when

an external energy source heats reactive, combustible gases, and therefore triggers

chemical reactions. This process continues without an external energy source when

the heat produced during the originally initiated endothermic reaction is sufficient

to further heat the unburned fraction of the reactive gases. The zone between the

unburned and burned fractions is called the combustion wave. The movement of the

combustion wave towards the unburned region is defined as propagation [65].

24 Chapter 3 Combustion of gaseous fuels

The analysis of the most basic and idealized mode of combustion wave propagation,

like in a laminar premixed flame, which is based on various assumptions, is presented

in Figure 3.2.

Figure 3.2: Schematic overview of a 1-D premixed planar combustion wave based on [10]

The subscript u indicates the conditions of the unburned gases before the wave, the

subscript b the conditions of the burned gases after the wave. This wave propagation

is adiabatic, steady, planar, based on ideal-gas laws, and relative to a combustible gas

mixture. The flame can be described via the Rankine-Hugoniot relations, while the

wave structure is ruled by the nonequilibrium processes of diffusion and reaction [10].

From the conservation equations, the following relations can be obtained [10]:

Mass :

ρuuu=ρbub=f(3.3)

Momentum :

ρuu2

u+pu=ρbu2

b+pb(3.4)

Energy :

hu+1

2u2

u=hb+1

2u2

b(3.5)

3.1 Laminar and turbulent combustion 25

Where ρuand ρbare the mass densities of the fluids upstream and downstream of

the wave, uuand ubthe velocities upstream and downstream of the wave, f the mass

flux through the wave, puand pbare the pressures upstream and downstream of the

wave and huand hbthe specific enthalpies in the two regions.

Because the flame propagates at a definite speed through the unburned mixture, it

can be regarded as an interface to the burned mixture of an infinitesimal thickness.

This flame sheet theory applies when there is a change of mixture properties from

the fuel concentration in the burned region yb=0 to the fuel concentration in the

unburned region yuand the burned gas temperature Tbto the unburned gas tem-

perature Tu. After the flame front, the temperature of the reactants has increased to

Tb, and the fuel concentration has dropped to approximately 0; the flame itself has

broadened to a finite thickness. This is due to convection and diffusion processes,

which will be further discussed in Section 3.1.2. Nonetheless, the chemical reaction

itself takes place in a very limited region, where the maximum temperature prevails

due to the high reaction rate. Towards the edge of the preheating zone δD, a finite

reaction zone δRcan be found, with δR«δD[10,11].

Starting with the earliest simplified analysis of laminar premixed flames by Mallard

and Le Chatelier in 1883, some significant assumptions were made to explain the

fundamental physical processes. This means, amongst others, to presume the non-di-

mensional Lewis number (Le) to be unity [11]. This is one of the non-dimensional

numbers associated with basic diffusion processes, thermal and mass diffusivity, and

the kinematic viscosity in laminar combustion [10].

The Lewis number Le can be defined as follows [10]:

Le =αh

Di,j

=λtc

ρcpDi,j

(3.6)

with λtc being the thermal conductivity, αhthe thermal diffusivity, ρthe density,

cpthe specific heat capacity at constant pressure, and Di,j the mixture-averaged

diffusion coefficient (mass diffusivity) [66].

The Schmidt number (Sc) brings the momentum diffusivity in relation to the mass

diffusivity [10]:

Sc =ν

Di,j

=µ

ρDi,j

(3.7)

26 Chapter 3 Combustion of gaseous fuels

also including µas the dynamic viscosity. The measure of the relative influence of

the momentum diffusivity to thermal diffusivity is considered the Prandtl number

(Pr) [10]:

Pr =ν

αh

=µcp

λtc

(3.8)

Le, Pr and Sc are related as follows [10]:

Le =Sc

Pr (3.9)

Equation 3.9 shows the equality of the Lewis number to the ratio of the Schmidt

number (Sc) and the Prandtl number (Pr), and therefore relevancy in a concurrent

heat and mass transfer. The Lewis number also has an influence on the growth rate

of the thermal and concentration boundary layers, which will concur when Le=1 [66].

The stoichiometry depends on the consumption of fuel and oxidizer by each other.

A fuel-rich environment is based on a surplus of fuel, a fuel-lean environment, on

the other hand, on the scarcity of fuel. Examples can be found in Table 3.1 [32].

Table 3.1: Examples of stoichiometric, fuel-rich and fuel-lean laminar premixed combustion [32]

Reaction

2 H2+ O2−−→ 2 H2O stoichiometric

3 H2+ O2−−→ 2 H2O + H2rich (H2left over)

CH4+ 3 O2−−→ 2 H2O + CO2+ O2lean (O2left over)

If a chemical reaction is made up of excatly 1 mol of fuel, the mole fraction of this

fuel can be calculated for a stoichiometric mixture as follows [32]:

χfuel,stoichiometric =1

1 + νmf

(3.10)

with νmf expressing the number of moles of O2for a complete reaction to CO2

and H2O. Premixed combustion processes of fuel and air as oxidizer must take the

existing N2into account and are therefore represented by the air equivalence ratio

λer, which can be presented by the fuel equivalence ratio Φer [10,32]:

Φer =1

λer

=χfuel/χair

χfuel,stoichiometric/χair,stoichiometric

(3.11)

3.1 Laminar and turbulent combustion 27

Corresponding to these relations, premixed laminar combustion processes are divided

into different categories, as shown in Table 3.2.

Table 3.2: Categories in premixed combustion based on the air equivalence ratio Φer and the fuel

equivalence ratio λer [32]

rich combustion Φer >1,λer <1

stoichiometric combustion Φer = 1,λer = 1

lean combustion Φer <1,λer >1

3.1.2 Laminar non-premixed combustion

In a non-premixed combustion process, the mixing of fuel and oxidizer and the

combustion itself occur simultaneously, as opposed to in a premixed environment.

Applications include candles, campfires, oil lamps, and laminar counter-flow and

laminar co-flow burners for research purposes. Due to the bigger range from 0

(being pure air) and ∞(being pure fuel) in regard to the equivalence ratio Φer, a

more complex chemistry develops in a non-premixed combustion process. The flame

front is located where the stoichiometric composition Φer = 1 prevails due to the

temperature distributions being surrounded by an area of rich combustion on the fuel

side and lean combustion on the air side. Historically, the types of laminar flames

were distinguished as premixed and diffusion flames. Since the diffusion process is

part of all these flame combustions, the term non-premixed is used, giving this type

of combustion a more unique notation [10,32].

The air and fuel are moved towards each other through a convective motion and

also via diffusion. During these transport processes, the temperature of fuel and air

increases, and finally, they meet in the reaction zone and get mixed up rapidly [10].

Finally, in comparison to premixed flames, a laminar burning velocity can not be

established for laminar non-premixed flames because of the missing propagation

phenomenon. In premixed combustion, diffusive processes induce propagation. In

the non-premixed laminar cases, the oxidizer and fuel diffuse to the flame front.

The position of the flame front is fixed because the flame can not propagate into

the oxidizer without fuel and vice versa. While oxidizer and fuel diffuse into the

flame front, energy is released when converting them into products through chemical

kinetics. Subsequently, the produced species and energy diffuse away into the lean

air side and rich fuel side [32].

28 Chapter 3 Combustion of gaseous fuels

A complete reaction can not be established due to the finite rate of the reaction

taking place and also the limited thickness of the reaction zone. Also, marginal

amounts of fuel and air perpetually leak away from the reaction zone. An infinite

reaction rate was assumed by Burke and Schumann [64], restricting the reaction

zone to a reaction sheet. In this case, the air and the fuel are restrained to their

particular zones, leading to marginal concentrations at the reaction sheet, preventing

a leakage. If this infinite reaction rate is presumed, the stoichiometric rates of air

and fuel transport control the combustion and, therefore, the rate of heat release.

The limiting and controlling factor, accordingly, is the rate of diffusion, as opposed

to the rate of the reaction [10].

One of the more straightforward methods to induce a laminar non-premixed flame

is via a co-flow flame or jet flame, as presented in Figure 3.3. This kind of burner

set-up merely includes two concentric tubes, with fuel flowing out of the center tube,

while the oxidizer is supplied by the outer ring. Therefore, the adjunct co-flow. On

the left image in Figure 3.3, a small divergence of the flame can be regarded at the

burner rim, following a continuous divergence onto the centerline. The fuel ejects

out of a vertical nozzle with a diameter of 2R and an axial velocity of Ve. The flow

velocity decreases after exiting the nozzle, and the flow strays slightly. At first, the

momentum of the fuel jet is preserved in the flowing field. Subsequently, as it enters

the conical zone, this momentum is conveyed to the oxidizer (air), decreasing the

velocity due to further reactions taking place. As the fuel jet proceeds downstream,

more and more air is being dragged into the mixture.

Figure 3.3: A non-premixed co-flow flame (left) and the scheme of a non-reacting fuel jet

(right) [11], reprinted with permission

3.1 Laminar and turbulent combustion 29

Also downstream of the flame, the jet edge (Figure 3.3, right side) expands gradually

in diameter as it moves away from the conical zone („potential core“) at the nozzle.

At the potential core, the velocity stays constant at Ve[11].

The height of the flame h can be approximated using the radius r, the velocity in

the direction of the jet Ve, and the mean diffusion coefficient Dmdc in the mixture,

as in the following correlation [32]:

h=r2Ve

2Dmdc (3.12)

Two horizontal levels are included at the edge of the nozzle and around the center of

the flame to examine the velocity distributions further. At the origin (nozzle), the

velocity is distributed in a top hat form, while finally, the velocity and fuel concen-

tration decrease down to zero constantly between the potential core when moving

towards the jet edge. Downstream of the potential core processes like diffusion and

viscous shear effects are effective all-over the area of the jet, with fuel molecules

diffusing outward based on Fick´s law [11].

Because the velocity and fuel mass fraction are decreasing when moving further along

the flame, the time for diffusion increases. The shape of the Burke-Schumann flame

can evolve tulip-like in case of under-ventilation, or if the air supply exceeds the

stoichiometric requirements and over-ventilates the flame, it forms a closed shape as

pictured in Figure 3.4.

Figure 3.4: Over- and under-ventilated Burke-Schumann flame [11], reprinted with permission

The tulip-like shape in an under-ventilated environment forms due to the fuel diffus-

ing outward seeking a higher amount of oxygen, resulting in an opened up shape with

a smaller flame height h = Hf. This laminar non-premixed flame is less applicable

for non-intrusive measurements due to its axisymmetric shape [11].

30 Chapter 3 Combustion of gaseous fuels

The Burke-Schumann flame is prone to buoyancy effects and is also impaired by the

stabilization process at the rim of the nozzles. [10]. Finally, when studying laminar

non-premixed combustion, counter-flow flames need to be regarded. This will be

done in detail in Section 3.2.

3.1.3 Turbulent combustion

Turbulent combustion is one of the most challenging problems in science due to its

increased complexity in the presence of multiphase flows and multiple and branching

chemical reactions. When regarding industrial processes, combustion usually takes

place in a turbulent rather than laminar manner. This is the case mainly due to

heat release and mixing processes. First, flow instabilities (i.e., gas expansion and

buoyancy) are induced by heat releases during combustion, leading to an increased

transition to turbulence. Second, combustion is enhanced by mixing processes, which

are accelerated by turbulence [38].

To gain an understanding of these kind of processes, a basic knowledge of the turbu-

lence-chemistry interactions in reactive systems is necessary [67]. The characteristics

of turbulent flames can be outlined mainly by length and time scales, non-dimen-

sional numbers, and, furthermore, by thermal and molecular diffusivities and fluc-

tuation intensity [68].

Turbulence induces a formation of eddies, leading to energy cascades from large-scale

to small-scale eddies. This kinetic energy transfer is finally weakened by viscous

dissipation, assumable only at the smallest length scale. Lengths widely agreed on

in literature include [68]:

•Integral length scale (ΛT): largest eddies in the flow, containing the most

kinetic energy.

•Taylor length scale (λT): most active eddies in the flow, receive energy from

larger eddies and dissipate to smaller ones.

•Kolmogorov length scale (ηT): smallest eddies in the flow.

Especially in numerical modelling of turbulent combustion, a determination of com-

bustion regimes is essential. By relating characteristic turbulent and chemical scales,

diffusivities, viscosities, and the previously defined length scales, non-dimensional

3.1 Laminar and turbulent combustion 31

numbers can be used to categorize these regimes. The definition of the case-rel-

evant time scales that control the flame is fundamental to describe a numerical

system [67,68].

The Reynolds number Re has been previously introduced as a way to categorize

effects of viscous and inertial forces on flow fields.

In turbulent combustion Re is based on the integral length scale [68]:

ReT=u0ΛT

ν(3.13)

Defined by the turbulent intensity (u0), integral length scale (ΛT), and additionally

by the kinematic viscosity (ν) [68].

The Damköhler number (Da) is used to relate the turbulent flow time-scale to the

chemical reaction time-scale (τT/τc), to define the nature of reaction and mixing in

a combustion system [68].

The chemical time scale is defined by the ratio of flame thickness to the propagation

speed (τc=δL/SL). The ratio of integral length scale to turbulent intensity describes

the turbulent flow time-scale (τT=ΛT/u0) [68].

Defining the Damköhler number in a turbulent environment as follows [68]:

DaT=τT

τc

=ΛT

δL

(3.14)

When regarding the smallest flow time-scale (Kolmogorov scale, τk), the Karlovitz

number (Ka) is needed to describe turbulent combustion processes [68].

Defining the Karlovitz number Ka in turbulent combustion as follows [68]:

KaT=τc

τk

(u0

)1.5

(δL

ΛT

(3.15)

Additional non-dimensional numbers in regard to diffusivities and viscosities were

presented previously in Section 3.1.1.

Re, Da and Ka in turbulent combustion are related as follows [12]:

ReT= (DaT)2(KaT)2(3.16)

Further dependencies in regard to premixed turbulent combustion processes can be

found in turbulent premixed combustion regime diagrams in literature such as [68].

32 Chapter 3 Combustion of gaseous fuels

In comparison to premixed combustion, it is difficult to define regimes for turbu-

lent non-premixed combustion processes. The computation of a reaction time-scale is

challenging due to the non-distinctiveness of the flame velocity. Multiple flow scales,

which can be spatially or temporally dependent, and additionally, further quantities

like the reaction zone and the diffusion layer need to be defined [67]. Eventually,

the separate introduction of fuel and oxidizer in non-premixed flames leads to an

entrainment by larger-scale eddies. This prompts the formation of fuel-lean and

fuel-rich cavities and also their local mixing, which can finally lead to a heat re-

lease due to quicker mixing at the molecular level than the chemical reaction. A

modification of the turbulent flow occurs, enhanced by the volume expansion [69].

Three versions of the Damköhler number can be calculated regarding non-premixed

flames. The differences are again based on the mixing time scales when deciding

between the integral, local, or Kolmogorov scales for the computation of Da. When

regarding sufficiently large or sufficiently small Damköhler numbers, two transitional

versions evolve. The former, when the laminar flamelet assumption (LFA) applies

leading to DaLFA, the latter, when extinction occurs introducing DaEXT [67].

Figure 3.5: Turbulent non-premixed combustion regime diagram based on [12]

An attempt to characterize non-premixed combustion regimes is presented in Figure

3.5 and the following outline ([67], [69]):

•Laminar flames (Re < 1): laminar diffusion flame, with no turbulence affecting

the combustion.

•Flamelets (Re > 1, Da ≥DaLFA): steady laminar flamelets can be preserved in

a turbulent regime. Reaction time scales are smaller than mixing time scales.

3.1 Laminar and turbulent combustion 33

•Unsteady effects (Re > 1, DaLFA > Da > DaEXT): mixing time scales cause

instabilities in the flame front and unsteady effects are dominant.

•Quenching (Re > 1, Da ≤DaEXT): when the Damköhler number is sufficiently

limited and small, the extinction of the flame occurs.

As discussed, combustion processes can generally be split into the classes of non-pre-

mixed or premixed combustion. In regard to turbulent technical procedures, promi-

nently diesel engines or furnaces combust under non-premixed conditions, as opposed

to homogeneous charge spark-ignition engines and gas turbines, which operate in a

premixed mode.

The diesel internal combustion engine auto-ignites when a liquid fuel spray is injected

into a cylinder with hot compressed air, therefore inducing partial mixing of the

evaporated fuel and the air. This partially premixed gas is consumed quickly, and

the following phase of the burnout takes place non-premixed. In gas furnaces, fuels

are injected into preheated air, which might be partially mixed with exhaust gases.

Following the ignition, the flame moves toward the nozzle until finally stabilizing

at a certain point, resulting in the lift-off height. An area of partial premixing

consequently develops between the nozzle and the lift-off height, followed by a region

of non-premixed combustion further downstream [70]. Evidently, partially premixed

combustion plays a role in non-premixed combustion, presenting a possibility of

simultaneously reducing NOxand soot formation, i.e., in diesel engines [71].

In a spark-ignition engine, fuel and air are premixed by turbulence and then com-

pressed into the cylinder for an adequate amount of time before the mixture is

combusted. An initial spark ignites the mixture, generating a flame that primar-

ily grows laminar and then by turbulent flame propagation. Making the turbulent

burning velocity a vital quantity in this type of combustion [70].

Details to gaseous combustion processes, in terms of oxidizer and fuel mixing, were

previously discussed in Sections 3.1.1 and Section 3.1.2 on the basis of laminar

combustion.

Another principle of subdividing turbulent combustion correlates to the ratio of tur-

bulent to chemical time scales, leading to slow or fast chemistry cases. Here the

focus lies on two mechanisms: autocatalytic reactions known as chain-branching

and recombination reactions known as chain-breaking.

Above a certain crossover temperature, hydrocarbon oxidation occurs due to

chain-branching reactions. At ambient pressure, the crossover temperature lies

34 Chapter 3 Combustion of gaseous fuels

between 950-1300 K, based on the reactive mixture. Below this temperature, the

chain-breaking dominates the chain-branching and leads to extinction [72]. To avoid

this extinction in a real-scale process, combustion operates at temperatures consid-

erably above the crossover temperature to assure rapid chemical reactions, referred

to as fast chemistry. Slow chemistry, on the other hand, is not as prone to turbu-

lence, as it operates dependent on local temperature and pressure and also chemical

composition [73].

3.2 Counter-flow combustion

The other main possibility to induce a laminar non-premixed flame, and the focus of

this study is based on a counter-flow set-up. A counter-flow flame can be established

by two opposed flows, one being fuel and the other oxidizer or via the Tsuji burner

([74]). Both can generally induce a „purer“ diffusion flame than a co-flow burner.

Variations of different combustion parameters can be studied thoroughly in these

kind of flames due to their one-dimensional and comparably simple behavior. This

has been done for the later parts of the past century to clarify topics such as transport

processes, extinction mechanisms, and complex chemical kinetics and correlations

of these flat flames. Especially Smooke et al. [75–79] have investigated these types

of flames in-depth experimentally and numerically [11].

Figure 3.6: Classification of counter-flow flames into Types I-IV from [13] commonly known

from [14], reprinted with permission

The counter-flow combustion was classified into two groups as previously established

and again separated into four types by Tsuji, as can be seen in Figure 3.6. Types

I and II are based on the counter-flow combustion, including two gaseous opposed

jets. While type I includes rectangular nozzles or circular tubes as burners to induce

flat flames, type II uses two matrix burners to eject the reactants. Burner types III

3.2 Counter-flow combustion 35

and IV are based on the forward stagnation region of a porous injector surrounded

by free-streaming oxidizer, with III in a spherical or hemispherical shape and IV in

a cylindrical shape [14].

When using a co-flow jet to induce a laminar non-premixed flame, a dead zone

near the rim of the nozzles is established. This phenomenon is caused by heat

loss close to the wall, decreasing temperatures, and therefore a decline of active

radicals. Furthermore, this leads to small spaces of premixed zones at the base of

the flame, effecting the structure and chemistry of this overall non-premixed flame.

Accordingly, the presented burners in Figure 3.6 are thought to be more suitable

to investigate certain fundamental processes of laminar diffusion combustion [14].

Moreover, experimental residence times are easier to control than with previously

discussed set-ups, the flames can be stabilized by an inert co-flow, and finally, the

transport of the reactants is controlled by convection, hence minimizing buoyancy

effects [10,13,80].

Figure 3.7: Type I counter-flow burner set-up with a steady strained 1-D diffusion flame [11],

reprinted with permission

The structure of a diffusion flame as pictured in Figure 3.7 can be defined by the

flame stretch K, which furthermore leads to the strain rate a. Despite its flat form

and stationarity, aerodynamic straining leads to a stretch in counter-flow flames.

Due to the symmetry in these flames, the strain rate can be related to the velocity

gradient found at the centerline leading from the top to bottom nozzle; this also

applies with unequal velocities of fuel and oxidizer [13,81].

36 Chapter 3 Combustion of gaseous fuels

The global strain rate a can be represented by the densities of fuel and oxidizer ρF

and ρO, the velocities of fuel and oxidizer VFand VO, and the burner separation

distance L [13,82]:

a=−2VO

L1 + VF

VOρF

ρO1/2(3.17)

The impact of strain on the combustion is not as well defined for non-premixed

flames, as opposed to premixed flames and the non-dimensional Lewis number Le.

Diffusion flames, though, are more inclined by flame extinction due to stretch. As the

velocities of fuel and oxidizer, and therefore the strain rate a, increase, the residence

time of the radicals decreases, leading to incomplete combustion. Ultimately, the

residence time equals the chemical time; the heat release of the flame is insufficient

in heating up the reactants, and the flame is blown off at this critical strain rate [13].

Following the ignition of the flame, the flow field acts aerodynamically complex.

Before the flame is reached, the flow slows down, though then speeding up again

when it is passed until the stagnation plane is reached. The radial acceleration is

constant, while the axial velocities are fastest at the nozzle exits [83–85].

To determine the position of the flame relative to the stagnation plane, either the

momentum of the streams or the volumetric flow of fuel and oxidizer and type of

reactants should be established. The physical position of the diffusion flame is always

located where the stoichiometric ratio is reached. The distance of the stagnation

plane to the bottom nozzle Zstagnation can be represented by the densities of fuel and

oxidizer ρFand ρO, the velocities of fuel and oxidizer VFand VO, and the burner

separation distance L [82,85]:

Zstagnation =L−1 + VO

VFρO

ρF1/2−1

L(3.18)

The flame front of the diffusion flame comprises the maximum temperature during

the combustion process, while the temperatures decrease constantly down- and up-

stream in the directions of the nozzles. The fuel concentration is consumed upon

reaching the flame front, while the oxidizer tries to diffuse to the fuel side persistently.

The combustion products are mostly in the proximity of the flame front [14, 84].

Contrary to co-flow jet flames, do counter-flow flames seem relatively resistant to

buoyancy instabilities after the positioning of the flame has taken place [80].

3.3 Laminar flamelets [in turbulent combustion] 37

3.3 Laminar flamelets [in turbulent combustion]

Under certain circumstances, reactions in turbulent combustion are known to include

thin sheets named flamelets that appear like narrow and steady laminar flames [86].

These flamelets are wrinkled and wrily due to the turbulent flow but seem to main-

tain the internal framework of a laminar flame, being physically one-dimensional

and stationary in time [32,87].

These approximations are, therefore, extensively used in large eddy simulations

(LES) and Reynolds-averaged Navier-Stokes equations (RANS) for turbulent com-

bustion process simulation. Flamelets can be steady or unsteady and non-premixed

or premixed leading to several different laminar flamelet models [87].

Flamelet assumption theories presume a fast enough combustion chemistry, sepa-

rating the flow into phases of burned gases and fresh gases that are parted by the

so-called flamelet elements. These theories are also based on the assumption that the

flamelet behaves explicitly like a laminar flame. The crucial information, though,

is the structure of the flow and the narrow continuous region where the chemical

reaction occurs. This region may also be thickened by limited turbulence without

impairing the flamelet assumption theories [88].

Central components of all flamelet models are the definition and the effect of the

flame stretch and the shape on the flame’s behavior. These two properties have an

extensive impact on the flames wrinkle and surface area A, as well as the local mass

burning rate of the flame front [89].

The flame stretch K can therefore be defined by the surface area [88]:

K=1

dt (3.19)

K changes the flame front spontaneously by increasing the flame surface via a high

stretch leading to flame quenching or a small to moderate stretch causing surface

production [88].

Amongst others, flamelet models either use a reaction progress variable C to describe

the combustion evolution in the flamelet regime or the level-set function G as a

starting point as introduced by Markstein in 1964 [32]. The G equation can be

deduced from the spontaneous flame surface, with G representing a scalar field,

with the flame front position located at G = G0, G > G0in the burned mixture and

G < G0in the unburned mixture [32,90].

38 Chapter 3 Combustion of gaseous fuels

G can be described with the time t and the flame front location xfand is presented

by [90]:

G(xf, t) = G0(3.20)

The progress variable C is typically described by a combination of reactive scalars

such as temperature, chemical species mass fractions, viscous-diffusive characteris-

tics or chemical source terms defined by the vector ψand can be calculated with

the use of the mixture fraction Z as follows [91]:

Ψ = Ψ(Z, C)(3.21)

The progress variable C has been interpreted variously, as found in [91].

3.4 Gaseous fuels

In this investigation, methane and diluted methane were considered for preliminary

studies. Methane has a significant fraction in the composition of natural gas, and

diluents are components in biomass-derived fuels. Furthermore, BDG, as presented

in the following subsection, were considered experimentally and numerically.

3.4.1 Methane and diluted Methane

One of the main components of natural gas is methane, with a volume fraction of

about 85 −90 %. An increased use of natural gas as an alternative fuel for internal

combustion engines such as spark-ignition engines, or industrial power plants, leads

to the need for further experimental and numerical investigation on the combustion

behavior of this hydrocarbon fuel [92,93].

A three-step mechanism, based on steady-state and partial equilibrium assumptions

for the oxidation of CH4, can be used to analyze the structure of a non-premixed

counter-flow methane-air flame as presented in Table 3.3.

3.4 Gaseous fuels 39

Table 3.3: Three-step methane reaction mechanism [33]

Reaction

I CH4+ O2−−→ CO + H2+ H2O

II CO + H2O←−→ CO2+ H2

III O2+ 2 H2−−→ 2 H2O

The rates for these steps can be associated with the rates of the elementary reac-

tions. Furthermore, the outer structure of this flame is based on the Burke-Schumann

structure and dependent on the overall one-step reaction as follows [33]:

CH4+ 2 O2−−→ CO2+ 2 H2O(3.22)

This mechanism is based on the four-step kinetical mechanism by K. Seshadri and

N. Peters [94,95], reducing the mechanism of hydrocarbon chemistry [96] of CH4to

the fundamental steps, though still leading to a reasonable flame structure. These

fundamental steps are connected to the rates of elementary reactions of the C1-chain

mechanism for the oxidation of methane [95,97].

The dilution of fuels advances the progress of investigating hydrocarbon combustion

processes and will therefore potentially improve technical industrial applications,

such as maintaining thermal stabilization in combustors or reducing nitrogen oxides

(NOx) by exhaust gas recirculation. The reacting flow field is affected by dilution

in fuel and/or oxidizer streams by altering reactions, the flame structure, flame

stability, and also emissions via thermal, dilution, or chemical effects [98]. A thermal

effect is induced by a change in physical properties due to an alteration of the

fuel composition. Furthermore, a dilution effect alters the carbon structure of the

fuel composition and the concentrations of reactive species, therefore, modifying

their frequency of colliding. Lastly, a chemical effect results from the chemical

participation of the diluents, leading to an increase or decrease of soot [99,100].

Additionally, a hydrodynamic effect can impact basic combustion phenomena like

mixing, entrainment, extinction due to strain, etc. A strong and diverging impact

bears the choice of either diluting the fuel or the oxidizer stream [98].

3.4.2 Biomass-based fuels

In the early 1920s, the utilization of biomass gasification to supply fuel for motor

vehicles evolved. Especially the government of Sweden promoted the conversion of

40 Chapter 3 Combustion of gaseous fuels

feedstock as fuel, establishing wood gas-fueled cars by the outbreak of World War

II. After the war, the use of low-priced hydrocarbons terminated the interest in

biomass-based fuels as engine fuel again [31].

Various products from gasification and pyrolysis, landfill gases, and syngas are con-

sidered to be BDG and have been reported in literature [47,101–108].

The compositions of selected BDG from gasification and pyrolysis processes are listed

in Table 3.4, where GG-X denotes products from gasification processes, and PG-X

denotes product combinations from pyrolysis processes. Also, a few undefined gases

with a volume fraction of <1 % were merged into the nitrogen fraction [47].

Table 3.4: BDG fuel mixtures from thermochemical conversion processes, found in [47,101–108]

BDG CO

[vol.-%]

CO2

[vol.-%]

CH4

[vol.-%]

C2H4

[vol.-%]

LHV

[MJ/kg]

Reaction

agent

Biomass

GG-H 35.50 27.00 28.70 6.5 0 2.30 9.20 Air-

Steam

Cellulose

GG-L1 27.92 30.11 35.39 4.36 0 2.22 8.60 Air Pine wood

GG-L2 37.65 28.89 27.17 4.78 0 1.51 8.40 Oxygen-

Steam

Pine wood

GG-W 20.00 12.00 18.00 2.00 0 48.00 4.60 Air Wood

GG-S 24.00 0 21.00 0 0 55.00 5.20 Air Biomass

wastes

GG-Vä 19.00 13.20 12.00 5.80 0 50.00 4.90 Air Wood

PG-D 56.80 4.20 21.60 10.20 5.10 2.10 15.80 - Sylvester

pine and

spruce

PG-Le 51.10 25.80 8.70 8.30 2.50 3.60 8.80 - Pressed oak

and beech

sawdusts

PG-Lu 26.00 61.50 1.60 8.30 0 2.60 3.90 - Pine bark

The GG- and PG- compositions change depending on the biomass type, reactor

parameters, process parameters, and gasifying agent. The main components of the

gaseous fuels are N2, H2, CO, CO2, and CH4, and possibly smaller C2-chained gases.

The most commonly used reaction agents are air, oxygen, steam or a combination

of these. The utilization of air as reaction agent leads to product gases with lower

heating values, but also decreases the operating costs. The use of oxygen, on the

other hand, leads to gaseous fuels with higher heating values to be used for the

manufacturing of fuel cells and chemicals [7,109].

3.5 Combustion properties 41

To improve the use and application of BDG as an energy source, it is essential to

gain a deeper insight into the combustion processes and further knowledge of flame

structures, species development, temperature distribution, and such.

3.5 Combustion properties

The fuel derived from BDG, especially its quality, and the grade of the following

combustion is based on many factors from the preceding thermochemical conversion

process. The composition of the fuel gas and also the derived amount can be influ-

enced by the type of gasifying agent, by the oxidizer that was utilized, a possible

addition of water or steam to the process, a heat surplus or heat loss during the

conversion, the development of PAH, soot and tar, biomass residence times, and

such. This last section gives a short insight on these topics.

3.5.1 Gasifying agents

As previously presented, a gasification process can be classified based on flow pat-

terns or methods of contact between biomass and gas, heating modes, type of reactor,

or finally on the gasifying agent. Here generally, steam, air, oxygen, or oxygen-en-

riched air are utilized. When air is used as the gasification agent, the operating

costs are low, but the deduced fuel gas has a relatively low LHV, as shown in table

3.4, with a high share of N2. Furthermore, this high nitrogen content leads to a

significant increase in volume of the product gas. Steam, on the other hand, leads

to a product gas with a high CO2content. Finally, oxygen as gasification agent leads

to a fuel with a higher-value, to be used in higher value utilization, such as fuel cells,

but the production includes more considerable costs. In sporadic cases and during

advanced gasification processes, hydrogen can be used as a gasification agent. This

leads to higher yields of methane and hydrogen to carbon ratios and substantially

higher heating values [7,31,110,111].

For pyrolysis processes, no gasification agents are used, but next to the particle size,

the inert gas flow has a big impact on the products and their yield. The inert gas

can, for example, minimize volatiles, and therefore avoid secondary reactions [109].

42 Chapter 3 Combustion of gaseous fuels

3.5.2 Enhancement with oxygen

When generating product gas via gasification, based on the gasifying agent, pro-

duced fuels can include combustibles like CH4, CO, and H2, but can also contain up

to 60 % or more of N2(or up to 30 % of CO2), which does not combust. In these

kinds of fuels, pollutant emissions can be noticeably decreased; however BDG bring

about other challenges. The burning velocity and heat release are reduced by the

large volume of inert gases, which finally leads to extinction or, in the worst case,

lead to explosions due to H2or CO induced by stability issues. To improve and

stabilize the processes when burning BDG, oxygen enhanced combustion is a bene-

ficial technique. Enriched air as the oxidizer increases flammability limits and also

the flame propagation velocity. This is due to the hydrogen in the syngas mixture

and the increasing reactivity of the fuel, based on the increased share of oxygen in

the combustion atmosphere. Furthermore, products of incomplete combustion and

greenhouse gases like CO2can be additionally reduced [8,112].

Also, the oxygen-enrichment of fuels can prompt the complete combustion of hydro-

carbons in practical systems and reduce pollutants like soot or particulates. Stable

components can be enhanced by small concentrations of oxygen to improve the pre-

viously mentioned difficulties. If used in combustion engines, an advantage of oxygen

addition is the possibility to utilize fuels of inferior qualities [113,114].

Chapter 4

Spectroscopic techniques

Spectroscopic techniques are based on the interaction of light with matter, such as,

i.e., ionization, absorption, and scattering processes [18]. In the following chapter

the emission of laser-induced fluorescence of formaldehyde and CH* chemilumines-

cence will be presented and investigated. Finally, the elastic Rayleigh scattering,

to determine temperature fields in combustion, is also introduced. These are the

three main spectroscopic techniques that were applied during this investigation of

synthetic BDG from thermochemical conversion processes.

4.1 Laser-induced fluorescence

Atoms or molecules from a lower energy level can be elevated to an excited state

of higher energy by chemical reactions, absorption of electromagnetic radiation, or

electron impact. Upon spontaneous transition from the excited state to the lower

energetic state, energy can be released, amongst others, in the form of luminescent

radiation, most commonly experienced as fluorescence, phosphorescence, and also

chemiluminescence [15,115,116].

As the excitation in both processes is induced by the absorption of photons, fluores-

cence and phosporescence can be summarized more generally as photoluminescence.

These techniques differ basically in their involvement of a change in electron spin

during the electronic energy transition [15]. Fluorescence is defined by a same spin

multiplicity of the emitting state and the final state, commonly singlet states, [117]

making them short-lived (<< 10−5s) [15]. A change in the electronic spin, hence

an emission from a triplet excited state to a singlet ground state, leads to phos-

phorescence. Here, the lifetimes of the excited states can take up to seconds or

minutes [15,117].

44 Chapter 4 Spectroscopic techniques

The physicist Alexander Jablonski suggested a three-energy-level diagram to de-

scribe the process of luminescence, the more complex version established over the

past decades including detailed mechanisms regarding absorption, fluorescence, and

phosphorescence, i.e., is commonly known as the Jablonski diagram [15,118]. Figure

4.1 illustrates the excitation and emission photophysics that can be induced by a

laser [117], elucidating the intra- and intermolecular electronic transitions [119].

Figure 4.1: Jablonski diagram [15], the basic version is commonly known from [16], reprinted

with permission

The initiating step when applying laser-induced fluorescence is the absorption of

light by the analyzed molecule. The energy E of a photon equals to the difference

between the lower and the upper energy levels [120]. The absorption spectra for

polyatomic molecules are far more complex than for those of isolated atoms due to

the various compositions of the rotational, vibrational, and electronic energies [15].

The energy E of a photon can hence be described as follows [15]:

E=Eelectronic +Evibrational +Erotational (4.1)

Eelectronic represents the energy states of the several bonding electrons (S0, S1, and

S2in Figure 4.1) of a specific molecule. Evibrational describes the total energy of

interatomic vibrations that take place in molecules. Finally, Erotational describes

4.1 Laser-induced fluorescence 45

the energy of the rotational states in a molecule. Typically a molecule has many

vibrational and rotational energy levels; therefore each electronic energy state can be

made of i = 1,2,3....n vibrational, and again each vibrational state of j = 1,2,3....m

rotational states, leading to a magnitude of solutions for multiatomic molecules [15].

Figure 4.1 is a graphical representation of a small amount of electrons and their

vibrational states. Potential absorption frequencies from the ground state S0to the

first electronic state S1can be described as follows [15]:

νi=1

(S1−e0

i−S0)(4.2)

where νistands for the absorption frequency, hPPlanck’s constant, and e’irepresents

the vibrational levels potentially ranging from i = 1,2,3....n. The energy differences

between vibrational levels typically differ by a factor of 10-100. As each vibrational

level includes multiple rotational levels, their differences in energy levels are much

smaller [15].

The type of excitation is based on the wavelength of the induced light and, therefore,

energy: electronic absorption is caused by ultra-violet or visible light, infrared light

excites the vibrational levels, and microwaves, on the other hand, the rotational

levels [120].

Figure 4.2: Electromagnetic spectrum, defining the regions of electronic, vibrational and rotational

absorption after light is induced [17], reprinted with permission

46 Chapter 4 Spectroscopic techniques

Due to Formulas 4.1 and 4.2, discrete values of energy levels and frequencies can be

determined. For the absorption spectra of atoms, this leads to discrete lines in the

electromagnetic spectrum presented in Figure 4.2. Molecular absorption spectra, on

the contrary, usually encompass a series of closely fit lines, which are only separated

by fractions of nanometers, that finally create an absorption band. This is based on

multiple electronic transitions and moreover on their numerous vibrational and also

rotational states, ultimately leading to an absorption region [15].

Upon being excited to higher energy levels via absorption of electromagnetic radi-

ation, atoms or molecules decay rapidly to a lower energetic level via several pro-

cesses, like nonradiative relaxation or luminescence. The principles of fluorescence

and phosphorescence can be explained by the blue arrows in Figure 4.1.

The fluorescence process begins after the absorption with a decay from the excited,

energetic State S1,2,3...x to the lowest vibrational level of S1, via internal conversion

and vibrational relaxation. An internal conversion can be understood as a transition

without radiation between levels of the same spin. Here, the transition of a higher

energetic level to a lower energetic level within a high vibrational state occurs when

both energetic states S1,2,3...x and S1have the same energy. Finally, fluorescence is

emitted when a molecule relaxes from the lowest vibrational level of S1to the ground

level S0and emits a photon [17,121].

Figure 4.3: Simultaneous electronic and vibrational change of energy levels with semi-stationary

nuclei during the absorption or emission of a photon, which can be explained by the Franck-Condon

principle [18], reprinted with permission

Electrons and nuclei in molecules show considerable discrepancies when regarding

their masses. On that account, electron clouds instantly adapt to changes of nuclear

structures during vibrations, which can be described by the Born-Oppenheimer ap-

proximation. On the basis of this semi-stationary nuclear configuration, vibronic

transitions, hence the concurrent change in electronic and vibrational energy lev-

4.1 Laser-induced fluorescence 47

els, can be explained with regard to Figure 4.3, using the Franck-Condon princi-

ple [18,122].

During the vibronic transition, the nuclear positioning remains unchanged in the

molecule, as the transition is much faster than the motion of the nuclei. The proceed-

ing vertical transition (upwards or downwards) occurs from the ground vibrational

state of the lower electronic level to the vibrational state in the excited electronic

state that coincides best energetically. This goes along with the Franck-Condon

principle of hardly changing the vibrational wavefunction and therefore retaining

the state of the nuclei. A wavefunction with a peak right above the one in the

ground state resembles each other most, leading to, i.e., 0 : 0 or 0 : 2 transitions as

pictured in Figure 4.3 [18,120,122].

Alternatively to fluorescence, intersystem crossing of a molecule from the excited

singlet state S1to an excited vibrational level of the excited triplet state T1can occur.

This is a radiationless transition between states of a different total electron spin,

typically favored when an atom with a high atomic number is present. The transition

from T1to the ground level S0while emitting a photon is called phosphorescence. In

this process, the lifetime is much longer in comparison to a fluorescence process due

to the change in quantum spin numbers. This can be based on the improbability

of the transition from two unpaired electrons (excited triplet state) to no unpaired

electrons (ground state), which is actually a forbidden transition.

The molecule structure, possible solvents, temperature fields, and pressure have

an influence on the rates of internal conversion, fluorescence, intersystem crossing,

and phosphorescence [17, 18, 120]. Multiple effects can influence the intensity of

emission during fluorescence or phosphorescence, also known as the quantum yield

or efficiency.

The quantum yield for any luminescence is simply the ratio of luminescing molecules

to those that were excited. This yield can lay between 0 and even reach an efficiency

close to unity for highly fluorescing molecules like fluorescein. The deactivation

processes of fluorescence and hence the extent of the fluorescence quantum yield

Φfare determined by the rate constants kx. Fluorescence kf, intersystem crossing

ki, external kec and internal kic conversion, predissociation kpd and dissociation kd

possibly lead to a deactivation of the lowest excited single state [15].

48 Chapter 4 Spectroscopic techniques

Φfcan be determined as follows, permiting a qualitative analysis of the influencing

factors of fluorescence [15]:

Φf=kf

kf+ki+kic +kec +kpd +kd

(4.3)

The chemical structure of molecules influences the dimension of kf, kpd, and kd, while

mainly the experimental environment dominates ki, kec, and kic. Laser-induced fluo-

rescence spectroscopy is a limited possibility of determining absolute concentrations

of the quantum yield due to the limiting effects of the previously mentioned quench-

ing processes [15,120].

Another possibility to decrease the fluorescence intensity is dynamic or collisional

quenching. This requires the contact between the excited molecule and a quenching

agent Q with a high concentration having a rate constant of kq[Q] [123].

In a quenching process with one quencher Φfcan be presented as [15]:

Φfq =kf

kf+ki+kic +kq[Q](4.4)

When regarding the ratio of the quantum yields without Φ0fq and with Φfq a

quencher, the Stern-Volmer equation is obtained, which describes the decrease in

intensity during dynamic quenching [15,123]:

Φ0

Φfq

= 1 + Kq[Q](4.5)

where [Q] illustrates the concentration of the quencher and the KqStern-Volmer

quenching constant defined as Kq= kq/ kf+ ki+ kic. Many molecules, for example,

oxygen, halogens, amines, i.e., can act as quenchers [15,123].

4.2 Rayleigh scattering

When electromagnetic radiation traverses matter, the greater fraction continues to

pass in the initial direction, while a small fraction is scattered in different directions.

The process can be separated into elastic Rayleigh scattering, induced by inho-

mogenities of the matter, elastic Mie scattering, and the inelastic Raman scattering,

due to interactions with vibrational and rotational levels in the matter [18,19,120].

In comparison to the previously discussed luminescence processes, there is no ex-

change of energy between the scattering molecules and the incident light photon,

4.2 Rayleigh scattering 49

leading to an elastic Rayleigh and Mie scattering. During both of these techniques,

the scattered light does not shift from its original frequency, hence wavelength.

During the inelastic Raman scattering (and also fluorescence), on the other hand,

photons induce a frequency-shift, as can be seen in Figure 4.4 [19].

Mie scattering occurs for molecules with larger diameters than the wavelength of

the incident light d > λ, [15], while Rayleigh scattering occurs in smaller molecules

with d < λ [124].

To advance the knowledge of combustion phenomena or complex flow fields of gases,

the use of laser-induced Rayleigh scattering measurements is a powerful diagnostic

tool. Difficulties can arise due to a collective scattering of multiple molecules, a

disturbance due to the inelastic scattering, background interference, or finally due

to anisotropy [125].

Figure 4.4: Principle of Rayleigh scattering, Raman scattering and fluorescence [19], reprinted

with permission

As a diagnostic tool, next to flow fields and flame structures, densities, and possibly

mixture fractions can be measured and, in case an ideal gas mixture can be presumed,

temperatures can be derived [125].

To deduce temperatures via the number density N, the total Rayleigh scattering

signal, which can be collected by a detector, is needed [34]:

S=ηINVRZ∆Ω δσ

δΩmix

δΩ(4.6)

with ηas the optical collection efficiency, I as the incident laser intensity, N the

number density of the scatterers, VRas the observation volume, ∆Ω as the solid

angle of the collection optics and (δσ/δΩ)mix as the differential scattering cross

section of the gas mixture [34].

50 Chapter 4 Spectroscopic techniques

The differential scattering cross section can be represented by the sum of all species

in the probe, weighed by mole fraction [34]:

δσ

δΩmix

n=0 δσi

δΩXi(4.7)

with Xias the mole fraction of the ith species.

The total Rayleigh cross section σhas to be depicted for a specific wavelength λand

can be presented for standard temperature and pressure (STP) as follows [34,125]:

σ(λ) = 24π3

λ4N2n(λ)2−1

n(λ)2+ 22

Fk(λ)(4.8)

with n(λ)as the refractive index and Fk(λ)as the King correction factor.

The King correction factor accounts for depolarization [126] and can be represented

for the depolarization ratio ρ0(λ)as follows [34]:

Fk(λ) = 6 + 3ρ0(λ)

6−7ρ0(λ)(4.9)

The depolarization rate ρ0(λ)has a finite value for asymmetric scatterers, but holds

the value zero for spherically symmetric molecules and represents the ratio of hor-

izontally to vertically polarized light ρh(λ)/ρv(λ)[34]. (The theory regarding this

topic will follow at the end of this chapter.)

With a(λ)defined as the mean molecular volume polarizability and γ(λ)the mean

molecular volume anisotropy, ρ0(λ)can be presented as follows [34]:

ρ0(λ) = 6γ2(λ)

45a2(λ) + 7γ2(λ)(4.10)

When finally introducing the molar refractivity RLas [34]:

RL=n2(λ)−1

n(λ) + 2

N(4.11)

with NAas the Avogadro constant.

4.2 Rayleigh scattering 51

With the inclusion of 4.8, 4.9, 4.10, and 4.11, and the assumption that the incident

light is polarized linearly, the solid angle is of a small finite value and the scattered

radiation is viewed perpendicular to the polarization vector, [34] the differential cross

section can be presented as [34]:

δσ(λ)

δΩ=9π2R2

L(λ)

λ4N2

6−7ρ0(λ)(4.12)

Table 4.1: Depolarization ratio and differential scattering cross sections of major BDG gases and

air [34,127,128] and own calculations

Species RL(λ)

(cm3mol−1)

ρ0(λ)δσ(λ)/δΩ

(E-28 cm2)

References (values

for calculation)

N24.484 0.02019402 6.22795 [127]

O24.065 0.05447471 5.24460 [127]

CO26.690 0.07525743 15.00290 [127] [128]

CO 5.034 0.01070242 7.84689 [127] [128]

CH46.630 0.00000000 13.43060 [127]

H22.086 0.01783944 1.34567 [127]

Air 4.393 0.02843004 6.10360 [127] [128]

For the major combustion species in BDG and air, the depolarization ratio, the molar

refractivity, and the calculated differential cross sections at STP for λ= 532 nm are

tabulated in Table 4.1.

Due to the temperature dependency of the mean molecular volume polarizability

a(λ)and the mean molecular volume anisotropy γ(λ), the differential Rayleigh cross

sections for gases can undergo a rise of 0.4−2.8 % per 1000 K for a λ= 532 nm [34].

The determination of temperatures based on Rayleigh scattering also depends on

the ideal gas law [129]:

pV =nRT or pV =NkT (4.13)

the development of heat also influences the number density of the fuels. An increase

of T from STP to about Tmax ∼2100 K (factor of about 7) in experimental combus-

52 Chapter 4 Spectroscopic techniques

tion of hydrocarbon flames as performed, deduces the Rayleigh scattering intensity

to about 1/7th.

As previously mentioned, the depolarization can be defined as the ratio of horizon-

tally to vertically polarized light or as the ratio of perpendicular to parallel light to

the original beam when the laser source is polarized, as shown in Figure 4.5 in the

yz plane [15].

Figure 4.5: Principle of polarization [15], reprinted with permission

Horizontally (perpendicular) and vertically (parallel) polarized experimental images

can be obtained by inserting a polarization filter between the sample (flame in be-

tween burners in the case of this work) and the ICCD camera and operating this

filter at a 90°or 0°angle, respectively. The light source should previously be con-

firmed to be vertically polarized as described in [130]. Details in regard to these

topics will be given in Section 5.1.

The Rayleigh scattering cross sections are generally a factor of 1000 larger than

the corresponding Raman cross sections, commonly leading to higher spatial and

temporal resolutions in Rayleigh measurements. However, due to a missing shift

in frequency after the light has been scattered by the molecule, experiments are

much more affected by the scattering processes of the background [124]. Finally,

non-premixed combustion makes a determination of temperature fields more com-

plicated, due to the incomplete mixing and also essentially due to the generation

of new species. A numerical model can anticipate this and calculate accurate cross

sections, to compare numerically and experimentally derived results. [124].

4.3 Chemiluminescence 53

4.3 Chemiluminescence

The phenomenon of chemiluminescence has already been mentioned as a variation

of luminescence in Section 4.1. An emission of light, as the relaxation of an ex-

cited molecule to its ground state, occurs. Originally identified and thought of as a

„curiosity“ in research laboratories, this technique is now considered a simple and

cheap but sensitive and selective measurement method. The types of luminescence

can be differentiated based on their source of energy, like electromagnetic radia-

tion leading to fluorescence or phosphorescence, heat leading to pyroluminescence,

frictional forces leading to triboluminescence, electron impact leading to cathodo-

luminescence, crystallization leading to crystalloluminescence, and finally, chemical

reactions leading to chemiluminescence [131,132].

Figure 4.6: CH*, OH* and C2* spectra of a methane-oxygen-hydrogen flame [20], reprinted with

permission

The process of chemiluminescence in chemical reactions is limited to a small number

of reactions only, therefore limiting the produced species. Most common in hydro-

carbon-air flames are CH*, OH*, C2*; their emission spectra are presented in Figure

4.6, and CO2* in Figure 4.7 [15,133].

The most basic type of reaction, where C* represents the excited state of species

C, hPPlanck’s constant, and νistands for the absorption frequency, is presented as

follows [15]:

A + B −−→ C∗+ D (4.14)

C∗ −−→ C + νih(4.15)

54 Chapter 4 Spectroscopic techniques

In combustion diagnostics for hydrocarbon flames, the technique of CH* chemilu-

minescence is frequently utilized to measure the heat release rate, to estimate the

equivalence ratio, and to use the radical as a marker for the flame location. CH,

as a short-lived molecule, prevails in a relatively narrow temperature and spatial

area. Therefore, detailed chemical kinetic mechanisms of pollutant formation like

NOxcan be advanced. The chemically excited CH* is responsible for the blue light

in low-sooting flames; the essential CH* emission occurs at 314 nm, 390 nm, and

431 nm [133–136]. About 80 % of the excited CH emission occurs at 431 nm, as

opposed to the residual excitation at 314 nm and 390 nm [137]. Lastly, all these

characteristics make an excellent basis to validate the numerical solutions with ex-

perimentally derived results.

Previous research [138–143] has focused on the following three major reaction path-

ways for CH* chemiluminescence [133,144]:

C2+ OH −−→ CH ∗+ CO(R1) (4.16)

C2H + O −−→ CH ∗+ CO(R2) (4.17)

C2H + O2−−→ CH ∗+ CO2(R3) (4.18)

It must be stated, though, CH* prediction and reaction pathways are still under

investigation and debate. This is mainly based on the uncertainty that remains

within the C2and C2H2chemistry [20].

With spontaneous emission of photons or destruction by collisional quenching fol-

lowing, as mentioned above. These pathways have been proposed [135,145]:

CH ∗ −−→ CH + νihP(4.19)

CH ∗+ N2←−→ CH + N2(4.20)

CH ∗+ O2←−→ CH + O2(4.21)

CH ∗+ H2O←−→ CH + H2O(4.22)

CH ∗+ H2←−→ CH + H2(4.23)

CH ∗+ CO2←−→ CH + CO2(4.24)

CH ∗+ CO ←−→ CH + CO (4.25)

CH ∗+ CH4←−→ CH + CH4(4.26)

4.3 Chemiluminescence 55

Furthermore, the CH* chemiluminescence signal Sem imaged onto a detector be

expressed as follows [146]:

Sem =1

4πA21τVemN∗Kem (4.27)

with A21 as the Einstein A (= 1.86 x1061/s) coefficient, τas the detectors exposure

time, Vem as the volume of the pixel, N* as the number density of the excited

species and Kem as the constant for the detector efficiency, and the solid collection

angle [146].

Electronically excited species generally have a low concentration in flames because

they are hardly produced, and when they are formed, they are quickly removed by

quenching processes. A limitation by the formation rate is therefore presumed [133].

If background radiation like soot is present in (diffusion) flames, CH* chemilumi-

nescence is weakened as a diagnostic tool. The incandescent soot particles can com-

promise luminescence measurements, even when using narrowband filters around

314 nm, 390 nm, or 431 nm, by inducing unwanted noise [144,147].

This limited the studies when using this technique to investigate non-premixed dif-

fusion flames due to the soot production in basic hydrocarbon-air flames. However,

non-sooting diffusion flames, especially with hydrogen-mixed fuels, oxygenated fuels,

or diluted fuels, can resolve this problem [147].

Figure 4.7: CH*, OH* with and without CO2* spectra in a hydrocarbon flame [21], reprinted

with permission

CH* chemiluminescence measurements can furthermore be influenced by „back-

ground“ radiation of CO2* as pictured in Figure 4.7. Mainly between 300 −600 nm

CO2* chemiluminescence is present as a broadband emission in hydrocarbon flames,

impacting the low concentration of CH* chemiluminescence peaks [21,148].

56 Chapter 4 Spectroscopic techniques

4.4 Techniques

During the investigation of combustion processes, formaldehyde plays a central role,

as it takes part in the oxidation of hydrocarbons, both in high and low-tempera-

ture regions [149]. Formaldehyde, being an essential intermediate species, makes

it the subject of many studies, detecting, i.e., the flame front. Formed as part of

the oxidation of hydrocarbon fuels, it is consumed during the ongoing combustion

chemistry [150,151].

Especially the excitation of this molecule via LIF or planar laser-induced fluorescence

(PLIF), in many conditions and surroundings, has sparked an interest in many

research groups. Some selected topics will be presented shortly in the following

section.

The electronic transition of the 401(A1A2←X1A2) vibronic band of excited

formaldehyde has been studied at elevated temperatures and pressures to inves-

tigate the sensitivity of the absorption spectra [152] and in decay rate investigations

in a gas cell [153]. Furthermore, has the saturation behavior of the formyl radi-

cal been considered for flame front imaging in hydrocarbons regarding methane and

dimethyl ether in the B-X band [154]. Laser-induced fluorescence was also utilized to

advance the topic of scalar dissipation. By determining flame markers that enable

estimates of the scalar dissipation rate at the stoichiometric surface of non-pre-

mixed flames [155], by the analysis of temperature and quenching corrections on the

thickness of the layers in diffusion flames [26], and by planer LIF in a counter-flow

diffusion flame with a stationary local extinction point [156]. Finally, formaldehyde

PLIF in a Nd:YAG laser system was utilized to investigate and understand the origin

of spectrally narrowband vs. broadband excitation in premixed and non-premixed

flames [157].

More profound knowledge regarding quenching processes via excitation of formalde-

hyde at λ= 355 nm were collected in a rich methane-oxygen flame for fluorescence

lifetime imaging to be utilized for quenching corrections [158] and during studies of

a dimethyl ether (DME)-air counter-flow diffusion flame to receive a representative

formaldehyde profile, also taking signal quenching and the Boltzmann distribution

into account [159]. Measurements at λ= 355 nm, utilizing an Nd:YAG laser system,

were also used to identify oxygenated species of combustion in the reaction zones.

A further understanding of the correlation between the presence and evolution of

these species depending on the operating conditions in an ethylene-air flame of a

McKenna burner was gained [160]. Lastly, also MILD combustion was monitored

4.4 Techniques 57

via formaldehyde LIF to examine the entrainment of air effecting the flame structure

when reducing the oxygen level [161]. The correlation of processes in cool flames to

understand chemistry and transport phenomena in engines was also advanced using

PLIF of formaldehyde. Partially premixed DME counter-flow cool flames were in-

vestigated after an addition of ozone [162] and subsequently also compared to hot

flames [163], to show the increased profile and indifference of the cool flame speed

to the equivalence ratio. Also, the development of formaldehyde in the pre-heat

zones of methane-air premixed flames after adding ozone to the combustion were

experimentally and numerically regarded [164].

Laser absorption diagnostics to sense formaldehyde in shock tube kinetic stud-

ies [165] or observe the thermal decomposition of formaldehyde behind a shock

wave [166], to investigate and update mechanisms were performed.

The understanding of the formaldehyde molecule during combustion was also fre-

quently explored directly in engine systems. The flame propagation in a spark-ig-

nition engine was studied via LIF of formaldehyde. Fluorescence images made the

tracking of the flame front possible due to the development of advanced program

routines, hence deriving velocity measurements [167]. Diverse other processes were

utilized to detect formaldehyde in other internal combustion engines (or exemplary

systems), like obtaining approximations of a saturation intensity and possible exis-

tence of photochemical effects [150], detecting background fluorescence of PAH or

other larger hydrocarbons simultaneously in a laminar counter-flow diffusion flame

and in a single stroke spark-ignition engine [168] or investigating the gasdynamical

and chemical processes that happen during the auto-ignition of endgas of an Otto

engine [169]. Oxygenated species were identified in a pressurized flow combustor,

based on a commercial diesel system [170], and low-temperature reactions during

fuel composition or forming of PAH in high-temperature regions using a high-pres-

sure cell [171], to advance the combustion processes in direct injection diesel engines,

were considered.

LIF formaldehyde studies in an HCCI engine were frequently executed. Iso-octane

and n-heptane were utilized to study the combustion process with formaldehyde as

a marker for the first stage of ignition [172], for the spatial signal distribution at

various injection timings [173], or to potentially use formaldehyde as fuel tracer in

low-temperature reactions [174]. Further, to extract the formaldehyde contribution

to the fluorescence spectra [175] during the ignition of diesel and n-heptane fuel

sprays in an optically-accessible diesel engine for studies to be utilized in an HCCI

58 Chapter 4 Spectroscopic techniques

engine. Finally, formaldehyde PLIF was applied to study DME cool flames experi-

mentally and numerically, playing a role in diesel engines, jet engines, spark-ignition

engines, and HCCI engines, in a CARAT burner [176] or to investigate the low- and

high-temperatured auto-ignition in high-pressure spray flames [177].

Furthermore, LIF techniques are extensively deployed to inspect BDG, and BDG

derived from thermochemical conversion processes as, i.e., in [178–181].

Experimental and numerical investigations of heat release (maximum position, pro-

file width, i.e.) in counter-flow flames [182,183] and in a bunsen configuration [184]

have been reported.

Though complexity is introduced by inelastic and elastic scattering induced by nu-

merous molecules and anisotropy phenomena, laser-induced Rayleigh scattering is

widely used to investigate complex flow fields and combustion [34,124,125].

Flow field and also flame structure, density, and mixture fraction [185,186] can be

determined by the intensity of Rayleigh scattering, furthermore temperature dis-

tributions [187]. To investigate complex flows [188] or the atmosphere [189], the

spectrum of Rayleigh scattering needs to be observed.

Rayleigh scattering at 532 nm can be applied in rapid-compression machines [190],

laminar and turbulent jet flames [191,192], premixed flames [193,194], for quantitaive

vapor-fuel imaging [195] and comparable applications.

Furthermore, this technology can also be applied in lower wavelengths like the deep

UV [196,197].

To decouple Rayleigh scattering signals from background scattering atomic or molec-

ular filters can be used, as presented from multiple research groups in [115]. This was

applied with the use of iodine vapor or molecular filters [198–200], also in flames.

Furthermore, atomic mercury filters can be used for ultraviolet filtered Rayleigh

scattering measurements [201,202].

To utilize Rayleigh scattering techniques in combustion or other applications, scat-

tering cross sections [203], depolization ratios [128,130], refractivities [127], polariz-

abilities [204] and anisotropy [205,206] were defined.

Finally, the applications of CH chemiluminescense as a diagnostic tool in combustion

are presented, while applications in other fields have been reported elsewhere [132].

Numerous research groups have evaluated CH chemiluminescence experimen-

tally [20, 137, 144] and numerically [133, 134, 145] in low-pressure, non-premixed,

premixed, i.e., flames.

4.4 Techniques 59

Furthermore, the behavior during CH chemiluminescence in axisymmetric lami-

nar [135], coflow diffusion [146], and counter-flow diffusion [136, 147] combustion

was studied.

No research is known, where the presented techniques of laser-induced fluorescence

of formaldehyde, laser-induced Rayleigh scattering, and CH chemiluminescence were

investigated during counter-flow laminar diffusion combustion of the multiple syn-

thetic biomass-based model fuels presented.

Chapter 5

Experimental set-up and conditions

As it is commonly used to induce stable, laminar diffusion flames, a counter-flow

set-up was chosen for the following investigations. Preliminary experimental stud-

ies with two opposed McKenna flat flame burners [22], also used in a counter-flow

set-up, have led to the need for a newly designed burner system as the centerpiece

of the test rig. The spectroscopic measurement system is based around a pulsed

Nd:YAG laser in combination with various optical and analytical elements and de-

vices. Subsequently, extensive data processing procedures need to be applied. The

application of experimental strategies and conclusions are presented in the following

chapter.

5.1 Experimental set-up

An introduction of the considerations and developments of the in-house counter-flow

burner system, as well as the Nd:YAG laser, the surrounding spectroscopic tools,

and analytical devices, will be given in the following sections.

5.1.1 Counter-flow burner

The different types of counter-flow burners were already presented in Chapter 3.2.

An opposed flow-induced laminar diffusion flame is initially achieved using two com-

mercial McKenna burners of the American company Holthius & Associates (Fig-

ure 5.1, showing one burner).

62 Chapter 5 Experimental set-up and conditions

Figure 5.1: Schematic sectional view of a McKenna flat flame burner [23]

Via a screwed connection at the bottom of the construction, a gaseous medium can

flow into the inner chamber of the burner and then stream through the porous burner

plate. This plate is made of stainless steel and is, according to the manufacturering

specifications, of a porosity of 54 %, with a pore size of 70 −120 µm [22]. Inside the

plate is a cooling water coil, which is attached by two connectors at the lower end

of the burner to a cold water circuit.

The inner chamber is surrounded by a shroud, which is also connected by a screw

connection to the bottom of the burner. The outer chamber can be used to further

insert (other) gaseous fluids into a separate area of the burner. These leave the

construction over an annular, porous disc, which encloses the burner plate.

During the preliminary testing, the use of these commercial burners led to three

different problems and, therefore, to the construction of an individual burner system

for the experimental investigations.

First, for a laminar diffusion flame, these two opposed burners need to be in perfect

opposite alignment. When having a slight misalignment of the nozzles, the flow

of one nozzle bends to one side, whereas the opposed flow bends to the opposite

side. Precision machining is necessary, so alignment is near perfect. Second, these

burners, in an opposed flow set-up, are not constructed for high enough flow rates

and, therefore, gas velocities. They are not long enough axially, so the flow does

not have enough time to fully develop and smooth out radially before leaving the

nozzle [13]. For the velocities and, therefore, strain rates that were to be investigated,

a stable flame could not be established with this construction.

5.1 Experimental set-up 63

Figure 5.2: Diffusion flames at L/D = 0.36 (left) and L/D = 0.83 (right) [23] induced by two

opposed McKenna flat flame burners

Finally, a distance between the burners of at least 18 mm needed to be reached while

sustaining a stable, laminar flame for the laser sheet to be guided through the two

opposed burners. The nozzle separation distance to nozzle diameter L/D needs to

be aimed at ratios close to 1 [207]. Below this, the flow profile is no longer flat,

making it difficult to interpret the achieved data without a flow quantification like

particle image velocimetry (PIV).

The diameter of a standard McKenna burner (inner ring) is 60.5 mm. An enlarge-

ment of the separation of the two opposed burners ratio of nozzle separation distance

to nozzle diameter (L/D) to 0.83 (Figure 5.2) during fundamental testing generated

a transient, irregularly shaped, and sooting flame. The physical appearance of these

flames indicated the cause of this being a matter of fluid dynamics, rather than

chemical reaction, based on a non-uniform gas distribution in the burners, which

induces irregular local strain rates [208].

The core components of the newly constructed stainless steel burner set-up are

two equivalent burners (Figure 5.3), with an extended axial length of 180 mm

in regard to the inner, main gas chamber. The main chamber (left (1)) is also

shrouded by a hollow body (left (2)) to induce an additional gaseous flow in

the form of a co-flow, as are the McKenna burners. Just above the nozzles of

the main gas chambers, a filling of a honeycomb monolith made of ceramics to

straighten out and distribute the incoming gas flow begins. The remainings of

the inner chambers are filled to capacity with steel wool for similar reasons.

64 Chapter 5 Experimental set-up and conditions

Figure 5.3: Scheme of (left) and actual constructed (right) counter-flow burner system. (1):

inner gas chamber top burner, (2): outer gas chamber top burner, (3): alignment system for top

and bottom burners, (4): stage to move bottom burner, (5): gas input top burner, (6): cooling

water input top burner.

With this design of the burner system, a L/D ratio of circa 0.8 was reached dur-

ing nearly all of the experimental investigations done, leading to laminar diffusion

flames. A system of movable and fixed poles (3) enabled an exact axial alignment

between the top and bottom burners. Furthermore, a motorized stage (4) made

an adjustment to varying L/D for differing experimental conditions possible and

precisely exact.

Finally, the top burner was equipped with a fresh water-cooling system (6) due to the

hotter surroundings at the top of the measuring cell, where the exhaust fumes gather.

A stainless steel cooling coil was wrapped around about 70 % of the top burner,

constantly being fed with water at ambient temperature. With these adjustments

in technical design and construction of the burner system, the difficulties in exact

alignment, top-nozzle to bottom-nozzle distance, and a developed flow of gaseous

fuels were overcome.

5.1.2 Nd:YAG laser

A laser (Light Amplification by Stimulated Emission of Radiation) is said to be

just a more complicated and unique flashlight, with energy in the form of electricity

entering and light exiting. That this is an understatement will be presented in the

5.1 Experimental set-up 65

following section [209]. Lasers emit coherent electromagnetic radiation at a defined

wavelength at a low divergence. Their wavelengths span from the electromagnetic

spectrum of infrared, visible, and ultraviolet up to the x-ray region; their emission

yields include continuous wave to extremely large energy or peak-power pulses, rang-

ing from very weak to terawatt power levels. This leads to a large variety of applica-

tions in scientific disciplines, material processing, and medical technology [210,211].

Some of the lasers that are utilized extensively are excimer lasers, liquid lasers like

dye lasers, gas lasers like He-Ne, argon ion, and CO2lasers, or solid-state lasers

like ruby, Nd:glass, and Nd:YAG lasers [211]. The latter technology will be further

considered due to the utilization of the spectroscopic techniques applied during this

study.

The first Nd:YAG laser was built and operated at Bell Labs in 1964 and has under-

gone tremendous development since then. This system’s flexibility and adaptability

make it a widely used tool in many research labs for both fundamental and techni-

cal investigations [210]. This makes them the favored of the solid-state lasers. An

yttrium aluminum garnet (short YAG) is the host medium; it is a Y3Al5O12 crystal,

where some of the Y3+ ions are substituted by Nd3+ ions [212].

Figure 5.4: Simplified four-level transition scheme of Nd:YAG energy levels based on [24]

Figure 5.4 shows the simplified energy level scheme for an Nd:YAG laser arising

from the three inner-shell 4f electrons of the Nd3+. The pumping from the ground

level (1) arrives at the two predominant absorption pump bands at (2), which lay at

∼730 nm and 800 nm. A nonradiative decay to the 4F3/2level (3), which therefore

assembles a large part of the pump power, makes this a potential candidate for the

upper level for laser action. Furthermore, the decay to the lower I-levels (4) takes

place [212]. The laser action typically occurs from the R2level of 4F3/2to a specific

66 Chapter 5 Experimental set-up and conditions

sublevel at 4I11/2with a transition of λ= 1.064 µm, making this the most frequently

used wavelength for Nd:YAG lasers at λ= 1064 nm [212].

Frequency conversion by nonlinear crystals of the fundamental λ= 1064 nm wave-

length (infrared) can produce other wavelengths to be emitted by the Nd:YAG laser.

Frequency doubling leads to the second-harmonic generation (SHG) at λ= 532 nm

(visible light) and those results can be doubled once more to create the fourth-

harmonic generation (FHG) at λ= 266 nm (ultraviolet). These wavelengths can

also be mixed with the fundamental λ= 1064 nm, leading to the third-harmonic

generation (THG) at λ= 355 nm (ultraviolet). Further specifications of the Lab-150

Nd:YAG that was used during these investigations can be found in Figure A.6 in

appendix A.

5.1.3 Spectroscopic set-up

The central part of the spectroscopic set-up is the pulsed Lab 150-10 Hertz Nd:YAG

laser from Spectra Physics (USA), as presented in Section 5.1.2.

Figure 5.5: Scheme of spectroscopic set-up. (1): Nd:YAG laser head, (2): periscope, (3):

periscope, (4): beam-shaping lenses, (5): counter-flow burner, (6): ICCD camera, (7): apertures,

(8): beam dump, (9): power unit Nd:YAG laser, (10): cooling water, (11): mass flow controllers,

(12): PTU [23]

5.1 Experimental set-up 67

These lasers are based on an oscillator-only system also featuring a pump cham-

ber and dual rod configuration, ensuring an excellent output beam quality. At the

wavelength of 1064 nm, the maximum output power that can be reached is 6.5 W,

decreasing with shorter wavelengths, as presented in Figure A.6. As seen in Fig-

ure 5.5, as a result of its size, the laser head of the Lab 150 Nd:YAG (1) had to be

placed on one of the longer sides of the measuring table.

The power unit (including the water cooling system) (9) was placed under the ex-

perimental set-up. The communication between the laser and additional elements

of the spectroscopic systems is done via a programmable timing unit (PTU) (12)

and will be further discussed in Section 5.3.

To investigate these counter-flow flames, a laser sheet was formed. Therefore, nu-

merous optical lenses needed to be installed in a fairly outstretched manner to

ultimately have their focal point precisely placed. To ensure a spatial alignment of

the laser beam onto the alternate side of the measuring table, a double periscope

system ((2), (3)) was installed. Periscopes are optical instruments used to direct the

laser beam on a higher or lower level. Crucial herewith was the use of two periscope

systems, including four mirrors. This ensured the initial vertical polarization of

the emitted light by the laser at the adjusted configuration of the crystals at the

measuring point. Mirrors with a degree of reflection of 99 % for the wavelength

range 350 −700 nm were used.

Figure 5.6: Schematic representation of the arrangement of the optical lenses used for beam

widening based on [23]

68 Chapter 5 Experimental set-up and conditions

The laser beam, now running downstream the experimental table, underwent an

expansion into a laser sheet as outlined in Figure 5.6. A system of plano-convex and

plano-concave lenses was used to first focus the laser beam on the centerline of the

diffusion flame and second, widen it into a sheet.

Following the second periscope, the laser beam enters a round, plano-convex lens.

The curved side of the optical element points in the direction of the laser source

in order to minimize spherical aberrations [213]. The distance of this first lens

corresponds to the focal point of the measurement and, therefore, to the focal length,

in this case being 1000 mm. The beam is bundled point-shaped at the target. By

transforming the beam into an elongated strip, the laser energy can be introduced

into a larger area.

The beam is first widened with the aid of a cylindrical plano-concave optic, followed

by a cylindrical converging lens which directs the beam path back to its original

shape. Whereas the diameter of the beam going in the vertical direction is being

considerably enlarged. The distance between the cylindrical lenses is determined by

the focal lengths of these optical elements and their common virtual focal point; the

following applies [23]:

ldistance =f1− |f2|(5.1)

where f1= focal length of the converging lens and f2= focal length of the diverging

lens. For the given experimental set-up, a plano-convex cylindrical lens with a focal

length of f1= 500 mm, as well as a plano-concave optic with f2= -150 mm were placed

so that the distance between the two elements corresponding to equation 5.1.3 adds

up to 350 mm. The laser sheet furthermore follows a Gaussian distribution profile,

as does the laser beam. In addition to these main components, necessary elements

like apertures (7) and beam dump (8) complete the spectroscopic set-up. Diffuse

light reflections due to metal surfaces in the laser lights pathway lead to aberrations

in the measured signals. This difficulty can be decreased by adding apertures and

blocking the surplus, unwanted and disorderly light downstream of the lens system.

Every experimental investigation using spectroscopic tools is in need of beam dumps

to capture the laser light in a closed system and therefore minimize the risk of injury.

Further details about the optical elements that were used during this investigation

can be found in Appendix A.

5.1 Experimental set-up 69

5.1.4 Analytical components

The selection of a suitable camera for a functional and flexible laser diagnostic

system must take all intended uses of the test bench into account. The differing

spectroscopic methods mentioned in Chapter 4 have high and specific demands in

regard to the camera:

•The phenomenon of fluorescence is extremely short-lived and often moves in

the time range of several millionths to billionths of a second after the exci-

tation of the species that is to be investigated [214]. Therefore, the camera

that records the fluorescence reaction must have very short and highly precise

reaction times and needs to be gated.

•The Rayleigh scattering signals and some of the fluorescence of specific fuel

mixtures are extremely weak and accordingly difficult to detect. Therefore,

the camera must exhibit a high sensitivity and is intensified.

•It must be possible to trigger the camera with the laser since the time inter-

vals between the emission of the laser light and the moment of the camera

recording must be defined clearly. This is the only way to ensure a reliable

and representative measurement.

All these requirements are met by the LaVision (Germany) Nanostar intensified

charge-coupled device ICCD camera. The basic structure and function of such a

camera are briefly described with the schematic representation in Figure 5.7.

Figure 5.7: Schematic representation of the technical operation of an ICCD camera [23]

70 Chapter 5 Experimental set-up and conditions

The captured photons reach the input window of the camera and then the photocath-

ode, generating electrons. With the help of a subsequently attached microchannel

plate (MCP), the number of electrons increases drastically, with the extent of the

reinforcement, based on the type of construction of the plate and the applied voltage.

The Nanostar camera reaches amplifications of up to 1000 [25] and can theoretically

achieve a reinforcement of 106when adding further MCP [215]. The increased num-

ber of electrons step onto a layer of phosphorus, creating new photons, which are

eventually registered by the charge-coupled device (CCD) sensor. This sensor pri-

marily acts as a light-sensitive photodiode. The photons now meet sensitive bitmaps,

are absorbed, and generate charge carriers, which are integrated into pixels.

This is followed by the readout phase, in which the charges of the pixels are shifted

into buffer memory and are then converted into a voltage with adjustable signal

amplification so that the image information can be read out [215]. The initially

analog signal is modified via an analog-to-digital converter (ADC) and sent to the

PCI interface board. In the last step, the digital data is stored as an image in

the buffer of the camera software and thus made accessible to the user of the test

rig. The Nanostar ICCD camera is controlled by the software DaVis, which enables

the reading, analysis, and processing of data and also allows external control over

the laser and energy monitors. This makes it possible to precisely define the time

window between a laser light pulse and the collection of photons on the CCD sensor.

Altogether, a number of parameters have to be determined in order to ensure an

optimal evaluation of the differing spectroscopic investigations that are carried out,

as seen in Table 5.1. In addition to these parameters, the DaVis program also offers

numerous possibilities for image postprocessing and evaluation, which will be looked

into further in Chapter 5.3.

5.1 Experimental set-up 71

Table 5.1: Technical parameters for experimental measurements with an ICCD camera [23]

Parameter Purpose

CCD Exposure Time Period of time in which the CCD chip is capable

of converting photons into digital signals and thus

into images.

Area of Interest Size of the examination area and thus definition of

the sensitive area of the CCD chip.

Delay Delay until the opening of the amplifying part of

the camera, corresponding to the time between the

activation of the Q-switch of the laser and the

beginning of the collection of photons in the

examination area by the camera.

Gate Period of time during which the amplifying part of

the camera converts photons into electrons and

amplifies them.

Gain Adjustment of the voltage strength at the

amplification part of the camera and thus

determination of the level of amplification of the

electrons.

Burst Count Number of integration and readout steps during

the CCD exposure time.

As pictured in Figures 5.5 and 5.8, the ICCD camera is installed behind the

counter-flow burner but still attached to the measurement table. With a system

of rails, the exact placement of the camera in regard to the focal point can be en-

sured. In this case, the laser sheet, which proceeds parallel to the camera’s lens,

cuts through the counter-flow disc-shaped flame. Due to the release of significant

heat in the area around the burners, caused by exhaust gases, the ICCD camera and

the diverging bandpass filters, etc., need to have a certain distance to the measuring

area. This was established by surrounding the counter-flow burner with a metal

enclosing, as confined as possible, and additionally by adding intermediate macro

rings between the camera and its lens. This distance changes the close-up limit, and

the potential magnification increases, making it possible to slide the camera on the

rail as far away from the exhaust heat as possible with the existing camera lens.

Significant when using an ICCD camera for differing spectroscopic measurement

techniques at different wavelengths are spectroscopic filters. Interference, absorp-

tion, or polarization filters, either as long-, short-, or bandpass, can transmit or

block a very specified and narrow range of wavelengths. The optics used during

these investigations can be found in the equipment list in Appendix A.

72 Chapter 5 Experimental set-up and conditions

As an additional component for the spectroscopic investigations, an energy monitor

also by LaVision (Germany) was installed. This instrument is capable of measuring

the relative energy of individual laser pulses simultaneously with the light detection.

A value for the relative energy is provided, therefore, for each image. With this

application, the accuracy of laser imaging increases based on an absolute quantifi-

cation [216]. The energy monitor can be operated with two measuring heads (4),

in this investigation being placed between the aperture (7) and before the burner

(5) and after the burner (5) (Figure 5.8). The laser sheet has to run through the

energy monitor heads completely in height. The results will be discussed further in

Chapter 5.3.

Figure 5.8: Top view scheme of spectroscopic set-up. (1): Nd:YAG laser head, (2): periscope,

(3a-c): beam-shaping lenses, (4): energy monitor, (5): counter-flow burner, (6): beam dump, (7):

aperture, (8): ICCD camera, (9): interference / absorption / polarization filters

To achieve a better overview and elaborate on the triggering process between the

parts of the spectroscopic set-up, a top view scheme is presented in Figure 5.8.

A laser beam with a specific wavelength and another beam with mixed (leftover)

wavelengths exit the head of the Nd:YAG laser; the positions depend on the arrange-

ment of the nonlinear crystals. In Figure 5.8, the actual emitted beam exits at the

top, while the residual radiation from other wavelengths goes straight into a beam

dump (6) at the bottom of the head. After exiting, the beam is directed towards the

flame with periscopes (2) and shaped into a sheet with the help of optical lenses, as

presented in Figure 5.6. (3a) illustrates the plano-convex lens, (3b) the cylindrical

plano-concave lens, and (3c) the cylindrical plano-convex lens. To minimize reflex-

ions, apertures (7) are added between the lenses. Before and after passing through

the flame in the counter-flow burner system, the laser sheet passes through energy

monitor heads (4). An ICCD camera (8) used as a diagnostic system and interfer-

ence / absorption / polarization filters (9) are also attached to the measurement

5.1 Experimental set-up 73

table. The attachment is crucial in case vibrations occur. The ICCD camera and

Nd:YAG laser are triggered, connected, and controlled by a PTU timing unit. The

Nd:YAG used emits bursts of light that are pulsed. After a specific delay, the ICCD

camera is internally triggered to take an image. The gate of the camera is opened for

a particular time, based on the specific spectroscopic set-up and experiment. The

technical parameters for the spectroscopic techniques investigated in this study are

presented in Table 5.2.

Table 5.2: Defining technical parameters of experimental fluorescence, Rayleigh scattering, and

chemiluminescence measurements

Experiment Exposure time

[µs]

Delay

[µs]

Gate

[µs]

Gain

[%]

Fluorescence 10.00 0.18 0.30 60.00

Rayleigh scattering 10.00 0.20 0.10 60.00

Chemiluminescence 10.00 / 0.10 30.00

A challenge to be conquered amid this experimental work was the extensive need

of partly dangerous gas compositions, especially when not combusted, and up to

nine simultaneous inputs per streamline via individual MFC. The detailed fuel

mixtures will be discussed further in the following Chapter 5.2.1, but in Figure 5.9,

the technical set-up in regard to the fuel blends is outlined.

Figure 5.9: Schematic representation of the burner set-up including gas inputs and mixing cham-

ber (MC), details of the gases used for the experimental work can be found in Appendix A

74 Chapter 5 Experimental set-up and conditions

Both shroud flows were established by the use of nitrogen, these consistently having

the same velocity as the main flows of fuel and oxidizer, to ensure a uniform flow field.

To establish a homogeneity among the fuels, first, a distance between the multiple

mass flow controllers and the burners had to be established, and second, a mixing

chamber (MC) had to be developed. Especially oxygen and hydrogen were partly

infused in small quantities; therefore a consistent distribution of the components in

the fuel mixture needed to be assured. The mixing chamber measures a length of

10 cm, a diameter of 4.3cm, and is filled with glass beads of 4mm diameter.

For safety reasons, multiple external non-return valves were installed before the

mixing chamber and before the final entry into the top burner. Every mass flow

controller includes an individual internal non-return valve for additional safety. Due

to the partial enrichment of the oxidizer air with varying amounts of oxygen during

some experimental investigations, an external non-return valve also had to be added

before the entrance into the main combustion chamber of the top burner.

All the mass flow controllers were operated by an additional computer unit using

the software get red-y by HTK Hamburg (Germany), which was delivered with the

MFC. Manual emergency closing valve systems were additionally installed after

each individual mass flow controller, as well as behind the mixing chamber. With

these fuel blends, the behavior regarding velocities, possible straining out, or dilution

effects needed to be absolutely familiar beforehand for various security reasons during

the experimental operations.

Not pictured for the sake of clarity are the two rotameters and further appliances and

chemicals needed for the flat field imagery when applying laser-induced fluorescence,

which were also part of the analytical components. For subsequent data processing,

all spectroscopic techniques require flat field correction imaging before and after the

combustion imagery. This procedure is standardized for the calibration of images,

and the objective is to eliminate artefacts from a picture that are caused by the

detector’s sensitivity or fixed-pattern noise [217].

5.2 Experimental conditions 75

5.2 Experimental conditions

The experimental design was carefully developed to account for possible compli-

cations during the main investigations. This is based on preliminary studies with

diluted methane mixtures at different velocities, with different compositions and

varying set-ups. The developed test rig, especially the counter-flow burner, and its

potential for unexpected heat development had to be thoroughly examined. Further-

more, spectroscopic techniques and their nuances, technical interaction, analytical

synergies, etc., had to be developed and be acquainted with before using explosive

and/or toxic fuel compositions. In the following chapters, the fuel mixtures and

experimental conditions will be presented.

5.2.1 Fuel mixtures - PG, GGL2, GGL3

Preliminary studies in preparation for the investigations of synthetic biomass-based

model fuels were done by using diluted methane mixtures as fuels and air as oxidizer

for the experimental work. As diluents, nitrogen and carbon dioxide were chosen,

both or at least one generally being a part of products resulting from biomass gasi-

fication or pyrolysis processes. An introduction to the use of methane as fuel and

the inert gases nitrogen and carbon dioxide as considerable components in BDG has

been given in Sections 3.4.1 and 3.4.2.

Especially due to the fact that the counter-flow burner system was an in-house con-

struction, the accuracy of the induced flow fields had to be verified. An uneven or

incorrect flow field would indicate incorrect input velocities and, therefore, flawed

straining-out behavior when comparing experimental and numerical solutions. The

specifications of all the gases used during these investigations can be found in Ap-

pendix A.

76 Chapter 5 Experimental set-up and conditions

The goal of verifying not only referenced solutions but also comparing experimental

solutions with their numerical counter-part led to the evolution of diluted methane

fuels, as seen in Table 5.3. The fuel was replaced gradually and in small steps

with either carbon dioxide or nitrogen, while finally leading to a displacement of

25 vol.-% of fuel to diluent when comparing CH4-001-X and CH4-007-X. The latter

could experimentally only be achieved with nitrogen as diluent at the investigated

velocity due to the straining out of methane flames with more than 68vol.-% of

carbon dioxide at a velocity of 60 cms-1 at the distinct burner configurations (like

nozzle separation, i.e.) regarded.

Table 5.3: Investigated diluted methane mixtures; X representing N2or CO2

Fuel Oxidizer

Name of methane

mixture

CH4

[vol.-%]

X(N2or CO2)

[vol.-%]

Air

[vol.-%]

CH4-001-X 50 50 100

CH4-002-X 40 60 100

CH4-003-X 37 63 100

CH4-004-X 34 66 100

CH4-005-X 31 69 100

CH4-006-X 28 72 100

CH4-007-X 25 75 100

5.2 Experimental conditions 77

The influence of oxygen enhancement of fuel or oxidizer on the combustion behavior

was already briefly discussed in Section 3.5.2. The impact on these diluted methane

mixtures was investigated by adding 5 vol.-% of oxygen to the fuel- and oxidizer sides,

with special regard to the formaldehyde evolution in the flame and temperature

distributions. These fuel and oxidizer compositions, based on Table 5.3, can be

found in Tables 5.4 and 5.5.

Table 5.4: Investigated diluted methane mixtures enriched with oxygen on the fuel side; X repre-

senting N2or CO2

Fuel Oxidizer

Name of methane

mixture

CH4

[vol.-%]

X(N2or CO2)

[vol.-%]

Air

[vol.-%]

CH4-001-X-O2-F 47.50 47.50 5.00 100.00

CH4-002-X-O2-F 38.00 57.00 5.00 100.00

CH4-003-X-O2-F 35.15 59.85 5.00 100.00

CH4-004-X-O2-F 32.30 62.70 5.00 100.00

CH4-005-X-O2-F 29.45 65.55 5.00 100.00

CH4-006-X-O2-F 26.60 68.40 5.00 100.00

CH4-007-X-O2-F 23.75 71.25 5.00 100.00

Table 5.5: Investigated diluted methane mixtures enriched with oxygen on the oxidizer side; X

representing N2or CO2

Fuel Oxidizer

Name of methane

mixture

CH4

[vol.-%]

X(N2or CO2)

[vol.-%]

CH4-001-X-O2-A 50.00 50.00 75.05 24.95

CH4-002-X-O2-A 40.00 60.00 75.05 24.95

CH4-003-X-O2-A 37.00 63.00 75.05 24.95

CH4-004-X-O2-A 34.00 66.00 75.05 24.95

CH4-005-X-O2-A 31.00 69.00 75.05 24.95

CH4-006-X-O2-A 28.00 72.00 75.05 24.95

CH4-007-X-O2-A 25.00 75.00 75.05 24.95

78 Chapter 5 Experimental set-up and conditions

Table 5.6: Investigated basic fuel mixtures from thermochemical conversion processes

Fuel Oxidizer

Name

fuel mix-

ture

CH4

[vol.-%]

CO2

[vol.-%]

LHV

[MJ/m3]

Air

[vol.-%]

GGL2 4.80 1.50 28.90 37.60 27.20 9.48 100.00

GGL3 1.60 61.00 7.90 24.50 5.00 4.23 100.00

PG 10.00 0 35.00 50.00 5.00 10.49 100.00

The fuel mixtures based on the thermochemical conversion processes gasification

and pyrolysis can be found in Table 5.6. GGL2 is based on a gasification process

with pine wood as biomass, with oxygen and steam as gasification agents. GGL3, on

the other hand, is based on gasification with air as gasifying agent and PG induced

via a pyrolysis process of woody biomass. For the primary investigations air has

been utilized as oxidizer, being induced at the same velocity as the fuel for every

experimental study.

Furthermore, as has been done during the preliminary investigations with diluted

methane mixtures, oxygen was added to the fuel- and also to the air-sides respec-

tively; leading to the compositions presented in Tables 5.7 and 5.8.

Table 5.7: Investigated fuel-oxidizer mixtures enriched with oxygen on the fuel side

Fuel Oxidizer

Name of

fuel-oxidizer

mixture

CH4

[vol.-%]

CO2

[vol.-%]

Air

[vol.-%]

GGL2-2.5-O2-F 4.68 1.46 28.18 36.66 26.52 2.50 100.00

GGL2-5-O2-F 4.56 1.43 27.45 35.72 25.84 5.00 100.00

GGL2-8.5-O2-F 4.39 1.37 26.44 34.41 24.89 8.50 100.00

GGL3-2.5-O2-F 1.56 59.47 7.70 23.89 4.88 2.50 100.00

GGL3-5-O2-F 1.52 57.95 7.50 23.28 4.75 5.00 100.00

GGL3-8.5-O2-F 1.46 55.81 7.23 22.42 4.58 8.50 100.00

PG-2.5-O2-F 9.75 0 34.13 48.75 4.88 2.50 100.00

PG-5-O2-F 9.50 0 33.25 47.50 4.75 5 100.00

PG-8.5-O2-F 9.15 0 32.02 45.75 4.58 8.50 100.00

5.3 Data processing 79

Table 5.8: Investigated fuel-oxidizer mixtures enriched with oxygen on the oxidizer side

Oxidizer Fuel

Name of fuel-oxidizer

mixture

[vol.-%]

Compositions as

presented in Table 5.6

GGL2-2.5-O2-A,

GGL3-2.5-O2-A,

PG-2.5-O2-A

77.03 22.97 -

GGL2-5-O2-A,

GGL3-5-O2-A,

PG-5-O2-A

75.05 24.95 -

GGL2-8.5-O2-A,

GGL3-8.5-O2-A,

PG-8.5-O2-A

72.28 27.72 -

5.2.2 Strain rates - Velocities and straining out

For the experimental investigation the fuel and oxidizer velocities 30 cms-1, 60 cms-1,

and 100 cms-1 were chosen for the BDG. For the counter-flow burner that was built,

with the fuels investigated at 30 cms-1, the first semi-stable flames could be induced.

Velocities of 100 cms-1 led to very stable flames that were still far enough away from

straining out for all the considered compositions.

For the dimension of the utilized burner system (Section 5.1.1), a strain out for

PG would be induced at about 443 cms-1, for GGL3 at 148cms-1, and for GGL2 at

around 2200 cms-1 due to the high amount of hydrogen present. An experimental

examination closer to strain out was not possible due to the laboratories technical

set-up in regard to the gas supply.

All the velocities and strain rates for the regarded compositions can be found in the

Appendix in Tables A.1 and A.2. The strain rates were calculated using Equation

3.17, the strain outs were determined via numerical solutions from DIFFLA.

5.3 Data processing

The processing for all of the spectroscopic data was preliminarily done with the

Davis Software, which is provided by LaVision (Germany) as a product for intelli-

gent laser imaging. This tool can be applied for imaging purposes, amongst other

things, in the fields of combustion, spray applications, and fluid dynamics in com-

bination with high-performance cameras. Being equipped with an ICCD camera

and additionally an energy monitor system all from LaVision, an unrestricted use

80 Chapter 5 Experimental set-up and conditions

of DaVis’s possibilities can be established with the current experimental test rig.

DaVis is set-up as a modular system, enabling an overall operation on a basic level

or, if necessary, applying the benefits of specific spectroscopic needs when added

as an upgrade package. Furthermore, the software itself can integrate most of the

experimental hardware and their regulations (such as, i.e., triggering) in one analysis

concept. A core functionality of DaVis is the processing work that can be done to

the raw experimental images collected. Usually collected in a bundle of hundreds of

unprocessed data, most prominent issues like artefacts as laser intensity gradients

and background noise can be eliminated or at least diminished. The hundreds of

raw experimental images are averaged to one raw data image to reduce random

noise. By combining images taken at different stages or applying filters, the desired

physical representation of the experimental work can be achieved and is constituted

in intensity counts [216].

For the LIF and Rayleigh measurements, one experimental run-through began and

ended with a flat field image as previously mentioned in Section 5.1.4. The concept

of flat fielding (spectroscopic) images is a way to calibrate experimental results

generated by cameras. Variations in pixel-to-pixel comparisons due to a slightly dif-

ferent quantum efficiency or marginally different gain can be removed by the use of

flat field images by flattening the relative response for each pixel to the approaching

radiation. A good and valuable flat field should be taken with the exact filters and

lenses that will be needed for the experimental investigations and also illuminated

by a strong light source, making it much brighter than the experimental image that

will be observed. This results in a calibration image of high signal-to-noise ratio

while not saturating the ICCD. The primary aspect of this projector flat field imag-

ing is to remove the variations from pixel-to-pixel, but also to compensate optical

fringes, vignetting, dust accumulation, or any other artefacts within the optical

path. As usual, when producing calibration images, 500 images were taken and

averaged to generate the final calibration images before and after each experimental

run-through [218].

For the LIF experiments, a 37 % formaldehyde in water solution (Carl Roth, Ger-

many) was diluted with nitrogen and pushed through a thoroughly investigated

flat-fielding device while excited at 355 nm with the Nd-YAG laser. In preliminary

investigations, problems of contaminated burner nozzles due to liquid droplets and

condensation of formaldehyde occurred, which led to the decision of an external ap-

5.3 Data processing 81

paratus. This device was carefully inserted sideways between the top and bottom

nozzles of the burner to flood the measurement area while not disturbing the course

of the laser beam and induce camera reflections. The volume flow of formaldehyde

and nitrogen was kept steady at volume flows of 1.5 lmin-1 and 2 lmin-1, respec-

tively, for all the experimental investigations that were conducted and regulated via

rotameters.

For the Rayleigh experiments, a constant flow of air was inserted into the top nozzle

and the Rayleigh scattering excited at 532 nm of the air was used for flat fielding

purposes. For the CH* chemiluminescence experiments, it was abstained from flat

fielding, among other things, due to the missing application of a spectroscopic ex-

citement source.

Figure 5.10: Diffusion flame (a) excited at 355 nm with fuel (PG-5-O2-F-100) coming from the

top and oxidizer (air) from the bottom nozzle at a velocity of 30 cms−1, formaldehyde flat fielding

image (b), flat fielded diffusion flame (c) with centerline from bottom to top nozzle and formaldehyde

signal (d) all at a burner separation distance of 18.9mm. An increase of intensity is represented

by the color change of dark blue, to blue, to green, to yellow, to red, to white.

In Figure 5.10, the data processing procedure for fluorescence measurements is elu-

cidated step by step. Before and after a diffusion flame image (a), the previously

discussed flat fielding images (b) are obtained and averaged. The laser sheet is

Gaussian in shape, achieving the highest excitation and therefore intensities in the

center between top and bottom nozzle, colored white in Figure 5.10.

82 Chapter 5 Experimental set-up and conditions

It is crucial to place this peak of power in the physical position where the formalde-

hyde develops in the flame, as can be seen via the red lines placed within images

(a) and (b). Secondly, the placement of the flat field needs to be precisely centered

around the vertical centerline, this being the area of interest in comparison to the

numerical data. In the next step, the two flat field images are averaged, and the

diffusion flame was divided by these ((a)/(b)) leading to image (c). The intensity

in counts measured alongside the centerline of the processed diffusion flame leads to

the preliminarily measured formaldehyde peak plotted against the burner separation

(d), this being 18.9mm for all the cases discussed.

The laser sheet is approximately 10 mm wide, thus making the intensity curves

4.45 mm directly next to each nozzle irrelevant. All images are corrected by the use

of an energy monitor, as discussed in Chapter 5.1.4.

Figure 5.11: Polarized diffusion flame (a) excited at 532 nm with fuel (GGL3-2.5-O2-F) com-

ing from the top and oxidizer (air) from the bottom nozzle at a velocity of 60 cms−1, depolarized

diffusion flame (b) air flat fielding image (c), flat fielded diffusion flame (d) with centerline from

bottom to top nozzle and Rayleigh signal (e) all at a burner separation distance of 18.9mm. An

increase of intensity is represented by the color change of dark blue, to blue, to green, to yellow, to

red, to white.

Accordingly, the data processing procedure for Rayleigh measurements is presented

in Figure 5.11. As already introduced above, for the LIF experiments, flat field

images (c) are being shot before and after the images of the diffusion flame ((a) and

(b)) under investigation. Since air has a strong Rayleigh scattering signal, it can be

used for flat fielding by introducing a constant flow through one of the nozzles. Due

to the polarized (a) and depolarized (b) Rayleigh scattering, as presented in Section

4.2, two averaged sets of images of the diffusion flame have to be shot consecutively.

The depolarized scattering (b) has to be subtracted from the polarized scattering

5.3 Data processing 83

(a), before it is divided by the two averaged flat field (c) images from before and

after. This leads to a corrected polarized Rayleigh scattering signal solely from the

examined flame (d). The scattering intensity is again measured along the vertical

centerline leading from the top to bottom nozzle and also plotted against the burner

separation in 18.9mm (e). Finally, all images taken were once again corrected by

the use of an energy monitor.

Due to the strong temperature dependence of the formaldehyde partition function, a

Boltzmann correction, as it was implemented in [26], was applied to the experimental

LIF data after the previously mentioned procedure.

Figure 5.12: The solid line represents

the temperature distribution and normalized

formaldehyde profiles in a CH4–O2counter-

flow and nitrogen diluted diffusion flame.

Solid squares represent the raw, uncorrected

profiles and correlate with profiles corrected

with the assumption Q12 ∼T−1(repre-

sented by the cross) and the assumption

Q12 ∼T−0.5(represented by the open cir-

cle). The thickness of the formaldehyde zone

is characterized as FWHM (full width half

maximum). Based on [26]

Figure 5.13: Normalized with respect to

their value at T= 300 Kcorrection co-

efficients were applied as a function of T,

with the cross representing a correction due

to the temperature dependence of the par-

tition function; the dashed line represent-

ing a quenching correction assuming Q12 ∼

T−0.5, the solid line representing a total cor-

rection assuming Q12 ∼T−0.5, the open

square representing a quenching correction

assuming Q12 ∼T−1and the total correc-

tion assuming Q12 ∼T−1. Based on [26]

As shown in Figure 5.12, the main quantity of formaldehyde forms in the lower tem-

perature region of the diffusion flame, as described here being between 700−1300 K.

At higher temperatures, 1800 K and increasing, a minimum Boltzmann correction

factor of three is necessary to process the experimental data correctly. For the in-

terpolation of the experimental data, a correction coefficient (cross)as a function

of T normalized corresponding to their value at T = 300 K as seen in Figure 5.13

was applied [26].

Chapter 6

Numerical investigations

The numerical investigations that were performed during this study are entirely

based on a model simulating one-dimensional counter-flow DIFfusion FLAmes

named DIFFLA, which will be presented in the following chapter. To support the

evaluation processes of the experimental solutions, further programs, namely CHFIT

and RAYFIT, were developed in-house and will also be discussed shortly.

6.1 DIFFLA

The Fortran 77 code DIFFLA was first developed by Frank Behrendt and Jürgen

Warnatz at the University of Heidelberg in Germany in 1989. The impact of different

fuel compositions, fuel or oxidizer velocities, temperatures, and further operating

conditions on stationary laminar diffusion flames can be derived from this modelling

code solving the governing equations for momentum, mass fraction, and energy [219].

This code can be used to numerically describe either premixed or non-premixed

combustion in counter-flow flames. The experimental work for this thesis was done

for comparison and validation of the DIFFLA code, which was not adapted for these

purposes.

6.1.1 Introduction to DIFFLA and comparable models

The program itself consists of two main modular parts, each composed of numer-

ous sub-routines; these enable the handling of sub-problems in the corresponding

sub-routine. Sub-programs, i.e., for analysing the reaction mechanism or for the

determination of the transport variables, can therefore be taken from other program

packages. This procedure reduces the number of possible sources of error [220].

86 Chapter 6 Numerical investigations

The first main program provides information from various files about the system

that is to be numerically represented, which is then transferred to the second main

program (integrator) by means of a temporary file. The first program reads in the

INPUT file, listing the species present in the system, the boundary conditions of the

integration, and a set of options that control the integration file. The MOLNEW

file provides the molecular data of the species and THERMO their thermochemical

data. Finally, the reaction mechanism is presented in the MECH file. The imported

data and files are checked for consistency. If inconsistencies occur, the program

responds with corresponding error messages and terminates [220].

Table 6.1: Considered species in the combustion model DIFF LA [220,221]

H2/O2/

CO/CO2-

systems

C1-

system

C2-

system

C3-

system

C4-

system

NO-

system

H C C2H C3H2C4H2N

H2CH C2H2C3H3C4H6N2

O1CH2C2H3C3H4n-C4H7NH

O23CH2C2H4C3H5i-C4H7NH2

OH CH3C2H5C3H61-C4H8NH3

HO2CH4C2H6n-C3H72-C4H8NO

H2O CHO HCCO i-C3H7c-2-C4H8NO2

H2O2CH2O CH2CO C3H8i-C4H8N2O

CO CH2OH CH3CO i-C4H9N2H

CO2CH3O CH2CHO p-C4H9CN

CH3OH CH3CHO s-C4H9C2N2

CH3O2C2H5OH t-C4H9HCN

CH3O2H C2H5O C4H10 H2CN

CH3CHOH i-C4H10 HNO

CH2CH2OH HNO2

CNO

HCNO

HOCN

HNCO

The thermal conductivity, diffusion, and toughness of individual species, their spe-

cific heat, and enthalpies are determined from the feeding-in of coefficients and

6.1 DIFFLA 87

molecular parameters. By means of an adjustment calculation, polynomials are fit

to the calculated values. These values are furthermore transferred to a temporary

intermediate data file. This file also includes a species-reaction-matrix regarding the

species involved in a specific reaction [220].

To solve combustion simulations of these premixed or non-premixed counter-flow

flames, 80 different chemical species, molecules, and radicals, are taken into account

(Table 6.1).

Finally, the second main program, the integrator, solves the modelling process by

reading in the temporary intermediate file, using these contained data to carry out

the integration of the conservation equations. The information is distinct for every

flame.

In cases where particle or temperature profiles from previous calculations are not yet

available for specific flames, the integration starts with arbitrary but wisely chosen

beginning profiles. Otherwise, previous calculations can be continued by taking their

start profiles from a file named OUTPUT [220].

As the integration progresses, profiles of the desired stationary flame are provided.

Whether a stationary solution is reasonable depends on the behavior of some of

the profiles when regarding the last 50 to 100 integration steps. Here, particular

attention is given to the temperature profile and that of hydrogen atoms since both

are quite sensitive to indicate the progress when achieving a stationary solution [220].

After a defined number of integration steps, the profiles obtained up to this point

are stored in the OUTPUT file. A continuation of the integration can be carried

out using this file, as previously mentioned. This technique allows for more time-

consuming integrations to be distributed over several program runs. Also, herewith

it is possible to use previously run solutions for integrations of flames, which differ

only slightly in their boundary conditions [220].

88 Chapter 6 Numerical investigations

The development of fast and powerful personal computers and, furthermore, the

increased availability of thermochemical and kinetic combustion data, lead to vari-

ous numerical modelling systems. The following represent alternatives for laminar

premixed and non-premixed numerical modelling [222]:

•PREMIX: developed at Sandia National Laboratories, a Fortran code mod-

elling steady laminar one-dimensional premixed flames

•RUN-1DL: developed in Cambridge, a Fortran code modelling steady lam-

inar one-dimensional and also quasi-one-dimensional premixed, unstrained,

strained, diffusion, partially premixed, two-phase flames, i.e.

•COSILAB: developed from the RUN-1DL code, simulating multiple flames

such as unstrained, premixed, diffusion, partially premixed flames, i.e.

•FlameMaster: developed at RWTH Aachen University, computing 0D and 1D

laminar flames

•OPPDIF: developed at Sandia National Laboratories, a Fortran code comput-

ing counter-flow diffusion flames

•CHEMKIN-II: developed at Sandia National Laboratories, intended for chem-

ical kinetics simulations, evolved from CHEMKIN-PRO

6.1.2 Governing equations

The modelling and simulation of combustion processes via numerical models offers

the opportunity to study flames that are challenging to induce experimentally and

also to validate simulated with experimental results and vice versa. The description

of a reactive flow with the help of a simulation program requires the ability to

describe the governing equations that control the process. As previously mentioned,

this incorporates the following [220,221]:

•The conservation of mass - the continuity equation

•The conservation of momentum - the motion equation

•The conservation of energy - the energy equation

•The conservation of chemical species - the conservation of chemical species

equation

6.1 DIFFLA 89

The flow of a fluid can be described by the conservation of mass, momentum, and

additionally as an equation of state. For the counter-flow flames considered, the

ideal gas law (Equation 4.13) can function as an equation of state with sufficient

precision when regarding the link between temperature, pressure, and density.

Conservation equations were made in the first half of the 19th century by Navier,

Poisson, St. Venant, and Stokes, while all these were basically only describing

isothermal flows of constant density. Finally, when investigating reactive flows,

the conservation of energy and mass in the system have to be included in the nu-

merical simulation. Especially the transition from isothermal to reactive flows can

increase the complexity of the model considerably, leading to the need for approxi-

mations [220,221].

The previously introduced conservation equations for fluids describing combustion

processes, and therefore the induced numerical model, are in need of some assump-

tions [220,221,223]:

•Classical mechanics are used to describe the impacts and general movement of

the atoms and molecules in the flame during the combustion process. Quantum

effects can be neglected due to the low pressure and high temperatures.

•The fluid can be considered continuous due to the mean free path of the atoms

and molecules being substantially smaller than the gradients of their velocity,

temperature, density, and concentrations.

•Under the pressure prevailing in the system, only two but not three impacts

of atoms or molecules are to be expected.

•The collisions of the atoms and molecules in the flame can be regarded as pre-

dominantly elastic, therefore conserving the kinetic energy and not changing

the internal degrees of freedom.

•The bulk viscosity for gases at low pressure, electromagnetic fields, gravita-

tion, buoyancy, energy flux due to a mass concentration gradient, the viscous

dissipation term during the conservation of energy, and the heat transport due

to thermal radiation are all neglected.

6.1.3 Boundary layer assumptions, equations, and conditions

To reduce the numerical simulation to one dimension and also specify the strain

field, the flame structure is simplified by employing boundary layer assumptions

90 Chapter 6 Numerical investigations

and by presuming an outer potential flow [224]. The approximation by Prandtl

says that the flow in the proximity of an object can be subdivided into two areas,

mainly influencing the viscosity of this flow. Inside the boundary layer, drag is

created and the viscosity is dominant; outside of the boundary layer, the viscosity

can be neglected [221]. These assumptions via the Prandtl approximation lead to

the following consequences [221,224]:

•The temperature T, density ρ, mass fractions of species i wi, and the transport

coefficients are solely functions of the radial coordinate y.

•The tangential velocity u is modified linearly with x.

•The radial velocity v is solely a function of y.

•The thermodynamic pressure p is granted constant in the flowing field.

The general boundary layer assumptions have been altered (more details in [220])

to be used for a Tsuji counter-flow configuration as presented in 3.2,

leading to the following relations [224]:

Continuity:

ρv = (ρv)w−Zy

0ρGdy (6.1)

Momentum:

δG

δt =1

δyµδG

δy −δG

δy −G2−H

ρ(6.2)

Temperature:

δT

δt =1

ρcp

δyλb

δG

δy −vδT

δy −

ρcpX

cp,iji

δT

δy −1

ρcpX

hirdi (6.3)

Mass of species i:

δwi

δt =1

δyDimδwi

δy −vδwi

δy +

δyDiTδlnT

δy +rdi

ρ(6.4)

6.1 DIFFLA 91

with t as time, v as the flow velocity in y-direction, y as the cartesian space coordi-

nates, p as pressure, H and G es eigenvalues, µas the dynamic viscosity, cpas the

specific heat capacity, λbas the heat conductivity, rdi as the mass scale chemical

rate of formation, jias the diffusive mass flux, Dimas the modified binary diffusion

coefficient, DiTas the thermodiffusion and the index i representing the ith species of

the system [224].

The boundary conditions at y = 0, the cylinder surface wall, can be presented as

follows [224]:

ρv(0) = (ρv)w(6.5)

G(0) = 0 (6.6)

T(0) = Tw(6.7)

wi(0) = wi,u −ji,w

(ρv)w

(6.8)

Mass fractions for hydrogen atoms experience a recombination at the wall surface,

leading to differing conditions, with w denoting the environment at the cylinder and

u denoting the conditions of the unburnt gas [224]:

wH(0) = 0 and wH2(0) = wH2,u −jH2,w

(ρv)w

−jH,w

(ρv)w

(6.9)

The boundary conditions at y = ye, the outer edge of the flame, are as follows [224]:

δG

δy = 0 (6.10)

T=Te(6.11)

wi=wi,e i= 1,2, ..., N (6.12)

with edenoting the up-streaming gas.

92 Chapter 6 Numerical investigations

6.1.4 Transport and thermodynamic data, reaction mechanisms

To solve all the equations from Sections 6.1.2 and 6.1.3, including the species previ-

ously mentioned in Table 6.1, thermodynamic and transport data, as well as reaction

mechanisms, need to be determined.

The macroscopic transport processes and properties can be traced back to micro-

scopic, especially molecular processes, by the kinetic theory of diluted gases. Limita-

tions that restrict or determine the intermolecular interaction lead to expressions for

diffusion, coefficients of thermal conductivity, and viscosity, describing the transport

processes during the combustion. Detailed derivations can be found in [220].

With the transport data listed in [221], the previously mentioned mixture diffusion

coefficients Di,m, the thermal diffusion coefficients DiT, the thermal conductivity λb,

and the viscosity µcan be calculated for the ith species as follows [220]:

Di,m =1−wi

Pj6=i

Di,j

(6.13)

i=kT

c2MiMj

ρDi,j (6.14)

λb=1

2X

xiλbi+X

λbi−1(6.15)

µ=1

2X

xiµi+X

µi−1(6.16)

with kias the thermal diffusion ratio, Mias the molar mass of the ith particle, and

Mjas the mean molar mass of the mixture in which a particle with the mass Mi

diffuses into.

Thermodynamic parameters like the absolute enthalpies hi(T), the absolute en-

tropies si(T), and specific heat capacities cp,i(T) at a given temperature T are in

need of known reference values at a standard temperature of T0= 25°C and a stan-

dard pressure of p0= 1 bar and represented as polynomial fits. These have been

multiply tabulated and were in use when establishing the program code DIFFLA

for a temperature range of 300 K to 5000 K [220].

6.1 DIFFLA 93

They can be represented as follows [220]:

hi(T) = h0

i,T0+ZT

T0cp,i(T)dT (6.17)

si(T, P) = s0

i,T0+ZT

cp,i(T)

TdT +Zp

pMi

dp (6.18)

cp,i(T) = a1+a2T+a3T2+a4T3+a5T4(6.19)

with R as the universal gas constant.

The reaction mechanism of the combustion can be presented by a set of elementary

reactions of the involved atoms and molecules to describe the participating chemical

processes. The mechanisms involved in the program code DIFFLA can be divided

into H2, O2, CO and CO2systems, the C1- C4hydrocarbon oxidation mechanisms,

the CH3OH mechanism, and the NO mechanism which can all be found in detail in

[220] and [221]. If kinetic parameters were not available, reverse reactions are not

included, but instead, they are computed by numerical simulation via the reaction

equilibrium constant [220,221].

In general, elementary reactions can be presented as follows [220]:

i=1

riAi→

i=1

v00

riAiwith r = 1, ..., E (6.20)

with v’r,i and v”r,i as the stoichiometric coefficients of the ith species of the rrd reaction

of the mechanism and rias the specific rate of formation of the ith species.

The specific rate of formation rican be presented as follows [220]:

ri=Mi

r=1

(v0

ri−v00

ri)kr

j=1

rjwith r = 1, ..., F (6.21)

The reaction rate coefficients krcan be presented as follows [220]:

kr(T) = ArTbrexp−Er

RT (6.22)

with Aras the collision factor, bras the temperature exponent, and Eraas the

activation energy of the rrd reaction. If k-1rfor the reverse reaction is not available

94 Chapter 6 Numerical investigations

the principle of microscopic reversibility, based on krof the initial reaction and the

equilibrium constant Kcr, can be applied as follows [220]:

r(T) = kr

k−1

(6.23)

Kcrcan be calculated based on the change in free enthalpy G0r,f as follows [220]:

r(T) = P

RT Piv0

ri−v00

riexp−G0

r,f

RT (6.24)

All further information in regard to the DIFFLA model, such as mechanisms and

reactions, can be found here [220].

6.2 CHFIT and RAYFIT

The numerical data treatment programs CHFIT and RAYFIT are designed to input

numerical data from DIFFLA to fit experimental CH* and CH2O fluorescence data

and polarized and depolarized Rayleigh data from counter-flow diffusion combus-

tion. These tools were developed in-house for comparison and validation reasons of

experimental and numerical data sets. The main purposes are the detection of the

difference in flame location between the modelled and experimentally obtained re-

sults and, furthermore, the comparability between the sets of data from the applied

spectroscopic techniques.

Both programs require the following subroutines:

•SPLINE and SPLINT: these routines calculate a SPLINE through six data

points around the target point, where the abscissa is then returned by SPLINT

for a given ordinate.

•FIT: this subroutine determines the scale and offset on the magnitude axis

and furthermore returns the difference in value between the modelled and the

experimental data.

Before CHFIT and RAYFIT read in their INPUT file, the provided experimental

file must be pre-processed. The experimental file results from the DaVis program

(as previously described in Chapter 5) in a matrix of 1024 rows by 1280 columns,

with each point representing a value of intensity in counts of a pixel for the two-di-

mensional image of a flame.

6.2 CHFIT and RAYFIT 95

The goal of the pre-processing program is to find the centerline between top and

bottom burners in the amount of 1024 x 1280 image pixels. To achieve this, the

pre-processing file has to be fed with the exact physical locations of the top and

bottom burner center points, of the laser sheet edges, and background positions in

pixel quantities. The outcome is the centerline of the experimental result, which can

now be directly compared to the DIFFLA numerical output files. The number of

grid points of the numerical data needs to be greater than the number of grid points

of the experimental data.

The simplest possibility to discretise a differential equation is via grid points that

have constant distances between them. However, areas of large gradients for the

temperature and concentrations of species can develop due to the rapid change

the counter-flow diffusion flames investigated undergo during the combustion. The

spatial resolution of such gradients requires the densest possible sequence of grid

points. Otherwise, there are also areas of low or no activity in the flames, especially

close to the nozzles, where there is only unburned, fresh gas to be found. The grid

points can be much further apart than in the previously mentioned areas due to

the extremely small gradients for T and the species concentrations. Therefore, an

equidistant grid with sufficiently small grid point spacing to resolve large gradients

would require too many grid points over the entire measuring area. Using grids that

are not equidistant can lead to much shorter computing times for the model. By

reducing the grid point spacing at locations of large gradients of the temperature

or species profile, higher resolution is achieved, compared to locations of smaller

gradients [220].

Figure 6.1: Schematic representation of the numerical and experimental data flows

96 Chapter 6 Numerical investigations

Both programs are fed input data of the experimentally derived and the modelled

diffusion flame, to finally output the comparison of the two. RAY FIT is additionally

in need of a file with ρ0λand RLλinformation to calculate the Rayleigh scattering

for the involved chemical species. The numerical and experimental data flows are

represented schematically in Figure 6.1 for a clearer overview.

Chapter 7

Results, discussion, and conclusions

The experimental investigations of counter-flow laminar diffusion flames with

biomass-based gaseous fuels and diluted methane mixtures at various velocities will

be discussed in the following chapter. The focus is on the spectroscopic investigations

discussed in Chapter 4 of these fuels during combustion. Furthermore, straining out

behavior, variation in dilution or fuel composition, oxygen addition, and other crit-

ical points of this investigation are considered and examined.

In all the following experimental and numerical curves presented, the fuel is induced

from the left and the oxidizer (air) from the right side.

7.1 Diluted methane

Preceding the study of biomass-based gaseous fuels during laminar counter-flow

combustion, an investigation of diluted methane mixtures with nitrogen and carbon

dioxide as diluents was conducted.

Figure 7.1: Sooting methane flame (left) and non-sooting diluted methane flame (right) with air

as oxidizer input from the top and fuel from the bottom nozzle

Numerical and experimental solutions could therefore be validated for more funda-

mental fuel compositions, as opposed to the investigated complex synthetic model

98 Chapter 7 Results, discussion, and conclusions

fuel mixtures based on thermochemical conversion processes presented in 5.6. Fur-

thermore, as previously mentioned, the experimental test rig could be internalized

and verified. The involvement of diluents during hydrocarbon combustion influences

the incandescence of fine soot particles, as presented in Figure 7.1.

7.1.1 Rayleigh measurements of diluted methane fuels

To establish the basis for the forthcoming Rayleigh intensity measurements and de-

velopments of complex gaseous mixtures and also to collect physical information

in regard to the flames, like flame width and positioning, Rayleigh scattering ex-

periments at 532 nm were executed. The combustion processes are more prone to

disturbance and extinction when diluting fuels with non-combusting gases like ni-

trogen and carbon dioxide, though also possibly leading to reduced emissions and

an increased control of the combustion. Dilution of hydrocarbons can be explained

by the thermal, dilution, and chemical effects.

The dilution effect reduces the concentration of the reactants and thus influences

the reaction rates. Furthermore, leading to the fact that the radical pool decreases.

This is reinforced by the thermal effect, as fewer radicals are present in flames

combusting at lower temperatures. The dilution effect when introducing carbon

dioxide as diluent is greater than when using nitrogen, not due to the triatomic vs.

diatomic structure of the gases, but based on the heat capacities. The chemical effect

of carbon dioxide in hydrocarbon combustion can shift the chemical equilibrium

towards the direction of the reactants. Nitrogen, in comparison, shows no chemical

effect on the combustion processes [99,225,226].

Figure 7.2: Experimentally derived and

modelled Rayleigh scattering and temperature

curves with a fuel mixture of CH4-002-N2 at

60 cms−1

Figure 7.3: Experimentally derived and

modelled Rayleigh scattering and temperature

curves with a fuel mixture of CH4-002-CO2 at

60 cms−1.

7.1 Diluted methane 99

To investigate the development of the combustion processes, temperature fields, flu-

orescence, and CH* when increasing the fraction of diluents on the fuel side, multiple

fuel compositions as illustrated in Table 5.3 were studied. At a fuel mixture of each

50 vol.-% methane and diluent as in CH4-001-X, there was still a clear sight of soot

as seen in Figure 7.1, which led to incandescence and therefore deviations in the ex-

perimental investigation of laser-induced fluorescence and CH* chemiluminescence.

Therefore, the Rayleigh scattering investigations were focused on fuel mixtures with

a diluent content of 60 vol.-% or higher, in which sooting and its chemical effects

could be decreased. Figures 7.2 - 7.6 show the studies of experimentally derived and

modelled Rayleigh scattering and the numerically deduced temperature distribution

in the flames.

The experimentally derived Rayleigh scattering curves are constantly circa 0.5mm

wider than the numerically calculated curves; this is still of good agreement as

can be seen in Figure 7.2 for a fuel with nitrogen diluted methane and Figure 7.3

for carbon dioxide diluted methane. This difference in width between modelled

and experimental solutions leads to variations in the positions of the minima of

<0.5mm, as presented in Figure 7.4. Experimental uncertainties are <2.47 %.

Figure 7.4: Minima of experimentally and numerically derived Rayleigh scattering curves with

fuel mixtures including varying fractions of nitrogen and carbon dioxide as diluents at 60 cms−1

Table 4.1 has already given an insight of the differential scattering cross section of

major BDG gases and air. Carbon dioxide exceeds nitrogen by a factor of about 2.5,

generally leading to lower temperatures when comparing the basic fuel mixtures with

both diluents due to the temperature and scattering being in a negative correlation

to each other.

100 Chapter 7 Results, discussion, and conclusions

Figure 7.5: Experimentally derived Rayleigh

scattering and temperature curves with fuel mix-

tures of CH-002-N2 in comparison to CH4-004-

N2 and CH4-007-N2 at 60 cms−1

Figure 7.6: Experimentally derived Rayleigh

scattering and temperature curves with fuel mix-

tures of CH-001-N2 in comparison to CH4-002-

CO2 and CH4-004-CO2 at 60 cms−1

Figure 7.5 shows the development of the flame geometry and temperature decrease

with an increase of 60 vol.-% (CH4-002-N2) of nitrogen to 75vol.-% (CH4-007-N2).

CH4-007-N2 at 60 cms-1 with the available burner set-up was the fuel with the max-

imum nitrogen dilution that led to a stable combustion. This was the case during

experimental and numerical investigation, again putting them in good agreement.

The flames width decreases more than 1 mm, accordingly >20 % in size. The tem-

perature declines from a maximum of 1837K to 1634 K, subsequently about >10 %.

Based on the heat capacity of carbon dioxide, a combustion with a fuel composition

of 75 vol.-% (CH4-007-CO2) diluent and 25 vol.-% methane is not feasible due to

the cool temperature and previous strain out of the flame at these conditions. As

presented in Figure 7.6, the highest fraction of carbon dioxide for a stable combusting

flame at 60 cms-1 was 66vol.-% (CH4-004-CO2), just before strain out. Between the

flames, CH4-001-CO2 and CH4-004-CO2 16 vol.-% of methane was replaced with

diluent. Leading to a temperature drop from a maximum of 1854 K to 1677 K. The

flame width decreases likewise as in the nitrogen cases.

When regarding the carbon dioxide cases, with decreasing methane and increasing

carbon dioxide in the fuel, the minimum Rayleigh intensity and, therefore, peak

temperatures move closer to the air side of the flame. In the cases with increasing

nitrogen in the compositions, the flame stays centered in the same position between

the nozzles. This is also based on the increasing molar mass on the fuel sides when

replacing methane with carbon dioxide, which is heavier by a factor of circa 2.75.

The Rayleigh scattering intensity on the air sides should be equal when regarding all

the experimental investigations. The finite of the laser sheet on both fuel- and air

sides and the divergent normalization factors with differing diluents had an impact

on the conclusive presentation of the experimental air side scattering.

7.1 Diluted methane 101

Figure 7.7: Minimum Rayleigh scattering intensities of experimentally and numerically derived

Rayleigh scattering curves with fuel mixtures including varying fractions of nitrogen and carbon

dioxide as diluents at 60 cms−1

The scattering intensity minima of the experimentally derived Rayleigh scattering

curves were consequently lower than their numerical counterparts, as was presented

in Figures 7.2 and 7.3. This was consistently the case, as presented on the overview

in Figure 7.7, showing all minimum Rayleigh scattering intensities of experimen-

tally and numerically derived Rayleigh scattering curves for diluted methane fuels.

This phenomenon also correlates with the deviating flame widths and can lead to a

deviation of up to 5 cm2(sr)−1. Experimental uncertainties are <3.99 %.

Figure 7.8: Experimentally derived Rayleigh

scattering and corresponding temperature

curves with fuel mixtures of CH-001-N2 and

CH4-004-N2 in comparison to CH-001-N2-O2-

F and CH4-004-N2-O2-F at 60 cms−1

Figure 7.9: Exp. derived Rayleigh scatter-

ing and corresponding temperature curves with

fuel mixtures of CH-001-CO2 and CH4-004-

CO2 in comparison to CH-001-CO2-O2-F and

CH4-004-CO2-O2-F at 60 cms−1

102 Chapter 7 Results, discussion, and conclusions

As discussed, the change in composition from CH4-001-CO2 to CH4-004-CO2 led to

a reduction in peak temperature from 1854 K to 1677 K. In comparison, the change

in composition from CH4-001-N2 to CH4-004-N2 lead to a drop from 1901 K to

1784 K. Subsequently, 5 vol.-% of oxygen were added to both the fuel- and air sides

of the previously mentioned compositions, the solutions are presented in Figures 7.8

- 7.11.

In fuel compositions with 50 vol.-% of hydrocarbon, such as methane, and 50 vol.-%

diluent, such as nitrogen and carbon dioxide, a displacement with oxygen of 5 vol.-%

in the fuel composition has a slight impact on the peak temperatures, such as 39 K

and 50 K, respectively. When increasing the share of diluent to 66 vol.-% the impact

of oxygen displacement leads to a rise in peak temperatures of 116 K and 73 K,

respectively.

This impact in peak temperatures increased when regarding Figures 7.10 and 7.11,

where 5 vol.-% of oxygen was added to the air sides during combustion. Here the peak

temperatures of fuel mixtures including 16 vol.-% more diluent than the 50vol.-%

hydrocarbon-diluent without oxygen addition mixtures are now quasi-equal. Also,

peak temperatures > 2000 K (2009K for nitrogen and 2020 K for carbon dioxide)

can be measured for both the CH4-004-X-O2-A cases.

Figure 7.10: Experimentally derived Rayleigh

scattering and corresponding temperature

curves with fuel mixtures of CH-001-N2 and

CH4-004-N2 in comparison to CH-001-N2-O2-

A and CH4-004-N2-O2-A at 60 cms−1

Figure 7.11: Exp. derived Rayleigh scatter-

ing and corresponding temperature curves with

fuel mixtures of CH-001-CO2 and CH4-004-

CO2 in comparison to CH-001-CO2-O2-A and

CH4-004-CO2-O2-A at 60 cms−1

7.1 Diluted methane 103

7.1.2 Laser-induced fluorescence of diluted methane fuels

Figure 7.12: Numerically derived formaldehyde

curves with fuel mixtures of CH-002-N2 - CH4-007-

N2 at 60 cms−1

As mentioned in the previous

chapters, formaldehyde is formed

as an intermediate species dur-

ing the combustion of hydrocar-

bons and can be detected via

laser-induced fluorescence exper-

iments at 355 nm. Formaldehyde

is mainly found when a hydrocar-

bon and air mixture reaches the

explosion limit, which can be as-

sociated with a ‚cool flame‘, this

helps in the detection of the inia-

tion and progression of the com-

bustion process [150].

Figure 7.13: Experimentally derived laser-induced

fluorescence curves at 355 nm with fuel mixtures of

CH-002-N2 - CH4-007-N2 at 60 cms−1

When applying fluorescence at

355 nm during combustion, the

difficulty of measuring actual

formaldehyde curves when com-

busting hydrocarbon mixtures as

fuels are a substantial challenge.

Since the formaldehyde fluores-

cence spectrum can be found

starting at around 350 nm over

a length of more than 200 nm,

multiple other hydrocarbon com-

pounds, which are also excited at

355 nm, lead to a greater extent

of fluorescence intensity.

In this investigation, the fluorescence measurements were focussed 30 nm above and

below 417 nm. When additionally following the steps for the data processing de-

scribed in Section 5.3, the diluted methane fuels using nitrogen represented in Fig-

ures 7.12 and 7.13 showed the numerical and experimental results to be in good

agreement. The physical position of the formaldehyde evolution between the fuel-

and air nozzles is in excellent agreement for both the experimentally derived and

the calculated curves.

104 Chapter 7 Results, discussion, and conclusions

Figure 7.14: Numerically derived formaldehyde

curves with fuel mixtures of CH-001-CO2 - CH4-004-

CO2 at 60 cms−1

The development of the bulgy

curvature on the fuel sides evolves

identical, when the fuels in-

crease in nitrogen and decrease

in methane. The peak intensi-

ties are not in ideal agreement

for the nitrogen cases, as opposed

to the carbon dioxide cases repre-

sented in Figures 7.14 and 7.15.

The physical positions and the

increase in peak intensity when

adding diluent to the fuel side are

Figure 7.15: Experimentally derived laser-induced

fluorescence curves at 355 nm with fuel mixtures of

CH-001-CO2 - CH4-004-CO2 at 60 cms−1

in good accordance in the nu-

merical and experimental cases.

In contrast to the nitrogen cases,

the experimental carbon diox-

ide cases seem to be generally

curvier than the numerical coun-

terparts. With both diluents, the

formaldehyde fluorescence curves

are wider in the experimental

than the numerical cases, which

is accordant to the Rayleigh scat-

tering solutions.

7.1.3 CH* chemiluminescence and CH of diluted methane fuels

Lastly, the experimental CH* chemiluminescence data are compared to numerically

calculated CH for diluted methane flames. As previously discussed, due to the

main emission of excited CH peaking near 431 nm, a very narrow filter around this

wavelength with only 5nm below and above was used for the experimental investiga-

tions. Figures 7.16 and 7.17 represent both solutions for the CH4-002-X fuel cases,

at fuel and air side velocities of 60 cms-1. In the numerical cases the CH radical is

presented; during the experimental investigations the intensities of the excited CH

radical during chemiluminescence is measured. This is based on the limitations of

the spectroscopic measuring set-up. The focus of these CH investigations, is the val-

idation of the modelled CH peak positioning, which is still feasible due to the close

relations of the positioning of the ground and excited states of the CH radical [136].

7.1 Diluted methane 105

Figure 7.16: Numerically derived CH and

experimentally derived CH* chemiluminescence

curves with fuel mixtures of CH-002-N2 at

60 cms−1

Figure 7.17: Numerically derived CH and

experimentally derived CH* chemiluminescence

curves with fuel mixtures of CH-002-CO2 at

60 cms−1

As expected, the experimental cases are broader in size than the modelled ground

state CH. This is based on background radiation from soot and/or CO2* chemilu-

minescence emissions, as presented in Section 4.3.

The peak position when using nitrogen as the diluent in the fuel mixtures is circa

0.5mm closer to the fuel sides during the experimental investigations, as presented

in Figures 7.18 and 7.19. This can still be referred to as a good agreement when

compared to the numerically calculated models. The developments of the peak

intensities of CH* and CH decrease when the share of nitrogen increases, which is

also in accordance when regarding all the results. The peak development of the

carbon dioxide cases can be observed in Figures 7.20 and 7.21. The peak position is

about 0.2mm off for the fuels with >50 % diluent and surprisingly further offsided

to the fuel side for the CH4-001-CO2 measurement. All these curves were normalized

with the maximum intensity of each figure.

Figure 7.18: Numerically derived CH curves

with fuel mixtures of CH-002-N2, CH-004-N2,

and CH-007-N2 at 60 cms−1

Figure 7.19: Experimentally derived CH*

chemiluminescence curves with CH-002-N2,

CH-004-N2, and CH-007-N2 fuels at 60 cms−1

106 Chapter 7 Results, discussion, and conclusions

Figure 7.20: Numerically derived CH curves

with fuel mixtures of CH-001-CO2, CH-002-

CO2, and CH-004-CO2 at 60 cms−1

Figure 7.21: Exp. derived CH* chemilumines-

cence curves with CH-001-CO2, CH-002-CO2,

and CH-004-CO2 fuels at 60 cms−1

In Figure 7.22, the peak positions of all methane fuels in regard to the fraction of

diluent in the fuel are presented. As was expected, based on the minima of the

Rayleigh scattering solutions presented in Section 7.1.1, the numerical maxima are

continuously slightly closer to the fuel side. The off-set in the peak positions are

<0.5mm for the nitrogen cases and <0.2mm for the carbon dioxide cases, the

experimentally and numerically derived CH* chemiluminescence and CH curves are

therefore regarded to be in very good agreement for these methane fuels. Experi-

mental uncertainties are <1.94 %.

Figure 7.22: Peak positions of experimentally and numerically derived CH* chemiluminescence

and CH curves with methane fuel mixtures including varying fractions of nitrogen and carbon

dioxide as diluents at 60 cms−1

7.2 Biomass-based gaseous fuels - GGL2, GGL3, and PG 107

7.2 Biomass-based gaseous fuels - GGL2, GGL3, and PG

In the previous section, the verification of the experimental test rig on the basis

of spectroscopic measurements and the numerical model DIFFLA resulted in good

agreements for numerous basic hydrocarbon fuel compositions. In the following

sections, the conclusions will be applied to more complex fuel based on synthetic

model product gases from biomass gasification and pyrolysis processes, which are

presented in Table 5.6.

7.2.1 Rayleigh measurements of biomass-based gaseous fuels

Figure 7.23: Experimentally derived Rayleigh scat-

tering and temperature curves with fuel mixtures of

GGL2 at 30 cms−1,60 cms−1, and 100 cms−1

In Figures 7.23 and 7.24 the

Rayleigh scattering and tempera-

ture developments for fuel based

on gasification processes GGL2

and GGL3 are presented for fuel

and air velocities of 30 cms-1,

60 cms-1 and 100 cms-1. For both

fuels, the physical positioning be-

tween the nozzles are adequate,

slight off-sets can be observed for

the slower cases, when compared

to the models. The flames widths

are represented best for the faster

case of 100 cms-1 velocities, as ex-

pected.

Figure 7.24: Experimentally derived Rayleigh scat-

tering and temperature curves with fuel mixtures of

GGL3 at 30 cms−1,60 cms−1, and 100 cms−1

As the velocities increase when

combusting GGL2, the flames are

pushed closer to the fuel side.

This might be based on the big

fraction of hydrogen, with a very

small molar mass, as opposed to

the oxidizer air. The temperature

evolutions differ immensely due

to the great inequality of diluents

vs. combustibles in the generated

fuel blends. As presented in Ta-

ble 5.6 30.4 vol.-% vs. 69.6 vol.-%

and 68.9 vol.-% vs. 31.1 vol.-% for

GGL2 and GGL3, respectively.

The maximum temperatures at

30 cms-1 for these fuels differ by

108 Chapter 7 Results, discussion, and conclusions

an amount of 530 K at 2076 K and 1546 K for GGL2 and GGL3, respectively. When

comparing to the fuels from pyrolysis presented in Figure 7.25 also at velocities of

30 cms-1, 60 cms-1, and 100 cms-1, more equalities are established towards fuel GGL2.

PG shows a diluent vs. combustibles ratio of 35vol.-% vs. 65 vol.-%, with no nitrogen

included. When fuel and air velocities are increased, the faster flames, therefore

peak temperatures, are also pushed towards the fuel side. The peak temperature

for 30 cms-1 is located at 1887 K, which is comparable to the diluted methane cases

with more than 50 vol.-% methane, though having a much lower LHV in total.

Figure 7.25: Experimentally derived Rayleigh scattering and temperature curves with fuel mix-

tures of PG at 30 cms−1,60 cms−1, and 100 cms−1

7.2.2 Laser-induced fluorescence of biomass-based gaseous fuels

During the preliminary investigations using diluted methane fuels, difficulties arose

in developing the relation of experimentally determined formaldehyde via laser-in-

duced fluorescence and the numerical model. These could be thoroughly worked

out with corresponding filters and data processing measures. The intensity of the

derived LIF formaldehyde scattering when investigating GGL2, GGL3, and PG at

multiple air and fuel velocities was very divergent, based on the composition of these

fuels and their potentials to produce this intermediate species. As the fuel and oxi-

dizer velocities increase from 30 cms-1 to 60 cms-1 to 100 cms-1 for the formaldehyde

measurements using GGL2 as fuel presented in Figure 7.26, the peak intensities also

rise. Also, smaller second peaks evolve closer to the fuel sides, where the tempera-

tures are close to 1000 K.

7.2 Biomass-based gaseous fuels - GGL2, GGL3, and PG 109

Figure 7.26: Numerically derived formaldehyde

curves with fuel mixtures of GGL2 at 30 cms−1,

60 cms−1, and 100 cms−1

This is a lower temperature area,

where formaldehyde is knowingly

prone to be formed. At GGL2-30

the second peak shows a smoother

transition to the maximum peak

at the fuel side, as opposed to

GGL2-100 where a distinct crater

forms before rising to the maxi-

mum formaldehyde intensity. As

presented in Figure 7.27, dur-

ing the experimental investiga-

tions two peaks emerged likewise

for all regarded velocities. GGL2-

30 is broader than the modelled

Figure 7.27: Experimentally derived laser-induced

fluorescence curves at 355 nm with fuel mixtures of

GGL2 at 30 cms−1,60 cms−1, and 100 cms−1

counterpart, as was shown in Fig.

7.23 during Rayleigh scattering.

The rise in main peak intensity

is in comparable agreement, while

the side peaks differ in their in-

creasing intensity, according to

the numerical model. When con-

necting these results to Figures

7.28 and 7.29, where GGL3 was

used as fuel, the development of

two peaks can also be observed.

For the modelled GGL3-100, the

second peak, closer to the fuel

side, even exceeds the main peak.

This is based on the temperature distribution, as was presented in the previous

section. The temperatures predominating during all of the combustion processes

investigated when using GGL3 as fuel are prevailing in the complete flame. These

cooler GGL3 flame serve as an ideal environment for an extensive formaldehyde evo-

lution, in spite of the major share of diluent nitrogen in the fuels. The experimental

investigations presented a two-peak distribution as well, but a much more limited

rise in intensities when comparing these. This could be based on the difference in

flame broadness when comparing the modelled and experimental data. The two

peaks in Figure 7.28 at each temperature, respectively, are about 1mm separated,

as opposed to Figure 7.29, where the separation distance is partially >2mm. An

additional 1mm in flame broadness leads to a big displacement when considering the

temperature gradient and therefore different actual temperatures and formaldehyde

evolutions, then calculated models.

110 Chapter 7 Results, discussion, and conclusions

Figure 7.28: Numerically derived formalde-

hyde curves with fuel mixtures of GGL3 at

30,60, and 100 cms−1

Figure 7.29: Experimentally derived LIF

curves at 355 nm with fuel mixtures of GGL3

at 30,60, and 100 cms−1

The best agreement between models and experiments is illustrated in Figures 7.30

and 7.31 for the evolution of the PG fuels. The major and the minor peaks are

in good alignment in all parts of the investigation. The modelled formaldehyde

curves correspond to the experimentally derived LIF counterparts in form and trend.

The rise of intensities with increasing velocities also correlates well. Finally, the

reduction of the flames width at raised fuel and air velocities coincides correctly

with the off-sets of the smaller peaks towards the air side. As previously seen in the

GGL2 cases, the transitions between the secondary to the main peaks show stronger

craters in the experimental cases in comparison to the numerical calculations. The

correlation between the GGL2, GGL3, and PG fuel mixtures at 100 cms-1 is pictured

in Figures 7.32 and 7.33, normalized in regard to the greatest peak at PG-100.

Figure 7.30: Numerically derived formalde-

hyde curves with fuel mixtures of PG at 30,60,

and 100 cms−1

Figure 7.31: Experimentally derived LIF

curves at 355 nm with fuel mixtures of PG at

30,60, and 100 cms−1

7.2 Biomass-based gaseous fuels - GGL2, GGL3, and PG 111

The relation between the curves fits very well in both modelled and experimental

investigations. As expected, GGL2-100 leads to a smaller formaldehyde intensity, in

comparison to the other fuels, due to considerably higher temperatures prevailing

throughout the flame [26]. PG fuels include the largest amount of methane in the

fuel composition; therefore the greatest overall intensity of formaldehyde throughout

the flame was measured, as expected [149,227].

Figure 7.32: Numerically derived formalde-

hyde curves with fuel mixtures of GGL2, GGL3,

and PG at 100 cms−1

Figure 7.33: Experimental LIF curves at

355 nm with fuel mixtures of GGL2, GGL3, and

PG at 100 cms−1

7.2.3 CH* chemiluminescence and CH of biomass-based fuels

Lastly, experimental CH* chemiluminescence measurements were compared to nu-

merically modelled CH curves in regard to intensity evolution, form, and peak posi-

tioning.

Figure 7.34: Numerically derived CH curves

with fuel mixtures of GGL2 at velocities of

30 cms−1,60 cms−1, and 100 cms−1

Figure 7.35: Experimentally derived CH*

chemiluminescence curves with fuel mixtures of

GGL2 at velocities of 30,60, and 100 cms−1

112 Chapter 7 Results, discussion, and conclusions

Flame GGL2-30 is slightly closer to the air side, as was expected from the solutions

of the Rayleigh scattering and fluorescence measurements. Furthermore, the peak

positioning and intensity development are in excellent agreement for GGL2 fuels, as

pictured in Figures 7.34 and 7.35, as is the case for PG fuels plotted in Figures 7.36

and 7.37. For all of these solutions, the experimental curves are broader in size than

the numerical counterpart, as was discussed in Section 7.1.3.

Figure 7.36: Numerically derived CH curves

with fuel mixtures of PG at velocities of

30 cms−1,60 cms−1, and 100 cms−1

Figure 7.37: Experimentally derived CH*

chemiluminescence curves with fuel mixtures of

PG at velocities of 30,60, and 100 cms−1

As witnessed in the previous section during the formaldehyde measurements, the

GGL3 fuels do not quite agree with the models, as presented in Figures 7.38 and

7.39. The LIF measurements did not agree with the increase of intensity for the peaks

closer to the fuel side at raised velocities; the chemiluminescence investigations do

not agree with the decrease of intensity at enlarged velocities.

Figure 7.38: Numerically derived CH curves

with fuel mixtures of GGL3 at velocities of

30 cms−1,60 cms−1, and 100 cms−1

Figure 7.39: Experimentally derived CH*

chemiluminescence curves with fuel mixtures of

GGL3 at velocities of 30,60, and 100 cms−1

An assumption could be that the oxygen of the oxidizer does not arrive at the flame

front in the quantity that is numerically modelled. Due to 61 vol.-% nitrogen present

in the fuel for the GGL3 cases, the formaldehyde evolution in the flames is strongly

7.3 Biomass-based gases with additional oxygen on the fuel side 113

in need of the oxygen from the oxidizer. This proposition is also strengthened by

the surplus of CH molecules measured during the experiments at raised velocities.

Illustrated in Figure 7.40 are the maximum peak positions of experimentally and

numerically derived CH* chemiluminescence and CH curves with fuel mixtures of

GGL2, GGL3, and PG at 30 cms−1,60 cms−1, and 100 cms−1. For all investigated

fuels, the peaks show a maximum displacement of <0.5mm. The best agreement

can be reached at the fastest velocity, 100 cms−1, as expected. The slowest velocity at

30 cms−1presents an edge case for stable combustion for the regarded experimental

set-up. Experimental uncertainties are <1.82 %.

Figure 7.40: Peak positions of experimentally and numerically derived CH* chemiluminescence

and CH curves with fuel mixtures of GGL2, GGL3, and PG at 30 cms−1,60 cms−1, and 100 cms−1

7.3 Biomass-based gases with additional oxygen on the fuel

side

This section investigates the combustion behavior of biomass-based gases with ad-

ditional oxygen supplied during the combustion process on the fuel sides for the

given fuel compositions GGL2, GGL3, and PG, as previously presented in Table

5.7. This should provide further insights into the utilization of these product gases

from biomass gasification or pyrolysis in industrial applications.

114 Chapter 7 Results, discussion, and conclusions

7.3.1 Temperature evolution with additional fuel side oxygen

Figure 7.41: Experimentally derived Rayleigh scat-

tering and temperature curves with fuel mixtures of

GGL2 at 60 cms−1and additional fuel side oxygen of

2.5,5, and 8.5vol. −%

As plotted in Figure 7.41, the

addition of 2.5, 5, and 8.5vol.-%

of oxygen to the fuel side when

combusting GGL2 fuel lead to

an increase of peak tempera-

tures of 27, 60, and 118 K respec-

tively in comparison to the case

without oxygen addition. Fur-

thermore, the flame width in-

creased markedly for a consider-

able amount of 8.5 vol.-% added,

which can not be observed for the

other GGL2 cases. The evolution

for the PG cases in Figure 7.43

displayed an analogous behavior.

Figure 7.42: Exp. derived Rayleigh scatter-

ing and temperature curves with fuel mixtures

of GGL3 at 60 cms−1and additional fuel side

oxygen of 2.5vol.−%,5vol.−%, and 8.5vol.−%

Figure 7.43: Experimentally derived Rayleigh

scattering and temperature curves with fuel mix-

tures of PG at 60 cms−1and additional fuel side

oxygen of 2.5vol.−%,5vol.−%, and 8.5vol.−%

Furthermore, especially for the PG cases, the Rayleigh scattering minima and, there-

fore, peak temperatures were pushed significantly in the direction of the fuel side.

The GGL3 experiments and models presented in Figure 7.42 encountered an ex-

tremer influence by the addition of oxygen in the fuel composition. At inputs

of 2.5 vol.-% and 5vol.-% an increase of peak temperatures of 67 and 148 K was

detected, at 8.5 vol.-% as much as 309K. This last case presented an increase of

> 20 vol.-% in temperature, but can not be regarded as a stable case. The numerical

model demonstrates the combustion instability with this amount of fuel side oxygen

appropriately, which could not be investigated experimentally due to its fluctuations.

Finally, a comparison of the minima of experimentally and numerically derived

Rayleigh scattering curves with with fuel mixtures of GGL2, GGL3, and PG at

7.3 Biomass-based gases with additional oxygen on the fuel side 115

60 cms−1and additional fuel side oxygen of 0vol.-%,2.5vol.-%,5vol.-%, and

8.5vol.-%are plotted in Figure 7.44. As previously mentioned, the experimental

PG case minima have moved up to 1 mm closer to the fuel side. The GGL3 cases

are in excellent agreement; the GGL2 cases are mostly in good agreement when

regarding the positioning of the experimentally and numerically derived scattering

minima. The evolution of the scattering when adding additional fuel side oxygen to

these combustion processes and the temperatures will be further discussed in Section

7.5.2. Experimental uncertainties are <3.48 %.

Figure 7.44: Minima of experimentally and numerically derived Rayleigh scattering curves with

fuel mixtures of GGL2, GGL3, and PG at 60 cms−1and additional fuel side oxygen of 0vol. −%,

2.5vol. −%,5vol. −%, and 8.5vol. −%

7.3.2 LIF of biomass-based gases with additional fuel side oxygen

An addition of oxygen to the combustion via the fuel composition had an over-

all drastic influence on the formaldehyde formation and will be presented in the

following section.

In Figures 7.45 and 7.46, the combustion of GGL2 at 60 cms-1 with and without

additional fuel side oxygen of 5 vol.-% is plotted. Showing a two-peak distribution

of formaldehyde. As expected, the addition of fuel side oxygen has a considerable

impact, especially on the before discussed smaller peak closer to the fuel side. This

great formaldehyde intensity peak can be allocated at a temperature distribution of

around 1000 K when consulting Figure 7.41 in the previous Section 7.3.1. Consider-

ing the continuously identified slightly increased flame widths for the

116 Chapter 7 Results, discussion, and conclusions

Figure 7.45: Numerically derived formalde-

hyde curves with fuel mixtures of GGL2 at

60 cms−1with and without additional fuel side

oxygen of 5vol. −%

Figure 7.46: Experimentally derived LIF

curves at 355 nm with fuel mixtures of GGL2 at

60 cms−1with and without additional fuel side

oxygen of 5vol. −%

experimentally derived results, the distances of the peaks throughout the flames seem

to be reasonable. The previously stated „main peaks“ show an enlarged increase in

intensity when compared to the numerical model. This could be traced back to

other, bigger hydrocarbon bonds also fluorescing at 355 nm, especially in regard to

GGL2-5-O2-F-60. The progression in regard to the formaldehyde intensity when

increasing fuel- and air side velocities is plotted in Figures 7.47 and 7.48. Striking is

once again the deep crater that develops for all fractions of oxygen added to the fuel

side compositions. The numerical model does show smaller shoulders developing at

the air side, next to the main intensity peaks. The experimental solutions though,

present two definite curves. As previously discussed, this could either be based on the

additional intensity of other hydrocarbons that are excited at 355 nm and fluoresce

between 387 −447 nm. Another cause could be an aberration of the interpolation

process described in 5.3. This mechanism was based on methane and oxygen flames,

partly diluted with nitrogen.

Figure 7.47: Numerically derived formalde-

hyde curves with fuel mixtures of GGL2 at

30,60, and 100 cms−1with additional fuel side

oxygen of 5vol. −%

Figure 7.48: Experimentally derived LIF

curves at 355 nm with fuel mixtures of GGL2

at 30,60, and 100 cms−1with additional fuel

side oxygen of 5vol. −%

7.3 Biomass-based gases with additional oxygen on the fuel side 117

Figure 7.49: Numerically derived formalde-

hyde curves with fuel mixtures of GGL3 at

30,60, and 100 cms−1with additional fuel side

oxygen of 5vol. −%

Figure 7.50: Experimentally derived LIF

curves at 355 nm with fuel mixtures of GGL3

at 30,60, and 100 cms−1with additional fuel

side oxygen of 5vol. −%

As presented in section 7.1, this procedure showed a very good agreement when

basing it on similar diluted methane flame types. Presumably, the temperature

curves have not been interpolated amply in regard to the high flame temperatures,

which are most prominent at GGL2, which created two peaks as opposed to one. Or

reversed: the model did not present the strong temperature gradient of formaldehyde

sufficiently. The physical positions of the flames between the nozzles continuously

show good agreement when incorporating the increased width of the experimental

solutions.

The GGL3 fuels with added oxygen exhibited in Figures 7.49 and 7.50 showed sim-

ilar results in comparison to the GGL2 fuels. The difference lies in the missing

shoulder in the curves at the air side when modelled. This missing shoulder in the

numerical solutions is represented by a much smaller second peak in the experimen-

tal investigation. Also included in this consideration should be the substantially

lower prevailing temperatures during GGL3 combustion.

Figure 7.51: Numerically derived formalde-

hyde curves with fuel mixtures of PG at

60 cms−1with additional fuel side oxygen of

2.5,5, and 8.5vol. −%

Figure 7.52: Experimentally derived LIF

curves at 355 nm with fuel mixtures of PG at

60 cms−1with additional fuel side oxygen of

2.5,5, and 8.5vol. −%

118 Chapter 7 Results, discussion, and conclusions

Decreased flame temperatures lead to reduced second peaks. However, it should

lead to an increased formaldehyde evolution due to its preference for cooler parts in

the flame. Figures 7.51 and 7.52 represent the PG formaldehyde and LIF curves.

As presented in Sections 7.2.1 and 7.3.1, the PG fuels represented mediocre temper-

atures, in comparison to both gasification fuels. As shown, also the formaldehyde

second peaks represent themselves as modest in relation to combusted formaldehyde

second peaks of GGL2 and GGL3 oxidated fuels.

7.3.3 Chemiluminescence with additional fuel side oxygen

Finally, a brief investigation regarding the CH* chemiluminescence and CH curves

is presented in Figures 7.53 and 7.54. As demonstrated in Section 7.2.2, the PG

fuels yielded the greatest peak intensity for a formaldehyde formation, which could

imply the biggest generation of CH* / CH in regard to the compared fuels, further-

more based on the largest fraction of hydrocarbon in the fuel. This was also the

case after the addition of oxygen; they are in good agreement concerning the peak

intensity measurement and physical positioning. GGL3 includes the most diluent of

the investigated fuels, but also 24.5vol.-%of CO.

Figure 7.53: Numerically derived CH curves

with fuel mixtures of GGL2, GGL3, and PG

at 60 cms−1with additional fuel side oxygen of

5vol. −%

Figure 7.54: Experimentally derived CH*

chemiluminescence with fuel mixtures of GGL2,

GGL3, and PG at 60 cms−1with additional fuel

side oxygen of 5vol. −%

A surplus of oxygen could have led to other chemiluminescence activities during the

combustion, as discussed in Section 4.3, and therefore to the greater experimental

peak. Figure 7.55 shows the peak positioning in excellent agreement for all fuels

with no surplus oxygen. With oxygen added, the PG and GGL3 experimental and

numerical peaks are furthermore in very good agreement. The GGL2 solutions show

an off-set of more than 1 mm; this could have its origin in the significantly greater

flame widths for these cases. Experimental uncertainties are <4.74 %.

7.4 Biomass-based gases with additional oxygen on the air side 119

Figure 7.55: Peak positions of experimentally and numerically derived CH* chemiluminescence

and CH curves with fuel mixtures of GGL2, GGL3, and PG at 60 cms−1with additional fuel side

oxygen of 0,2.5,5, and 8.5vol. −%

7.4 Biomass-based gases with additional oxygen on the air side

In this last section of the experimental and numerical investigations, the combustion

behavior of biomass-based gases with additional oxygen supplied during the com-

bustion process on the air sides for the presented fuel compositions GGL2, GGL3,

and PG is looked into, as shown in Table 5.8. Due to the experimental time-limita-

tion in the use of the laser-laboratories of institute EVUR at Technische Universität

Berlin before a relocation, some of the following investigations are entirely based on

numerically derived results.

7.4.1 Temperature evolution with additional air side oxygen

The enhancement of the combustion processes, and therefore temperature evolu-

tions, with GGL2 fuels using additional oxygen fractions of 2.5,5, and 8.5 vol.-%

in the oxidizer is plotted in Figure 7.56. The maximum temperature at 60 cms-1

without oxygen enhancement was 2018 K, with fuel side oxygen enhancement of the

mentioned fractions 2046, 2077, and 2135K respectively and finally, with oxygen

enrichment on the air side 2096, 2167, and 2253 K respectively.

120 Chapter 7 Results, discussion, and conclusions

Figure 7.56: Numerically derived Rayleigh scattering

and temperature curves with fuel mixtures of GGL2 at

60 cms−1with additional air side oxygen of 2.5,5, and

8.5vol. −%

Clearly, the additional oxygen in

the oxidizer had a more severe im-

pact in terms of heat production,

leading to a maximum increase of

T of 11.6 % at 8.5 vol.-%. The

large fraction of hydrogen in this

fuel makes GGL2 prone to react

willingly with any addition of ox-

idizer. A different reaction ten-

dency was exhibited by the addi-

tion of oxygen to the GGL3 fuels,

as pictured in Figure 7.57.

Figure 7.57: Numerically derived Rayleigh scattering

and temperature curves with fuel mixtures of GGL3 at

60 cms−1with additional air side oxygen of 2.5,5, and

8.5vol. −%

Which was expected due to the

much lower peak temperature of

1484 K at 60 cms-1, amongst oth-

ers due to the 61 vol.-% of ni-

trogen in the composition. A

development of 1552, 1633, and

1793 K at air side enhancement

can be compared to 1530, 1571,

and 1620 K at the corresponding

fuel side counterpart for 2.5, 5 and

8.5 vol.-% added oxygen. GGL3

was the only fuel, where the al-

tered fuel side had a greater

impact on the temperature development. Finally, illustrated in Figure 7.58 are the

corresponding temperature and Rayleigh scattering curves with additional air side

oxygen of 2.5, 5, and 8.5 vol.-% for PG fuels. With an 1867, 1908, and 1977 K vs. an

1898, 1960, and 2041 K temperature evolution, in comparison to a start at 1829 K

at 60 cms-1 without oxygen enhancement. This shows a maximum increase of T of

11.5-% at 8.5 vol.-%, possibly due to the significant fraction of carbon monoxide in

the fuel.

7.4 Biomass-based gases with additional oxygen on the air side 121

Figure 7.58: Numerically derived Rayleigh scattering and temperature curves with fuel mixtures

of PG at 60 cms−1with additional air side oxygen of 2.5,5, and 8.5vol. −%

7.4.2 Formaldehyde of biomass gases with additional air side oxygen

As previously mentioned, due to the move of the institute EVUR where these in-

vestigations took place, especially the experimental analysis could not be fulfilled as

planned. This is not ideal due to the off-sets in the comparisons of the experimental

and numerical solutions, as was the case in Section 7.3.2 for additional fuel side

oxygen.

Remarkably, as presented in Figures 7.59 and 7.60, did the input of additional

air side oxygen not have the outstanding effect on the formaldehyde evolution, as

expected. In comparison to all three cases at 60 cms-1 without any fuel- or air side

additions, the formaldehyde curves have only increased slightly and developed their

peaks mainly on the fuel sides. In regard to the temperature evolution presented

in the previous section, a decrease of the formaldehyde peak development was at

least predicted for the GGL2 and PG cases due to the rise of peak temperatures

for all velocities and oxygen fraction inputs investigated. The GGL3 fuels showed

an increase in peak temperatures, though these were all <1800 K. Therefore, an

increase of the formaldehyde evolution was expected with an oxygen addition to the

oxidizer.

122 Chapter 7 Results, discussion, and conclusions

Figure 7.59: Numerically derived formalde-

hyde curves with fuel mixtures of GGL2, GGL3,

and PG at 60 cms−1with additional air side

oxygen of 5vol. −%

Figure 7.60: Numerically derived formalde-

hyde curves with fuel mixtures of GGL2, GGL3,

and PG at 60 cms−1with no additional air side

oxygen

7.4.3 Chemiluminescence with additional air side oxygen

The strongest quantitative chemiluminescence signal was measured during these

additional air side enhancement experiments for the PG cases at 8.5vol.-% oxygen

feed-in. In Figures 7.61 and 7.62, once again, the numerically derived CH curves and

the experimentally collected CH* chemiluminescence signals can be well compared.

As expected, the outcome for the GGL2 and PG gas compositions are in good agree-

ment, both for the modelled and the experimental cases. The physical positioning

for GGL2 is in excellent agreement; PG appears to be slightly closer to the air side.

As previously presented in section 7.3.3, once again, the CH* chemiluminescence

signal is multiple factors greater than its modelled counterpart for GGL3 fuels with

oxygen enhancement on the air side. The physical positioning is satisfactory.

Figure 7.61: Numerically derived CH curves

with fuel mixtures of GGL2, GGL3, and PG

at 60 cms−1with additional air side oxygen of

5vol. −%

Figure 7.62: Experimentally derived CH*

chemiluminescence with fuel mixtures of GGL2,

GGL3, and PG at 60 cms−1with additional air

side oxygen of 5vol. −%

7.5 Combustion behavior 123

7.5 Combustion behavior

The previously in this chapter presented investigations provided a suitable method

to verify the structure of diluted hydrocarbon diffusion flames to furthermore show

the overall agreement between numerical and experimental solutions for the newly

planned and built counter-flow burner set-up. Moreover, the combustion behavior of

more complex, synthetic model fuel combinations, also examining the effectiveness of

the flame inhibitor oxygen, was investigated and will be summarized in the following

sections.

7.5.1 Combustion behavior of diluted methane flames

By using several spectroscopic techniques, multiple possibilities arose to review the

combustion behavior of diluted methane flames. Already visible to the naked eye in

the laboratory was the non-stability of the combustion in regard to an adjustment of

fuel- and air side velocities or dilution of the hydrocarbon fuels. These phenomena

were presented by small wave-like motions of the counter-flow flames between top

and bottom nozzle or straining out. Furthermore, soot production and decrease

by an increase of carbon dioxide or nitrogen to the fuel composition were clearly

evident, as presented in Figure 7.1.

The increase of minimum Rayleigh scattering signals for further induced diluents

(up to 15 % of nitrogen or 16 % of carbon dioxide) into the fuels was expected due

to the chemical, diluent, and thermal effects previously mentioned. When regarding

the numerical solutions presented in Figure 7.7, an increase of scattering signal of

<17 % was expected. The experimental investigations led to solutions of up to

<25 %. This correlates to the divergent widths of the flames and will be discussed

further in Section 7.5.2.

Moreover, the evolution of formaldehyde fluorescence was clearly presented in all fig-

ures of Section 7.1.2 by a consistent trade-off of small quantities of 3 vol.-% between

methane and diluent. This increase of diluent in the fuel has a more significant

impact on carbon dioxide, as opposed to nitrogen, due to the fall in temperature

based on the heat capacity and small contributions from radiative heat loss, which

is consistent with literature.

The investigation of CH* chemiluminescence was foremost used to consider the phys-

ical positioning between fuel- and oxidizer nozzles for the experimental solutions in

124 Chapter 7 Results, discussion, and conclusions

comparison to their numerical models. Overall, this led to very satisfying solutions

for both diluents, as presented in Figure 7.22.

7.5.2 Combustion behavior of biomass-based flames

In the following section, the combustion behavior of the regarded biomass-based

flames, induced with synthetic model fuels, and the influence of oxygen enhancement

on the overall performance will be the main focus.

Figure 7.63: Minimum Rayleigh scattering intensities of experimentally and numerically derived

Rayleigh scattering curves with fuel mixtures of GGL2, GGL3, and PG at 60 cms−1with additional

fuel side oxygen of 0,2.5,5, and 8.5vol. −%

Figure 7.63 presents the evolution of the minimum Rayleigh scattering intensities

of experimentally and numerically derived Rayleigh scattering curves with fuel mix-

tures of GGL2, GGL3, and PG at 60 cms−1with additional fuel side oxygen of

0,2.5,5, and 8.5vol.-%. For all of the regarded fuels, the minimum peak intensities

of the experimental studies were higher than their numerical counterparts, leading

to deviations up to 30 %. Furthermore, the experimental uncertainties are <3.66 %.

The potential large errors can have at least three origins: the inconsistencies in the

flames widths, the large differences in Rayleigh cross sections for different species,

and the spectroscopic set-up. The numerically modelled Rayleigh scattering is based

on the expected number of different chemical species in every point of the centerline

of the flame. If the numerically calculated and the experimentally measured flame

widths show an aberration, the compositions of chemical species at each specific

7.5 Combustion behavior 125

point will have an off-set and will lead to an error in scattering intensities. Further-

more, the large differences in Rayleigh scattering cross sections for chemical species,

as presented in Table 4.1, can then lead to an even larger error for a specific point in

the centerline. Lastly, though the experimental uncertainties are on an acceptable

level, it is difficult to detect a possible fundamental deviation in the experimental

set-up due to the large number of components. The discussion of the temperature

evolution during these investigations, especially in regard to the influence of oxygen

enhancement on the overall behavior, will regardlessly be based on the numerical

Rayleigh scattering solutions and the resulting maximum temperatures during these

combustions.

The temperature evolution of the basic (no surplus oxygen) composition of

biomass-based fuels was measured and modelled to be GGL2 > PG > GGL3, as

presented in Figure 7.64. GGL2 included 27.2 vol.-% of hydrogen in the composition

and, on the other hand, included the least amount of diluents (overall 30.4 vol.-%)

in the mixture leading to 2018.20 K as peak temperature. The pyrolysis gas PG fol-

lowed in regard to maximum combustion temperature with 1829.25 K, with a carbon

dioxide fraction of 35 vol.-% (and no nitrogen) and 50 vol.-% of carbon monoxide.

In comparison, PG has a slightly higher LHV than GGL2, as presented in 5.6.

GGL3 consisted of more than double the share in diluents compared to GGL2 with

61 vol.-% of nitrogen, leading to the lowest maximum temperature at 1484.07 K.

Figure 7.64: Numerically derived maximum temperatures with fuel mixtures of GGL2, GGL3,

and PG at 60 cms−1with additional fuel- and air side oxygen of 0,2.5,5, and 8.5vol. −%

126 Chapter 7 Results, discussion, and conclusions

The oxygen enhancement of the GGL2, GGL3, and PG fuels had a significant effect

on the temperature evolutions as expected, though with diverse consequences. An

increase in heat development could be observed for all velocities and volume fractions

of oxygen added to the fuels. A drawback of fuels from biomass-based processes is the

commonly lower combustion temperatures and, therefore, heat release reductions in

comparison to other (hydrocarbon) fuels. An alternative to pure oxygen as oxidizer,

in terms of upgrading costs, seems to be the inclusion of small quantities of air to

either the fuel- or oxidizer-sides.

Table 7.1: Temperature evolution of GGL2 fuels with additional oxygen on fuel or air side

GGL2 Peak T

fuel side [K]

T increase

fuel side [%]

Peak T

air side [K]

T increase

air side [%]

0% oxygen 2018.20 / 2018.20 /

5% oxygen 2077.82 2.95 2167.04 7.37

8.5% oxygen 2135.62 5.82 2252.95 11.63

Table 7.2: Temperature evolution of GGL3 fuels with additional oxygen on fuel or air side

GGL3 Peak T

fuel side [K]

T increase

fuel side [%]

Peak T

air side [K]

T increase

air side [%]

0% oxygen 1484.07 / 1484.07 /

5% oxygen 1632.41 9.99 1570.28 5.80

8.5% oxygen / / 1619.96 9.16

Table 7.3: Temperature evolution of PG fuels with additional oxygen on fuel or air side

PG Peak T

fuel side [K]

T increase

fuel side [%]

Peak T

air side [K]

T increase

air side [%]

0% oxygen 1829.25 / 1829.25 /

5% oxygen 1908.04 4.30 1960.26 7.16

8.5% oxygen 1977.26 8.09 2040.08 11.53

When considering the GGL2 fuels, the oxygen enrichment on both sides led to

constantly stable flames, in regard to a strain out, with a maximum temperature

increase of 11.63 % at 8.5 vol.-% of oxygen enhancement on the air side. Only the

lower volume fraction of oxygen added to the air side induced a slight movement of

the flames towards the air side nozzle. Table 7.1 presents the temperature increase

for both fuel and air side enhancement. The impact when adding oxygen to the

air side was most distinctive for this fuel (T increase air side), as opposed to an

7.5 Combustion behavior 127

addition to the fuel side (T increase fuel side). The combustion with GGL3 fuels

showed different reactions to an oxygen enhancement, as opposed to the two other

fuels. First, the 8.5 vol.-% of oxygen enrichment only leads to a stable combustion for

GGL3 flames when being added to the air side, as opposed to the fuel side. Second,

the temperature increase is notably less when adding oxygen to the air side (7.2).

PG fuels combusted very well and smooth at all oxygen enhancements, leading to

slight increases in regard to the flames widths. The temperature increase on the

air side is comparable to the GGL2 results; on the fuel side, the impact was much

higher in comparison to GGL2, though having 4.6vol.-% less combustible gases in

the fuel, presented in Table 7.3.

For all the basic biomass-based fuels investigated, the laser-induced fluorescence

measurements led to two peaks of formaldehyde in the flames. A smaller one closer

to the fuel side nozzle and a larger one closer to the air side nozzle. This was

confirmed in good agreement by the numerical model investigations, leading only to

a dissent for the GGL3 fuels. Here the smaller peaks on the fuel side were numerically

modelled to be more extensive, especially in regard to the 100 cms-1 velocity cases.

The diluted methane flames were thoroughly investigated with the help of a spectro-

graph, to demonstrate the sole detection of formaldehyde in the physical positions

of the flame where it was assumed to be. These examinations hence excluded other

chemical compounds at the physical positions in the flames where formaldehyde

is suspected and, furthermore, confirmed the agreement between experimental and

numerical data.

The oxygen enhancement in fuels and oxidizers led to consistently higher temper-

atures and heat distributions throughout the flames. This had evident influences

on the formaldehyde positioning, as shown in sections 7.3.2 and 7.4.2. For the fuel

side addition for GGL2, GGL3, and PG, a very definite region closer to the fuel side

nozzle in parts of the flames with T around 1000 K, was preferred. Typically for

formaldehyde, once the hotter combustion regions set in, it quickly disappears. To

appear again on the other side of the flame, where cooler temperatures prevail, next

to the air side nozzle.

Next to an increase of formaldehyde intensity for fuels with enhanced oxygen, the

CH* chemiluminescence measurements and the CH modelling led to the rise in CH*

and also CH. In comparison to the air side enhancement, which had a severe influence

on the temperature in most investigated cases and also on the CH* / CH intensities,

but only very moderately on the formaldehyde distribution.

128 Chapter 7 Results, discussion, and conclusions

7.6 Critical points

The utilization of biomass-based gases from gasification or pyrolysis for combustion

processes is accompanied by multiple critical points. These are mainly based on

efficiency or economic issues and are shortly presented in the following sections.

7.6.1 Gas cleaning and composition

For the experimental work done during these investigations, fuels were mixed with

the use of bottled gases, a system of mass flow controllers, and a mixing chamber as

presented in Section 5.1.4 and constituted as synthetic biomass-based fuels.

The mixing of precisely predefined fractions of compositions was a necessary pre-

condition during the experimental work. When taking multiple images of flames,

an experimental cycle has to run stable for at least 120 seconds, preferably longer.

In Chapter 2, the different steps for differing gaseous yields during gasification or

pyrolysis processes are described as dependent on process conditions. Contaminants

and byproducts can primarily be minimized by optimizing the operation procedures

of the system and choosing sensible properties in regard to the feedstock. In any

case, this will impact the subsequent application, such as consecutive combustion

processes. Furthermore, before a practical utilization of the product gases from

gasification or pyrolysis of lignocellulosic biomass can be initiated, downstream gas

cleaning steps have to be conducted. Particulate or gaseous impurities such as tar,

nitrogenic aggregates, sulphuric compounds, particulate matters, or alkali metals

need to be physically removed. Possibilities are extensive cold gas cleaning, lead-

ing to hot gas cleaning, filtration, the use of scrubbers, thermal cracking, or the

utilization of sorbents or catalysts [228,229].

Critical in these matters are the comprehensive cleaning steps, the extent of clean-

liness the product gases can achieve, the attainment of a continuous product gas

composition, and economic expenses that can be substantial.

7.6.2 LHV

The lower heating values of the investigated fuel compositions were presented in

Chapter 5, with PG > GGL2 > GGL3 being 10.49 MJ/m3>9.48 MJ/m3>

4.23 MJ/m3. Representing the maximum amount of heat which can be obtained

from combusting these particular fuels. In comparison, the LHV of diesel, petrol, and

7.6 Critical points 129

kerosene, are at least by factors 3-6 greater. A practical application of biomass-based

gases instead of the mentioned fuels is therefore in question, due to the decrease in

performance and progress, also power output and thermal efficiency.

7.6.3 Low strain rates

Lastly, the utilization of low strain rate flames in combustion will be presented

shortly. Commonly known, a non-premixed flame loses heat to the lean and the rich

side, as opposed to only the preheating zone, as is the case for a premixed flame.

So in comparison, at the same fuel- and air side compositions and velocities, the

diffusion flame extinguishes sooner due to an inflicted strain rate.

GGL2 has the maximum strain rate of the investigated fuel mixtures. For the basic

mixture, a maximum for velocities of 2200 cms-1 of 4612.1 s−1was numerically de-

rived. This is mainly based on the high fraction of hydrogen in the fuel composition.

GGL3 only has, due to the significant proportion of nitrogen, a maximum strain rate

of 318.1 s−1at fuel- and air side velocities of 148 cms-1. PG lays, as it has been for all

the investigations, moderately between the two gasification gases with a maximum

strain rate of 918.1 s−1at 442 cms-1. These correspondingly low strain rates lead to

issues regarding quenching and premature extinction in industrial utilizations.

Chapter 8

Summary and outlook

The aim of this investigation was to validate a model for laminar one-dimensional

counter-flow diffusion flames combusting biomass-based synthetic fuels composed of

N2, H2, CO, CO2, CH4, and O2and, furthermore, to gain a deeper understanding

of the combustion behavior of these fuels with and without oxygen enhancement.

A counter-flow burner system was designed and built to perform combustion ex-

periments. To analyze the flames, several spectroscopic techniques were applied,

such as laser-induced fluorescence measurements of formaldehyde and laser-induced

Rayleigh scattering. The latter was used to deduce temperature evolutions for differ-

ent fuels, velocities, and compositions. Finally, CH* chemiluminescence experiments

were performed, illuminating the combustion mechanisms of these biomass-based

mixtures from multiple perspectives.

Moreover, multiple individual post-processing techniques were established, also

within the group, and applied to finally facilitate an opportunity to practically com-

pare the numerically calculated data with raw experimental results accurately.

Laser-induced Rayleigh scattering data were collected experimentally and processed

via RAYFIT to fit the numerical data that was calculated. DIFFLA provided the

chemical composition at each point of the centerline of the flames. When includ-

ing the depolarization ratios and scattering cross sections for all involved chemical

species, the numerically predicted Rayleigh scattering signals at the centerlines could

be calculated. To calculate temperature evolutions from Rayleigh scattering data,

the exact chemical composition at a measuring point is necessary. With the available

experimental set-up, reproducible and accurate laser-induced Raman measurements

of such kind were not feasible; thus the comparability of the numerical and exper-

imental Rayleigh data was used for validation purposes. Finally, the temperature

curves derived from the numerical model were considered.

132 Chapter 8 Summary and outlook

Positioning deviations in minimum intensities were <0.5mm for the hydrocarbon

cases, which presents an excellent agreement between numerical and experimental

data. The same was the case for the biomass-based synthetic fuels, with one excep-

tion for a PG and GGL2 fuel case. The scattering intensities were consistently higher

when comparing the experimental results to the model for all fuels and velocities

regarded, leading to deviations of up to 30 %. None of the other experimental results

from formaldehyde or chemiluminescence measurements showed such disagreements

in just one direction, steadily. The root causes are based on the inconsistencies in the

flames widths, the large differences in Rayleigh cross sections for different species,

and the spectroscopic set-up. The flames widths were consistently up to <1mm

broader when regarding the experimental results, leading to differences in chemical

composition when comparing to numerical solutions at specific measuring points.

The previously mentioned large differences in cross sections lead, furthermore, to

significant discrepancies in scattering intensities between model and experiment. It

is presumed that these deviations are much smaller; this can be based on the exper-

imental results from laser-induced fluorescence measurements.

A possibility to collect the signal intensity for varying fluorescence intensities during

laser-induced fluorescence measurements of formaldehyde for various fuel composi-

tions was investigated during this investigation. Challenges occurred due to smaller

and larger hydrocarbon bonds emitting fluorescence at similar wavelengths and, in

addition, formaldehyde emitting not only at specific wavelengths but creating a band

across >100 nm. A Boltzmann correction, as presented in Chapter 5, and a fitting

of the experimental fluorescence data via CHFIT had to be carried out. DIFFLA

provided the numerical solution for the formaldehyde intensity at each measuring

point.

The peak developments and intensities for diluted hydrocarbon cases in regard to

formaldehyde intensity, presented an excellent agreement between numerical and

experimental data. Moreover, the flames widths were consistently up to <1mm

broader, which agrees with the previous observations during the Rayleigh scatter-

ing investigations. The biomass-based synthetic fuels showed excellent agreement

for the PG and GGL2 cases without an oxygen enhancement, a poorer agreement

for the GGL3 cases. GGL3 included the largest fraction of diluents in the fuel,

in comparison to the two other fuels, actually with an ideal temperature distribu-

tion for a formaldehyde evolution. Deviations could again have their origin in the

flame widths: differing Rayleigh scattering results due to deviating widths lead to

differences in temperature at a measuring point. This leads to deviations in the

133

Boltzmann corrections and, therefore, experimental fluorescence solutions. This in-

terpretation could have also led to the LIF solutions for the oxygen-enhanced BDG

fuels, with consistently more distinct peak separations between the multiple peaks

for the experimental solutions.

In general, the laser-induced fluorescence measurements showed good agreement.

Accordingly, one can assume that the experimental and numerical Rayleigh scatter-

ing solutions are not deviating up to 30 %. If the temperature evolution throughout

the flames during the experiments were so deviant from the modelled solutions, the

fluorescence measurements and calculations would not be so well related.

Lastly, the peak positions of the experimental CH* chemiluminescence data were

compared to numerically calculated CH curves for all flames at all parameters men-

tioned. An off-set of physical positioning <0.5mm again showed a very good

agreement for both hydrocarbon and biomass-based synthetic fuels. Peak widths

could only be regarded cautiously due to limitations of the spectroscopic measuring

set-up and the therefore measured CH* solutions. The focus of the CH investiga-

tions was the validation of the modelled CH peak positioning, which is still feasible

due to the close relations of the positioning of the ground and excited states of the

CH radical.

Furthermore, the presented system comprised a range of flames combusting at strain

rates starting at around 60 s-1, leading up to circa 250 s-1. Therefore incorporating

the boundary condition at very low fuel- and oxidizer velocities for combustion,

though not including the other limiting condition, at a final strain out. Diluted

hydrocarbon fuels, as well as biomass-based synthetic fuels composed of multiple

components, were examined based on physical positioning between the burner noz-

zles for all spectroscopic techniques, Rayleigh scattering minima and intensity evolu-

tions, and laser-induced fluorescence peak maxima and intensity evolutions, leading

to good agreements between modelled and experimentally derived data and there-

fore validating the numerical model previously developed. As far as the author is

aware, such extensive characterizations between experimental and numerical inves-

tigations for these biomass-based synthetic fuels were presented for the first time in

the present work.

The combustion behavior of the GGL2, GGL3, and PG fuels in regard to formalde-

hyde distribution, CH*/CH evolution, and Rayleigh scattering, and therefore tem-

perature development, has been thoroughly discussed in Chapter 7. Moreover,

the oxygen enhancement on fuel or oxidizer sides had a significant effect on the

Rayleigh scattering intensities and formaldehyde distribution within the flames. Of

134 Chapter 8 Summary and outlook

special interest were the temperature evolutions. A maximum temperature increase

of 11.63 % at 8.5 vol.-% of oxygen enhancement on the air side could be reached for

a GGL2 fuel. A GGL3 fuel, with the lowest LHV of all fuels regarded, reached a

temperature increase of 9.99 % at 5 vol.-% of oxygen enhancement on the fuel side.

An enhancement with pure oxygen or less expensive air, while accepting a higher

fraction of nitrogen in the fuel, is strongly suggested for these biomass-based fu-

els. Leading to further utilization in industrial applications or to refine potentially

problematic fuels by reducing corrosive elements.

8.1 Future work

An outlook regarding future developments of this investigation is based on multiple

aspects, primarily based on the experimental set-up and the variations of process

conditions.

Multiple drawbacks, based on the limitations of the available experimental test rig,

can be advanced to further develop the experimental investigations. The diameters

of the burner nozzles could be expanded, leading to stabilized flames and further

possibilities of fuel- and oxidizer velocity increases. Conjoined with this change in

burner dimension, an enlargement of the laser laboratory gas provisioning systems

and safety installations would need to be undertaken.

All in all, this would present the possibility of investigating combustion processes

closer to strain out circumstances. Furthermore, this could enlarge the variety of

biomass-based synthetic fuels that should be investigated, for example, with higher

fractions of carbon monoxide or otherwise higher simultaneous portions of oxygen

and hydrogen in the fuels. An analysis of an unknown fuel mixture or possibly

an online measurement directly after an industrial application seems difficult when

applying measurement techniques at the current state.

The experimental spectroscopic system could be extended with further laser sys-

tems, including the possible utilization of other wavelengths to excite more chemical

compounds in these hydrocarbon flames, such as OH, and therefore explore the

combustion more in-depth or improve the current measurements.

The numerical solution could be expanded to include more data in regard to the

buoyancy behavior of counter-flow flames and, furthermore, elaborate quenching

assumptions.

Appendix A

Appendix

In the following Appendix A, additional information regarding the investigated fuels,

including fuel- and oxidizer-velocities and strain rates, gases, images of the set-up,

and equipment lists are presented.

136 Chapter A Appendix

A.1 Fuels

In the following section, additional information regarding the investigated fuels is

listed.

Table A.1: Investigated fuels with fuel- and oxidizer-velocities and strain rates for biomass-based

mixtures

Name of fuel Velocities

[cms−1]

Strain rates

[s−1]

GGL2 30, 60, 100 66.1, 132.1, 220.1

GGL2-2.5-O2-F 30, 60, 100 66.1, 132.1, 220.1

GGL2-5-O2-F 30, 60, 100 66.1, 131.1, 219.1

GGL2-8.5-O2-F 30, 60, 100 65.1, 131.1, 218.1

GGL2-2.5-O2-A 30, 60, 100 66.1, 132.1, 220.1

GGL2-5-O2-A 30, 60, 100 66.1, 132.1, 220.1

GGL2-8.5-O2-A 30, 60, 100 66.1, 132.1, 220.1

GGL3 30, 60, 100 64.1, 128.1, 214.1

GGL3-2.5-O2-F 30, 60, 100 64.1, 128.1, 213.1

GGL3-5-O2-F 30, 60, 100 64.1, 128.1, 213.1

GGL3-8.5-O2-F 30, 60, 100 64.1, 128.1, 213.1

GGL3-2.5-O2-A 30, 60, 100 64.1, 129.1, 214.1

GGL3-5-O2-A 30, 60, 100 64.1, 129.1, 214.1

GGL3-8.5-O2-A 30, 60, 100 64.1, 129.1, 214.1

PG 30, 60, 100 62.1, 125.1, 208.1

PG-2.5-O2-F 30, 60, 100 62.1, 125.1, 208.1

PG-5-O2-F 30, 60, 100 62.1, 125.1, 208.1

PG-8.5-O2-F 30, 60, 100 62.1, 125.1, 208.1

PG-2.5-O2-A 30, 60, 100 62.1, 125.1, 208.1

PG-5-O2-A 30, 60, 100 62.1, 125.1, 208.1

PG-8.5-O2-F 30, 60, 100 62.1, 125.1, 208.1

A.1 Fuels 137

Table A.2: Investigated fuels with fuel- and oxidizer-velocities and strain rates for diluted methane

mixtures

Name of fuel Velocities

[cms−1]

Strain rates

[s−1]

CH4-001-CO2 60 126.1

CH4-002-CO2 60 123.1

CH4-003-CO2 60 122.1

CH4-004-CO2 60 122.1

CH4-001-CO2-O2-F 60 126.1

CH4-002-CO2-O2-F 60 123.1

CH4-003-CO2-O2-F 60 122.1

CH4-004-CO2-O2-F 60 121.1

CH4-001-CO2-O2-A 60 126.1

CH4-002-CO2-O2-A 60 123.1

CH4-003-CO2-O2-A 60 122.1

CH4-004-CO2-O2-A 60 122.1

CH4-002-N2 60 134.1

CH4-003-N2 60 134.1

CH4-004-N2 60 133.1

CH4-005-N2 60 133.1

CH4-006-N2 60 132.1

CH4-007-N2 60 132.1

CH4-002-N2-O2-F 60 134.1

CH4-003-N2-O2-F 60 133.1

CH4-004-N2-O2-F 60 133.1

CH4-005-N2-O2-F 60 132.1

CH4-006-N2-O2-F 60 132.1

CH4-007-N2-O2-F 60 131.1

CH4-002-N2-O2-A 60 135.1

CH4-003-N2-O2-A 60 134.1

CH4-004-N2-O2-A 60 133.1

CH4-005-N2-O2-A 60 133.1

CH4-006-N2-O2-A 60 132.1

CH4-007-N2-O2-A 60 132.1

138 Chapter A Appendix

A.2 Experimental set-up

Figure A.1: Side view of experimental set-up, including laser, periscopes, and lenses [23]

Figure A.2: Arrangement of the lenses as schematically repre-

sented in Figure 5.6 [23]

Figure A.3: Periscopes

for spectroscopic measure-

ments [23]

A.3 Equipment list 139

A.3 Equipment list

In the following section, the main equipment and components of the experimental

measurement system, and the gases are listed.

A.3.1 Camera, energy monitor, and electrical equipment

Table A.3: List of the main components for the spectroscopic electrical system

Equipment Type Manufacturer

Nanostar ICCD camera SN 1010630203 LaVision

Telephoto lens EF 100mm f/2 USM Canon

Lens NL-3 B+W

Macro intermediate ring DG-C Viltrox

Online energy monitor 1108005 LaVision

System computer 1104009 LaVision

Synchronisation PTU X 1108092 LaVision

Software DaVis 8 1105106 LaVision

Powermeter P/N 7Z01200 Ophir

A.3.2 Gas mixing

Table A.4: List of the main components for the gas mixing system

Equipment Type Manufacturer

Gas mixing chamber Custom made EVUR

Red-y MFC Air GSC-C9TA-BB12 Vögtlin

Red-y MFC Nitrogen GSC-C9TA-BB12 Vögtlin

Red-y MFC Nitrogen GSC-C9TA-BB26 Vögtlin

Red-y MFC Methane GSC-C9TA-BB26 Vögtlin

Red-y MFC Carbon monoxide GSC-C9TA-BB12 Vögtlin

Red-y MFC Carbon monoxide GSC-C9KA-BB26 Vögtlin

Red-y MFC Carbon dioxide GSC-C9TA-BB12 Vögtlin

Red-y MFC Carbon dioxide GSC-C9TA-BB26 Vögtlin

Red-y MFC Hydrogen GSC-B9SA-BB23 Vögtlin

Red-y MFC Oxygen GSC-C9TA-BB12 Vögtlin

140 Chapter A Appendix

A.3.3 Optical lenses, mirrors, filters, and equipment

Table A.5: List of the main components for the spectroscopic system

Equipment Type Manufacturer

Broadband dielectric mirror 10Q20BB.1 Newport

Planoconvexe round lens LA4663-UV Thorlabs

Planoconvexe round lens LA4184-UV Thorlabs

Planoconcave cylindrical lens LK1743L1-A Thorlabs

Planoconcave cylindrical lens LJ4147-UV Thorlabs

Planoconcave cylindrical lens LK1419L1 Thorlabs

Heat absorption glass filter KG-3/R500013 5101-14237 pgo/Schott

Optical glass filter GG 385/14600171764 pgo/Schott

Neutral density filter NE2R01A Thorlabs

Neutral density filter NE2R05A Thorlabs

Neutral density filter NE2R10A Thorlabs

Neutral density filter NE2R20A Thorlabs

Brightline HC Filter FF01-417/60 Semrock

Brightline HC Filter FF01-523/3 Semrock

Brightline HC Filter FF01-430/10 Semrock

Brightline HC Filter FF01-415/10 Semrock

Periscope RS99 Thorlabs

Kinematic holder for optics KM100 Thorlabs

Adjustable holder for round optics LH160C Thorlabs

Adjustable holder for cyl. optics CYLCP Thorlabs

Rotatable holder for optics RSP2C Thorlabs

Photodetector DET10A Thorlabs

Polarizer LPVISE200-A Thorlabs

A.3.4 Gases

Table A.6: List of the utilized gases

Gas Type Manufacturer

Nitrogen Technical, Alphagaz 2 Air Liquide

Air Alphagaz 2 Air Liquide

Carbon monoxide N25 Air Liquide

Hydrogen N50 Air Liquide

Carbon dioxide N45 Air Liquide

Methane N55 Air Liquide

Oxygen N55 Air Liquide

A.3 Equipment list 141

A.3.5 Counter-flow burner

Figure A.4: Inside and outside view of in-house built counter-flow diffusion burner

Figure A.5: Inside view of in-house built counter-flow diffusion burner with dimensions

Table A.7: List of the main components for the counter-flow burner system

Equipment Type Manufacturer

Motorized Translation Stage MLJ150/M Thorlabs

142 Chapter A Appendix

A.3.6 Laser

The Spectra Physics pulsed Nd:YAG laser Lab 150-10 was used for these investiga-

tions,

Figure A.6: Specifications of Spectra Physics pulsed Nd:YAG lasers [23]

A.3 Equipment list 143

Figure A.7: Further specifications of Spectra Physics pulsed Nd:YAG lasers [23]

Appendix B

Publications

In the following Appendix B, publications and presentations in regard to these in-

vestigations are listed.

146 Chapter B Publications

B.1 Peer-reviewed journals

Scharl, M.-T.; Greenhalgh, D.; Dieguez-Alonso, A.; Behrendt, F.; Numerical and Ex-

perimental Investigation of Laminar One-Dimensional Counter-Flow Flames Using

Product Gas From Pyrolysis and Gasification of Woody Biomass, Eurasian Chemico-

Technological Journal 20 (2018).

B.2 Presentations

Scharl, M.-T.; Greenhalgh, D.; Dieguez-Alonso, A.; Behrendt, F.; Numerical and

experimental studies of laminar counter-flow diffusion flames using biomass-based

gaseous fuels; DGMK-Fachbereichstagung Thermochemische Konversion – Schlüs-

selbaustein für zukünftige Energie-und Rohstoffsysteme, Dresden (Deutschland), 23-

24. Mai 2019 - Awarded as best scientific poster of the conference

Scharl, M.-T.; Greenhalgh, D.; Dieguez-Alonso, A.; Behrendt, F.; Numerical and

experimental studies of laminar counter-flow diffusion flames using biomass-based

gaseous fuels; 41st Meeting of the Italian Section of the Combustion Institute, Sor-

rento (Italien), 23.-26. Mai 2018

Scharl, M.-T.; Greenhalgh, D.; Dieguez-Alonso, A.; Behrendt, F.; Numerical and

experimental studies of laminar counter-flow diffusion flames using low-entalphy

fuels; 17th International Symposium on Transport Phenomena and Dynamics of

Rotating Machinery, Maui (USA), 16.-21. Dezember 2017

Bibliography

[1] S. N. Naik, V. V. Goud, P. K. Rout, and A. K. Dalai. Production of first and

second generation biofuels: A comprehensive review. Renewable and Sustain-

able Energy Reviews, 14:578–597, 2010.

[2] T. Damartzis and A. Zabaniotou. Thermochemical conversion of biomass to

second generation biofuels through integrated process design—A review. Re-

newable and Sustainable Energy Reviews, 15:366–378, 2011.

[3] P. McKendry. Energy production from biomass (part 2): conversion technolo-

gies. Bioresource Technology, 83:47–54, 2002.

[4] A. Dieguez Alonso. Fixed-bed biomass pyrolysis: mechanisms and biochar

production. PhD thesis, Technische Universität Berlin, 2015.

[5] M. Balat. Mechanisms of Thermochemical Biomass Conversion Processes. Part

1: Reactions of Pyrolysis. Energy Sources, Part A: Recovery, Utilization, and

Environmental Effects, 30:620–635, 2008.

[6] C. A. Koufopanos, N. Papayannakos, G. Maschio, and A. Lucchesi. Modelling

of the Pyrolysis of Biomass Particles. Studies on Kinetics, Thermal and Heat

Transfer Effects. The Canadian Journal of Chemical Engineering, 69(4):907–

915, 1991.

[7] J. Cheng, editor. Biomass to Renewable Energy Processes. CRC Press, 2017.

[8] R. C. Brown, editor. Thermochemical Processing of Biomass: Conversion into

Fuels, Chemicals and Power, volume 2. John Wiley and Sons, Inc., 2019.

[9] N. Peters. Fifteen Lectures on Laminar and Turbulent Combustion. Ercoftac

Summer School; Aachen, Germany, 14.-28. September 1992.

[10] C. K. Law. Combustion Physics. Cambridge University Press, 2006.

[11] A. Mukhopadhyay and S. Sen. Fundamentals of Combustion Engineering.

CRC Press, 2019.

[12] D. Veynante and L. Vervisch. Turbulent combustion modeling. Progress in

Energy and Combustion Science, 28:193–266, 2002.

148 Bibliography

[13] A. E. Karatas. High-Pressure Soot Formation and Diffusion Flame Extinc-

tion Characteristics of Gaseous and Liquid Fuels. PhD thesis, University of

Toronto, 2014.

[14] H. Tsuji. Counterflow Diffusion Flames. Progress in Energy and Combustion

Science, 8:93–119, 1982.

[15] D. A. Skoog, F. J. Holler, and S. R. Crouch, editors. Principles of Intrumental

Analysis. Cengage Learning, 7 edition, 2007.

[16] A. Jablonski. Efficiency of Anti-Stokes Fluorescence in Dyes. Nature, 131:839–

840, 1933.

[17] D. C. Harris, editor. Quantitaive Chemical Analysis. W. H. Freeman and

Company, 8 edition, 2010.

[18] B. Valeur and M. N. Berberan-Santos, editors. Molecular Fluorescence: Prin-

ciples and Applications. Wiley-VCH Verlag, 2 edition, 2012.

[19] D. F. G. Durao, M. V. Heitor, J. H. Whitelaw, and P. O. Witze, editors. Com-

busting Flow Diagnostics. NATO ASI Series. Springer, Dordrecht, 1 edition,

1992.

[20] T. Kathrotia, U. Riedel, A. Seipel, K. Moshammer, and A. Brockhinke. Ex-

perimental and numerical study of chemiluminescent species in low-pressure

flames. Applied Physics B, 107:571–584, 2012.

[21] M. M. Kamal. Two-line (CH*/CO2*) chemiluminescence technique for equiv-

alence ratio mapping in turbulent stratified flames. Energy, 192, 2020.

[22] Holthuis and Associates Flat Flame Burners. Website: www.flatflame.com;

last accessed: 25.07.2021. 2013.

[23] A. Kobusinski. Aufbau und Einsatz eines Laser-Diagnostik-Systems zur

spektroskopischen Untersuchung des Verbrennungsverhaltens gasförmiger

Brennstoffe in laminaren Flammen. Master’s thesis, Technische Universität

Berlin, 2015.

[24] W. Radloff. Laser in Wissenschaft und Technik. Spektrum Akademischer

Verlag, 2010.

[25] LaVision GmbH. Nanostar Camera System Operation Manual. LaVision

GmbH, 1999.

[26] D. C. Kyritsis, V. S. Santoro, and A. Gomez. The effect of temperature

correction on the measured thickness of formaldehyde zones in diffusion flames

for 355 nm excitation. Experiments in Fluids, (37):769–772, 2004.

Bibliography 149

[27] Spectra Physics. Quanta-Ray Lab-Series - pulsed Nd:YAG Lasers. Spectra

Physics, 1335 Terra Bella Avanue; Mountain View, CA 94043, June 2003.

[28] P. Basu. Biomass Gasification and Pyrolysis: Practical Design. Academic

Press, 2010.

[29] D. L. Klass. Biomass for Renewable Energy, Fuels, and Chemicals. Academic

Press, 1 edition, 1998.

[30] M. Brebu and C. Vasile. Thermal Degradation of Lignin — A Review. Cellu-

lose Chemistry and Technology, 44(9):353–363, 2010.

[31] P. McKendry. Energy production from biomass (part 3): gasification tech-

nologies. Bioresource Technology, 83:55–63, 2002.

[32] J. Warnatz, U. Maas, and R.W. Dibble. Combustion: Physical and Chemical

Fundamentals, Modeling and Simulation, Experiments, Pollutant Formation.

Springer, 4 edition, 2006.

[33] K. Seshadri and N. Peters. Asymptotic Structure and Extinction of Methane-

Air Diffusion Flames. Combustion and Flame, 73:23–44, 1988.

[34] G. Sutton, A. Levick, G. Edwards, and D. Greenhalgh. A combustion temper-

ature and species standard for the calibration of laser diagnostic techniques.

Combustion and Flame, 147:39–48, 2006.

[35] Energy Union. Website: https://ec.europa.eu/energy/topics/energy-

strategy/energy-union; last accessed: 25.07.2021. 06 2020.

[36] M. Ringel. Fostering the use of renewable energies in the European Union: the

race between feed-in tariffs and green certificates. Renewable Energy, 31:1–17,

2006.

[37] R. F. Sawyer. Science based policy for addressing energy and environmental

problems. Proceedings of the Combustion Institute, 32:45–56, 2009.

[38] N. Peters. Turbulent Combustion. Cambridge University Press, 2000.

[39] N. Peters. Laminar flamelet concepts in turbulent combustion. Twenty-first

Symposium (International) on Combustion / The Combustion Institute, pages

1231–1250, 1988.

[40] E. N. Kalogirou. Waste-to-Energy Technologies and Global Applications. CRC

Press by Taylor and Francis Group, 2018.

[41] P. McKendry. Energy production from biomass (part 1): overview of biomass.

Bioresource Technology, 83:37–46, 2002.

150 Bibliography

[42] A. Pandey, C. Larroche, S. C. Ricke, C.-G. Dussap, and E. Gnansounou,

editors. Biofuels, Alternative Feedstocks and Conversion Processes. Academic

Press, 1 edition, 2011.

[43] A. Mohr and S. Raman. Lessons from first generation biofuels and implications

for the sustainability appraisal of second generation biofuels. Energy Policy,

63:114–122, 2013.

[44] M. A. Carriquiry, X. Du, and G. R. Timilsina. Second generation biofuels:

Economics and policies. Energy Policy, 39:4222–4234, 2011.

[45] V. B. Agbor, N. Cicek, R. Sparling, A. Berlin, and D. B. Levin. Biomass

pretreatment: Fundamentals toward application. Biotechnology Advances,

29:675–685, 2011.

[46] L. C. R. Sá, L. M. E. F. Loureiro, L. J. R. Nunes, and A. M. M. Mendes.

Torrefaction as a Pretreatment Technology for Chlorine Elimination from

Biomass: A Case Study Using Eucalyptus globulus Labill. Resources, 9(54),

2020.

[47] C. Liu, B. Yan, G. Chen, and X. S. Bai. Structures and burning velocity

of biomass derived gas flames. International Journal of Hydrogen Energy,

35(542-555), 2010.

[48] M. A. Leon, Md. M. Rahman, and S. C. Bhattacharya. A Study on Improved

Biomass Briquetting. Energy for Sustainable Development, 6:67–71, 2002.

[49] R. Razuan, K. N. Finney, Q. Chen, V. N. Sharifi, and J. Swithenbank. Pel-

letised fuel production from palm kernel cake. Fuel Processing Technology,

92:609–615, 2011.

[50] A. Faaij. Modern Biomass Conversion Technologies. Mitigation and Adapta-

tion Strategies for Global Change, 11:343–375, 2006.

[51] H. B. Goyal, D. Seal, and R. C. Saxena. Bio-fuels from thermochemical con-

version of renewable sources: A review. Renewable and Sustainable Energy

Reviews, 12:504–517, 2008.

[52] M. Crocker, editor. Thermochemical Conversion of Biomass to Liquid Fuels

and Chemicals. Royal Society of Chemistry, 2010.

[53] H. Chen, editor. Lignocellulose Biorefinery Engineering: Principles and Ap-

plications. Woodhead Publishing Limited, 2015.

[54] D. Mohan, C. U. Pittman Jr., and P. H. Steele. Pyrolysis of Wood/Biomass

for Bio-oil: A Critical Review. Energy and Fuels, 20:848–889, 2006.

[55] P. A. Horne and P. T. Williams. Influence of temperature on the products

from the flash pyrolysis of biomass. Fuel, 75:1051–1059, 1996.

Bibliography 151

[56] L. Zhang, C. Xu, and P. Champagne. Overview of recent advances in thermo-

chemical conversion of biomass. Energy Conversion and Management, 51:969–

982, 2010.

[57] A. Demirbas. Current Technologies for the Thermo-Conversion of Biomass

into Fuels and Chemicals. Energy Sources, 26:715–730, 2004.

[58] M. Balat. Mechanisms of Thermochemical Biomass Conversion Processes. Part

2: Reactions of Gasification. Energy Sources, Part A: Recovery, Utilization,

and Environmental Effects, 30:636–648, 2008.

[59] H. Juentgen. Reactivities of carbon to steam and hydrogen and applications

to technical processes — A review. Carbon, 19:167–173, 1981.

[60] J. Mathieu and J. Scott. An Introduction to Turbulent Flow. Cambridge

University Press, 2000.

[61] P. A. Davidson. Turbulence, An Introduction for Scientists and Engineers.

Oxford University Press, 2004.

[62] B. Rehm, J. Schubert, A. Haghshenas, A. S. Paknejad, and J. Hughes, editors.

Managed Pressure Drilling. Gulf Publishing Company, 2008.

[63] S. R. Turns. An Introduction to Combustion: Concepts and Applications.

Mcgraw-Hill Professional, 2000.

[64] K. K. Kuo. Principles of Combustion. John Wiley and Sons, Inc., 2005.

[65] N. Kubota. Propellants and Explosives: Thermochemical Aspects of Combus-

tion. Wiley-VCH, 2015.

[66] J. C. Kloppers and D. G. Kroeger. The Lewis factor and its influence on

the performance prediction of wet-cooling towers. International Journal of

Thermal Sciences, pages 879–884, 2005.

[67] B. J. Isaac, A. Parente, C. Galletti, J. N. Thornock, P. J. Smith, and L. Tog-

notti. A Novel Methodology for Chemical Time Scale Evaluation with Detailed

Chemical Reaction Kinetics. Energy and Fuels, 27:2255–2265, 2013.

[68] F. Tat Cheong Yuen. Experimental Investigation of the Dynamics and Struc-

ture of Lean-Premixed Turbulent Combustion. PhD thesis, University of

Toronto, 2009.

[69] K. K. Kuo and R. Acharya. Fundamentals of Turbulent and Multiphase Com-

bustion. John Wiley and Sons, Inc., 2012.

[70] N. Peters. Four Lectures on Turbulent Combustion. Ercoftac Summer School;

Aachen, Germany, 15.-19. September 1997.

152 Bibliography

[71] C. Noehre, M. Andersson, B. Johansson, and A. Hultqvist. Characterization

of Partially Premixed Combustion. Powertrain and Fluid Systems Conference

and Exhibition; Toronto, Canada, 16.-19. October 2006.

[72] P. Clavin and G. Searby. Combustion Waves and Fronts in Flows: Flames,

Shocks, Detonations, Ablation Fronts and Explosion of Stars. Cambridge Uni-

versity Press, 2016.

[73] F. A. Williams and S. C. Li. Some Basic Considerations of Pollutant Emis-

sion and Knock in Internal Combustion Engines. SAE 2000 World Congress;

Detroit, USA, 6.-9. March 2000.

[74] H. Tsuji and I. Yamaoka. The counterflow diffusion flame in the forward stag-

nation region of a porous cylinder. Symposium (International) on Combustion,

11:979–984, 1967.

[75] M. D. Smooke, I. K. Puri, and K. Seshadri. A comparison between numerical

calculations and experimental measurements of the structure of a counterflow

diffusion flame burning diluted methane in diluted air. Symposium (Interna-

tional) on Combustion, 21:1783–1792, 1988.

[76] M. D. Smooke, R. A. Yetter, T. P. Parr, D. M. Hanson-Parr, M. A. Tanoff,

M. B. Colket, and R. J. Hall. Computational and experimental study of am-

monium perchlorate/ethylene counterflow diffusion flames. Proceedings of the

Combustion Institute, 28:2013–2020, 2000.

[77] I. K. Puri, K. Seshadri, M. D. Smooke, and D. Keyes. A Comparison Between

Numerical Calculations and Experimental Measurements of the Structure of a

Counterflow Methane-Air Diffusion Flame. Combustion Science and Technol-

ogy, 56:1–22, 1987.

[78] K. Seshadri, C. Trevino, and M. D. Smooke. Analysis of the structure and

mechanisms of extinction of a counterflow methanol-air diffusion flame. Com-

bustion and Flame, 76:111–132, 1989.

[79] G. Amantini, J. H. Frank, M. D. Smooke, and A. Gomez. Computational and

experimental study of steady axisymmetric non-premixed methane counterflow

flames. Combustion Theory and Modelling, 11:47–72, 2007.

[80] L. Figura and A. Gomez. Laminar counterflow steady diffusion flames under

high pressure (P≤3 MPa) conditions. Combustion and Flame, 159:142–150,

2012.

[81] T. Poinsot and D. Veynante. Theoretical and Numerical Combustion. R.T.

Edwards Inc., 2005.

[82] K. Seshadri and F. A.Williams. Laminar flow between parallel plates with

injection of a reactant at high reynolds number. International Journal of Heat

and Mass Transfer, 21:251–253, 1978.

Bibliography 153

[83] H. Tsuji and I. Yamaoka. Structure analysis of counterflow diffusion flames in

the forward stagnation region of a porous cylinder. Symposium (International)

on Combustion, 13:723–731, 1971.

[84] T. P. Pandya and F. J. Weinberg. The structure of flat, counter-flow diffusion

flames. Proceedings of the Royal Society of London, 279:544–561, 1964.

[85] A. E. Karatas. Soot Formation in Co-flow and Counterflow Laminar Diffusion

Flames of Fuel Mixtures. Master’s thesis, University of Toronto, 2009.

[86] R. Borghi and S. N. B. Murthy. Turbulent Reactive Flows. Springer-Verlag

New York Inc., 1989.

[87] K. Bray. Laminar Flamelets in Turbulent Combustion Modeling. Combustion

Science and Technology, 188(9):1372–1375, June 2016.

[88] C. Meneveau and T. Poinsot. Stretching and Quenching of Flamelets in Pre-

mixed Turbulent Combustion. Combustion and Flame, pages 311–332, 1991.

[89] J. A. van Oijen, R. J. M. Bastiaans, G. R. A. Groot, and L. P. H. de Goey.

Direct Numerical Simulations of Premixed Turbulent Flames with Reduced

Chemistry: Validation and Flamelet Analysis. Flow, Turbulence and Combus-

tion, 75:67–84, 2005.

[90] H. Pitsch. A consistent level set formation for large-eddy simulation of pre-

mixed turbulent combustion. Combustion and Flame, 143:587–598, 2005.

[91] M. Ihme, L. Shunn, and J. Zhang. Regularization of reaction progress variable

for application to flamelet-based combustion models. Journal of Computa-

tional Physics, 231:7715–7721, 2012.

[92] F. Halter, F. Foucher, L. Landry, and C. Mounaim-Rousselle. Effect of Dilution

by Nitrogen and/or Carbon Dioxide on Methane and Iso-Octane Air Flames.

Combustion Science and Technology, 181:813–827, 2009.

[93] M. Elia, M. Ulinski, and M. Metghalchi. Laminar Burning Velocity of

Methane-Air-Diluent Mixtures. Journal of Engineering for Gas Turbines and

Power, pages 190–196, 2001.

[94] R. Glowinski, B. Larrouturou, and R. Temam. Numerical Simulation of Com-

bustion Phenomena. Springer, May 1985.

[95] K. Seshadri and N. Peters. The inner structure of methane-air flames. Com-

bustion and Flame, 81:96–118, 1990.

[96] W. P. Jones and R. P. Lindstedt. Global reaction schemes for hydrocarbon

combustion. Combustion and Flame, 73:233–249, 1988.

154 Bibliography

[97] D. J. Hucknall. Chemistry of Hydrocarbon Combustion. Chapman and Hall,

1985.

[98] C.-H. Hwang, C. B. Oh, and C.-E. Lee. Effects of CO2dilution on the inter-

actions of a CH4-air nonpremixed jet flame with a single vortex. International

Journal of Thermal Sciences, 48:1423–1431, 2009.

[99] A. E. Karatas and Ö. L. Gülder. Effects of carbon dioxide and nitrogen ad-

dition on soot processes in laminar diffusion flames of ethylene-air at high

pressures. Fuel, 200:76–80, 2017.

[100] D. X. Du, R. L. Axelbaum, and C. K. Law. The influence of carbon dioxide

and oxygen as additives on soot formation in diffusion flames. Twenty-Third

Symposium (International) on Combustion / The Combustion Institute, pages

1501–1507, 1990.

[101] C. Serrano, J. J. Hernandez, C. Mandilas, C. G. W. Sheppard, and R. Wool-

ley. Lamniar burning behaviour of biomass gasification-derived producer gas.

International Journal of Hydrogen Energy, 33:851–862, 2008.

[102] T. Hanaoka, S. Inoue, S. Uno, T. Ogi, and T. Minowa. Effect of woody biomass

components on air-steam gasification. Biomass and Bioenergy, 28:69–76, 2005.

[103] P. Lv, Z. Yuan, L. Ma, C. Wu, Y. Chen, and J. Zhu. Hydrogen-rich gas

production from biomass air and oxygen/steam gasification in a downdraft

gasifier. Renewable Energy, 32:2173–2185, 2007.

[104] K. Stahl and M. Neergaard. IGCC power plant for biomass utilisation, var-

namo, sweden. Biomass and Bioenergy, 15:205–211, 1998.

[105] FAO (Food and Agricultural Organization of the United Nations). Wood gas

as engine fuel. FAO Forestry Papers 72, 1986.

[106] C. Dupont, J.-M. Commandre, P. Gauthier, G. Boissonnet, S. Salvador, and

D. Schweich. Biomass pyrolysis experiments in an analytical entrained flow

reactor between 1073 K and 1273 K. Fuel, 87:1155–1164, 2008.

[107] J. Lede, F. Broust, F.-T. Ndiaye, and M. Ferrer. Properties of bio-oils produced

by biomass fast pyrolysis in a cyclone reactor. Fuel, 86:1800–1810, 2007.

[108] H. Luik, I. Johannes, V. Palu, L. Luik, and K. Kruusement. Transformations

of biomass internal oxygen at varied pyrolysis conditions. Journal of Analytical

and Applied Pyrolysis, 79:121–127, 2007.

[109] A. K. Hossain and P. A. Davies. Pyrolysis liquids and gases as alternative

fuels in internal combustion engines — A review. Renewable and Sustainable

Energy Reviews, 21:165–189, 2013.

Bibliography 155

[110] J. Rezaiyan and N. P. Cheremisinoff, editors. Gasification Technologies: A

Primer for Engineers and Scientists. CRC Press by Taylor and Francis Group,

2005.

[111] M. Balat. Gasification of Biomass to Produce Gaseous Products. Energy

Sources, Part A: Recovery, Utilization, and Environmental Effects, 31:516–

526, 2009.

[112] H. A. Yepes and A. A. Amell. Laminar burning velocity with oxygen-enriched

air of syngas produced from biomass gasification. International Journal of

Hydrogen Energy, 38:7519–7527, 2013.

[113] H. Bockhorn, F. Fetting, and H. W. Wenz. Investigation of the Formation

of High Molecular Hydrocarbons and Soot in Premixed Hydrocarbon-Oxygen

Flames. Berichte der Bunsengesellschaft Für Physikalische Chemie, 87:1067–

1073, 1983.

[114] P. Baskar and A. Senthilkumar. Effects of oxygen enriched combustion on pol-

lution and performance characteristics of a diesel engine. Engineering Science

and Technology, an International Journal, 2015.

[115] M. Alden, J. Bood, Z. Li, and M. Richter. Visualization and understanding of

combustion processes using spatially and temporally resolved laser diagnostic

techniques. Proceedings of the Combustion Institute, 33:69–97, 2011.

[116] Z. Yang, X. Yu, J. Peng, and J. Zhang. Laser Technology and its Applications,

chapter 5. IntechOpen, 2019.

[117] M. Baudelet, editor. Laser spectroscopy for sensing: Fundamentals, Techniques

and Applications. Woodhead Publishing Limited, 2014.

[118] D. Frackowiak. The Jablonski diagram. Journal of Photochemistry and Pho-

tobioligy, 2:399–408, 1988.

[119] J. Zimmermann, A. Zeug, and B. Röder. A generalization of the Jablonski

diagram to account for polarization and anisotropy effects in time-resolved

experiments. Physical Chemistry Chemical Physics, 5:2964–2969, 2003.

[120] H. H. Telle, A. Gonzalez Urena, and R. J. Donovan, editors. Laser Chemistry:

Spectroscopy, Dynamics and Applications. John Wiley and Sons, Inc., 2007.

[121] W. W. Parson, editor. Modern Optical Spectroscopy: With Exercises and

Examples from Biophysics and Biochemistry. Springer, 2 edition, 2015.

[122] P. Atkins and R. Friedman, editors. Molecular Quantum Mechanics. Oxford

University Press, 4 edition, 2005.

[123] J. R. Lakowicz, editor. Principles of Fluorescence Spectroscopy. Springer, 3

edition, 2006.

156 Bibliography

[124] F.-Q. Zhao and H. Hiroyasu. The applications of laser Rayleigh scattering to

combustion diagnostics. Progress in Energy and Combustion Science, 19:447–

485, 1993.

[125] R. B. Miles, W. R. Lempert, and J. N. Forkey. Laser Rayleigh scattering.

Measurement Science and Technology, 12:33–51, 2001.

[126] R. Thalman, K. J. Zarzana, M. A. Tolbert, and R. Volkamer. Rayleigh scat-

tering cross-section measurements of nitrogen, argon, oxygen and air. Journal

of Quantitative Spectroscopy and Radiative Transfer, 2014.

[127] W. C. Gardiner Jr., Y. Hidaka, and T. Tanzawa. Refractivity of Combustion

Gases. Combustion and Flame, 40:213–219, 1981.

[128] J. Fielding, J. H. Frank, S. A. Kaiser, M. D. Smooke, and M. B. Long. Po-

larized/depolarized Rayleigh scattering for determining fuel concentrations in

flames. Proceedings of the Combustion Institute, 29:2703–2709, 2002.

[129] C. H. Kautz, P. R. L. Heron, M. E. Loverude, and L. C. McDermott. Stu-

dent understanding of the ideal gas law, Part I: A macroscopic perspective.

American Journal of Physics, 73:1055–1063, 2005.

[130] J. A. Sutton and J. F. Driscoll. Rayleigh scattering cross sections of combustion

species at 266, 355, and 532 nm for thermometry applications. Optics Letters,

29(22):2620–2622, 2004.

[131] A. M. Garcia-Campana and W. R. G. Baeyens, editors. Chemiluminescence

in Analytical Chemistry. Marcel Dekker Inc., 2001.

[132] C. Dodeigne, L. Thunus, and R. Lejeune. Chemiluminescence as diagnostic

tool. A review. Talanta, 51:415–439, 2000.

[133] V. N. Nori and J. M. Seitzman. CH* chemiluminescence modeling for combus-

tion diagnostics. Proceedings of the Combustion Institute, 32:895–903, 2009.

[134] V. N. Nori and J. M. Seitzman. Evaluation of Chemiluminescence as a

Combustion Diagnostic under Varying Operating Conditions. 46th AIAA

Aerospace Sciences Meeting and Exhibit; Reno, USA, 7.-10. January 2008.

[135] K. T. Walsh, M. B. Long, M. A. Tanoff, and M. D. Smooke. Experimental

and computational study of CH, CH*, and OH* in an axisymmetric laminar

diffusion flame. Symposium (International) on Combustion, 27:615–623, 1998.

[136] M. De Leo, A. Saveliev, L. A. Kennedy, and S. A. Zelepouga. OH and CH

luminescence in opposed flow methane oxy-flames. Combustion and Flame,

149:435–447, 2007.

Bibliography 157

[137] P. Nau, J. Krüger, A. Lackner, M. Letzgus, and A. Brockhinke. On the

quantification of OH*, CH*, and C2* chemiluminescence in flames. Applied

Physics B, 107:551–559, 2012.

[138] A. G. Gaydon and H. G. Wolfhard. Flames: Their Structure, Radiation, and

Temperature. Chapman and Hall, 4 edition, 1979.

[139] G. P. Glass, G. B. Kistiakowsky, J. V. Michael, and H. Niki. The oxidation re-

actions of acetylene and methane. Symposium (International) on Combustion,

10:513–522, 1965.

[140] J. M. Hall, J. de Vries, A. R. Amadio, and E. L. Petersen. Towards a Kinetics

Model of CH Chemiluminescence. 43rd AIAA Aerospace Sciences Meeting and

Exhibit; Reno, USA, 10.-13. January 2005.

[141] R. M. I. Elsamra, S. Vranckx, and S. A. Carl. CH(A2Delta) Formation in

hydrocarbon combustion: the temperature dependence of the rate constant

of the reaction C2H + O2→CH(A2Delta) + CO2.The Journal of Physical

Chemistry A, 109, 2005.

[142] G. P. Smith; J. Luque; C. Park; J. B. Jeffries; D. R. Crosley. Low pressure flame

determinations of rate constants for OH(A) and CH(A) chemiluminescence.

Combustion and Flame, 131:59–69, 2002.

[143] G. P. Smith, C. Park, and J. Luque. A note on chemiluminescence in low-

pressure hydrogen and methane–nitrous oxide flames. Combustion and Flame,

140:385–389, 2005.

[144] S. Karnani and D. Dunn-Rankin. Visualizing CH* chemiluminescence in soot-

ing flames. Combustion and Flame, 160:2275–2278, 2013.

[145] C. S. Panoutsos, Y. Hardalupas, and A. M. K. P. Taylor. Numerical evalua-

tion of equivalence ratio measurement using OH* and CH* chemiluminescence

in premixed and non-premixed methane–air flames. Combustion and Flame,

156:273–291, 2009.

[146] D. Giassi, S. Cao, B. A. V. Bennett, D. P. Stocker, F. Takahashi, M. D.

Smooke, and M. B. Long. Analysis of CH* concentration and flame heat

release rate in laminar coflow diffusion flames under microgravity and normal

gravity. Combustion and Flame, 167:198–206, 2016.

[147] A. Hossain and Y. Nakamura. A numerical study on the ability to predict

the heat release rate using CH* chemiluminescence in non-sooting counterflow

diffusion flames. Combustion and Flame, 161:162–172, 2014.

[148] T. F. Guiberti, D. Durox, and T. Schuller. Flame chemiluminescence from CO2

- and N2-diluted laminar CH4/air premixed flames. Combustion and Flame,

181:110–122, 2017.

158 Bibliography

[149] R. Klein and L. J. Schoen. Role of Formaldehyde in Combustion. Literature

of the Combustion of Petroleum, pages 58–68, 1958.

[150] C. Brackmann, J. Nygren, X. Bai, Z. Li, H. Bladh, B. Axelsson, I. Denbratt,

L. Koopmans, P.-E. Bengtsson, and M. Alden. Laser-induced fluorescence of

formaldehyde in combustion using third harmonic Nd:YAG laser excitation.

Spectrochimica Acta Part A, 59:3347–3356, 2003.

[151] B. Lewis and G. von Elbe, editors. Combustion, Flames and Explosions of

Gases. Academic Press Inc., 3 edition, 1987.

[152] X. Bai, T. Metz, F. Ossler, and M. Alden. Absorption of formaldehyde (H2CO)

in the ˜

A1A2←˜

X1A1band system at elevated temperatures and pressures.

Spectrochimica Acta Part A, 60:821–828, 2004.

[153] T. Metz, X. Bai, F. Ossler, and M. Alden. Fluorescence lifetimes of formalde-

hyde (H2CO) in the ˜

A1A2→˜

X1A1band system at elevated temperatures and

pressures. Spectrochimica Acta Part A, 60:1043–1053, 2004.

[154] J. Kiefer, Z. S. Li, T. Seeger, A. Leipertz, and M. Alden. Planar laser-induced

fluorescence of HCO for instantaneous flame front imaging in hydrocarbon

flames. Proceedings of the Combustion Institute, 32:921–928, 2009.

[155] K. Bijjula and D. C. Kyritsis. Experimental evaluation of flame observables for

simplified scalar dissipation rate measurements in laminar diffusion flamelets.

Proceedings of the Combustion Institute, 30:493–500, 2005.

[156] K. R. Gosselin, W. F. Carnell Jr., and M. W. Renfro. Formaldehyde Fluores-

cence as a Marker for Scalar Dissipation Through Local Extinction. Combus-

tion Science and Technology, 187:1742–1758, 2015.

[157] K. N. Gabet and J. A. Sutton. Narrowband versus broadband excitation

for CH2O PLIF imaging in flames using a frequency-tripled Nd:YAG laser.

Experiments in Fluids, 55, 2014.

[158] A. Ehn, O. Johansson, J. Bood, A. Arvidsson, B. Li, and M. Alden. Fluo-

rescence lifetime imaging in a flame. Proceedings of the Combustion Institute,

33:807–813, 2011.

[159] C. Brackmann, J. Bood, and M. Alden. Quantitative measurements of species

and temperature in a DME-air counterflow diffusion flame using laser diag-

nostic methods. Combustion Science and Technology, 178:1165–1184, 2006.

[160] M. de Joannon, A. Ciajolo, R. Ragucci, A. Tregrossi, and A. Cavaliere.

Spectroscopic behavior of oxygenated combustion by-products. Chemosphere,

51:1071–1077, 2003.

Bibliography 159

[161] P. R. Medwell, P. A. M. Kalt, and B. B. Dally. Simultaneous imaging of

OH, formaldehyde, and temperature of turbulent nonpremixed jet flames in a

heated and diluted coflow. Combustion and Flame, 148:48–61, 2007.

[162] C. B. Reuter, S. H. Won, and Y. Ju. Flame structure and ignition limit of

partially premixed cool flames in a counterflow burner. Proceedings of the

Combustion Institute, 36, 2017.

[163] C. B. Reuter, S. H. Won, and Y. Ju. Experimental study of the dynamics and

structure of self-sustaining premixed cool flames using a counterflow burner.

Combustion and Flame, 166, 2016.

[164] W. Weng, E. Nilsson, A. Ehn, J. Zhu, Y. Zhou, Z. Wang, Z. Li, M. Alden,

and K. Cen. Investigation of formaldehyde enhancement by ozone addition in

CH4/air premixed flames. Combustion and Flame, 2014.

[165] S. Wang, D. F. Davidson, and R. K. Hanson. High-temperature laser absorp-

tion diagnostics for CH2O and CH3CHO and their application to shock tube

kinetic studies. Combustion and Flame, 160:1930–1938, 2013.

[166] G. Friedrichs, D. F. Davidson, and R. K. Hanson. Validation of a Thermal

Decomposition Mechanism of Formaldehyde by Detection oh CH2O and HCO

Behind Shock Waves. International Journal of Chemical Kinetics, 2004.

[167] H. Bladh, C. Brackmann, P. Dahlander, I. Denbratt, and P.-E Bengtsson.

Flame propagation visualization in a spark-ignition engine using laser-induced

fluorescence of cool-flame species. Measurement Science and Technology,

16:1083–1091, 2005.

[168] C. Brackmann, Z. Li, M. Rupinski, N. Docquier, G. Pengloan, and M. Alden.

Strategies for Formaldehyde Detection in Flames and Engines Using a Single-

Mode Nd:YAG/OPO Laser System. Applied Spectroscopy, 59(6), 2005.

[169] R. Schiessl, P. Pixner, A. Dreizler, and U. Maas. Formaldehyde formation

in the endgas of Otto engines: Numerical simulations and quantitative con-

centration measurements. Combustion Science and Technology, 149:339–360,

1999.

[170] M. de Joannon, R. Ragucci, A. Cavaliere, and A. Ciajolo. Identification of

oxygenated compounds in combustion systems. Chemosphere, 42:843–851,

2001.

[171] G. Bruneaux. Combustion structure of free and wall-impinging diesel jets by

simultaneous laser-induced fluorescence of formaldehyde, poly-aromatic hydro-

carbons, and hydroxides. International Journal of Engine Research, 9, 2008.

[172] R. Collin, J. Nygren, M. Richter, M. Alden, L. Hildingsson, and B. Johansson.

Simultaneous OH- and Formaldehyde-LIF Measurements in an HCCI Engine.

SAE Transactions, Journal of Fuels and Lubricants, 112:2479–2486, 2003.

160 Bibliography

[173] M. Richter, R. Collin, J. Nygren, M. Alden, L. Hildingsson, and B. Johnasson.

Studies of the Combustion Process with Simultaneous Formaldehyde and OH

PLIF in a Direct-Injected HCCI Engine. JSME International Journal, 48(4),

2005.

[174] G. Särner, M. Richter, M. Alden, L. Hildingsson, A. Hultqvist, and B. Johans-

son. Simultaneous PLIF Measurements for Visualization of Formaldehyde- and

Fuel- Distributions in a DI HCCI Engine. SAE Technical Papers 2005-01-3869,

2005.

[175] A. J. Donkerbroek, A. P. van Vliet, L. M. T. Somers, P. J. M. Frijters, R. J. H.

Klein-Douwel, N. J. Dam, W. L. Meerts, and J. J. ter Meulen. Time- and

space-resolved quantitative LIF measurements of formaldehyde in a heavy-

duty diesel engine. Combustion and Flame, 157:155–166, 2010.

[176] A. G. Novoselov, C. B. Reuter, O. R. Yehia, S. H. Won, M. K. Fu, K. Kok-

manian, M. Hultmark, Y. Ju, and M. E. Mueller. Turbulent nonpremixed

cool flames: Experimental measurements, Direct Numerical Simulation, and

manifold-based combustion modeling. Combustion and Flame, 209:144–154,

2019.

[177] S. A. Skeen, J. Manin, and L. M. Pickett. Simultaneous formaldehyde PLIF

and high-speed schlieren imaging for ignition visualization in high-pressure

spray flames. Proceedings of the Combustion Institute, 35:3167–3174, 2015.

[178] T. Shimizu, H. Nakamura, T. Tezuka, S. Hasegawa, and K. Maruta. OH-

and CH2O-LIF measurements for hydrogen flames and methane, n-butane

and dimethyl ether weak flames in a micro flow reactor with a controlled

temperature profile. Energy and Fuels, 2017.

[179] N. Zobel and A. Anca-Couce. Slow pyrolysis of wood particles: Characteriza-

tion of volatiles by Laser-Induced Fluorescence. Proceedings of the Combustion

Institute, 34:2355–2362, 2013.

[180] C. Brackmann, M. Alden, P.-E. Bengtsson, K. O. Davidsson, and J. B. C.

Pettersson. Optical and Mass Spectrometric Study of the Pyrolysis Gas of

Wood Particles. Applied Spectroscopy, 57(2), 2003.

[181] M. J. Prins, Z. S. Li, R. J. M. Bastiaans, J. A. van Oijen, M. Alden, and

L. P. H. de Goey. Biomass pyrolysis in a heated-grid reactor: Visualization of

carbon monoxide and formaldehyde using Laser-Induced Fluorescence. Journal

of Analytical and Applied Pyrolysis, 92:280–286, 2011.

[182] A. Fayoux, K. Zähringer, O. Gicquel, and J. C. Rolon. Experimental and nu-

merical determination of heat release in counterflow premixed laminar flames.

Proceedings of the Combustion Institute, 30:251–257, 2005.

Bibliography 161

[183] R. Yuan, J. Kariuki, A. Dowlut, R. Balachandran, and E. Mastorakos. Reac-

tion zone visualisation in swirling spray n-heptane flames. Proceedings of the

Combustion Institute, 35:1649–1656, 2015.

[184] I. A. Mulla, A. Dowlut, T. Hussain, Z. M. Nikolaou, S. R. Chakravarthy,

N. Swaminathan, and R. Balachandran. Heat release rate estimation in lam-

inar premixed flames using laser-induced fluorescence of CH2O and H-atom.

Combustion and Flame, 165:373–383, 2016.

[185] T.M. Dyer. Rayleigh scattering measurements of time-resolved concentration

in a turbulent propane jet. AIAA, 17, 1979.

[186] M. C. Escoda and M. B. Long. Rayleigh scattering measurements of the gas

concentration field in turbulent jets. AIAA, 21, 1983.

[187] R. W. Dibble and R. E. Hollenbach. Laser Rayleigh thermometry in turbulent

flames. 18th Symp. (Int.) on Combustion (Combustion Institute), 1981.

[188] B. R. Miles and W. R. Lempert. Quantitative flow visualization in unseeded

flows. Annu. Rev. Fluid. Mech., 29:285–326, 1997.

[189] H. Shimizu, S. A. Lee, and C. Y. She. High spectral resolution lidar system

with atomic blocking filters for measuring atmospheric parameters. Appl. Opt.,

22, 1983.

[190] P. Desgroux, L. Gasnot, and L. R. Sochet. Instantaneous temperature mea-

surement in a rapid-compression machine using laser Rayleigh scattering. Ap-

plied Physics B, 61:69–72, 1995.

[191] F. Fuest, R. S. Barlow, J.-Y. Chen, and A. Dreizler. Raman/Rayleigh scatter-

ing and CO-LIF measurements in laminar and turbulent jet flames of dimethyl

ether. Combustion and Flame, 159:2533–2562, 2012.

[192] R. L. Gordon, A. R. Masri, and E. Mastorakos. Simultaneous Rayleigh tem-

perature, OH- and CH2O-LIF imaging of methane jets in a vitiated coflow.

Combustion and Flame, 155:181–195, 2008.

[193] F. T. C. Yuen and Ö. L. Gülder. Premixed turbulent flame front structure

investigation by Rayleigh scattering in the thin reaction zone regime. Proceed-

ings of the Combustion Institute, 32:1747–1754, 2009.

[194] H. Schwarz, L. Zimmer, D. Durox, and S. Candel. Detailed measurements of

equivalence ratio modulations in premixed flames using laser Rayleigh scat-

tering and absorption spectroscopy. Experiments, 49:809–821, 2010.

[195] C. Espey, J. E. Dec, T. A. Litzinger, and D. A. Santavicca. Planar Laser

Rayleigh Scattering for Quantitative Vapor-Fuel Imaging in a Diesel Jet. Com-

bustion and Flame, 109:65–86, 1997.

162 Bibliography

[196] D. Ityaksov, H. Linnartz, and W. Ubachs. Deep-UV Rayleigh Scattering of

N2, CH4and SF6.Molecular Physics, 106:2471–2479, 2008.

[197] S.-H. Jin and G.-S. Kim. Simultaneous measurements of burning velocity and

temperature distribution of combustion using UV laser Rayleigh scattering.

Measurement, 169:177–184, 2021.

[198] W. R. Lempert, P.-F. Wu, and R. B. Miles. Filtered Rayleigh scattering

measurements using a MHz rate pulse-burst laser. 35th Aerospace Sciences

Meeting and Exhibit, 1997.

[199] D. Müller, R. Pagel, A. Burkert, V. Wagner, and W. Paa. Two-dimensional

temperature measurements in particle loaded technical flames by filtered

Rayleigh scattering. Applied Optics, 53(9), 2014.

[200] G. S. Elliott, N. Glumac, and C. D. Carter. Molecular filtered Rayleigh scatter-

ing applied to combustion. Measurement Science and Technology, 12:452–466,

2001.

[201] A. P. Yalin and R. B. Miles. Temperature Measurements by Ultraviolet Fil-

tered Rayleigh Scattering Using a Mercury Filter. Journal of Thermophysics

and Heat Transfer, 14(2), 2000.

[202] A. P. Yalin, Y. Z. Ionikh, and R. B. Miles. Gas temperature measurements

in weakly ionized glow discharges with filtered Rayleigh scattering. Applied

Optics, 41(18):3753–3762, 2002.

[203] M. Sneep and W. Ubachs. Direct measurement of the Rayleigh scattering cross

section in various gases. Journal of Quantitative Spectroscopy and Radiative

Transfer, 92:293–310, 2005.

[204] U. Hohm and K. Kerl. Temperature dependence of mean molecular polariz-

ability of gas molecules. Molecular Physics, 58:541–550, 1986.

[205] R. D. Sharma. Contribution of the polarizability anisotropy to Rayleigh scat-

tering. Journal of Geophysical Research, 112, 2007.

[206] M. P. Bogaard, A. D. Buckingham, R. K. Pierens, and A. H. White. Rayleigh

Scattering Depolarization Ratio and Molecular Polarizability Anisotropy for

Gases. Journal of the Chemical Society, Faraday Transactions1: Physical

Chemistry in Condensed Phases, 74, 1978.

[207] C. G. Fotache, T. G. Kreutz, D. L. Zhu, and C. K. Law. An Experimental

Study of Ignition in Nonpremixed Counterflowing Hydrogen versus Heated

Air. Combustion Science and Technology, 109:373–393, 1995.

[208] C. B. Oh, A. Hamins, M. Bundy, and J. Park. The two-dimensional structure

of low strain rate counterflow nonpremixed-methane flames in normal and

microgravity. Combustion Theory and Modelling, 12:283–302, 2008.

Bibliography 163

[209] C. Breck Hitz, J. Ewing, and J. Hecht, editors. Introduction to Laser Tech-

nology. John Wiley and Sons, Inc., 4 edition, 2012.

[210] D. C. Dumitras, editor. Nd YAG Laser. InTech, 2012.

[211] K. Thyagarajan and A. Ghatak, editors. Lasers: Fundamentals and Applica-

tions. Springer, 2 edition, 2010.

[212] O. Svelto and D. C. Hanna, editors. Principles of Lasers. Springer Science

and Business Media, LLC, 4 edition, 1998.

[213] F. Durst, A. Melling, and J. H. Whitelaw. Theorie und Praxis der Laser-

Doppler-Anemometrie. Braun-Verlag, 1987.

[214] D. A. Skoog, F. J. Holler, and S. R. Crouch. Instrumentelle Analytik: Grund-

lagen - Geräte - Anwendungen. Springer Spektrum, 6 edition, 2013.

[215] W. Zinth and U. Zinth. Optik: Lichstrahlen - Wellen - Photonen. De Gruyter,

2018.

[216] LaVision GmbH. Website: www.lavision.de; last accessed: 25.07.2021. 2021.

[217] V. Van Nieuwenhove, J. De Beenhouwer, F. De Carlo, L. Mancini, F. Marone,

and J. Sijbers. Dynamic intensity normalization using eigen flat fields in X-ray

imaging. Optics Express, 23:27975–27989, 2015.

[218] S. B. Howell. Handbook of CCD Astronomy. Cambridge University Press, 2

edition, 2006.

[219] J. A. M. Withag, J. B. W. Kok, and K. Syed. Transient Combustion Modeling

of an Oscilating Lean Premixed Methane/Air Flame. Proceedings of the ASME

Turbo Expo, 2008.

[220] F. Behrendt. Simulation laminarer Gegenstromdiffusionsflammen unter Ver-

wendung detaillierter Reaktionsmechanismen. PhD thesis, Ruprecht-Karls-

Universität Heidelberg, 1989.

[221] V. Cuervo Pinera. Numerical Investigation of Laminar One-Dimensional

Counter-Flow Flames from Product Gas of Woody Biomass Pyrolisis and Gasi-

fication. Master’s thesis, Technische Universität Berlin, 2015.

[222] I. Glassman, R. A. Yetter, and N. G. Glumac, editors. Combustion. Academic

Press, 5 edition, 2015.

[223] R. Byron Bird, W. E. Stewart, and E. N. Lightfoot, editors. Transport Phe-

nomena. John Wiley and Sons, Inc., 2 edition, 2002.

164 Bibliography

[224] V. Sick, A. Arnold, E. Dießel, T. Dreier, W. Ketterle, B. Lange, J. Wol-

frum, K. U. Thiele, F. Behrendt, and J. Warnatz. Two-Dimensional Laser

Diagnostics and Modeling of Counterflow Diffusion Flames. Symposium (In-

ternational) on Combustion, 23:495–501, 1990.

[225] I. Glassman. Sooting laminar diffusion flames: Effect of dilution, addi-

tives, pressure, and microgravity. Symposium (International) on Combustion,

27:1589–1596, 1998.

[226] F. Liu, A. E. Karatas, Ö. L. Gülder, and M. Gu. Numerical and experimental

study of the influence of CO2and N2dilution on soot formation in laminar

coflow C2H4/air diffusion flames at pressures between 5 and 20 atm. Combus-

tion and Flame, 162:2231–2247, 2015.

[227] R. G. W. Norrish. A Theory of the Combustion of Hydrocarbons. Proceedings

of the Royal Society of London, 150:36–57, 1935.

[228] W. Zhang, H. Liu, I. U. Hai, Y. Neubauer, P. Schröder, H. Oldenburg,

A. Seilkopf, and A. Kölling. Gas cleaning strategies for biomass gasification

product gas. International Journal of Low-Carbon Technologies, 7:69–74, 2012.

[229] M. Asadullah. Biomass gasification gas cleaning for downstream applications:

A comparative critical review. Renewable and Sustainable Energy Reviews,

40:118–132, 2014.