improved understanding of anaerobic digester...
TRANSCRIPT
IMPROVED UNDERSTANDING OF ANAEROBIC DIGESTER
PROCESSES BY STABLE ISOTOPE TECHNIQUES
DANIEL GIRMA MULAT PhD THESIS
. SCIENCE AND TECHNOLOGY
. 2015
Aarhus University
Department of Engineering
Science and Technology
Hangøvej 2
8200 Aarhus N
Denmark
i
Preface
This PhD dissertation has been submitted to Aarhus University in partial fulfillment of the
requirements of the degree Doctor of Philosophy at the graduate school of science and technology
(GSST). My main supervisor is Anders Feilberg, Associate Professor at Department of Engineering,
Aarhus University. My co-supervisors are Anders Peter S. Adamsen, Senior Scientist and Alastair
James Ward, Assistant Professor at Department of Engineering, Aarhus University. This study was
conducted from December 1st 2011 until January 29
th 2015 at the Department of Engineering,
Aarhus University, located at Research Centre Foulum, Denmark. I also spent about 5 months
abroad between January 2012 and June 2012 at Department of Environmental Microbiology,
Helmholtz Centre for Environmental Research-UFZ, Leipzig, Germany in collaboration with
Department of Biochemical Conversion, Deutsches Biomasseforschungszentrum (DBFZ), Leipzig,
Germany. I conducted an experiment at UFZ and DBFZ aimed at investigating the effect of
different operating condition on process performance, isotope signatures, methanogenic pathways
and microbial community composition in lab-scale continuous tank reactor (CSTR) and have gained
hands on experience with operating CSTR, isotope ratio mass spectrometer (IRMS) and some
molecular biology techniques. My immediate supervisors were Dr. Marcell Nikolausz, Scientist at
UFZ and Dr. H. Fabian Jacobi, Head of process monitoring and simulation group at DBFZ.
The primary focus of this Ph.D. project was the development of stable isotope techniques and its
application for better understanding of key intermediates (acetate and hydrogen) and metabolic
pathways involved during conversion of organic material to methane in biogas reactor. This thesis
consists of ten chapters. Chapter 1 is a general introduction about the motivation for conducting this
PhD study and its objectives. A literature review on biogas technology as a source of renewable
energy, biochemistry of anaerobic digestion and operating conditions regulating biogas processes is
presented in chapter 2. General methods that were used and developed in this PhD study are
presented in chapter 3. The key results of the experimental work are presented in chapters 4 to 9 as
published papers, manuscripts under review and manuscripts in preparation. In chapter 10, the
general findings of this thesis are summarized and discussed as well as conclusions and future
perspectives are provided.
Some of the results of this PhD study were presented at two international conferences and one
national conference. I gave oral presentations at International Conference on Biogas Microbiology
ICBM, held on 10th
to 12th
June 2014 in Uppsala, Sweden and at International Conference on
Anaerobic Digestion, BiogasScience 2014, held on 26th
to 30th
October 2014 in Vienna, Austria.
Poster was presented at a local conference Energy and Environment for the future-sustainable
energy for a fossil free society and environmental friendly technologies, November 24- 25,
Copenhagen, Denmark. This thesis is based on the following published papers, manuscripts
submitted and manuscripts in preparation.
Peer-Reviewed Journal Articles
1. Mulat, D. G., Ward, A. J., Adamsen, A. P. S., Voigt, N. V., Nielsen, J. L., & Feilberg, A.
(2014). Quantifying contribution of synthrophic acetate oxidation to methane production in
thermophilic anaerobic reactors by membrane inlet mass spectrometry. Environmental
science & technology 48(4), 2505-2511.
ii
2. Mosbæk, F., Kjeldal, H., Mulat, D. G., Albertsen, M., Ward, A. J., Feilberg, A. & Nielsen,
J. L., (2015). Acetate oxidizing microbial communities during acid accumulation in
anaerobic digestion. In preparation for peer-reviewed journal.
3. Mulat, D. G., Mosbæk, F., Ward, A. J., Polag, D., Greule, M., Keppler, F., Nielsen, J. L., &
Feilberg, A. (2015). Effect of exogenous hydrogen addition on process performance,
methanogenesis and homo-acetogenesis pathways during an in situ biogas upgrading. In
preparation for peer-reviewed journal.
4. Mulat, D. G., & Feilberg, A. (2014). GC/MS method for determining carbon isotope
enrichment and concentration of underivatized short-chain fatty acids by direct aqueous
solution injection of biogas digester samples. Submitted to Talanta. Under review.
5. Mulat, D. G., Jacobi, F., Feilberg, A., Adamsen, A. P. S., & Nikolausz, M. (2015).
Changing feeding regimes to demonstrate flexible biogas production: effects on process
performance, microbial community structure and methanogenesis pathways. In submission
for Bioresource Technology Journal.
6. Mulat, D. G., Feilberg, A., Jacobi, F., & Nikolausz, M. (2015). Stable isotope techniques as
a tool for process monitoring of biogas reactors operating under different condition. In
preparation for peer-reviewed journal.
Conference and poster presentations
7. Mulat, D. G., Ward, A. J., Adamsen, A. P. S., Voigt, N. V., Nielsen, J. L., & Feilberg, A.
(2014). Application of online membrane inlet mass spectrometry (MIMS) combined with
isotope labelling of substrates for quantifying methanogenesis pathway in anaerobic
reactors. Paper presented at International Conference on Biogas Microbiology ICBM, June
10-12, Uppsala, Sweden.
8. Mulat, D. G., Jacobi, F., Feilberg, A., Adamsen, A. P. S., Richnow, H. H., & Nikolausz, M.
(2014). Shifts in methanogenic pathways in response to change in substrate feeding pattern
studied by stable isotope techniques. Conference proceedings for the International
conference on anaerobic digestion, BiogasScience 2014, October 26-30, Vienna, Austria.
9. Mulat, D. G., Ward, A. J., Adamsen, A. P. S., Voigt, N. V., Nielsen, J. L., & Feilberg,
A.(2014). Application of on-line mass spectrometry with stable isotope pairing to study the
pathway of methane production from acetate in anaerobic reactor. Poster session at Energy
and Environment for the future-sustainable energy for a fossil free society and
environmental friendly technologies, November 24- 25, Copenhagen, Denmark.
iii
Acknowledgements
I would like to thank God for blessing my family, giving me the strength, guidance and resources to
complete my thesis. I am grateful for the numerous people who have helped along my journey here
at Aarhus University and made possible in completing this study. A special thanks goes to my main
PhD supervisor, Associate Professor Anders Feilberg for sharing his experience and helpful
suggestion as well as inspiration and continued encouragement throughout my PhD study. I would
also like to thank my PhD co-supervisors: Senior Scientist Anders Peter S. Adamsen and Assistant
Professor Alastair James Ward for their support and scientific advice.
The project would not have been possible without financial support from the Danish Strategic
Research Council (Grant No. 10-093944). I express gratitude to the external project partners:
Professor Jeppe Lund Nielsen and PhD student Freya Mosbæk for the help with the microbial
community analysis of our samples; Sabine Lindholst for the great help with MIMS and micro-GC
instruments; Dr. Niels Vinther Voigt for the great discussion about MIMS and hydrogen addition
experiment set up; Dr. Daniela Polag, Dr. Markus Greule and Prof. Frank Keppler for analyzing the
isotope composition of our biogas samples. My gratitude also goes to those who supported my
scientific work and assist me with administrative stuffs during my stay at the Department of
Environmental Microbiology, Helmholtz Centre for Environmental Research (UFZ), Leipzig,
Germany in collaboration with Department of Biochemical Conversion, Deutsches
Biomasseforschungszentrum (DBFZ), Leipzig, Germany. In particular, I would like to thank my
immediate supervisors Dr. Marcell Nikolausz and Dr. H. Fabian Jacobi for the fruitful scientific
discussion and great help during my stay in Leipzig; Dr. Sabine Kleinsteuber for allowing me to
join her research group meeting and work in the group’s lab; Dr. Hans-Hermann Richnow for
allowing me to work in his isotope biogeochemistry lab; PhD students Zuopeng Lv and Athaydes
Francisco Leite Junior as well as the lab technicians at UFZ and DBFZ for their invaluable support.
Thanks to the many supports from the current and former lab technicians at the biogas plant in
Foulum: Heidi Grønbæk Christiansen, Britt Amby Malthesen, Claudia Nagy and Patricia De Sousa
as well the Foulum biogas plant manager Mogens Møller Hansen.
Also many thanks to our research group members for creating such a wonderful working and
learning environment. The support from the head and secretary of our section, the PhD partners at
the Graduate School of Science and Technology (GSST) is greatly appreciated and special mention
to Anja Torup Hansen and Morten Dam Rasmussen. I am indebted to all my friends who have been
encouraging along the way and special mention to Setegn for our good friendship and all awesome
dinners we have had together in Foulum.
I want to thank my family. Many thanks to my wife Fre without whose love, encouragement,
understanding and support I would not have completed this work and to my baby girl Yohanna for
being part of our life, source of inspiration and energy.
Lastly, and most importantly, I wish to thank my mother, brothers and close relatives who have
supported me and kept me focused with their love, prayers and words of encouragement. I am
dedicating this thesis to my mom Itenesh who raised me, supported me, taught me and love me.
iv
Abstract
Anaerobic digestion (AD) of organic matter to methane-rich biogas is carried out by diverse
consortia of anaerobic bacteria and archaea for the purpose of waste management and renewable
energy production. However, the advantages of AD for treating organic wastes have not been fully
realized. Full-scale biogas plants are often operated at suboptimal organic loading rates (OLR) to
avoid process imbalance and failure of the plants. This shows that the process is still far from
optimized due to incomplete process understanding. Developing a comprehensive understanding of
biogas process is the key to employing appropriate strategies that allow stable operation of biogas
plants at optimum OLR, which in turn increases productivity and economy. Therefore, research
aimed at generating in-depth knowledge of the degradation mechanisms of key intermediates and
most important methanogenic pathways for the formation of biogas in AD is highly desired. In this
regard, the main focus of this thesis was the development and application of stable isotope
techniques for better understanding of biogas process. The specific objectives of the PhD study
were: to develop an online membrane inlet quadrupole mass spectrometry (MIMS) methodology for
monitoring of the isotopic distribution of CO2 and CH4 in AD to be used in combination with
isotopically labeled substrates; to quantity the relative contribution of acetoclastic methanogenesis
(AM) and syntrophic acetate oxidation coupled to hydrogenotrophic methanogenesis pathway
(SAO-HM) to total CH4 production from degradation of acetate; to investigate the effect of
exogenous hydrogen addition on process performance, methanogenesis, homo-acetogenesis and
microbial community structure; to develop GC/MS method for simultaneous determination of the
concentration of underivatized volatile fatty acid (VFA) and isotope ratio of underivatized acetate
by direct injection of aqueous biogas digester samples; and to investigate possible use of stable
isotope techniques as a tool for monitoring the actual state of biogas process and as an early
warning tool to process instability.
In this study, the experiments involved lab-scale continuous stirred tank reactors (CSTRs) and batch
incubation with 13
C labeled tracer substrates and specific inhibitor to acetoclastic methanogens. The
stable isotope composition of products and reactants at natural abundance and 13
C enriched
substances were measured and molecular biology techniques and basic analytical methods were
used to support the results of the isotope analysis.
An online MIMS method was developed to trace the incorporation of 13
C into the produced CO2
and CH4 in real time when incubated with [2-13
C]acetate in a thermophilic anaerobic reactor. This
novel approach was applied for quantification of the relative contribution of SAO-HM to methane
production from acetate, which was demonstrated to reach a high degree of contribution. Protein
based stable isotope probing (protein-SIP) and metagenome analysis showed that peptides from the
bacteria class Clostridia, the hydrogenotrophs Methanoculleus and the mixotrophic
Methanosarcina were labelled with 13
C during degradation of high concentration of 13
C labeled
acetate (100 mM), indicating Clostridia possibly oxidizes acetate to CO2 in syntrophic association
with the hydrogenotrophs. Another very simple, accurate, reproducible and rapid method based on
GC/MS was developed for determining both the isotope enrichment of acetate and concentration of
underivatized VFA in a biogas digester sample by direct liquid injection of acidified aqueous
v
samples. As an example of application of this method to a biogas process, it was demonstrated that
a stable isotope tracer experiment in combination with tracer-to-tracee ratio (TTR) determination by
the GC/MS method proved that carbon dioxide was reduced to acetate under high H2 partial
pressure, indicating the activity of homoacetogens. The results of exogenous H2 gas addition for an
in situ biogas upgrading showed that all the added H2 in the presence of stoichiometric amount of
CO2 was almost completely utilized and the methane content of the biogas reached up to 90% with
concomitant decrease in the CO2 content. Unlike the control reactors, the degradation of acetate and
other VFA decreased in the H2 reactors and finally accumulated. Acetate degradation resumed after
the concentration of H2 in the reactor headspace decreased by flushing with helium. The observed
lower carbon isotope fractionation between CO2 and CH4 in the H2 reactors is possibly explained by
the differential reversibility concept, indicating exogenous H2 addition may have led to high H2
concentration within micro-aggregates of methanogens. In addition, it was shown that different
operating conditions (change in feeding interval and continuous increase in OLR) had influenced
the process performance, methanogenesis pathways and bacterial community composition in lab-
scale CSTRs fed with distillers dried grains with solubles (DDGS). Unlike shorter feeding interval
(every 2 hours), longer feeding interval (daily and every 2 days) led to a dynamic process, as
depicted in the short term changes of biogas production rate, biogas composition (CH4, CO2 and
H2), isotope composition of methane (δ13
C-CH4 and δD-CH4), total VFA, acetate and propionate.
The δ13
C-CO2 remained relatively stable under different feeding intervals during steady state
operation but changed slightly during an increase in OLR. Longer feeding interval allows the
flexibility to produce more biogas at times of high energy demand due to the possibility to increase
production shortly after a feeding event whereas feeding did not change the biogas production rate
in the CSTR fed at shorter intervals. Interestingly, the CSTRs fed at longer intervals when
compared to those fed at shorter intervals, demonstrated significantly higher methane yield by about
14% and were less susceptible to stress condition. The bacterial community structure varied
between CSTRs fed under different feeding intervals whereas methanogens remained stable with
higher abundance of the genera Methanosarcina and Methanoculleus. Our observation of the
dominating methanogens community was supported with isotope analysis, indicating HM and AM
contributed almost equally to the produced methane from each feeding event. In addition, among
the studied process monitoring tools, a combination of parameters based on the measurement of the
isotope composition of CH4 and CO2 at natural abundance, biogas production rate and biogas
composition would indicate the actual state and performance of the process as well as process
imbalance at early stage.
In conclusion, the measurement of the stable isotope composition of CH4, CO2 and acetate
improved our understanding about methanogenesis, homo-acetogenesis and the degradation
mechanisms of key intermediates in AD. The dominating role of SAO-HM and the abundance of
the Methanosarcina in biogas reactors suggested that a re-evaluation of biogas process optimization
and operating conditions should be employed with the consideration of the importance of the SAO-
HM and the Methanosarcina in AD.
vi
Dansk Resume
Anaerob nedbrydning af organisk stof til metan-rig biogas foregår ved hjælp af flere grupper af
anaerobe bakterier og arkebakterier for at håndtere biprodukter og producere biogas. Hidtil er
fordelene ved anaerob nedbrydning ikke blevet fuldt ud udnyttet, da fuldskala biogasanlæg ofte
køres lidt under det optimale indfødningsniveau for at undgå procesubalance og reaktornedbrud.
Udvikling af en udtømmende forståelse af biogasprocessen er nøglen til at anvende driftsstrategier
der sikrer en stabil drift ved optimal indfødning, og dermed forøge produktivitet og lønsomhed. Det
betyder at det er vigtigt med forskning der sikrer en dybere forståelse af omsætning af
nøglemellemprodukter og produktion af metan i biogasreaktoren. Hovedformålet med denne
afhandling var at udvikle og anvende stabile isotopteknikker for at få en dybere forståelse af
biogasprocessen. Mere specifikt var formålet at udvikle en online membran-inlet
massespektrometrisk (MIMS) metode til at monitere fordelingen af CO2 og CH4-isotoper i
biogasprocessen ved anvendelse af substrater mærket med stabile isotoper; at kvantificere de
relative bidrag fra den acetoklastiske metandannelse (AM) og den syntrofe acetatoxidation koblet
med hydrogenotrof metandannelse (SAO-HM) i biogasprocessen; at undersøge effekten af ekstern
tilførsel af hydrogen på metandannelsen og strukturen af de mikrobielle samfund; at udvikle en GC-
MS metode for samtidig bestemmelse af koncentrationer af flygtige fede syrer (VFA) og
isotopforholdet ved direkte injicering af biogasvæske; og at undersøge mulig anvendelse af stabile
isotopteknikker til udvikling af et værktøj til at monitere aktuel tilstand af biogasprocessen og
forudsige kommende procesproblemer.
Denne afhandling beskriver forsøg i laboratorie-skala med kontinuerligt omrørte reaktorer (CSTR)
og batch inkuberet med 13
C-mærket substrat og specifikke inhibitorers effekt på acetoklastisk
metandannelse. Fordelingen af stabile isotoper i reaktanter og produkter af naturligt forekommende
og 13
C-berigede substrater blev målt og understøttet af molekylærbiologiske teknikker og basale
analytiske metoder. En online MIMS-metode er blevet udviklet til at spore indbygning af 13
C i den
producerede CO2 og CH4 i realtime ved inkubering med [2-13
C]acetat i en termofil anaerob reaktor.
Denne nye metode blev anvendt til at kvantificere den relative fordeling af SAO-HM i forhold til
produktion fra acetat. Det blev demonstreret at førstnævnte kunne nå en høj andel. En
proteinbaseret stabil isotop-metode (protein-SIP) and metagenom-analyse viste at peptider fra
bakteriegrupper Clostridia, den hydrogenotrofe Methanoculleus og den mixotrofe Methanosarcina
blev mærket med 13
C ved omsætning af høje koncentrationer af 13
C-mærket acetat (100 mM). Det
indikerer at Clostridia muligvis oxiderer acetat til CO2 i syntrofisk samarbejde med de
hydrogenotrofe bakterier.
En anden enkel, nøjagtig, reproducerbar og hurtig metode baseret på GC-MS blev udviklet for at
bestemme både den isotopiske berigelse af acetat og koncentrationer af uderivatiseret VFA i en
biogas reaktorprøve ved direkte injicering af forsurede væskeprøver. Som et eksempel på
anvendelse af denne metode i en biogasproces blev det demonstreret i en stabil isotop sporstof-
forsøg at CO2 blev reduceret til acetat under høje H2 partieltryk, hvilket indikerer homoacetogen
aktivitet. Tilførsel af ekstern H2 i en in situ opgradering viste at alt tilsat H2 ved tilstedeværelse af
støkiometrisk mængde af CO2 blev næsten komplet forbrugt, og at metanindholdet i biogassen
kunne nå op på en koncentration på 90% med tilsvarende reduktion i CO2-indholdet. Sammenlignet
med en kontrolreaktor uden tilførsel af ekstern H2 vistes en lavere omsætning af acetat og andre
VFA og som senere akkumuleredes i reaktoren med tilført H2. Acetatomsætning blev genoptaget
efter at H2 i gasfasen blev fjernet ved at skylle med helium. Den observerede lave kulstof-isotop
fraktionering imellem CO2 og CH4 i H2-reaktoren kan muligvis forklares med konceptet differentiel
vii
reversibilitet, som indikerer at ekstern tilførsel af H2 kan have ledt til høje H2-koncentrationer inde
i mikro-aggregater af metanogene bakterier. Det blev også vist at forskellige procesbetingelser
(ændring i indfødningsinterval og kontinuerlig forøgelse af indfødt organisk stof) påvirkede
processen og fordelingen af de mikrobielle samfund i laboratorieskalaforsøg.
En kontinuerlig omrørt reaktor (CSTR) blev fodret med DDGS (tørret kornbærme; distillers dried
grains with solubles). Modsat kortere indfødningsintervaller (hver 2. time) ledte længere
indfødningsintervaller (dagligt eller hver 2. dag) til en mere dynamisk proces påvist ved hurtige
skift i biogasproduktionsrater, fordelingen af CH4, CO2 og H2, isotopfordelingen af (δ13C-CH4 og
δD-CH4) samt total VFA, acetat og propionate. δ13C-CO2 forblev relativt stabilt under forskellige
indfødningsintervaller ved stabile driftsforhold, men ændredes lidt ved højere indfødningsmængder.
Længere indfødningsintervaller tillod en fleksibilitet til at producere mere biogas på tidspunkter
med høj efterspørgsel, da produktionen stiger kort efter en indfødning, hvorimod indfødning med
kortere intervaller ikke ændrede biogasproduktionen i en CSTR-reaktor. CSTR med længere
intervaller mellem indfødning af substrat viste en signifikant højere metanudbytte på 14%
sammenlignet med tilsvarende indfødning med kortere intervaller og var mindre sårbar overfor
stressforhold. Strukturen af bakteriesamfundene varierede imellem CSTR indfødt med forskellige
intervaller, hvorimod de metanogene bakterier forblev stabile med høje mængder af slægterne
Methanosarcina og Methanoculleus. Vores observation af de dominerende metanogene samfund
blev understøttet med isotopanalyser som indikerede, at de hydrogenotrofe og acetatotrofe
metanogene bakterier bidrog næsten ligeligt til den producerede CH4 fra de enkelte indfødninger.
Ud fra de anvendte teknikker vil det være muligt at udvikle en metode til at indikere aktuel status og
forudsige begyndende ustabilitet i biogasprocessen.
Det konkluderes at målinger af sammensætningen af de stabile CH4, CO2 og acetat har forøget
vores forståelse af dannelse af metan og homo-acetat og omsætningsmekanismer for de vigtigste
mellemprodukter i biogasprocessen. Den dominerende rolle af SAO-HM og tilstedeværelse af
Methanosarcina i biogasreaktorer antyder at en revurdering af metoder til at optimere og køre
biogasreaktorer bør foretages med fokus på betydningen af SAOHM og Methanosarcina i
biogasprocessen.
viii
Lists of abbreviations
AD Anaerobic digestion
AM Acetoclastic methanogenesis
AU Aarhus University
BP Base pair
CF Continuous flow
CRDS Cavity ring-down spectroscopy
CSIA Compound specific isotope analysis
CSTR Continuous stirred tank reactor
DBFZ Deutsches Biomasseforschungszentrum
DDGS Distillers dried grains with solubles
DI Dual-inlet
DNA Deoxyribonucleic acid
DNA-SIP Deoxyribonucleic acid stable isotope probing
EA/IRMS Elemental analyzer isotope ratio mass spectrometry
EI Electron impact
fmc The fraction of CH4 produced from the reduction of CO2
GC Gas chromatography
GC/C/IRMS Gas chromatography combustion isotope ratio mass spectrometry
GC/MS Gas chromatography mass spectrometry
GSST Graduate school of science and technology
HM Hydrogenotrophic methanogenesis
HRT Hydraulic retention time
HYCon HYdrogen Control for optimization of methane production from livestock waste
ICBM International Conference on Biogas Microbiology
IRMS Isotope ratio mass spectrometry
KE Kinetic energy
LC-MS/MS Liquid chromatography tandem mass spectrometry
LF Longer feeding interval
mcrA Alpha subunit of methyl coenzyme M reductase gene
MFC Mass flow controller
MIMS Membrane inlet quadrupole mass spectrometry
MRC Methyl coenzyme M reductase
mRNA Messenger ribonucleic acid
nMDS Non-metric multidimensional scaling
OLR Organic loading rate
PCR Polymerase chain reaction
Protein-SIP Protein based stable isotope probing
QMS Quadrupole mass spectrometer
Rd1 Every day fed reactor
Rd2 Every 2 d fed reactor
Rh2 Every 2 h fed reactor
RNA Ribonucleic acid
rRNA Ribosomal ribonucleic acid
SAO Syntrophic acetate oxidation
SAOB Syntrophic acetate-oxidizing bacteria
SBP Specific biogas production
ix
SCFA Short chain fatty acids
SF Shorter feeding interval
SIP Stable isotope probing
SMP Specific methane production
TAN Total ammonia nitrogen
TDLAS Tunable diode laser absorption spectroscopy
TIC Total inorganic carbon
T-RFLP Terminal restriction fragment length polymorphism
T-RFs Terminal restriction fragments
TS Total solid
TTR Tracer-to-tracee ratio
UFZ Helmholtz Centre for Environmental Research
VFA Volatile fatty acids
V-PDB Vienna Pee Dee Belemnite
VS Volatile solid
V-SMOW Vienna Standard Mean Ocean Water
VSR Volatile solid reduction
WP Work package
α Isotope fractionation factor
δ13
C-CH4 13
C isotopic signature of CH4
δ13
C-CO2 13
C isotopic signature of CO2
δD-CH4 Hydrogen isotope signature of CH4
δma Acetate-derived CH4 through acetoclastic methanogenesis
δmc CO2-derived CH4 through hydrogenotrophic methanogenesis
ε Isotope enrichment
1
Table of Contents
Chapter 1: Introduction ...................................................................................................................................... 3
1.1 Background information .......................................................................................................................... 3
1.2. Objectives ............................................................................................................................................... 4
Chapter 2: Literature review .............................................................................................................................. 5
2.1 Biogas as a source of renewable energy .................................................................................................. 5
2.2 Biochemistry of anaerobic digestion ....................................................................................................... 5
2.3 Operating conditions and factors regulating biogas process ................................................................... 9
2.3.1Temperature ...................................................................................................................................... 9
2.3.2 pH ................................................................................................................................................... 10
2.3.3 Short chain fatty acids (SCFA) ....................................................................................................... 10
2.3.4 Ammonia ......................................................................................................................................... 11
2.3.5 Organic loading rate (OLR) ........................................................................................................... 11
2.3.6 Hydraulic retention time (HRT) ..................................................................................................... 12
2.3.7 Feeding interval .............................................................................................................................. 12
Chapter 3: General analytical and molecular biology methods ....................................................................... 13
3.1 Stable isotope techniques....................................................................................................................... 13
3.1.1 Basics, notations and terminology .................................................................................................. 13
3.1.2 Tracer experiment........................................................................................................................... 15
3.1.3 Stable isotope analysis at natural abundance and fractionation factor ......................................... 15
3.1.4 Gas chromatography (GC/MS) ...................................................................................................... 18
3.1.5 Membrane inlet quadruple mass spectrometry (MIMS) ................................................................. 19
3.1.6 Isotope ratio mass spectrometry (IRMS) ........................................................................................ 20
3.2 Molecular biology methods ................................................................................................................... 21
3.2.1 T-RFLP analysis ............................................................................................................................. 21
3.2.2 Protein-SIP ..................................................................................................................................... 23
References ....................................................................................................................................................... 24
Chapter 4: Paper I- Quantifying contribution of synthrophic acetate oxidation to methane production in
thermophilic anaerobic reactors by membrane inlet mass spectrometry ......................................................... 30
Chapter 5: Paper II-Acetate oxidizing microbial communities during acid accumulation in anaerobic
digestion .......................................................................................................................................................... 54
2
Chapter 6: Paper III- Effect of exogenous hydrogen addition on process performance, methanogenesis and
homo-acetogenesis pathways during an in situ biogas upgrading ................................................................... 81
Chapter 7: Paper IV- GC/MS method for determining carbon isotope enrichment and concentration of
underivatized short-chain fatty acids by direct aqueous solution injection of biogas digester samples .......... 99
Chapter 8: Paper V- Changing feeding regimes to demonstrate flexible biogas production: effects on process
performance, microbial community structure and methanogenesis pathways .............................................. 119
Chapter 9: Paper VI- Stable isotope techniques as a tool for process monitoring of biogas reactors operating
under different condition ............................................................................................................................... 146
Chapter 10: General discussion and conclusion ............................................................................................ 178
10.1 Developing online MIMS method for quantifying methanogenesis pathway-Paper I ...................... 178
10.2 Role of SAO-HM to methane production from degradation of acetate-Paper II ............................... 179
10.3 Influence of H2 on methanogenesis and homoacetogenesis -Papers III and IV ................................ 181
10.4 Factors regulating biogas process and process monitoring tool-papers V and VI ............................. 183
10.5 General conclusion ............................................................................................................................ 186
10.6 Perspectives ....................................................................................................................................... 187
References ................................................................................................................................................. 190
3
Chapter 1: Introduction
1.1 Background information
Anaerobic digestion (AD) of organic material to biogas is carried out by a complex community
consisting of hydrolytic, fermentative, acetogenic and methanogenic microorganisms (Conrad,
2005). Biogas contains mainly methane (55-75%) and carbon dioxide (25-45). The main metabolic
pathways mediating the production of biogas from organic material consist of four main steps:
hydrolysis, acidogenesis, acetogenesis and methanogenesis (Burke, 1993; Sasaki et al., 2011a).
Hydrolysis involves the conversion of polymers such as polysaccharides, proteins and lipids to
monomers and oligomers and further fermented to organic acids, alcohols, CO2 and H2 via
acidogenesis. The alcohols and organic acids other than acetate are fermented to acetate, CO2, H2 or
formate through acetogenesis. Eventually, methane is produced from acetate and H2/CO2 via
methanogenesis pathway. In the hydrogenotrophic methanogenesis (HM) pathway, hydrogen and/or
formate are used to reduce carbon dioxide to methane. Acetate can be directly cleaved to methane
via acetoclastic methanogenesis (AM). Depending on the operating conditions of the reactor,
acetate can also mediate syntrophic acetate oxidation (SAO) pathway, whereby acetate is first
oxidized to CO2 and then the CO2 is reduced to CH4 via HM. Depending on the environmental
condition of the system, other biochemical reactions such as homo-acetogenesis can also take place
in AD. Homo-acetogenesis involves the production of acetate by the reduction of CO2 with H2.
The production of biogas from agricultural and industrial wastes in AD has been used as pollution
control and for energy recovery purposes. However, the advantages of AD for treating organic
wastes are presently challenged by suboptimal process conditions. Full-scale biogas plants are often
operated at suboptimal organic loading rates (OLR) to avoid process imbalance and failure of the
plants (Kleyböcker et al., 2012). This process is still far from optimized due to incomplete process
understanding(Madsen et al., 2011). Developing a comprehensive understanding of biogas process
is the key to employ appropriate strategies that allow stable operation of biogas plants at optimum
OLR, which in turn increases productivity and the economy. Therefore, research aimed at
generating in-depth knowledge of the degradation mechanisms of key intermediates and most
important methanogenic pathways for the formation of CH4 in AD is highly desired.
In this regard, the main focus of this thesis was the development and application of stable isotope
techniques for better understanding of biogas process. Stable isotope approaches in conjunction
with stable isotope tracer and inhibition experiments have been extensively used in environmental
ecology studies for quantifying biochemical pathways in natural environments (Conrad, 2005).
However, the application of stable isotope analysis at natural abundance in engineered biogas
reactors is in its infancy (Keppler et al., 2010; Nikolausz et al., 2013). Therefore, we extended the
application of stable isotope approaches to biogas reactors for better understanding of
methanogenesis pathways and degradation mechanisms of key intermediates in AD. Moreover, we
investigated the application of stable isotope techniques for indicating the actual state of biogas
process and as an early warning tool to process imbalance.
4
1.2. Objectives
This PhD project is a part of a big project named hydrogen control for optimization of methane
production from livestock waste (HYCON), which was financially supported by funding from the
Danish Strategic Research Council. This interdisciplinary project involves the participation of
several institutes and industry, namely, Department of Engineering, Aarhus University, Denmark;
Department of Biotechnology, Chemistry and Environmental Engineering, Aalborg University,
Denmark; Danish Technological Institute, Denmark; Advanced Water Management Centre, the
University of Queensland, Australia; and Xergi, Denmark. The overall aim of the HYCON project
is to investigate the influence and distribution of hydrogen in the anaerobic digestion of organic
waste and develop new methods for controlling hydrogen in order to make biogas production more
efficient. There are seven work packages (WP) under HYCON project and my PhD project is a part
of the “WP1- sensor development, MIMS and micro-gas chromatography techniques, stable isotope
techniques”.
The focus of this PhD study was the analysis of stable isotopic composition of substances consumed
and produced in AD in conjunction with stable isotope tracer and specific inhibition experiments to
investigate the relative contribution of methanogenesis pathways to methane production and the role
of key intermediates such as hydrogen and acetate in anaerobic biogas reactors. In addition,
molecular biology techniques were used to support the results of the isotope analysis. The outcome
of this research was to generate new knowledge that deepens our understanding of methanogenesis
pathways and degradation of key intermediates in AD which can be used to devise some strategies
for optimization of biogas process and monitoring the actual state of biogas process.
The specific objectives of the PhD study were:
1. To develop a MIMS methodology for monitoring of the isotopic distribution of CO2 and
CH4 in AD to be used in combination with isotopically labeled substrates.
2. To quantity the relative contribution of AM and SAO-HM to total CH4 production from
degradation of acetate.
3. To investigate the effect of exogenous hydrogen addition on process performance,
methanogenic pathways, microbial community composition and homo-acetogenesis
pathway.
4. To develop GC/MS method for simultaneous determination of the concentration of
underivatized SCFA and isotope ratio of underivatized acetate by direct injection of aqueous
biogas digester samples.
5. To investigate possible use of stable isotope techniques as a tool for monitoring the actual
state of biogas process and as an early warning tool to process instability.
In order to attain the objectives of this study, several experiments were carried out. The experiments
are generally divided into four parts. The first part (chapter 4) was aimed at quantifying
methanogenic pathways that contributed to the production of methane from degradation of acetate.
Moreover, we aimed at optimizing a novel method based on membrane inlet quadrupole mass
spectrometry (MIMS) in conjunction with 13
C labeled acetate for determining the isotope
composition of dissolved CO2 and CH4 in a fermentation broth. In the second part (chapter 5), we
investigated the effects of different concentrations of acetate on methanogenesis pathways. In the
third part (chapter 6), we used 13
C fully labeled acetate as substrate with and without hydrogen
addition, to investigate the effect of exogenous hydrogen addition on methanogenesis, homo-
acetogenesis, microbial community composition and process performance. We also aimed at
5
developing a GC/MS method for simultaneously determining the isotopic ratio and concentration of
underivatized SCFA by direct injection of acidified biogas digester liquid samples (chapter 7). The
last part of the study (chapters 8 and 9) was mainly focused on the application of stable isotope
techniques in combination with molecular fingerprinting techniques for monitoring the actual state
of the biogas process in lab-scale continuous stirred tank reactor (CSTR) operating at different
conditions (feeding interval and OLR). The possible application of stable isotope techniques as an
early warning tool to process disturbance was investigated in CSTRs fed with distillers dried grains
with solubles (DDGS).
Chapter 2: Literature review
2.1 Biogas as a source of renewable energy
The modern society generates large amount of waste products that pose major risk to human health
and environment. Several technologies are available to treat waste products. Among these,
anaerobic digestion (AD) of waste materials to biogas has gaining significant importance for
production of renewable energy source, preventing environmental pollution and reduction of green
gas emissions. Moreover, the digestate can be used as a bio-fertilizer. Biogas consists of mainly
methane (55-75%) and carbon dioxide (25-45%) as well as trace amount of ammonia and hydrogen
sulfide depending on substrates and operating conditions. Biogas can be can be directly burned in a
combined and heat power (CHP) unit for the generation of heat and electricity or can be further
upgraded to natural gas quality to be used a transportation fuel or injected into a natural gas grid
system for storage (Karellas et al., 2010).
The feedstocks used for biogas production constitute of monosubstrate or codigestion of mixture of
several organic materials depending on the reactor technology, availability of feedstock, economic
consideration etc. Commonly used feedstocks consist of organic wastes (pig and cattle manure and
sludge from wastewater treatment plants), energy crops (sweet sorghum, miscanthus, rape,
sunflower etc), conventional crops (maize, wheat, sugar beet etc) and other organic feedstocks
(glycerol) (Karellas et al., 2010). Most of full-scale anaerobic digesters in Denmark are designed
mainly for the codigestion of manure with a smaller fraction of other wastes as a supplemental
substrate (Karakashev et al., 2005). On the other hand, maize as monosubstrate is the mainly used
feedstock in the agricultural biogas plants in Germany, providing about 60% of the biogas energy
(Lebuhn et al., 2014). In developing countries like China and India, thousands of small-scale biogas
plants fed with animal manure and food waste products are installed in the rural areas for electricity
generation and cooking purposes (Lebuhn et al., 2014).
2.2 Biochemistry of anaerobic digestion
AD of organic materials to biogas is characterized by the four major steps: hydrolysis, acidogenesis,
acetogenesis and methanogenesis (Figure 1). Several microorganisms including, fermentative
bacteria, acetogenic bacteria and methanogens are involved for the degradation of organic material
to the most oxidized and reduced forms, CO2 and CH4, respectively. A well balanced system is
essential for the steady biogas process (Schink & Stams, 2006).
Polymers such as polysaccharides, proteins, and lipids are first hydrolyzed to simpler monomers
and oligomers (sugars, amino acids, fatty acids and glycerol), typically by the action of extracellular
hydrolytic enzymes. These enzymes are produced by primary fermenting (acidogenic) bacteria
which ferment the monomers further to short-chain fatty acids (SCFA), alcohols, succinate, lactate,
6
carbon dioxide and hydrogen. Then syntrophic acetogenic bacteria convert the alcohols and SCFA
other than acetate to acetate, carbon dioxide and hydrogen or formate, which are subsequently used
by methanogens for the production of methane. The methanogens produce methane from the direct
cleavage of acetate through the acetoclastic methanogenesis pathway as well as the reduction of
carbon dioxide with hydrogen via the hydrogenotrophic methanogenesis pathway (Schink & Stams,
2006). Acetate can be first oxidized to carbon dioxide by syntrophic acetate oxidation bacteria
(SAOB) and subsequently reduced to methane by hydrogenotrophic methanogens (Hattori, 2008).
The conversion of organic polymers to soluble monomers is carried out by anaerobic bacteria such
as Bacterioides, Clostridium and Actinomycetales etc. Carbohydrates (cellulose and hemicellulose)
are hydrolyzed to monomers and oligomers by enzymes such as cellulases, xylanases or amylases
whereas proteins and lipids are enzymatically hydrolyzed to their corresponding monomers by
peptidase and lipases, respectively. However, the enzymatic hydrolysis of the major fraction of
lignocellulosic biomass (i.e, cellulose and hemicellulose) is hindered due to its complex structure.
Therefore, the biomass needs to be treated to expose the hydrolysable fractions to the enzymes
using different chemical, biological and physical pretreatment methods (Kumar et al., 2009).
In the second stage (acidogenesis), acidogenic bacteria transform the products of the first reaction
into SCFA, alcohols, hydrogen and carbon dioxide. Different facultative and obligatory anaerobic
bacteria are known to carry out acidogenesis pathways. For instance, Petrimonas sulfuriphila
ferments sugars to acetate and Paludibacter propionicigenes produces propionate, acetate, and
succinate during sugar fermentation (Ziganshin et al., 2011).
In the third stage, known as acetogenesis, the alcohols and SCFA except acetate are transformed by
acetogenic bacteria into hydrogen, carbon dioxide and acetic acid (Table 1). Some of the known
acetogenic bacteria include Acetobacterium spp., Sporomusa spp. and Ruminococcus spp.
Hydrogen or formate plays an important intermediary role in this process, as the reaction will only
occur if the hydrogen partial pressure (or formate concentration) is low enough to
thermodynamically allow the conversion of all the acids. It is generally assumed that H2 and
formate are in thermodynamic equilibrium due to the inter-conversion of H2 to formate through
formate-hydrogen lyase according to equation 7 (Table 1). From thermodynamic prediction, the H2
partial pressure as low as 10 Pa and 100 Pa are necessary for oxidation of propionate and butyrate,
respectively (Dolfing et al., 2008; Schmidt & Ahring, 1993). Such low H2 partial pressure is
achieved by syntrophic transfer of hydrogen from hydrogen-producing bacteria to H2-consuming
methanogens (Schmidt & Ahring, 1993). Therefore, close contact of acetogenic bacteria with
hydrogenotrophs in syntrophic associations is required for lowering the partial pressure of hydrogen
in the system. This syntrophic relationship between syntrophic bacterial and hydrogenotrophs is
termed as interspecies hydrogen transfer.
7
Figure 1 Simplified diagram showing major steps involved during the degradation of organic
matter to biogas. Adapted from (Schink & Stams, 2006).
The final stage of anaerobic digestion was performed by methanogens that produce methane as an
end product of their anaerobic respiration. All methanogens belong to the phylum Euryarchaeota,
which are strictly anaerobic archaea. They are relatively small and less diverse group compared to
the bacteria involved in the first three steps of AD. There are only a few substrates utilized by
methanogens, including CO2, acetate and methyl-group containing compounds. Hydrogenotrophs
reduce CO2 to methane via H2 as the primary electron donor known as hydrogenotrophic
methanogenesis pathway (HM; equation 5; Table 1). There are also members of hydrogenotrophic
communities that utilize formate as electron donor (Liu & Whitman, 2008). HM are carried out by
the members of the order Methanobacteriales, Methanomicrobiales and Methanococcales.
Acetate is a main intermediate during anaerobic degradation of organic material and hence, it is
assumed that two thirds of the methane produced is generated from direct cleavage of acetate
through acetoclastic methanogenesis (AM; equation 3; Table1) by acetoclastic methanogens
(Batstone et al., 2006). Within the order of Methanosarcinales, only members of the genera
Organic Matter
(Complex polymers)
Short chain Fatty Acids
alcohols, lactate, succinate
Monomers & Oligomers
(sugars, amino acids, long chain fatty
acids)
AcetateH2, CO2,
formate
BIOGAS
(CH4, CO2)
2) Acidogenesis
1) Hydrolysis
3) Acetogenesis
4) Methanogenesis
Syntrophic acetate
oxidation
Acetoclastic
methanogenesis
Hydrogenotrophic
methanogenesis
Ammonia
H2S
Homo-acetogenesis
8
Methanosaeta and Methanosarcina are known to undertake AM. Members of Methanosarcina are
mixotrophic, which are capable of utilizing several substrates such as CO2/H2, acetate, methyl-
containing compounds as well as capable of carry out syntrophic acetate oxidation of acetate to
CO2. Methanosaeta is strictly acetoclastic methanogens that uses only acetate.
Table 1 Standard Gibbs free-energy changes (∆G´) at 25 and 55 °C for some of the major
acetogenesis and methanogenesis pathways at pH 7a
Equation Reaction
∆G°´
(kJ/mol)
∆G´55
(kJ/mol)
1 CH3CH2CH2COO- + 2H2O → 2CH3COO
- + 2H2 + H
+ 48.1 37.9
CH3CH2CH2COO- + 2HCO3
- → 2CH3COO
- + 2HCOO
- + H
+ 45.5 36.1
2 CH3CH2COO- + 3H2O → CH3COO
- + HCO3- + 3H2 + H
+ 76.1 62.3
CH3CH2COO- + 2HCO3
- → CH3COO
- + 3HCOO
- + H
+ 72.2 59.7
3 CH3COO- + H2O → CH4 + HCO3- -31.0 -34.7
4 CH3COO- + 4H2O → 3HCO3
- + 4H2 + H
+ +104.1
5 4H2 + HCO3- + H
+ → CH4 + 3H2O -135.6 -122.5
4HCOO- + H2O + H+ → CH4 + 3HCO3- -130.0 -118.9
6 HCO3- + 2H2 + 0.5 H
+ → 0.5CH3COO
- + 2H2O -55
7 H2 + HCO3- → HCOO
- + H2O -1.3
aData were obtained from literatures (Kotsyurbenko et al., 2001; Schmidt & Ahring, 1993; Thiele &
Zeikus, 1988)
Acetate can also follow a two-step reaction pathway whereby acetate is first oxidized to CO2 and H2
through syntrophic acetate oxidation pathway (SAO; equation 4; Table 1) and the CO2 is
subsequently reduced to CH4 by the HM. Under standard condition, SAO is not thermodynamically
favorable (∆G°´ = +104.1) and can only be feasible if the H2 partial pressure is kept low by
coupling with H2-consuming methanogens (Hattori, 2008). As a result of this it is generally
assumed that AM is a dominant pathway for methane production from acetate. Therefore, most
biogas reactor operation and optimization is based on maintaining favorable condition for AM, with
little consideration to the importance of the SAO pathway. However, some studies recently found
that SAO coupled to HM is a dominant pathway in thermophilic methanogenic reactors (Goberna et
al., 2009; Hori et al., 2006; Karakashev et al., 2006; Krakat et al., 2010; Ryan et al., 2010; Sasaki et
al., 2011a). The first microbe found to perform acetate oxidation was a thermophilic bacterium
belonging to the group of homo-acetogenic bacteria capable of reversing the acetate-forming
reaction from hydrogen and carbon dioxide [17]. To date five syntrophic acetate oxidizing bacteria
(SAOB) have been isolated and characterized (Thermacetogenium phaeum, Thermotoga
lettingae, Tepidanaerobacter acetatoxydans, Clostridium ultunense, Syntrophaceticus schinkii)
(Sun et al., 2014).
9
There are other groups of methanogens named methylotrophic methanogens, which are generally
less dominant in biogas digesters fed with agricultural waste products and energy crops and hence,
less studied in AD. This group is capable of disproportionation of methyl-group containing
compounds, including methanol, methylated amines and organic sulfur compounds (methanethiol
and dimethylsulfide) to methane and carbon dioxide. They are limited to the order of
Methanosarcinales, except for Methanosphaera species, which belong to the order of
Methanobacteriales (Liu & Whitman, 2008).
In addition to the main pathways discussed above, homo-acetogenic bacteria can utilize H2 and CO2
for the production of acetate via the homo-acetogenesis pathway according to equation 6 (Table 6).
The function of homo-acetogenic bacteria in the overall process of biogas production is less
understood. Members of homo-acetogenic bacteria are also capable of participating in sugar
fermentation and degradation of special substrates such as methyl-group containing compounds
(Schink & Stams, 2006). Their activity to perform homo-acetogenesis pathway is highly sensitive to
the amount of hydrogen in the system. In the presence of H2-consuming microorganisms (e.g.
hydrogenotrophic methanogens), the amount of hydrogen is kept very low. In such environment,
homo-acetogenic bacteria are outcompeted by the hydrogenotrophs due to the latter having a lower
hydrogen threshold than the former (Hoehler et al., 1999). However, homo-acetogenic bacteria have
been shown to be stimulated in a high hydrogen concentration environment such as a biohydrogen
producing digester (Siriwongrungson et al., 2007) as well as land-fill (Chen et al., 2003). Their
activity in actual biogas digester is less clear and needs further research.
2.3 Operating conditions and factors regulating biogas process
As discussed in the previous section, steady state operation of biogas process depends on the
coordinated activity of a complex microbial association. Since the growth rates and the sensitivity
toward environmental changes differ widely between the different groups, any change in operating
and environmental conditions could lead to disturbances in the balance between the different
microbial groups, which might lead to reactor failure. As a result, biogas plants are often operated
with sub-optimal organic loading rates (OLR) to avoid these problems, which in turn reduces the
productivity of the biogas plants (Ahring, 2003). Therefore, it is important to use an appropriate
monitoring and control tool to maintain process stability at optimum OLR with maximum
productivity.
In the following section some of the most important environmental and operating conditions such
as temperature, pH, SCFA, ammonia, OLR, hydraulic retention time (HRT) and feeding interval
that regulate biogas process are discussed. Moreover, some of the process parameters that indicate
process imbalance are discussed.
2.3.1Temperature
AD of organic matter to methane has been documented under different temperatures ranges and
these ranges are classified into three: psychrophilic (0-20 °C), mesophilic (25-40 °C) and
thermophilic (50-65 °C) conditions. Most industrial biogas plants are operated either at mesophilic
or thermophilic temperatures (Angelidaki & Ahring, 1994).
Temperature is the most important parameter since it influences microbial activity, hydrolysis
kinetics, solubility of different chemicals in a digester, equilibrium reactions and the microbial
community structure. The reaction rate of many chemical reactions increased with an increase in
10
temperature according to Arrhenius equation. In AD in particular, hydrolysis rate and activity of
microorganisms are increased with an increase in temperature (Ahring, 2003). Therefore,
thermophilic digestion leads to a higher efficiency in the degradation of organic material than
mesophilic digestion and hence, operate at a lower hydraulic retention time (HRT) (Levén et al.,
2007). Another advantage of thermophilic digestion is that it provides sanitation because of
pathogen destruction is more effective at higher temperature (Arthurson, 2008). Despite these
advantages, the thermophilic digestion is more sensitive to environmental disturbances than
mesophilic digestion. For instance, mesophilic digester are commonly less affected by inhibitory
effects of ammonia released during the degradation of protein rich substrate (Angelidaki & Ahring,
1994). It should be noted that the need for heating digester and substrate is greater in thermophilic
digestion in comparison to the mesophilic digestion. Temperature has also significant influence on
microbial community since members of certain microorganisms have specific temperature range for
optimum growth. In general the microbial communities in thermophilic digestion are more diverse
than the one in mesophilic digestion (Levén et al., 2007). Therefore, the decision to run reactors at
thermophilic or mesophilic temperature depends on several factors such as whether there is a need
for heating digester, substrate composition, relevant environmental regulation and others.
2.3.2 pH
The pH requirements that are optimal for the growth of microorganisms vary for different groups of
microorganisms. Low pH (5 to 6.5) is generally optimum for the growth of fermentative bacteria,
which are responsible for enzymatic hydrolysis of polymers to monomers and subsequent
conversion to acids. Neutral pH is optimum for the growth of methanogens. Methanogens are
known to be more sensitive to pH changes than the fermentative bacteria. pH is also an important
parameter since the toxicity of intermediates such as ammonia and SCFA is a function of the pH of
the system. In general, a pH range between 6.8 and 8 is suggested to be an optimum condition for
operating biogas plants.
The buffer capacity of the system that is expressed in terms of alkalinity is an important parameter
that provides resistance to significant and rapid changes in pH. A kinetic uncoupling of acid
producer and consumer is usually associated with an accumulation of SCFA and thus, pH decreases
in less buffered system (Ahring et al., 1995). Since pH reduction is associated with process
imbalance, alkalinity or pH is used as a tool for monitoring process imbalance. However, in highly
buffered systems, pH changes can be small, even when the process is extremely stressed, suggesting
pH is less important to indicate process imbalance in this condition (Angelidaki & Ahring, 1994).
This shows that the use of pH as a tool for monitoring process depends on the specific reactor
system and operating condition.
2.3.3 Short chain fatty acids (SCFA)
SCFA including acetate are key intermediates in an anaerobic digestion of organic matters to
methane (Boe & Angelidaki, 2012; Boe et al., 2007; Diamantis et al., 2006; Mulat et al., 2014;
Wagner et al., 2014). Accidental reactor overload could lead to reactor acidification. This is
associated with the accumulation of SCFA as a consequence of kinetic uncoupling between acid
producing and acid consuming microorganisms as discussed above (Ahring et al., 1995). In such
condition, the accumulation of SCFA could reach the level that may have a direct toxic effect and
could also lower the pH to suboptimal values that may further reduce the activity of methanogens
(Franke-Whittle et al., 2014a; Weiland, 2010). Since the concentration and dynamics of SCFA
reflect process imbalance, individual and total SCFA concentrations are the most commonly
11
monitored parameters in AD. However, how these parameters should be used for process control is
still less clear. Nevertheless, several studies have underlined that monitoring the relative changes of
individual and total VFA levels over time is more important than their absolute concentrations to
indicate process imbalance (Angelidaki et al., 1993; Franke-Whittle et al., 2014b; Ward et al.,
2011b).
The microbial community structure can also be influenced by the concentration of SCFA. A
previous study showed that the concentration of acetate has a limiting factor on the growth of
certain members of acetoclastic methanogens (Karakashev et al., 2005). Methanosaeta spp. has a
lower acetate threshold whereas Methanosarcina spp. have higher acetate threshold. Therefore, in
the presence of high acetate concentration, Methanosarcina spp. has competitive advantage to
dominate the methanogenic communities (Karakashev et al., 2005).
2.3.4 Ammonia
Ammonia is released during AD of substrates rich in proteins. Ammonia is necessary for the growth
of microorganism in AD but when the concentration exceeds a certain level it has inhibitory effect
on the microorganisms. Methanogens are the most sensitive to ammonia level. It is the free
ammonia that has inhibitory effect on methanogens. Since free ammonia is in equilibrium with
ammonium ion, the level of ammonia is a function of the pH, ammonium concentration and
temperature of the system. For the same amount of ammonium concentration, the proportion of
free ammonia to ammonium concentration increases with an increase in pH and temperature. As
discussed before, thermophilic digestion is susceptible to ammonia inhibition compared to
mesophilic temperature. This is because of an increase in free ammonia level with an increase in
temperature. Temperature also indirectly increases the pH due to the decrease in solubility of
carbon dioxide at higher temperature and thereby further increases the level of free ammonia in the
system (Angelidaki & Ahring, 1994).
Ammonia has been shown to be a selective agent for certain groups of methanogens to dominate the
microbial community partly due to their higher tolerance to the inhibitory level of ammonia than the
others. It is well known that hydrogenotrophic methanogens and the genus Methanosarcina is more
tolerant to higher ammonia level whereas members of the genus Methanosaeta are sensitive to
ammonia (Karakashev et al., 2005) and it may no longer detected at total ammonia nitrogen (TAN)
concentrations exceeding 2.5 gNH4+-N L
-1 (De Vrieze et al., 2012; Nettmann et al., 2010). Another
study demonstrated that Methanosarcina dominated distillers dried grains with solubles (DDGS)
fed reactors operated at high OLR of 5 gVS L-1
d-1
with TAN concentration of 2.94 gNH4+-N L
-1
whereas Methanosaeta dominated at low OLR of 2 gVS L-1
d-1
with TAN concentration of 1.82
gNH4+-N L
-1 (Ziganshin et al., 2011).
2.3.5 Organic loading rate (OLR)
Organic loading rate (OLR) represent the amount of feed added into a digester per unit of time.
Depending on substrate, temperature and reactor design, different range of OLR are employed.
Typical well-functioning thermophilic digester can be loaded in the range of 4-5 kg VS m-3
d-1
whereas mesophilic digester has a load of 2-3 kg VS m-3
d-1
.
An accidental increase in organic loading is the most common disturbance, which could lead to
process instability and process failure in the worst case (Ahring et al., 1995). Depending on the
substrate, parameters such as SCFA and hydrogen concentration have been suggested as tools for
12
monitoring process imbalance so that corrective action can be employed before the process
collapses. Hydrogen and SCFA have been suggested as a good parameter for a digester treating
carbohydrate rich substrate. Hydrogen closely follows SCFA accumulation in this digester (Boe et
al., 2010). For a digester treating sewage sludge and rape seed oil, the concentration ratio of
volatile fatty acids to calcium acted as an early warning indicator (Kleyböcker et al., 2012).
2.3.6 Hydraulic retention time (HRT)
The hydraulic retention time (HRT) is a term commonly used to represent the statistically average
residence time of the soluble substrate in the digester. The HRT, which depends on the
characteristics of feedstock, reactor design temperature of the digester and environmental
conditions, should be long enough to allow metabolism by organisms for the degradation of organic
material to biogas. For slowly degradable substrates, the HRT is normally longer to allow the
solubilization of the organic material efficiently and in this case, hydrolysis is considered as a rate-
limiting step. Continuous stirred tank reactor (CSTR) is operated at longer HRT (10-60 days).
Anaerobic digestion is sensitive to change in HRT. A change in HRT could lead to shift in
methanogenic pathways and in the worst case process failure. Previous study showed that for
acetate fed digester under mesophilic condition, SAO was reported as the primary pathway at low
dilution rate (0.025 day-1
) whereas AM dominated at a higher dilution rate (0.6 day-1
) (Shigematsu
et al., 2004).
2.3.7 Feeding interval
Biogas plants are traditionally operated with a continuous and constant substrate feed to achieve
nearly the same amount of biogas and electricity generation throughout the day. Recently the
importance of flexible biogas production to balance the supply of electricity generated from
fluctuating sources such as solar and wind has been emphasized (Hahn et al., 2014; Szarka et al.,
2013). Flexible biogas production can be attained by feeding substrate at different intervals in order
to produce more biogas during high energy demand periods (Lv et al., 2014b; Mauky et al., 2014).
Previous studies showed that feeding at different intervals have an effect on process dynamics,
depicted from different SCFA accumulation rate, pH and biogas production as well as influencing
the methanogenic community and methanogenic pathways. Once per day fed reactor with maize
silage showed a transient accumulation of VFA after feeding event and subsequent utilization of the
VFA afterwards. The methanogenesis pathway was highly dynamic in this reactor which correlated
with the change in concentration of VFA. On the other hand, the reactor that was fed twice per day
did not show significant change in concentration of VFA and gas production between feeding
events (Lv et al., 2014b). Another study of acetate-fed reactors demonstrated that an hourly fed
reactor was dominated by the strictly acetoclastic Methanosaeta, whereas a daily fed reactor was
dominated by the Methanosarcina (Conklin et al., 2006). The Methanosarcina dominated reactor
was more tolerant to environmental perturbation (organic overloading) compared to the
Methanosaeta dominated reactor. Another study based on synthetic feed compared the effect of
daily and every 2 days feeding on bacterial community dynamics and the results showed that the
latter has higher tolerance to organic shock load of 8 gCOD L-1
and high total ammonia nitrogen
(TAN) levels up to 8000 mgNH4+-N L
-1 (De Vrieze et al., 2013). The dominancy of one or the other
methanogens in a biogas digester depends on operating and environmental conditions (Liu &
Whitman, 2008).
13
Chapter 3: General analytical and molecular biology methods
Several analytical and molecular biology methods can be employed for better understanding of key
intermediates, degradation pathways, methanogenesis and microbial community structure. The main
focus of this thesis was the development and application of stable isotope techniques for better
understanding of biogas process. Therefore, basic chemistry of stable isotope and the analytical
methods used to determine isotope composition of biogas and methane precursors are discussed in
this chapter. Other experimental details specific to particular aspects of research and routine
methods used for the measurement of basic process parameters are given in the relevant chapters
(chapters 4-9). As a complementary to isotope techniques, different molecular biology techniques
were employed during this project. Since the main focus of this work was stable isotope technique,
detail description of the molecular methods is not provided. Nevertheless, a brief introduction to the
molecular biology techniques used during this PhD study is also discussed.
3.1 Stable isotope techniques
3.1.1 Basics, notations and terminology
Isotopes are atoms that contain the same number of protons and electrons but different number of
neutrons and hence have different atomic mass. Elements of the periodic table are represented with
“atomic formula”, which is expressed as where Z, A and X represent the atomic number,
atomic mass and the symbol of the element. For instance, the most common isotope of carbon is
represented as , where 12 is the atomic mass, 6 is the atomic number (number of
proton/electron).
Isotopes are generally classified into two: stable and radioactive isotope. Radioactive isotopes are
unstable nuclei and spontaneously disintegrate and disperse extra energy by discharging radiation as
alpha, beta and gamma rays. Stable isotopes are energetically stable and do not decay (Michener &
Lajtha, 2008a). Carbon has 3 isotopes: 12
C is a light stable isotope, 13
C is a heavy stable isotope and 14
C is a radioactive isotope. Hydrogen has 3 isotopes: 1H (protium) is a light stable isotope,
2H (D,
deuterium) is a heavy stable isotope and 3H is a radioactive isotope. Table 2 lists the relative
abundances of the stable isotopes used in our study (C and H). The focus of this PhD study was on
stable isotope measurements as discussed below. We did not consider using radio isotope tracer
experiment as several difficulties are associated working with radioisotope tracer experiments
because of the requirement for strict health and safety regulations for handing radioisotopes, and
high cost associated radioactive material training, regulation, and waste disposal (Pack et al., 2011).
Another advantage of stable isotope is that multiple isotope labelling of the same or different
substrates in a single experiment can be employed to simultaneously follow the specific labeled
isotope in one or several products.
14
Table 2 Relative abundances of carbon and hydrogen stable isotopes (Michener & Lajtha, 2008a)
Element Isotope
Abundance
(%)
Relative
mass
difference
(%)
International
standard
Absolute abundance
of the standard (Rstandard)
Hydrogen 1H 99.985 100
Vienna Standard
Mean Ocean
Water (V-SMOW) 2H:
1H = 0.00015576
2H (D) 0.0155
Carbon 12
C 98.892 8.3
Vienna Pee Dee
Belemnite (V-PDB) 13
C:12
C = 0.0112372
13
C 1.108
The difference between the isotopic ratio of a reactant and the product at natural abundance is very
small, so isotopic composition of a substance is reported relative to internationally accepted
standard (Table 2) and expressed in units of parts per thousand. For instance, the stable carbon and
hydrogen isotope data was reported in delta notation (δ13
C and δD) in parts per thousand (‰) unit
versus the Vienna Pee Dee Belemnite (V-PDB) and Vienna Standard Mean Ocean Water (V-
SMOW), respectively:
δx= [(R)sample/(R)standard - 1]*103 (‰) (1)
where δx is the δ13
C or δD; R is the 13
C/12
C or D/H ratios (Whiticar, 1999).
The process of isotopic fractionation between substrate (A) and product (B) is expressed as apparent
fractionation factor, α, which can be defined as follows:
αA-B = (δA+ 1000)/(δB + 1000) (2)
Values of α for carbon isotope usually are near 1.00. If an α value is >1, it means that the
instantaneous product is enriched in the heavier isotope relative to the reactant (or substrate). Since
the isotope effects are generally very small (i.e., α~ 1), the deviation of the fractionation factor can
be expressed in terms of isotope enrichment, ε, in units of per mil (‰) as follows: (Conrad, 2005)
εA-B = (αA-B - 1 )*1000 (‰) (3)
In biogas field, the contribution of hydrogenotrophic methanogenesis to the produced methane can
be calculated by re-arranging the following mass balance equation (Conrad, 2005):
δ13
C-CH4 = fmc* δmc + (1- fmc)* δma (4)
where δ13
C-CH4, δma and δmc are the 13
C isotopic signature of total CH4, CH4 produced from direct
cleavage of acetate via acetoclastic methanogenesis (AM) and CH4 produced from the reduction of
CO2 via hydrogenotrophic methanogenesis (HM), respectively while fmc is the fraction of CH4
produced via HM.
15
In tracer experiment, as the tracer substrate is enriched with heavy isotope, the isotope data is often
reported as atom%:
Atom%=[(R)sample/(R)sample + 1]*100 (5)
R is the ratio of heavy-to-light isotope (Michener & Lajtha, 2008b).
In this PhD project, we employed two approaches based on stable isotope techniques. One is the
determination of the stable isotope composition of labeled compounds in conjunction with tracer
(labeled substances) experiments and the other is the measurement of stable isotope composition of
substances at natural abundance (unlabeled substances). In both cases the stable composition of a
substance (or more often the ratio of the heavy to light isotope) was measured by a mass
spectrometer. These two approaches are discussed in detail as follows.
3.1.2 Tracer experiment
This approach involves applying trace amount of labeled substances (usually heavy isotope labeled)
and subsequent measuring of the flows and fate of the heavy isotope by either low resolution or
high resolution mass spectrometer. Labeled substances are usually enriched with higher proportion
of heavy isotope and hence they have isotope range outside of the natural isotope found in the
unlabeled substances. The labeled isotope is sometimes termed as “tracer” and the naturally
abundant isotope (unlabeled isotope) is known as “tracee”. Tracer experiments are widely applied
for understanding processes in several fields of science and biogas field in particular (Conrad,
2005). One of the applications is to follow the incorporation of heavy isotope into the produced
products when labeled isotope is used as a substrate (reactant). Depending on which products would
incorporate the heavy isotope and the level of the heavy isotope incorporation, the results can be
used for identifying the reaction pathway or determining reaction kinetics.
In particular to biogas field, incubation of 13
C labeled acetate under anaerobic condition and the
subsequent monitoring of the incorporation of the 13
C into the produced CH4 and CO2 with gas
chromatography mass spectrometry (GC/MS), is widely employed for identifying specific
degradation kinetics and methanogenic pathways (Sasaki et al., 2011b). More recently, we
employed an on-line method based on membrane inlet quadrupole mass spectrometry (MIMS) for
measuring the isotope composition of dissolved CH4 and CO2 in conjunction with 13
C labeled
acetate (Mulat et al., 2014). The principle of MIMS method is further discussed in the next section.
Another application is the use of 13
C labeled CO2 for monitoring the incorporation of 13
C into
acetate in order to prove the activity of H2-dependent homo-acetogenesis pathway under the studied
anaerobic digestion. This application is demonstrated in the paper IV (chapter 7) and the principle
of GC/MS method is discussed in the next section.
3.1.3 Stable isotope analysis at natural abundance and fractionation factor
In this approach, labeled substances are not used. Instead the isotope ratio of a substance at natural
abundance is measured. The difference of the stable isotope ratio of substrate and product can be
used to identify which chemical or biological or physical process is involved during the conversion
of substrate to products. An important phenomenon known as fractionation factor is used to
understand the isotope difference resulting from certain processes.
16
The number of electrons determines the chemical reactions. Since isotopes of an element have the
same number of electrons, the chemical behavior of isotopes is generally regarded as “qualitatively”
similar. However, the atomic mass controls the vibrational energy of the nucleus and hence
difference in atomic mass number as in isotopes lead to difference in both reaction rate and bond
strength. Moreover, the kinetic energy (KE) of a molecule expressed as, KE=1/2mv2, where m is the
mass and v is the velocity, is also affected by the mass number. This leads to different velocities for
the molecules with different masses known as isotopomers (e.g. H216
O vs. H218
O). Another
important physical law that governs the behavior of isotopes is the frequency of vibration. Since
heavy atoms vibrate more slowly than the light isotope, molecules with heavy isotope have lower
energy and hence, more stable and stronger bond. The differences in velocity and bond strength
among isotopes and isotopomers lead to different distribution of light and heavy isotopes between
source substrate or reactant and product during chemical, physical or biological transformations.
This difference in isotopic distribution is known as isotope fractionation (Michener & Lajtha,
2008a).
Several mechanisms could lead to isotope fractionations and the most important one are equilibrium
(or thermodynamic) and kinetic effects. Chemical equilibrium fractionation occurs when the
distribution of isotopes in a reactant and product involved in a (thermodynamic) equilibrium
reaction differs. Kinetic isotope fractionation is determined by both bond strength and isotope
kinetics as discussed above. This mechanism is expressed in several processes such as diffusion,
evaporation, biological reactions etc. It results from irreversible i.e. one-way physical, chemical or
biological processes. During these processes, the kinetic isotope effect leads to the incorporation of
light isotope in the product leaving the heavy isotope in the reactant. Most often, the kinetic isotope
effects lead to larger fractionation factor than the equilibrium isotope effects (Johnson et al., 2004;
Michener & Lajtha, 2008a). Figure 2 shows the relative changes in the ratio of heavy-to-light
isotopes of substrate, instantaneous product and the cumulative product during the unidirectional
kinetic fractionation processes.
Figure 2 Relative changes in δ values of substrate, instantaneous product, and cumulative product
during unidirectional kinetic fractionation processes. The solid curve (upper) represents the
substrate, the thick dotted curve (middle) represents the instantaneous product, and the dashed curve
(lowest δ values) represents the cumulative product. The horizontal line is drawn to highlight the
17
fact that initial substrate and cumulative product have the same isotopic composition if the reaction
goes to completion (i.e., all substrate is consumed). The fractionation factor (ε) is constant
(Michener & Lajtha, 2008a).
In the field of anaerobic digestion, measurement of stable isotope composition of CH4 and methane
precursors such as CO2 and acetate has been extensively used for quantifying metabolic pathways in
environmental research fields (Conrad, 2005). However, its application in the engineered biogas
digesters is in its infancy (Keppler et al., 2010; Nikolausz et al., 2013). The isotope signature of the
produced biogas (CH4 and CO2) in anaerobic digesters can be used to identify methanogenic
pathways because different methanogenic pathways result in a distinct variation in isotope
abundances of the produced biogas (Conrad, 2005). HM leads to higher fractionation than AM.
Therefore, the 13
C isotope signatures of CH4 (δ13
C-CH4) produced through HM is more depleted in 13
C than those produced through AM. These distinct isotope signatures of CH4 due to the different
methanogenic pathways can be used to identify and quantify methanogenic pathways in AD.
Similar to the δ13
C-CH4, the hydrogen isotope signature of CH4 (δD-CH4) may provide additional
information about the dominant methanogenesis pathways during AD (Nikolausz et al., 2013).
Since the stable isotope ratio between the heavy and light stable isotopes differs by only few
percent, a highly precise analytical techniques such as isotope ratio mass spectrometry (IRMS)
(Meier-Augenstein, 1999), cavity ring-down spectroscopy (CRDS) (Zare et al., 2009) and tunable
diode laser absorption spectroscopy (TDLAS) are required (Keppler et al., 2010). Gas phase
samples (e.g., CO2 and CH4) are measured offline with the former technique whereas the latter two
enables on-line measurement of the carbon isotope ratios of CH4 at higher time resolution (Keppler
et al., 2010; Zare et al., 2009). Compared to TDLAS, cavity ring-down spectroscopy (CRDS)
offers an improved analytical sensitivity due to the use of extremely long path lengths when making
absorption measurement.
Stable isotope analysis of biogas at natural abundance can be used not only for identifying
methanogenic pathways, but also as a tool for process monitoring. Measurement of stable isotope
ratio of biogas at natural abundance has been suggested as a tool for process monitoring (Lv et al.,
2014b; Nikolausz et al., 2013; Polag et al., 2014). More recently, TDLAS was applied for real-time
monitoring of stable carbon isotopes of methane (δ13
CH4) in a pilot-scale biogas digester fed with
maize silage and the results suggested that the δ13
CH4 responded earlier than other basic process
parameters to process perturbation with high organic loading rate (Polag et al., 2014). In another
study, the stable isotope signatures of biogas showed a temporal variation in a less frequently fed
reactor which was highly correlated with the availability of methane precursors and the change in
the activity of methanogens (Lv et al., 2014b). Early warning tool based on the measurement of the
stable isotope signature of the produced biogas has been demonstrated for a reactor fed with co-
digestion of dry chicken waste and maize silage to ammonia inhibition (Lv et al., 2014a).
In our study, we used GC/C/IRMS for measuring the carbon and hydrogen isotope ratios at natural
abundance and hence, brief introduction about this method is described in the following section.
The main focus of our study was to demonstrate how the knowledge of isotope signatures of biogas
can be used as a tool for process monitoring. The use of GC/C/IRMS is less suitable for process
monitoring at full-scale biogas plant since it is expensive, bulky and heavy, and therefore generally
confined to laboratory settings. Optical spectrometers such as TDLAS are cheaper, more compact
and easier to operate at ambient condition compared to IRMS (Keppler et al., 2010). Therefore,
18
future application of isotope measurement as a process monitoring tool requires a development of
much cheaper optical spectromer than those existing today.
3.1.4 Gas chromatography (GC/MS)
Compared to IRMS, conventional gas chromatograph mass spectrometry (GC/MS) has low mass
resolution and thus, has limited accuracy and resolution. Therefore, it is typically used to measure
the isotope ratio of labeled substance. The GC/MS system contains five main parts: inlet system,
column, ion source, mass analyzer and ion detector. The inlet system allows an introduction of
sample (either gas or liquid sample depending on the instrument). In most cases, helium flow
carries the gaseous molecules into a column where separation of mixture of compounds takes place.
The choice of an appropriate column is very important to get a well separated chromatogram that
allows sufficiently accurate measurement of the isotopes. Then the individual separated gaseous
molecules are transferred into an ion source where neutral molecules are converted to ions. In a
conventional GC/MS, electron impact (EI) or chemical ionization can be used. The choice of
ionization source depends on factors such as the nature of the sample and the type of information
required from the analysis and others. We discuss briefly the principle of EI, since our work was
based on this ionization system. In the EI, gaseous molecules are bombarded with high energy
electrons (usually 70 eV) in order to generate ions. Then the ions are transferred into the mass
analyzer where they are separated according to their m/z. In our study, we used two different mass
spectrometer equipped with different mass analyzer: quadrupole mass filter for the MIMS system
(Paper I; in chapter 4) whereas the GC/MS used for the isotope measurement of acetate is equipped
with a quadrupole ion trap (Paper V; in chapter 7). The diagram that illustrates the configuration of
quadrupole ion trap and quadrupole mass filter is shown in Figure 3. The detail working principle
of the quadrupole mass filter and quadruple ion trap is discussed elsewhere (March, 2009). After the
separation of the ions, they are detected by an ion detector. Low resolution MS is normally
equipped with a single detector and therefore cannot simultaneously detect particular isotope pairs
for isotope ratio measurement. Finally the isotope abundance of individual ion is collected by a
computer and then presented as atom% as discussed before.
Figure 3 (A) A quadrupole mass filter. The ions enter and travel in the z-direction, while oscillating
in the x–y plane. The oscillation is controlled by the direct current (U) and radio frequency (V)
potentials applied to each pair of rods. Only those ions with stable trajectories at the selected U and
V values will travel the length of the quadrupole mass filter and be detected (March, 2009); (B) A
cross-section of an ion trap showing a mass-selective ion ejection and mass analysis occurs. (1) a
(A)(B) 1
2
3
19
filament which generates electrons; (2) an ion trap with enclosed EI source; (3) faraday cup which
detects ion (March, 1997).
3.1.5 Membrane inlet quadruple mass spectrometry (MIMS)
In this PhD study, we demonstrated for the first time the use of MIMS method for monitoring the
temporal variation in isotopic distribution of dissolved CH4 and CO2 in anaerobic digestion in
conjunction with 13
C labeled acetate (Paper I; in chapter 4) (Mulat et al., 2014). The MIMS system
contains two parts: a membrane probe and a quadrupole mass spectromer (QMS). A schematic
picture of the experimental set-up for the MIMS measurement of dissolved CO2 and CH4 in
fermentation broth under anaerobic condition is presented in Figure 4. In our study, the membrane
probe is directly submerged into a fermentation broth to allow the transfer of dissolved gaseous
molecules into the high vacuum section of the QMS. The processes that governs the transfer of the
permeant (gaseous molecules) from the sample into the QMS is known as pervaporation, which
follows three steps: (1) permeant adsorption and solubilization in the membrane ; (2) permeate
through the membrane ; and (3) evaporation into the mass spectrometer section (C. B. Silva et al.,
1999). A membrane probe is generally made from a thin semi-permeable membrane (mostly
silicone) supported by a thin perforated stainless steel plate. The principle of ion generation, ion
separation and detection in the QMS has been described in the previous section. In our study, the
transfer line between the membrane probe and the QMS is placed in a cold trap (dry ice) system to
effectively condense the water vapor that otherwise enters the QMS and interferes with the
measurement of 13
CH4 (Mulat et al., 2014).
Figure 4 A schematic representation of MIMS measurement set-up for measuring the carbon
isotope ratio of dissolved CO2 and CH4 in a fermentation broth under anaerobic condition. The
picture on the top is a membrane probe. Adapted from (Shevela & Messinger, 2013)
The application of MIMS for the detection of dissolved gases and volatile organic compounds in
fermentation reactors (Bohátka, 1997; Lloyd et al., 1985; Mulat & Feilberg, 2005; Tarkiainen et al.,
2005) and specifically in a biogas process (Bastidas-Oyanedel et al., 2010; Lloyd et al., 1985; Ward
et al., 2011a) was reported earlier. MIMS is a highly sensitive, rapid, simple, accurate method that
does not require a sample preparation step and provides high time resolution and sample frequency
(Bastidas-Oyanedel et al., 2010; Davey et al., 2011; Tarkiainen et al., 2005). Despite these
advantages, the application of MIMS based on quadrupole mass analyzer for measuring the isotopic
ratio of substances is limited. MIMS was used for measuring nitrogen isotope ratio of N2 gas for
studying denitrification and nitrogen fixation in conjunction with 15
NO3- tracer experiment in
20
aquatic systems (An et al., 2001; Steingruber et al., 2001). The MIMS method with 15
N isotope
pairing was successfully applied for estimating the denitrification and nitrogen fixation
simultaneously in aquatic systems.
3.1.6 Isotope ratio mass spectrometry (IRMS)
Stable isotope ratios are typically measured by isotope ratio mass spectrometry (IRMS), which was
invented by J.J. Thompson in 1910. In contrast to the GC/MS and MIMS described in previous
sections, IRMS can measure isotopic composition at low enrichment and natural abundance level.
At natural abundance measurement means minute variations in less abundant isotope (usually
heavier isotope) are detected in the presence of large amounts of the lighter isotope. Precision and
accuracy of carbon isotope measurement is better than 0.1 ‰ when sufficient material is available
for analysis. For D/H, the precision is much lower because of the low D abundance, and accuracy is
often less than 5 ‰ due to instrumental artifacts (Ireland, 2013). Recently the use of continuous
flow (CF) IRMS allows introduction of a sample containing a mixture of compounds that can be
separated into individual components for subsequent isotope analysis of individual components
compared to the traditional dual-inlet (DI) IRMS (Michener & Lajtha, 2008a).
The basic components of the IRMS contain inlet system, ion source, mass analyzer and ion detector
(Figure 5). Samples are introduced into the IRMS system through different inlet systems depending
on the nature of the analyte (solid or gaseous mixture samples). If the sample is solid, it has to be
first combusted in an elemental analyzer (EA) prior to gas chromatograph (GC) and then carried
through the GC by using helium flows. This type of configuration is known as elemental analyzer
isotope ratio mass spectrometry, EA/IRMS. Then the gas molecules pass to the ion source where
they are converted to ions. Electron impact source is the most commonly used ion source for
isotope analysis by IRMS. In the ion source under vacuum (~10-8
torr), electrons are generated and
accelerate to collide with gaseous analytes and form ionized molecules. Then the ion beam enters
into the mass analyzer, where they are separated based on their m/z. The mass analyzer used in
IRMS is a magnetic sector which provides a very high resolution to accurately determine the
isotope ratio. Then the separated ions are detected by the ion detector. IRMS is equipped with 3 or
more faraday cups to detect specific ions (e.g., m/z 44, 45 and 46 in case of CO2) simultaneously.
The ion flowing through a resistor that is coupled with the faraday cup creates a voltage that is
registered by a computer. A software is used to convert the relative abundance of the ions to isotope
ratio and then presented as a δ as discussed before (Michener & Lajtha, 2008a).
21
Figure 5 Schematic representation of the continuous flow elemental analyzer interfaces to the ion
source of an isotope ratio mass spectrometer (center).
If the sample is a mixture of gases, a different configuration than the one described above should be
used. In this case, GC is coupled with IRMS to allow separation of individual components of the
sample mixture prior to isotope ratio measurement. For the analysis of carbon and hydrogen isotope
ratio, additional interfaces such as combustion and pyrolysis units are normally used between GC
and IRMS when the individual components are different from CO2 and H2, respectively. Individual
analytes should be oxidized to CO2 in the combustion unit whereas hydrogen atom-containing
analytes are converted to H2 in the pyrolysis unit. The former is referred as gas chromatography
combustion isotope ratio mass spectrometry (GC/C/IRMS). This method allows compound specific
stable isotope analysis (CSIA) where the isotopes of individual components in chemically complex
samples can be determined due to the separation of analytes in the GC system.
3.2 Molecular biology methods
In our study, terminal restriction fragment length polymorphism (T-RFLP) and protein based stable
isotope probing (protein-SIP) were used to characterize the structure and function of the microbial
community in AD. A brief introduction of these methods is provided as follows.
3.2.1 T-RFLP analysis
To understand the structure of microbial community in natural environment, methods based on
cultivation-independent molecular biology has been used frequently for their rapid response that
allows monitoring of community structure of large number of samples in short time. These
22
techniques include fingerprinting techniques such as terminal restriction fragment length
polymorphism (T-RFLP) analysis, denaturing gradient gel electrophoresis and automated ribosomal
intergenic spacer analysis. Due to its simplicity, T-RFLP analysis is the most frequently used high-
throughput fingerprinting methods in diverse environment such as soil, marine, activated sludge and
biogas digesters. It should be noted that T-RFLP has limitations which could arise due to
incomplete restriction digestion and problems inherent to any PCR-based method (Nocker et al.,
2007; Schütte et al., 2008). Despite its limitations, T-RFLP become a valuable method for rapidly
monitoring changes of the dominant microorganisms in AD on temporal scales and in response to
different operating condition (Nikolausz et al., 2013; Nocker et al., 2007; Ziganshin et al., 2011).
We limited our discussion to T-RFLP analysis, since we employed this method for identification of
microbial communities at DNA and transcript level (mRNA) in AD during the PhD project.
In T-RFLP analysis, rDNA (16S rRNA gene) or mRNA transcript are used as a biological marker to
provide some important insight into the microbial community with respect to presence, diversity,
activity and phylogenetic relationships of microorganisms in AD in particular and complex habitat
in general. It needs to be emphasized that DNA is known to persist in the environment after cell
death as extracellular DNA, which makes it difficult to determine the functional role of the
microorganisms from which the 16S rRNA gene sequences are recovered. In contrast, RNA is labile
and total ribosome numbers have been correlated with cellular activity and hence, mRNA at
transcript level might reflect the diversity of the metabolically active members of the community
(Mengoni et al., 2005). As an alternative to rDNA and mRNA biological markers, functional
marker can be used to identify specific group of microorganisms. One example is the use of the
alpha subunit of methyl coenzyme M reductase gene (mcrA) in methanogen phylogenies. Since all
known methanogens express the methyl coenzyme M reductase (MCR) that catalyzes the terminal
step in methanogenesis pathway, mcrA based T-RFLP analysis is a specific method for
identification of methanogens in complex environmental samples (Friedrich, 2005).
Typical steps in the analysis of the microbial community composition based on T-RFLP profiles of
16S rRNA genes, mRNA transcripts and functional genes of mcrA is shown in Figure 6. It involves
extraction of nucleic acids (either total DNA or mRNA) followed by polymerase chain reaction
(PCR) amplification of the target genes (either 16S rRNA genes or mcrA genes) from the total
community DNA using appropriate fluorescently-labeled primers. For mRNA transcript analysis,
cDNA should be generated by reverse transcription of mRNA prior to PCR implication step. Then,
the resulting mixture of rRNA or mcrA gene amplicons are purified and digested with restriction
enzymes that have usually 4 base pair (bp) recognition sites. The fluorescently-labeled terminal
restriction fragments (T-RFs) generated in the previous step was further separated and detected by
capillary electrophoresis on an automated DNA sequencer. The resulting peaks on a typical
electropherogram represent the T-RFs of various sizes and heights that reflect the composition of
the numerically dominant populations in the community. Further information on the identity of the
microbial community can be obtained when coupled with 16S rRNA or mcrA clone library
construction and close sequencing (Schütte et al., 2008). For cloning and sequencing, the target
genes are amplified with PCR using appropriate fluorescently-unlabeled primers.
23
Figure 6 General scheme showing T-RFLP analysis of the microbial community composition of
anaerobic digester samples at DNA and mRNA transcript level. Clone library construction and
clone sequencing for phylogenetic analysis of the microbial community composition is depicted
with red arrows. Adapted from (Schütte et al., 2008).
3.2.2 Protein-SIP
Microbial community compositions studied by molecular techniques are hampered by the reduced
link between phylogenetic information and the specific functions carried out by these
microorganisms. It is often difficult to link which microorganisms are carrying out a specific step of
AD. Recently stable isotope probing (SIP) is increasingly being used in attempts to link the identity
of microorganisms to their functions (Jehmlich et al., 2008). With regards to biogas field, SIP
involves incubating inoculum obtained from anaerobic biogas digester with a specific stable isotope
labeled substrate (e.g 13
C-labeled acetate or 13
C-labeled lignocellulosic biomass). Then, the 13
C
enriched cellular components of the microorganism such as protein and nucleic acids (RNA or
DNA) are extracted and analyzed to determine which community members incorporated the 13
C
isotope into their cellular components. In this way it is possible to identify the biologically active
microorganisms that have incorporated the 13
C isotope and hence can be used to identify the
microorganisms that carried out the specific step of the AD process (Dumont & Murrell, 2005).
Depending on which cellular components of the microorganisms are analyzed, it is known as DNA-
SIP, RNA-SIP and protein- SIP method. The latter is a very sensitive and accurate method which
requires only a small amount of isotope incorporation (a minimum of 2% of 13
C incorporation)
whereas nucleic acid based SIP approaches require an incorporation of at least 20% of 13
C
DNA isolation
PCR (fluorescently-labeled
primers)
Cloning of unlabeled
PCR amplicons
Sequencing
Phylogenetic
analysis
Microbial communities
analysis
Anaerobic digester samples
RNA isolation
Reverse transcription
to generate cDNA
Digestion of labeled PCR amplicons
with restriction enzymes
Electropherogram
of separated T-RFs
24
(Jehmlich et al., 2010; Seifert et al., 2012). Since protein-SIP was used in experiments included in
this PhD project, this method is discussed briefly in the following.
Protein based SIP involves incubating inoculum sourced from biogas digester with stable isotope
labeled substrate (13
C labeled acetate or 13
C fully labeled maize leaf in our study) followed by
extraction of protein (Figure 1, chapter 5). Then it is tryptic digested and afterwards the peptides are
separated and detected with liquid chromatography tandem mass spectrometry (LC-MS/MS)
method. In parallel to this, the total DNA of the microbial community in the AD is extracted and
used for the construction of a metagenome which serves as the basis for identification of the
peptides. Peptides are identified by matching peptide sequence against the metagenome (merged
with a subspace of the NCBI database) using OpenMS. Additionally the peptide sequence is used to
assign proteins. Since proteins are involved in certain enzymatic activities and characteristics of
some groups of microorganisms, identification of proteins enables tracking of the phylogenic
affiliation of microorganisms as well as their function. Moreover, the degree of incorporation of the
labeled isotope can be used to measure the metabolic activity of a specific physiological group
(Jehmlich et al., 2010).
Previously this method has been used for identifying the species responsible for anoxic toluene
degradation in an artificial mixed culture fed with gluconate and [13
C7]-toluene under denitrifying
conditions (Jehmlich et al., 2008). To our knowledge, this method has not yet used for
characterization of active microorganisms in biogas digesters. In this context, we applied protein-
SIP aimed at identification of the functional microorganisms involved in recovery of accumulated
acid in AD (paper 2; chapter 5).
References
Ahring, B. 2003. Perspectives for Anaerobic Digestion. in: Biomethanation I, (Eds.) B. Ahring, I.
Angelidaki, E.C. de Macario, H.N. Gavala, J. Hofman-Bang, A.J.L. Macario, S.J.W.H.O.
Elferink, L. Raskin, A.J.M. Stams, P. Westermann, D. Zheng, Vol. 81, Springer Berlin
Heidelberg, pp. 1-30.
Ahring, B.K., Sandberg, M., Angelidaki, I. 1995. Volatile fatty acids as indicators of process
imbalance in anaerobic digestors. Applied Microbiology and Biotechnology, 43(3), 559-565.
An, S., Gardner, W.S., Kana, T. 2001. Simultaneous measurement of denitrification and nitrogen
fixation using isotope pairing with membrane inlet mass spectrometry analysis. Appl.
Environ. Microbiol., 67(3), 1171-1178.
Angelidaki, I., Ahring, B.K. 1994. Anaerobic thermophilic digestion of manure at different
ammonia loads: Effect of temperature. Water Research, 28(3), 727-731.
Angelidaki, I., Ellegaard, L., Ahring, B.K. 1993. A mathematical model for dynamic simulation of
anaerobic digestion of complex substrates: focusing on ammonia inhibition. Biotechnology
and Bioengineering, 42(2), 159-166.
Arthurson, V. 2008. Proper sanitization of sewage sludge: a critical issue for a sustainable society.
Applied and environmental microbiology, 74(17), 5267-5275.
Bastidas-Oyanedel, J.-R., Mohd-Zaki, Z., Pratt, S., Steyer, J.-P., Batstone, D.J. 2010. Development
of membrane inlet mass spectrometry for examination of fermentation processes. Talanta,
83(2), 482-492.
Batstone, D.J., Kelle, J., Steyer, J. 2006. A review of ADM1 extensions, applications, and analysis:
2002-2005. Water Sci. Technol., 54(4), 1-10.
25
Boe, K., Angelidaki, I. 2012. Pilot-scale application of an online VFA sensor for monitoring and
control of a manure digester. Water Science & Technology, 66(11), 2496-2503.
Boe, K., Batstone, D.J., Angelidaki, I. 2007. An innovative online VFA monitoring system for the
anerobic process, based on headspace gas chromatography. Biotechnology and
bioengineering, 96(4), 712-721.
Boe, K., Batstone, D.J., Steyer, J.-P., Angelidaki, I. 2010. State indicators for monitoring the
anaerobic digestion process. Water research, 44(20), 5973-5980.
Bohátka, S. 1997. Process monitoring in fermentors and living plants by membrane inlet mass
spectrometry. Rapid Commun. Mass Sp., 11(6), 656-661.
Burke, R.A. 1993. Possible influence of hydrogen concentration on microbial methane stable
hydrogen isotopic composition. Chemosphere, 26(1-4), 55-67.
C. B. Silva, A., Augusti, R., Dalmazio, I., Windmoller, D., M. Lago, R. 1999. MIMS evaluation of
pervaporation processes. Physical Chemistry Chemical Physics, 1(10), 2501-2504.
Chen, A.C., Ohashi, A., Harada, H. 2003. Acetate synthesis from H2/CO2 in simulated and actual
landfill samples. Environmental technology, 24(4), 435-443.
Conklin, A., Stensel, H.D., Ferguson, J. 2006. Growth kinetics and competition between
Methanosarcina and Methanosaeta in mesophilic anaerobic digestion. Water Environ Res,
78(5), 486-96.
Conrad, R. 2005. Quantification of methanogenic pathways using stable carbon isotopic signatures:
a review and a proposal. Organic Geochemistry, 36(5), 739-752.
Davey, N.G., Krogh, E.T., Gill, C.G. 2011. Membrane-introduction mass spectrometry (MIMS).
Trac-Trend Anal. Chem., 30(9), 1477-1485.
De Vrieze, J., Hennebel, T., Boon, N., Verstraete, W. 2012. Methanosarchina: The rediscovered
methanogen for heavy duty biomethanation. Bioresour. Technol. , 112, 1-9.
De Vrieze, J., Verstraete, W., Boon, N. 2013. Repeated pulse feeding induces functional stability in
anaerobic digestion. Microb Biotechnol, 6(4), 414-24.
Diamantis, V., Melidis, P., Aivasidis, A. 2006. Continuous determination of volatile products in
anaerobic fermenters by on-line capillary gas chromatography. Analytica Chimica Acta,
573–574(0), 189-194.
Dolfing, J., Jiang, B., Henstra, A.M., Stams, A.J., Plugge, C.M. 2008. Syntrophic growth on
formate: a new microbial niche in anoxic environments. Applied and environmental
microbiology, 74(19), 6126-6131.
Dumont, M.G., Murrell, J.C. 2005. Stable isotope probing—linking microbial identity to function.
Nature Reviews Microbiology, 3(6), 499-504.
Franke-Whittle, I.H., Walter, A., Ebner, C., Insam, H. 2014a. Investigation into the effect of high
concentrations of volatile fatty acids in anaerobic digestion on methanogenic communities.
Waste Management(0).
Franke-Whittle, I.H., Walter, A., Ebner, C., Insam, H. 2014b. Investigation into the effect of high
concentrations of volatile fatty acids in anaerobic digestion on methanogenic communities.
Waste Management, 34(11), 2080-2089.
Friedrich, M.W. 2005. Methyl-coenzyme M reductase genes: Unique functional markers for
methanogenic and anaerobic methane-oxidizing Archaea. in: Methods in Enzymology, Vol.
397, pp. 428-442.
Goberna, M., Insam, H., Franke-Whittle, I. 2009. Effect of biowaste sludge maturation on the
diversity of thermophilic bacteria and archaea in an anaerobic reactor. Applied and
environmental microbiology, 75(8), 2566-2572.
26
Hahn, H., Krautkremer, B., Hartmann, K., Wachendorf, M. 2014. Review of concepts for a
demand-driven biogas supply for flexible power generation. Renewable and Sustainable
Energy Reviews, 29(0), 383-393.
Hattori, S. 2008. Syntrophic acetate-oxidizing microbes in methanogenic environments. Microbes
Environ., 23(2), 118-127.
Hoehler, T.M., Albert, D.B., Alperin, M.J., Martens, C.S. 1999. Acetogenesis from CO2 in an
anoxic marine sediment. Limnol. Oceanogr, 44(3), 662-667.
Hori, T., Haruta, S., Ueno, Y., Ishii, M., Igarashi, Y. 2006. Dynamic transition of a methanogenic
population in response to the concentration of volatile fatty acids in a thermophilic
anaerobic digester. Appl. Environ. Microbiol., 72(2), 1623-1630.
Ireland, T.R. 2013. Invited Review Article: Recent developments in isotope-ratio mass
spectrometry for geochemistry and cosmochemistry. Review of Scientific Instruments, 84(1),
-.
Jehmlich, N., Schmidt, F., Taubert, M., Seifert, J., Bastida, F., von Bergen, M., Richnow, H.-H.,
Vogt, C. 2010. Protein-based stable isotope probing. Nat. Protocols, 5(12), 1957-1966.
Jehmlich, N., Schmidt, F., von Bergen, M., Richnow, H.-H., Vogt, C. 2008. Protein-based stable
isotope probing (Protein-SIP) reveals active species within anoxic mixed cultures. ISME J,
2(11), 1122-1133.
Johnson, C.M., Beard, B.L., Albarède, F. 2004. Overview and general concepts. Reviews in
Mineralogy and geochemistry, 55(1), 1-24.
Karakashev, D., Batstone, D.J., Angelidaki, I. 2005. Influence of environmental conditions on
methanogenic compositions in anaerobic biogas reactors. Appl Environ Microbiol, 71(1),
331-8.
Karakashev, D., Batstone, D.J., Trably, E., Angelidaki, I. 2006. Acetate oxidation is the dominant
methanogenic pathway from acetate in the absence of Methanosaetaceae. Appl. Environ.
Microbiol., 72(7), 5138-5141.
Karellas, S., Boukis, I., Kontopoulos, G. 2010. Development of an investment decision tool for
biogas production from agricultural waste. Renewable and Sustainable Energy Reviews,
14(4), 1273-1282.
Keppler, F., Laukenmann, S., Rinne, J., Heuwinkel, H., Greule, M., Whiticar, M., Lelieveld, J.
2010. Measurements of 13C/12C methane from anaerobic digesters: Comparison of optical
spectrometry with continuous-flow isotope ratio mass spectrometry. Environmental science
& technology, 44(13), 5067-5073.
Kleyböcker, A., Liebrich, M., Verstraete, W., Kraume, M., Würdemann, H. 2012. Early warning
indicators for process failure due to organic overloading by rapeseed oil in one-stage
continuously stirred tank reactor, sewage sludge and waste digesters. Bioresource
Technology, 123(0), 534-541.
Kotsyurbenko, O.R., Glagolev, M.V., Nozhevnikova, A.N., Conrad, R. 2001. Competition between
homoacetogenic bacteria and methanogenic archaea for hydrogen at low temperature. FEMS
microbiology ecology, 38(2‐3), 153-159.
Krakat, N., Westphal, A., Schmidt, S., Scherer, P. 2010. Anaerobic digestion of renewable biomass:
thermophilic temperature governs methanogen population dynamics. Appl. Environ.
Microbiol., 76(6), 1842-1850.
Kumar, P., Barrett, D.M., Delwiche, M.J., Stroeve, P. 2009. Methods for pretreatment of
lignocellulosic biomass for efficient hydrolysis and biofuel production. Industrial &
Engineering Chemistry Research, 48(8), 3713-3729.
Lebuhn, M., Munk, B., Effenberger, M. 2014. Agricultural biogas production in Germany-from
practice to microbiology basics. Energy, Sustainability and Society, 4(1), 10.
27
Levén, L., Eriksson, A.R.B., Schnürer, A. 2007. Effect of process temperature on bacterial and
archaeal communities in two methanogenic bioreactors treating organic household waste.
FEMS Microbiology Ecology, 59(3), 683-693.
Liu, Y., Whitman, W.B. 2008. Metabolic, phylogenetic, and ecological diversity of the
methanogenic archaea. Annals of the New York Academy of Sciences, 1125(1), 171-189.
Lloyd, D., Bohátka, S., Szilágyi, J. 1985. Quadrupole mass-spectrometery in the monitoring and
control of fermentations. Biosensors, 1(2), 179-212.
Lv, Z., Hu, M., Harms, H., Richnow, H.H., Liebetrau, J., Nikolausz, M. 2014a. Stable isotope
composition of biogas allows early warning of complete process failure as a result of
ammonia inhibition in anaerobic digesters. Bioresource technology, 167, 251-259.
Lv, Z., Leite, A.F., Harms, H., Richnow, H.H., Liebetrau, J., Nikolausz, M. 2014b. Influences of the
substrate feeding regime on methanogenic activity in biogas reactors approached by
molecular and stable isotope methods. Anaerobe, 29, 91-9.
Madsen, M., Holm-Nielsen, J.B., Esbensen, K.H. 2011. Monitoring of anaerobic digestion
processes: A review perspective. Renewable and sustainable energy reviews, 15(6), 3141-
3155.
March, R.E. 1997. An introduction to quadrupole ion trap mass spectrometry. Journal of mass
spectrometry, 32(4), 351-369.
March, R.E. 2009. Quadrupole ion traps. Mass spectrometry reviews, 28(6), 961-989.
Mauky, E., Jacobi, H.F., Liebetrau, J., Nelles, M. 2014. Flexible biogas production for demand-
driven energy supply - Feeding strategies and types of substrates. Bioresour Technol.
Meier-Augenstein, W. 1999. Applied gas chromatography coupled to isotope ratio mass
spectrometry. J Chromatogr A, 842(1-2), 351-71.
Mengoni, A., Tatti, E., Decorosi, F., Viti, C., Bazzicalupo, M., Giovannetti, L. 2005. Comparison of
16S rRNA and 16S rDNA T-RFLP Approaches to Study Bacterial Communities in Soil
Microcosms Treated with Chromate as Perturbing Agent. Microbial Ecology, 50(3), 375-
384.
Michener, R., Lajtha, K. 2008a. Stable isotopes in ecology and environmental science. 2nd ed. John
Wiley & Sons, Malden, USA.
Michener, R., Lajtha, K. 2008b. Stable isotopes in ecology and environmental science. John Wiley
& Sons.
Mulat, D.G., Feilberg, A. 2005. GC/MS method for determining carbon isotope enrichment and
concentration of underivatized short-chain fatty acids by direct aqueous solution injection of
biogas digester samples Under review in Talanta, X(X), P.X.
Mulat, D.G., Ward, A.J., Adamsen, A.P.S., Voigt, N.V., Nielsen, J.L., Feilberg, A. 2014.
Quantifying contribution of synthrophic acetate oxidation to methane production in
thermophilic anaerobic reactors by membrane inlet mass spectrometry. Environmental
science & technology, 48(4), 2505-2511.
Nettmann, E., Bergmann, I., Pramschufer, S., Mundt, K., Plogsties, V., Herrmann, C., Klocke, M.
2010. Polyphasic analyses of methanogenic archaeal communities in agricultural biogas
plants. Appl Environ Microbiol, 76(8), 2540-8.
Nikolausz, M., Walter, R.F.H., Sträuber, H., Liebetrau, J., Schmidt, T., Kleinsteuber, S., Bratfisch,
F., Günther, U., Richnow, H.H. 2013. Evaluation of stable isotope fingerprinting techniques
for the assessment of the predominant methanogenic pathways in anaerobic digesters.
Applied Microbiology and Biotechnology, 97(5), 2251-2262.
Nocker, A., Burr, M., Camper, A.K. 2007. Genotypic microbial community profiling: a critical
technical review. Microbial ecology, 54(2), 276-289.
28
Pack, M.A., Heintz, M.B., Reeburgh, W.S., Trumbore, S.E., Valentine, D.L., Xu, X., Druffel, E.R.
2011. A method for measuring methane oxidation rates using low-levels of 14C-labeled
methane and accelerator mass spectrometry. Limnol. Oceanogr.-Methods, 9, 245-260.
Polag, D., Krapf , L.C., Heuwinkel, H., Laukenmann, S., Lelieveld, J., Keppler, F. 2014. Stable
carbon isotopes of methane for real-time process monitoring in anaerobic digesters.
Engineering in Life Sciences, 14(2), 153-160.
Ryan, P., Forbes, C., McHugh, S., O’Reilly, C., Fleming, G., Colleran, E. 2010. Enrichment of
acetogenic bacteria in high rate anaerobic reactors under mesophilic and thermophilic
conditions. Water Research, 44(14), 4261-4269.
Sasaki, D., Hori, T., Haruta, S., Ueno, Y., Ishii, M., Igarashi, Y. 2011a. Methanogenic pathway and
community structure in a thermophilic anaerobic digestion process of organic solid waste. J.
Biosci. Bioeng., 111(1), 41-46.
Sasaki, D., Hori, T., Haruta, S., Ueno, Y., Ishii, M., Igarashi, Y. 2011b. Methanogenic pathway and
community structure in a thermophilic anaerobic digestion process of organic solid waste. J
Biosci Bioeng, 111(1), 41-6.
Schink, B., Stams, A.J. 2006. Syntrophism among prokaryotes. Springer.
Schmidt, J.E., Ahring, B.K. 1993. Effects of hydrogen and formate on the degradation of propionate
and butyrate in thermophilic granules from an upflow anaerobic sludge blanket reactor.
Applied and environmental microbiology, 59(8), 2546-2551.
Schütte, U.E., Abdo, Z., Bent, S., Shyu, C., Williams, C., Pierson, J., Forney, L. 2008. Advances in
the use of terminal restriction fragment length polymorphism (T-RFLP) analysis of 16S
rRNA genes to characterize microbial communities. Applied Microbiology and
Biotechnology, 80(3), 365-380.
Seifert, J., Taubert, M., Jehmlich, N., Schmidt, F., Völker, U., Vogt, C., Richnow, H.-H., von
Bergen, M. 2012. Protein-based stable isotope probing (protein-SIP) in functional
metaproteomics. Mass Spectrometry Reviews, 31(6), 683-697.
Shevela, D., Messinger, J. 2013. Studying the oxidation of water to molecular oxygen in
photosynthetic and artificial systems by time-resolved membrane-inlet mass spectrometry.
Frontiers in Plant Science, 4, 473.
Shigematsu, T., Tang, Y., Kobayashi, T., Kawaguchi, H., Morimura, S., Kida, K. 2004. Effect of
dilution rate on metabolic pathway shift between aceticlastic and nonaceticlastic
methanogenesis in chemostat cultivation. Applied and environmental microbiology, 70(7),
4048-4052.
Siriwongrungson, V., Zeng, R.J., Angelidaki, I. 2007. Homoacetogenesis as the alternative pathway
for H< sub> 2</sub> sink during thermophilic anaerobic degradation of butyrate under
suppressed methanogenesis. Water research, 41(18), 4204-4210.
Steingruber, S.M., Friedrich, J., Gächter, R., Wehrli, B. 2001. Measurement of denitrification in
sediments with the 15N isotope pairing technique. Appl. Environ. Microbiol., 67(9), 3771-
3778.
Sun, L., Müller, B., Westerholm, M., Schnürer, A. 2014. Syntrophic acetate oxidation in industrial
CSTR biogas digesters. Journal of Biotechnology, 171(0), 39-44.
Szarka, N., Scholwin, F., Trommler, M., Fabian Jacobi, H., Eichhorn, M., Ortwein, A., Thrän, D.
2013. A novel role for bioenergy: A flexible, demand-oriented power supply. Energy, 61(0),
18-26.
Tarkiainen, V., Kotiaho, T., Mattila, I., Virkajarvi, L., Aristidou, A., Ketola, R.A. 2005. On-line
monitoring of continuous beer fermentation process using automatic membrane inlet mass
spectrometric system. Talanta, 65(5), 1254-1263.
29
Thiele, J.H., Zeikus, J.G. 1988. Control of interspecies electron flow during anaerobic digestion:
significance of formate transfer versus hydrogen transfer during syntrophic methanogenesis
in flocs. Applied and environmental microbiology, 54(1), 20-29.
Wagner, A.O., Reitschuler, C., Illmer, P. 2014. Effect of different acetate:propionate ratios on the
methanogenic community during thermophilic anaerobic digestion in batch experiments.
Biochemical Engineering Journal, 90(0), 154-161.
Ward, A.J., Bruni, E., Lykkegaard, M.K., Feilberg, A., Adamsen, A.P., Jensen, A.P., Poulsen, A.K.
2011a. Real time monitoring of a biogas digester with gas chromatography, near-infrared
spectroscopy, and membrane-inlet mass spectrometry. Bioresour. Technol. , 102(5), 4098-
4103.
Ward, A.J., Bruni, E., Lykkegaard, M.K., Feilberg, A., Adamsen, A.P.S., Jensen, A.P., Poulsen,
A.K. 2011b. Real time monitoring of a biogas digester with gas chromatography, near-
infrared spectroscopy, and membrane-inlet mass spectrometry. Bioresource Technology,
102(5), 4098-4103.
Weiland, P. 2010. Biogas production: current state and perspectives. Applied Microbiology and
Biotechnology, 85(4), 849-860.
Whiticar, M.J. 1999. Carbon and hydrogen isotope systematics of bacterial formation and oxidation
of methane. Chemical Geology, 161(1–3), 291-314.
Zare, R.N., Kuramoto, D.S., Haase, C., Tan, S.M., Crosson, E.R., Saad, N.M. 2009. High-precision
optical measurements of 13C/12C isotope ratios in organic compounds at natural abundance.
Proceedings of the National Academy of Sciences, 106(27), 10928-10932.
Ziganshin, A., Schmidt, T., Scholwin, F., Il’inskaya, O., Harms, H., Kleinsteuber, S. 2011. Bacteria
and archaea involved in anaerobic digestion of distillers grains with solubles. Applied
Microbiology and Biotechnology, 89(6), 2039-2052.
30
Chapter 4: Paper I- Quantifying contribution of synthrophic acetate oxidation to
methane production in thermophilic anaerobic reactors by membrane inlet mass
spectrometry
Used with permission from Environmental science & technology 48(4), 2505-2511.
Quantifying Contribution of Synthrophic Acetate Oxidation toMethane Production in Thermophilic Anaerobic Reactors byMembrane Inlet Mass SpectrometryDaniel Girma Mulat,† Alastair James Ward,† Anders Peter S. Adamsen,† Niels Vinther Voigt,‡
Jeppe Lund Nielsen,§ and Anders Feilberg*,†
†Department of Engineering, Aarhus University, Hangøvej 2, DK-8200 Aarhus N, Denmark‡Danish Technological Institute, Kongsvang Alle 29, DK-8000 Aarhus C, Denmark§Center for Microbial Communities, Department of Biotechnology, Chemistry and Environmental Engineering, Aalborg University,Sohngaardsholmsvej 49, DK-9000 Aalborg, Denmark
*S Supporting Information
ABSTRACT: A unique method was developed and applied for monitoring methanogenesis pathways based on isotope labeledsubstrates combined with online membrane inlet quadrupole mass spectrometry (MIMS). In our study, a fermentation samplefrom a full-scale biogas plant fed with pig and cattle manure, maize silage, and deep litter was incubated with 100 mM of [2-13C]sodium acetate under thermophilic anaerobic conditions. MIMS was used to measure the isotopic distribution of dissolved CO2and CH4 during the degradation of acetate, while excluding interference from water by applying a cold trap. After 6 days ofincubation, the proportion of methane derived from reduction of CO2 had increased significantly and reached up to 87% of totalmethane, suggesting that synthrophic acetate oxidation coupled to hydrogenotrophic methanogenesis (SAO-HM) played animportant role in the degradation of acetate. This study provided a new approach for online quantification of the relativecontribution of methanogenesis pathways to methane production with a time resolution shorter than one minute. The observedcontribution of SAO-HM to methane production under the tested conditions challenges the current widely accepted anaerobicdigestion model (ADM1), which strongly emphasizes the importance of the acetoclastic methanogenesis.
1. INTRODUCTION
Acetate is a key intermediate in anaerobic digestion of organicmatter, whereby it is converted to methane by two differentpathways.1−3 In acetoclastic methanogenesis (AM), acetate isdirectly cleaved in such a way that the methyl group of acetateis converted primarily to CH4 while the carboxyl group isconverted to CO2 (Supporting Information (SI) Section 1).Acetate can also follow a two-step reaction pathway wherebyacetate is first oxidized to CO2 and H2 by synthrophic acetateoxidation (SAO) and subsequently the CO2 is reduced to CH4by hydrogenotrophic methanogenesis (HM) (SI Section 1).Under standard conditions, SAO is not thermodynamicallyfavorable (ΔG°′ = +104.1 kJ/mol) and can only be feasible ifthe H2 partial pressure is kept low by coupling with H2-consuming methanogens.3 It is generally assumed thatacetoclastic methanogenesis is a dominant pathway for methane
production from acetate. Therefore, biogas reactor operationand optimization is mostly based on maintaining favorablecondition for acetoclastic methanogenesis, with little consid-eration to the importance of the SAO pathway. However, somestudies recently found that SAO coupled to HM is a dominantpathway in thermophilic methanogenic reactors.2−7 In most ofthe previous studies, the inoculum was sourced from anaerobicdigesters treating sewage sludge2,7 and synthetic wastewater.6
The role of SAO-HM and its quantitative contribution tomethane production in a mixed culture sourced from anaerobicdigesters treating common agriculture wastes such as pig and
Received: January 15, 2014Accepted: January 17, 2014Published: January 17, 2014
Article
pubs.acs.org/est
© 2014 American Chemical Society 2505 dx.doi.org/10.1021/es403144e | Environ. Sci. Technol. 2014, 48, 2505−2511
cattle manure, maize silage, and deep litter manure is stillunclear.The relative contribution of methanogenesis pathways can be
identified and quantified by measuring the isotopic distributionof CO2 and CH4
1,8 with gas chromatography−combustion−isotope ratio mass spectrometry (GC-C-IRMS)9 and gaschromatography−mass spectrometry (GC-MS) in combinationwith 13C labeled acetate.2 In addition, radiometric measurementof 14CH4 and 14CO2 production is routinely employed toidentify methanogenesis pathways with 14C labeled substrates.4
However, the offline gas sampling procedure in thesetechniques limits the knowledge on temporal variation of theisotopic distribution of CO2 and CH4. Moreover, the exactisotopic composition of dissolved gases in a liquid is notnecessarily the same as that of gases in the headspace, since thelatter is most probably modified by additional isotope effectsduring phase transition.8
Membrane inlet quadrupole mass spectrometry (MIMS) isan alternative technique to GC-MS, GC-C-IRMS, andradiometric approach for online measurement of the isotopiccomposition of dissolved CO2 and CH4 directly in afermentation broth. Previous work has shown the suitabilityof MIMS based on a quadrupole analyzer for rapid detection ofdissolved gases and volatile organic compounds in fermentationreactors10−12 and specifically in a biogas process.10,13,14 Inaddition to being simple, accurate, and fast, with response ofseconds to minutes, MIMS is highly sensitive and does notrequire sample preparation.12,14,15
There have been only a few applications of MIMS formeasuring isotopic distribution. One example is the measure-ment of the nitrogen isotopic distribution of N2 gas with MIMSduring incubation experiments after the addition of 15NO3
−
tracer.16 The MIMS method with 15N isotope pairing wassuccessfully applied for estimating the denitrification andnitrogen fixation simultaneously in aquatic systems.17 Althoughthe quadrupole mass spectrometry has a low resolution thatlimits its use for measuring isotopic distribution at naturalabundance, it was demonstrated successfully for measuring thenitrogen isotopic distribution in conjunction with isotopepairing.17 However, there has not been any report so far on itsapplication for determining the carbon isotopic distribution ofdissolved compounds in anaerobic reactors. To our knowledge,this is the first application of MIMS based on a quadrupoleanalyzer for monitoring the temporal variation in isotopicdistribution of dissolved CH4 and CO2 in anaerobic digestionwith 13C labeled acetate.In the present study, 13C labeled acetate was used as a
substrate in thermophilic anaerobic digestion and theincorporation of 13C into CH4 and CO2 was monitored withMIMS. The source of the inoculum was a full-scale biogasdigester working with a mixture of pig and cattle manure, maizesilage, and deep litter manure. The aim of this study was two-fold. One aim was to investigate the capability of MIMS formonitoring the temporal variation in isotopic distribution ofdissolved CH4 and CO2 in a fermentation broth. The secondaim was to quantify the relative contribution of SAO-HMversus AM to methane production from high acetateconcentration (100 mM) by following the temporal isotopicdistribution of CH4 and CO2 with MIMS. Effects oftemperature and cold trap were also investigated, which areimportant for optimization of the instrument. Moreover, wereported the linearity, response time, and detection limits of theMIMS methodology.
2. EXPERIMENTAL SECTION
Sources of Inoculum. Inoculum was obtained from acommercial full-scale biogas digester at research center Foulum,Denmark. The digester works with a mixture of pig and cattlemanure, maize silage, and deep litter manure. It runs underthermophilic condition ca. 52 °C. The total solid (TS), volatilesolid (VS), pH value, and total ammoniacal nitrogen (TAN) ofthe inoculum were 75.8 g/L, 60.6 g/L, 7.70, and 1.84 g/L,respectively. The inoculum was preincubated under anaerobiccondition at 52 °C for 2 weeks prior to the main tracerexperiment to reduce the background contribution of 12CO2and 12CH4 from the original substrates.
Operation of Anaerobic Digestion. Glass serum bottles(500-mL) were used for preparing the anaerobic incubationassay. Aliquots (192 mL) of inoculum were transferred into the500-mL serum bottles, which were then sealed with butylrubber stoppers and aluminum crimps. Four treatments wereprepared: (i) 13C treatment with 13C methyl labeled acetate,[2-13C] sodium acetate, as the substrate, (ii) 13C treatment with13C fully labeled acetate, [U-13C] sodium acetate, as thesubstrate, (iii) unlabeled treatment with unlabeled sodiumacetate as the substrate, and (iv) blank reactor with distilledwater instead of substrate. The appropriate substrate was addedonce to each serum bottle to give a final concentration of 100mM. Each 13C treatment was run in duplicate, whereasunlabeled and blank treatments were run in quadruplicateunder static incubation condition for 20 days at 52 °C.
MIMS Measurement. A schematic picture of theexperimental setup for the MIMS measurement of theanaerobic incubation assay is presented in SI Figure S1.TheMIMS system consisted of a quadrupole mass spectrometer(QMS) connected to a silicone membrane probe for themeasurement of dissolved carbon dioxide and methane directlyin a fermentation broth. Characteristic ions of each compoundwere monitored in multiple ion detection (MID) mode: m/z17 for 13C methane (13CH4), m/z 15 for
12C methane (12CH4),m/z 45 for 13C carbon dioxide (13CO2), m/z 44 for
12C carbondioxide (12CO2) and m/z 18 for water (H2O). A dry ice coldtrap was used in the vacuum line to minimize interference ofwater on methane measurements.18 Two types of MIMSmeasurements were employed:
(i) MIMS measurement of standard solution: Aqueousstandard solutions of carbon dioxide and methane atdifferent concentrations were prepared from a standardgas mixture of CO2 and CH4. The dissolved concen-tration of methane and carbon dioxide in water wascalculated from the product of gas partial pressure andtemperature-corrected Henry’s constants.19 The temper-ature-corrected Henry’s constants (kH) for CH4 and CO2
at 52 °C are 0.00225 and 0.0663 M/atm, respectively,which were calculated from the Henry’s constants atstandard condition (koH), the temperature dependence ofthe Henry’s constants and the actual temperature of theincubation experiment.19
(ii) MIMS measurement of anaerobic digestion: MIMSmeasurement was carried out for duplicate experimentsfor all of the four treatments, giving a total of eightreactors to be measured every day. Because the MIMSsystem was equipped with only one membrane probe,the MIMS probe was switched among the eight reactorsand the MIMS measurement was conducted for 4 min ineach reactor. To sum up, the MIMS measurements of
Environmental Science & Technology Article
dx.doi.org/10.1021/es403144e | Environ. Sci. Technol. 2014, 48, 2505−25112506
eight reactors were conducted every 4 min for the total of20 incubation days. The details of the MIMS setup aregiven in SI Section 2.
Analytical Methods and Proteome Analysis. Gascomposition (CO2 and CH4) and volume were measuredperiodically. The concentrations of VFAs and pH weremeasured almost every day. Liquid samples for proteomeanalysis were collected from all reactors at the middle of theincubation experiment. A detail description of the method isprovided in SI Section 2.Calculation of Isotope Distribution. The proportion of
12CO2,13CO2,
12CH4, and13CH4 during the degradation of
[2-13C]acetate depends on the methanogenesis pathways(SAO-HM and AM). Moreover, the proportion of 13CH4 and12CH4 during SAO-HM pathway is affected by the kineticisotope effect, whereby the production of 12CH4 is slightlyfaster than 13CH4 during the reduction of 12CO2 and 13CO2,respectively (i.e., 12k/13k = 1.065 on average under thermophiliccondition).8,20
The proportion of 13CO2 to total carbon dioxide, 13CO2(atom%), indicates the contribution level of SAO pathway tomethane production due to the fact that 13CO2 will beproduced only via SAO oxidation of 13CH3COO
− (SI eq S2).On the other hand, a significant proportion of 13CH4 to totalmethane will be produced if 13CH3COO
− is directly cleaved byAM (SI eq S1). Hence, the proportion of 13CH4 to totalmethane production, 13CH4 (atom%), during the degradationof 13CH3COO
− ascribed to the contribution of AM to methaneproduction. Therefore, the contribution (%) that SAO coupledto HM makes to total methane production can be expressed interms of the ratio of 13CO2 (atom%) to 13CH4 (atom%). Anincrease in the ratio of 13CO2 (atom%) to 13CH4 (atom%)represents the increase in the contribution of SAO-HM tomethane production and vice versa. The terms “SAO-HM” and“SAO” are used interchangeably in our study when describingthe quantitative contribution of SAO pathway to methaneproduction.An example of a schematic representation of calculated mass
balance of 13CH4,12CH4,
13CO2, and12CO2 by assuming 50%
methane production from acetoclastic methanogenesis and theremaining 50% from SAO is shown in Figure S2 (SI Section 2).The kinetics isotope effect during the reduction of carbondioxide to methane, 12k/13k = 1.065,8,20 was included in thecalculation. Using this principle, the mass balance of eachcarbon isotope species assuming 0% up to 100% SAO pathwaywas calculated. In Figure 1, the quantitative contribution ofSAO (%) to methane production is shown as a function of thecalculated ratio of atom% of 13CO2 to
13CH4, which was linear.The linear equation obtained in Figure 1 was rewritten as eq 1to estimate the contribution of SAO to methane production asa function of 13CO2 (atom%)/13CH4 (atom%). The overalluncertainty associated with estimation of SAO (%) using eq 1as a function of MIMS data of 13CO2 (atom%)/13CH4 (atom%)was 5.9% (SI Section 7).
=+y
SAO(%)0.01
0.01 (1)
where y is 13CO2 (atom%)/13CH4 (atom%).Experimental Methods for the Quantification of SAO-
HM Pathway Using MIMS Data. In our study, MIMS wasused to measure the isotopic distribution of dissolved carbondioxide (12CO2 and
13CO2) and methane (12CH4 and13CH4)
during the anaerobic digestion of [2-13C]acetate and [U−13C]-acetate. The [2-13C]acetate substrate was used to studymethanogenesis pathways. On the other hand, the [U−13C]-acetate substrate was used to correct the background totalinorganic carbon (TIC) arising from the inoculum. Theunlabeled products (12CO2 and
12CH4) during the degradationof [U−13C]acetate were considered as the background TIC thatcomes from the inoculum and they were subtracted from theMIMS values of 12CO2 and 12CH4 measured during thedegradation of [2-13C]acetate. The background TIC wasaccounted in this way for reporting the TIC-corrected 13CO2(atom%)/13CH4 (atom%) of [2-13C]acetate reactor. Thequantitative contribution (%) that SAO makes to total methaneproduction was determined by solving eq 1 with the TIC-corrected 13CO2 (atom%)/13CH4 (atom%) of [2-13C]acetatereactor. The details of MIMS data presentation is described inSI Section 2.
3. RESULTS AND DISCUSSIONCharacteristic Performance of MIMS for CO2 and CH4
Measurement. With our constructed anaerobic reactor andapplication of the cold trap, dissolved CO2 and CH4 weredirectly measured by MIMS without interference from water.The cold trap not only effectively condensed the water vaporthat otherwise interferes with the measurement of 13CH4 (SIFigure S3), but it also reduced the pressure inside the massspectrometer (data not shown). Tests of the effect of sampletemperature on the MIMS instrument showed that safeoperation of a MIMS system is achievable for typical anaerobicdigestion temperatures (30−52 °C). It is important to maintaina constant sample temperature while measuring dissolved CO2and CH4 with MIMS; otherwise the temperature differenceleads to different signal intensities (SI Figure S4). The responsetime of the instrument was faster than 1 min for both CH4 andCO2 (SI Figure S5). The observed fast response time showedthat MIMS can be used for online and onsite measurements ofdissolved carbon dioxide and methane in real time.
External Standard Calibration and Detection Limit.Standard solutions of methane and carbon dioxide wereprepared by continuously purging a standard gas mixturethrough deionized water until equilibrium concentrations werereached. Calibration curves were obtained by plotting the peakarea of standard solutions versus the known concentrations ofthe dissolved gases. Although the cold trap was an effectivemeans for condensing the water vapor that otherwise interferes
Figure 1. Quantitative contribution of SAO (%) to methaneproduction as a function of the calculated ratio of atom% of 13CO2to 13CH4.
Environmental Science & Technology Article
dx.doi.org/10.1021/es403144e | Environ. Sci. Technol. 2014, 48, 2505−25112507
with the measurement of 13CH4, the diffusion of a smallamount of water into the mass spectrometry is unavoidable.The presence of water in the mass spectrometry affects theintensity of the ions measured by the MIMS. Therefore, theabsolute intensities were normalized with the intensity of m/z18 (H2O
+) to obtain comparable data. All the intensitiespresented in the manuscript are relative (normalized)intensities (Figures 2 and SI S6). Dissolved carbon dioxideand methane showed a linear dynamic range in theconcentration range of 2.7−27.1 mM (r2 = 0.994) and 0.14−1.4 mM (r2 = 0.998), respectively (Figure 2). The detectionlimits of CO2 and CH4 (calculated from three times thestandard deviation of the replicate blank samples) were 0.07and 0.28 mM, respectively. By implementing the cold trap, thedetection limit of CH4 was improved by a factor of 3 (from 0.84to 0.28 mM) due to the reduction of the pressure inside themass spectrometer.Degradation of Acetate and Methane Production. In
our experiment, a small concentration of acetate was detectedin the blank (inoculum only) reactors (less than 5 mM andestimated to be less than 5% of the added acetate in the otherreactors). The methane production from the blank reactor wasvery small (less than 5% of the acetate-fed reactors). Themethane production of 13C labeled reactors was corrected forthe methane produced by the inoculum. The accumulatedmethane production for the [2-13C]acetate reactor was 234.7mL/g COD and variations in methane production for theduplicate assays were less than (±3.5) mL/g COD, which isvery small (Figure 3). In addition, acetate degradation alsoexhibited very small variation (±1 to ±4 mM) for the first 14days for the duplicate assays (Figure 3). The last 6 days of theincubation period showed slightly higher variation of acetatedegradation (±5 to ±6 mM) for the duplicate assays.Methane production from degradation of 100 mM [2-13C]-
acetate followed a typical batch experiment whereby methaneproduction rate was slower at the beginning of the incubation,followed by exponential increase of methane production and alater stationary phase (Figure 3). It is clear from Figure 3 thatboth acetate degradation and accumulated methane productionshowed similar patterns for the whole incubation period. Theacetate degradation was very slow in the first 6 days of theincubation except for the first 48 h (Figure 3). Acetateconcentration was reduced by 8 mM and methane productionwas increased by 20 mL/g COD in the first 48 h. The reasonfor the high acetate degradation rate for 48 h followed by aslower degradation rate from day 3 until day 6 was apparentlyunknown. However, the slow acetate degradation rate from day
3 until day 6 and a subsequent increase in acetate degradationrate afterward until day 18 is a typical lag phase for this type ofreactor (Figure 3). A decrease in acetate concentration from 88to 23 mM from day 6 until day 18 was accompanied by asignificant increase in methane production from 40 to 288 mL/g COD (Figure 3). Almost 72% of acetate was degraded duringthis period. When the acetate concentration was less than 23mM on the last 2 days of the incubation, the accumulatedmethane production was reduced slightly (Figure 3).
Estimation of Methanogenic Pathways to TotalMethane Production from [2-13C]Acetate. The evolutionof 13CH4,
12CH4,12CO2, and
13CO2 during the degradation of[2-13C]acetate was followed using the developed MIMSmethod. The contribution of the background total inorganiccarbon (TIC) from the inoculum was corrected from themeasured [2-13C]acetate incubation and the TIC-correctedMIMS data are presented as 13CO2 (atom%)/
13CH4 (atom%)(Figure 4a) and 13CO2 (atom%) (Figure 4b). The ratio of TIC-corrected 13CO2 (atom%) to 13CH4 (atom%) was used toestimate the contribution of SAO-HM pathway to methaneproduction according to eq 1 and it is shown in Figure 4a.The proportion of 13CO2 to total carbon dioxide, represented
as 13CO2 (atom%), is a clear qualitative demonstration of theSAO pathway. The evolution of the higher amount of 13CO2compared to total carbon dioxide production can only happendue to continuous production of 13CO2 from [2-13C]acetate bySAO according to SI equation S2. Therefore, the observedincrease in 13CO2 (atom%) is a qualitative indication of the
Figure 2. Calibration curves for (a) methane (ion current m/z 15 to m/z 18 ratio) and (b) carbon dioxide (ion current m/z 44 to m/z 18 ratio)using MIMS.
Figure 3. Temporal change in the degradation of 100 mM[2-13C]acetate (▲) and accumulated methane production (■) from100 mM [2-13C]acetate. The lines represent mean values (n = 2) anderror bars denote the data range.
Environmental Science & Technology Article
dx.doi.org/10.1021/es403144e | Environ. Sci. Technol. 2014, 48, 2505−25112508
increase in contribution of SAO to acetate degradation overtime (Figure 4b). The subsequent discussion is an attempt toquantify the contribution (%) of SAO-HM during thedegradation of acetate to methane.In Figure 4a and b the temporal evolution of the ratio of
13CO2 (atom%) to 13CH4 (atom%) and 13CO2 (atom%)followed three distinct trends which can be roughly groupedaccording to the incubation time (the first 5 days, day 11−18,and the last 2 days). MIMS measurement from day 6 until day10 was not conducted due to the failure of the instrument tomaintain the required high vacuum (∼10−6 mbar). MIMSmeasurement was resumed on day 11 after the leak wasrepaired.A very small change in the proportion of 13CO2 (atom%)
from 1.7 to 5.9 was observed during the first 5 days of theincubation period (Figure 4b). Similar to the change in 13CO2(atom%), the change in the ratio of 13CO2 (atom%) to 13CH4
(atom%) was very slow (Figure 4a), i.e. it changes from 0.02 to0.09. The calculated contributions (%) of SAO-HM to totalmethane production range from 3% to 10% (Figure 4a). Thissmall contribution of SAO-HM corroborates well with the slowmethane production rate and acetate degradation rate duringthe first 6 days of the incubation period (Figure 3).As shown from TIC-corrected MIMS data for day 5 to 11,
the proportion of 13CO2 (atom%) was increased from 5.9 to10.5 (Figure 4b) and subsequently the ratio of 13CO2 (atom%)to 13CH4 (atom%) was increased from 0.09 to 0.35 (Figure 4a).The contribution of SAO-HM to methane production wasincreased by 26% from day 5 to 11 (Figure 4a). Although wemissed the MIMS measurement from day 6 until day 10, it isobvious that SAO-HM is taking place at a higher rate as shownfrom MIMS data on days 5 and 11. Accumulated methaneproduction was also increased by 54.7 mL/g COD and acetatewas reduced by 23 mM between day 6 and 11 (Figure 3).The ratio of 13CO2 to total carbon dioxide was increased
significantly from 10.5 to 26.2 atom % from day 11 to 18(Figure 4b). Similarly, the ratio of 13CO2 (atom%) to 13CH4(atom%) was increased significantly from 0.35 to 0.86 (Figure4a). The SAO-HM contribution to methane production wasincreased from 36 to 87% of the methane production in thisperiod (Figure 4a). The increase in SAO-HM over time showedthat there is a shift of methanogenic pathways from acetoclasticmethanogenesis to SAO-HM pathways during the course of theincubation of [2-13C]acetate. Accumulated methane productionwas increased significantly by 133.3 mL/g COD and acetate
was reduced by 49 mM from day 11 until day 18 (Figure 3).The shift in methanogenic pathway was accompanied by fasteracetate degradation and high methane production rate.The evolution rate of 13CO2 to total carbon dioxide was
slightly slower in the last 2 days of the incubation and remainedat 27.1 atom% (Figure 4b). Similarly, the ratio of 13CO2 (atom%) to 13CH4 (atom%) had started to decrease slightly andreached 0.79 (Figure 4a). As shown in Figure 3, the methaneproduction rate decreased as a consequence of the smallconcentration of acetate available in the reactor (Figure 3). Theslight decrease in the ratio of 13CO2 (atom%) to
13CH4 (atom%) could be attributed to a decrease in production of carbondioxide derived from the oxidation of acetate at low acetateconcentration. Despite the decrease in the ratio, SAO-HM wasstill the dominant acetate degradation pathway, whichcontributed up to 80% of the methane production from acetateeven in the presence of a small concentration of acetate. Bothacetoclastic methanogenesis and SAO-HM contributed to thedegradation of acetate at different proportion and thecontribution of the latter was increased during the course ofthe incubation. All in all, the SAO-HM contributed to 49% ofthe total methane production (147 mL of methane out of thetotal 300 mL of methane) whereas the remaining 51% of thetotal methane (153 mL of methane) was produced via theacetoclastic methanogenesis.
Microbial Community Composition. The microbialcommunity composition were analyzed by using a proteomeanalysis and revealed a heterogeneous distribution representing18 phyla, 30 classes, and 107 genera. Bacterial proteinsrepresented 74.5% of the proteins, while archaeal proteinsaccounted for most of the remaining (∼25%). The mostabundant phyla were Firmicutes (30.9%), Euryarchaeota(23.3%), Proteobacteria (16.5%), and Bacteroidetes (5.7%),while the most abundant classes were Methanomicrobia(23%), Clostridia (19.7%), and Bacilli (8%) (SI Table S2).Within the Archaea, the mixotrophic Methanosarcina (10.8%)and hydrogenotrophic Methanoculleus (8.7%) were the mostabundant methanogens. Methanoculleus sp. and Methanosarcinasp. have been reported as widely distributed in thermophilicanaerobic reactors, especially those treating manure andagricultural wastes.21,22 Similar diversity of Bacteria and Archaeacommunities were present in all acetate-fed and blank reactors(data not shown).
Merits of MIMS Measurement. In our study, the MIMSprobe was submerged in a fermentation broth and the MIMSmeasurement was taken for few minutes. The MIMS probe was
Figure 4. Temporal evolution of (a) 13CO2 (atom%)/13CH4 (atom%) (■) and the contribution of SAO (%) to methane production (□); (b) 13CO2
(atom%) (▲). The lines represent mean values (n = 2) and error bars denote the data range.
Environmental Science & Technology Article
dx.doi.org/10.1021/es403144e | Environ. Sci. Technol. 2014, 48, 2505−25112509
not submerged in one sample for the whole incubation period(20 days), since only one membrane probe was available andthere was a need to run many samples per day. In some cases,MIMS probe multiplexing can be used to continuously measuredissolved gases in several fermentors whereby multiplemembrane probes are coupled to one quadruple massspectrometer (QMS) through sufficiently long vacuum pipe-lines and a multivalve system.11 Hence, it is more ademonstration of the principle, which can be used for onlineand continuous monitoring in one or several reactors if multiplemembrane probes that work with high vacuum system areconstructed.In previous studies where GC-MS was used in combination
with 13C labeled substrates, the headspace gas samples weremeasured offline once usually at the end of the incubation.2,23
Consequently, the quantitative contribution of each meth-anogenetic pathway to methane production during theincubation period, where acetate concentration changedsignificantly, was not provided. This study, however, provideda new approach for online quantification of the relativecontribution of methanogenesis pathways to methane produc-tion with a time resolution of shorter than 1 minute. Thisshorter time resolution and continuous isotope measurementdemonstrated that an increase in the proportion of SAO-HM tomethane production was accompanied with a rapid acetatedegradation and high methane production rate in the studypresented here. Another disadvantage of the headspace gasmeasurement by GC-MS is that the isotopic composition of theheadspace gas is not necessarily the same as the dissolved gas,since the former is most probably modified by additionalisotope effects during phase transition.8 In this study, however,MIMS was applied for rapidly monitoring the isotopicdistribution of dissolved gases in a liquid.There are also inconsistencies in the way GC-MS data were
used to quantify the contribution (%) of SAO to methaneproduction. For instance, Hori et al.23 reported the proportionof methane produced via SAO pathway was equal to the valueof atom% of 13CH4 whereas Sasaki et al.
2 multiplied the value ofatom% of 13CH4 by 2 to calculate the contribution of SAO (%)to total methane production. In our study, however, thecalculation of the isotopic distribution of carbon dioxide andmethane was thoroughly executed and the kinetic isotope effectwas also included in the calculation. The mathematical equation(eq 1) we derived from the isotopic calculation was shown toprovide a good estimation to the quantitative contribution ofSAO pathway to methane production.With regard to the established 14C radioisotope tracer
experiment, only qualitative information can be obtained inorder to identify which pathway dominates the degradation ofacetate. In radioisotope analysis, if 14CO2/
14CH4 > 1, SAO isjudged to be a dominant pathway during the degradation of[2-14C]acetate, otherwise acetoclastic methanogenesis predom-inates.4 There are also several difficulties working withradioisotope tracer experiments because of the requirementfor strict health and safety regulations for handing radio-isotopes, and high cost associated radioactive material training,regulation, and waste disposal.24
SAO As a Key Acetate Degradation Pathway. In thisstudy, it has been shown that synthrophic acetate oxidationcoupled to hydrogenotrophic methanogenesis is an importantmethanogenic pathway which contributes from 3% to 87% ofmethane production on average during the course of acetatedegradation. Similar findings were reported in a few studies,
where SAO contributed 35−89%,7 80%,2 and 13.1−21.3%.23The difference in the contribution level of SAO to methaneproduction may originate in the source of inoculum andconcentration of acetate used as a substrate. For instance, 100mM7 and 4 mM2 acetate were inoculated with mixed culturesobtained from anaerobic digesters treating sewage sludge, and0.5 mM acetate23 was inoculated with a mixed culture obtainedfrom an anaerobic digester treating synthetic wastewater. Thefindings from previous studies2,7,23 and our study indicate thatSAO is favored at high acetate concentration (4−100 mM).However, further research is required to understand the role ofSAO at different environmental conditions such as acetateconcentration, ammonia, inoculum sources, and temperature.Analysis of the community structure of the thermophilic
anaerobic digester sludge showed that methane production wasstable and efficient without acetate accumulation even in theabsence of strict acetoclastic methanogens (Methanosaetaceae).The digesters were dominated by synthrophic acetate oxidizingbacteria (SAOB) in synthrophic association with stricthydrogenotrophic methanogens (often Methanobacteriales orMethanomicrobiales).4 Abundant synthrophic acetate oxidizingbacteria affiliate into the class Clostridia within the phylumFirmicutes as well as the family Thermotogaceae within thephylum Thermotogae.3 In our study, it appears that bacteriabelonging to the phyla Firmicutes are those mainly responsiblefor acetate oxidation with a subsequent methane production bythe hydrogenotrophic Methanoculleus. Moreover, species withinthe mixotrophic Methanosarcinaceae able to utilize severalsubstrates such as acetate, hydrogen and carbon dioxide,methanol, and methylamines were also prevalent in thedigesters.4,25 Given their ability to shift between the twometabolisms depending on growth conditions26 and being amixotrophic microorganism, members of the Methanosarcinamay play an important role for the conversion of acetate tomethane via both acetoclastic and synthrophic acetate oxidationcoupled to hydrogenotrophic methanogenesis in the studypresented here.Microorganisms mediating SAO-HM pathway have shown to
tolerate environmental stress such as high organic loading andhigh levels of ammonium up to 7000 mg TAN/L.5,27 Furtherexperiments into the role of SAO-HM pathways could open anew possibility to optimize biogas production by maintainingfavorable environmental condition for these stress-tolerantmicroorganisms. Moreover, kinetic parameters will be requiredin order to modify the currently widely accepted anaerobicdigestion model (ADM1) which emphasizes acetoclasticmethanogenesis as a dominant pathway of methane productionfrom acetate.28 In this regard, the rapid and simple techniqueemployed in our study can be used in future experiments tofacilitate the insight into the kinetics and quantitativeinformation of SAO-HM pathway to methane production.
■ ASSOCIATED CONTENT
*S Supporting InformationAcetate degradation reactions, details of experimental part,additional results, and other details. This information isavailable free of charge via the Internet at http://pubs.acs.org/.
■ AUTHOR INFORMATION
Corresponding Author*Phone: +45 30896099; e-mail: [email protected].
Environmental Science & Technology Article
dx.doi.org/10.1021/es403144e | Environ. Sci. Technol. 2014, 48, 2505−25112510
NotesThe authors declare no competing financial interest.
■ ACKNOWLEDGMENTSThis research was financially supported by the Danish StrategicResearch Council (Grant 10-093944).
■ REFERENCES(1) Zinder, S. H.; Koch, M. Non-aceticlastic methanogenesis fromacetate: Acetate oxidation by a thermophilic syntrophic coculture.Arch. Microbiol. 1984, 138, 263−272.(2) Sasaki, D.; Hori, T.; Haruta, S.; Ueno, Y.; Ishii, M.; Igarashi, Y.Methanogenic pathway and community structure in a thermophilicanaerobic digestion process of organic solid waste. J. Biosci. Bioeng.2011, 111, 41−46.(3) Hattori, S. Syntrophic acetate-oxidizing microbes in methano-genic environments. Microbes Environ. 2008, 23, 118−127.(4) Karakashev, D.; Batstone, D. J.; Trably, E.; Angelidaki, I. Acetateoxidation is the dominant methanogenic pathway from acetate in theabsence of Methanosaetaceae. Appl. Environ. Microbiol. 2006, 72,5138−5141.(5) De Vrieze, J.; Hennebel, T.; Boon, N.; Verstraete, W.Methanosarchina: The rediscovered methanogen for heavy dutybiomethanation. Bioresour. Technol. 2012, 112, 1−9.(6) Hori, T.; Haruta, S.; Ueno, Y.; Ishii, M.; Igarashi, Y. Dynamictransition of a methanogenic population in response to theconcentration of volatile fatty acids in a thermophilic anaerobicdigester. Appl. Environ. Microbiol. 2006, 72, 1623−1630.(7) Hao, L. P.; Lu, F.; He, P. J.; Li, L.; Shao, L. M. Predominantcontribution of syntrophic acetate oxidation to thermophilic methaneformation at high acetate concentrations. Environ. Sci. Technol. 2010,45, 508−513.(8) Conrad, R. Quantification of methanogenic pathways using stablecarbon isotopic signatures: A review and a proposal. Org. Geochem.2005, 36, 739−752.(9) Meier-Augenstein, W. Applied gas chromatography coupled toisotope ratio mass spectrometry. J. Chromatogr. A 1999, 842, 351−371.(10) Lloyd, D.; Bohatka, S.; Szilagyi, J. Quadrupole mass-spectrometery in the monitoring and control of fermentations.Biosensors 1985, 1, 179−212.(11) Bohatka, S. Process monitoring in fermentors and living plantsby membrane inlet mass spectrometry. Rapid Commun. Mass Spectrom.1997, 11, 656−661.(12) Tarkiainen, V.; Kotiaho, T.; Mattila, I.; Virkajarvi, L.; Aristidou,A.; Ketola, R. A. On-line monitoring of continuous beer fermentationprocess using automatic membrane inlet mass spectrometric system.Talanta 2005, 65, 1254−1263.(13) Ward, A. J.; Bruni, E.; Lykkegaard, M. K.; Feilberg, A.; Adamsen,A. P. S.; Jensen, A. P.; Poulsen, A. K. Real time monitoring of a biogasdigester with gas chromatography, near-infrared spectroscopy, andmembrane-inlet mass spectrometry. Bioresour. Technol. 2011, 102,4098−4103.(14) Bastidas-Oyanedel, J.-R.; Mohd-Zaki, Z.; Pratt, S.; Steyer, J.-P.;Batstone, D. J. Development of membrane inlet mass spectrometry forexamination of fermentation processes. Talanta 2010, 83, 482−492.(15) Davey, N. G.; Krogh, E. T.; Gill, C. G. Membrane-introductionmass spectrometry (MIMS). TrAC, Trends Anal. Chem. 2011, 30,1477−1485.(16) An, S.; Gardner, W. S.; Kana, T. Simultaneous measurement ofdenitrification and nitrogen fixation using isotope pairing withmembrane inlet mass spectrometry analysis. Appl. Environ. Microbiol.2001, 67, 1171−1178.(17) Steingruber, S. M.; Friedrich, J.; Gachter, R.; Wehrli, B.Measurement of denitrification in sediments with the 15N isotopepairing technique. Appl. Environ. Microbiol. 2001, 67, 3771−3778.(18) Schluter, M.; Gentz, T. Application of membrane inlet massspectrometry for online and in situ analysis of methane in aquaticenvironments. J. Am. Soc. Mass Spectrom. 2008, 19, 1395−1402.
(19) Sander, R. Compilation of Henry’s Law Constants for Inorganicand Organic Species of Potential Importance in Environmental Chemistry;Max-Planck Institute of Chemistry, Air Chemistry Department: Mainz,Germany, 1999; pp 1−107.(20) Pohlman, J.; Kaneko, M.; Heuer, V.; Coffin, R.; Whiticar, M.Methane sources and production in the northern Cascadia margin gashydrate system. Earth Planet. Sci. Lett. 2009, 287, 504−512.(21) Demirel, B.; Scherer, P. The roles of acetotrophic andhydrogenotrophic methanogens during anaerobic conversion ofbiomass to methane: A review. Rev. Environ. Sci. Biotechnol. 2008, 7,173−190.(22) Krober, M.; Bekel, T.; Diaz, N. N.; Goesmann, A.; Jaenicke, S.;Krause, L.; Miller, D.; Runte, K. J.; Viehover, P.; Puhler, A.Phylogenetic characterization of a biogas plant microbial communityintegrating clone library 16S-rDNA sequences and metagenomesequence data obtained by 454-pyrosequencing. J. Biotechnol. 2009,142, 38−49.(23) Hori, T.; Sasaki, D.; Haruta, S.; Shigematsu, T.; Ueno, Y.; Ishii,M.; Igarashi, Y. Detection of active, potentially acetate-oxidizingsyntrophs in an anaerobic digester by flux measurement andformyltetrahydrofolate synthetase (FTHFS) expression profiling.Microbiology 2011, 157, 1980−1989.(24) Pack, M. A.; Heintz, M. B.; Reeburgh, W. S.; Trumbore, S. E.;Valentine, D. L.; Xu, X.; Druffel, E. R. A method for measuringmethane oxidation rates using low-levels of 14C-labeled methane andaccelerator mass spectrometry. Limnol. Oceanogr.-Methods 2011, 9,245−260.(25) Krakat, N.; Westphal, A.; Schmidt, S.; Scherer, P. Anaerobicdigestion of renewable biomass: Thermophilic temperature governsmethanogen population dynamics. Appl. Environ. Microbiol. 2010, 76,1842−1850.(26) Qu, X.; Mazeas, L.; Vavilin, V. A.; Epissard, J.; Lemunier, M.;Mouchel, J. M.; He, P. j.; Bouchez, T. Combined monitoring ofchanges in δ13CH4 and archaeal community structure duringmesophilic methanization of municipal solid waste. FEMS Microbiol.Ecol. 2009, 68, 236−245.(27) Schnurer, A.; Nordberg, A. Ammonia, a selective agent formethane production by syntrophic acetate oxidation at mesophilictemperature. Water Sci. Technol. 2008, 57, 735−740.(28) Batstone, D. J.; Kelle, J.; Steyer, J. A review of ADM1 extensions,applications, and analysis: 2002−2005. Water Sci. Technol. 2006, 54,1−10.
Environmental Science & Technology Article
dx.doi.org/10.1021/es403144e | Environ. Sci. Technol. 2014, 48, 2505−25112511
38
Supporting Information for paper 1
Quantifying contribution of synthrophic acetate oxidation to methane production
in thermophilic anaerobic reactors by membrane inlet mass spectrometry
Daniel Girma Mulat1, Alastair James Ward
1, Anders Peter S. Adamsen
1, Niels Vinther Voigt
2, Jeppe
Lund Nielsen3, Anders Feilberg
1*
1Department of Engineering, Aarhus University, Hangøvej 2, DK-8200 Aarhus N, Denmark
2Danish Technological Institute, Kongsvang Allé 29, DK-8000 Aarhus C, Denmark
3Center for Microbial Communities, Department of Biotechnology, Chemistry and Environmental
Engineering, Aalborg University, Sohngaardsholmsvej 49, DK-9000 Aalborg, Denmark
*Corresponding author:
Anders Feilberg; phone: +45 30896099; e-mail: [email protected]
16 pages, 6 figures, 2 tables, 8 sections were included in the supporting materials.
39
1 Reactions involved for the biochemical conversions of acetate to methane
Acetoclastic methanogenesis (AM): *CH3COO
- + H2O →
*CH4 + HCO3
- ∆G°´= -31.0 kJ mol
-1 (S1)
Syntrophic acetate oxidation coupled to hydrogenotrophic methanogenesis (SAO-HM): *CH3COO
- + 4H2O → H
*CO3
- + HCO3
- + 4H2 + H
+ ∆G°´= +104.1 kJ mol
-1 (S2)
H*CO3
- (or HCO3
-) + 4H2 + H
+ → *CH4 (or CH4) + 3H2O ∆G°´= -135.6 kJ mol
-1
(S3)
(asterisks represent the carbon of the methyl group in acetate) (1, 2)
2 Experimental
MIMS measurement
A schematic picture of the experimental set-up for the MIMS measurement of anaerobic digestion and
standard dissolved gas is presented in Figure S1. The quadrupole mass spectrometer (Balzers QMG
420) was equipped with a membrane pump for rough pumping and a turbo pump for generating high
vacuum pumping down to ~10-6
mbar. The analytes were ionized by electron ionization (70 eV) and
the ions were separated according to their mass to charge ratio (m/z) by a quadrupole mass filter and
finally detected with a secondary electron multiplier. QuadstarTM version 6.02 software was used for
acquisition of data. Mass spectra were acquired at 5 s amu−1
. Characteristic ions of each compound
were monitored in multiple ion detection (MID) mode.
N21
2
3
Exhaust gas
reservior
MFC
QMS
Cold trap
MFC
MIMS
probe
Magnetic
stirrer
Anaerobic reactor
in a water bath
CO2/CH4
Figure S1 MIMS measurement set-up for calibration of standard dissolved gases and for measurement
of dissolved gases in anaerobic reactor. During calibration of dissolved gases in water, the three gas
lines (1, 2 and 3) were connected. During the MIMS measurement of anaerobic digestion products, the
two gas lines (2 and 3) were disconnected. MFC is mass flow controller; QMS is quadrupole mass
spectrometer.
40
The anaerobic reactor system was constructed by connecting three gas lines to a gas-tight serum bottle
(see Figure S1). Gas line 1 was made by connecting stainless steel and capillary tubing as follows. One
end of the 0.32 cm o.d. stainless steel tubing was connected to the high vacuum section of the mass
spectrometer through a bellows-sealed valve (Swagelok, Denmark) for safety purpose and the other end
was connected to 0.32 cm o.d. capillary tubing. The stainless tubing was bent to a U-shape and kept in
a 1 L wide-mouth Dewar flask (Sigma Aldrich, Denmark) filled with dry ice (~ -80°C) to make a cold
trap system. The other end of the 0.32 cm o.d. capillary tubing was connected to a membrane probe.
The membrane probe was made from 127 μm thick silicon membrane (SIL-TEC Sheeting, Technical
Products Inc., USA) inlet and supported by a thin perforated stainless steel plate. Before connecting the
capillary tubing to the membrane probe, a hole was drilled in a butyl rubber stopper (30 mm o.d.,
Apodan Nordic, Denmark) and the capillary tubing was inserted directly in this. A serum bottle
containing fermentation broth was sealed with the butyl rubber stopper and secured with an aluminum
crimp. The stated MIMS setup was used for measurement of anaerobic digestion process.
A small modification was made to the MIMS setup during preparation of standard gas solution.
Additional two gas lines (2 and 3) were made (see Figure S1). Two holes were drilled in the butyl
rubber stopper and two lines of PTFE tubing (0.32 cm o.d.) were inserted directly in these. Gas line 2
was constructed by connecting one of the PTFE tubing from the serum bottle through a three-way
connector to two mass flow controllers (EL-FLOW®Select, Bronkhorst) for adding a specified flow of
CO2/CH4 and N2 gases. One mass flow controller (MFC) was connected to a gas mixture cylinder (60
vol% CH4 and 40 vol% CO2) and the other MFC was connected to a gas cylinder of N2 (purity > 99.99
vol%) with PTFE tubing (0.64 cm o.d.). Gas line 3 was based on submerging the other end of the PTFE
tubing from the serum bottle into a beaker filled with water for removing exhaust gas.
Two types of MIMS measurements were employed:
(i) MIMS calibration: Aqueous standard solutions of carbon dioxide and methane at different
concentrations were prepared in deionized water from the standard gas mixture CH4 (60 vol%) and CO2
(40 vol%). A N2 gas (purity > 99.99 vol%) was used for dilution. The dilution was carried out by
adjusting the volumetric flow of the component gases in the mixture by using a MFC. A gas mixture
(CO2 and CH4) and N2 gas with defined concentration (vol%) was bubbled into a serum bottle (500
mL) filled with 200 mL of deionized water. The gas was continuously bubbled into water until the
concentration of dissolved gas was stable. The stabilization took 40-45 minutes and it was monitored
with MIMS by recording the characteristics ions of methane at m/z 15 and carbon dioxide at m/z 44.
(ii) MIMS measurement of anaerobic digestion: A serum bottle (500-mL) containing a fermentation
solution was quickly transferred from the incubation chamber to a water bath (52°C) for MIMS
measurement. The serum bottle was uncapped and the MIMS probe was immediately submerged into
the fermentation broth and the bottle was capped again. During capping and uncapping the bottle, N2
gas was used for flushing across the mouth of the bottle. The effectiveness of N2 flushing for avoiding
the diffusion of air into a reactor was tested by incubating two parallel sets of unlabeled acetate
reactors. One set of reactors was capped all the time until the end of the experiment. The other set was
used for MIMS measurement where the bottles were capped and uncapped during the measurement as
explained earlier. The acetate degradation rate and methane production profile of both sets of control
reactors were similar (data not shown). N2 flushing across the mouth of the bottles is an effective
technique in order to avoid the diffusion of air into the fermentation bottle. Sufficient mixing during
MIMS measurement was provided with a magnetic stirrer. After every MIMS measurement, the
41
membrane probe was first rinsed with acidified water (0.1 M H2SO4) and then deionized water in order
to avoid contamination of the inlet system and to provide a constant background signal. After the
MIMS measurement, the serum bottle was kept in the incubation chamber until the next measurement
day.
Analytical methods
The volume of a produced gas was measured using an acidified water displacement method at room
temperature and atmospheric pressure. Samples of headspace gas were taken using a gas tight syringe
with a needle through a septum and transferred into a 20 mL headspace vial. The compositions of CH4
and CO2 in the headspace samples were analyzed using Agilent technologies 7890A gas
chromatograph equipped with a thermal conductivity detector and an Agilent technology GC sampler
80. Methane and carbon dioxide were isolated using Alltech’s CTR I stainless steel column. It is
essentially a column within a column which was packed with different materials: outer column (1.8 m
x 0.64 cm) was packed with an activated molecular sieve and inner column (1.8 m x 0.32 cm) was
packed with a porous polymer mixture. The carrier gas was helium at 30 mL/min. The temperatures of
injection port, column oven and detector were set at 110, 40 and 150 °C, respectively.
Liquid samples for VFA analysis were withdrawn with a syringe equipped with a needle through a
septum and its pH was measured immediately. The liquid samples (1.000 g) were first acidified with 4
mL of 0.3 M oxalic acid containing the internal standard dimethylpropanoic acid, then centrifuged at
4,500 rpm for 12 min and filtered through a 0.45m GHP membrane. Finally an aliquot of the
supernatant solution was transferred into a vial. The concentrations of volatile fatty acids were
determined by Agilent Technologies 7890A gas chromatograph equipped with flame ionization
detector (FID). A polar phase capillary column, HP-INNOWax (30m x 0.25 mm x 0.25 μm), was used
for separation. Helium was used as a carrier gas at 1.8 ml/min flow rate. The analyses were performed
using a temperature programme: 5 min at 100°C, a linear gradient from 100-120°C at the rate of 10
°C/min, 5 min at 120°C, a linear gradient from 120°C to the final temperature of 220°C at the rate of
30°C/min and final hold at
220°C for 3 min. The temperatures of injection port and detector were set at
285°C and 300°C, respectively.
Proteome analysis
Liquid samples for proteome analysis were collected from all reactors at the middle of the incubation
experiment and stored in a freezer (~ -20 °C) until the analysis. Proteins were extracted followed by
tryptic digestion as described elsewhere (3) and mass spectrometry analysis by an automated LC-ESI-
MS/MS with an UltiMate 3000 RSLCnano system on-line coupled to a Q Exactive mass spectrometer
via a Nanospray Flex ion source (Thermo Fisher Scientific). The resulting mass spectra were used to
search for protein homologs using an in-house Mascot database search engine and the NCBI and the
UniProt database as described in details elsewhere (4).
42
MIMS data presentation
The characteristic ion signal intensities were used for calculating the isotopic composition of methane
and carbon dioxide. Peaks at m/z 45 and 44 were regarded as signals from 13
CO2 and 12
CO2,
respectively. A peak at m/z 17 was regarded as a molecular ion for 13
CH4. A peak at m/z 15 includes the
signals from 12
CH3+ and
13CH2
+. According to the fragment ions of
13CH4 mass spectrum, the signal
from 13
CH2+ represented 22.8% of the peak intensity of m/z 17 (data not shown). The signal intensity of
12CH3
+ at m/z 15 was corrected by subtracting the calculated signal intensity for
13CH2
+ from the total
signal intensity at m/z 15. The corrected signal intensity at m/z 15 represents 12
CH3+ which is a
fragment ion of 12
CH4. An example of MIMS data correction and presentation is given in Table S1. The
data was taken from MIMS measurement of [2-13
C] acetate and [U-13
C] acetate reactors on the first day
of the incubation time.
The inoculum used to incubate [2-13
C] acetate substrate was not free from inorganic carbon species
(CO2 (dissolved), HCO3- and CO3
2-). The sum of these inorganic carbon species in a solution of the
inoculum were represented as background total inorganic carbon (TIC). The background TIC of the
inoculum was calculated according to equation S4 using the CO2(dissolved) measured with MIMS,
carbonates (HCO3- and CO3
2-) equilibrium constants and pH of the inoculum (5).
CO2(total) = CO2(dissolved) {[𝐻+]2+𝐾𝑎1[𝐻
+]+𝐾𝑎1𝐾𝑎2
⌊𝐻+⌋2} (S4)
The MIMS results of the [2-13
C] reactor are presented after correcting the contribution of the
background TIC to m/z 44 (12
CO2) and m/z 15 (12
CH4) signals of the [2-13
C] sodium acetate reactor. The
MIMS measurement data for [U-13
C] acetate reactor was used to make the TIC correction. The
unlabeled products (12
CO2 and 12
CH4) during the degradation of [U-13
C] acetate were considered as the
background TIC that comes from the inoculum and they were subtracted from the MIMS values of 12
CO2 and 12
CH4 measured during the degradation of [2-13
C] acetate. An example of TIC correction is
given in Table S1.
Table S1 Isotopic distribution of methane and carbon dioxide from 13
C labeled acetate
Ion current
Substrate m/z 15 (12
CH4) in nA
m/z 17 (13
CH4)
in nA 13
CH4 (atom%)
Actual
13CH2
+
subtracteda
Background
TIC-correctedb Actual
Background
TIC-correctedc
13CH3COO
- 4.0 2.4 1.4 7.0 83.7
13CH3
13COO
- 3.3 1.0 9.9
Ion current
Substrate m/z 44 (12
CO2) in nA
m/z 45 (13
CO2)
in nA 13
CO2 (atom%) 13
CO2 (atom%)/13
CH4(atom%)
Actual
Background
TIC-correctedd Actual
Background
TIC-corrected
Background
TIC-corrected 13
CH3COO- 49 45.1 0.7 1.6 0.02
43
13CH3
13COO
- 3.9
a. The signal from 13
CH2+ represented 22.8% of the peak intensity of m/z 17.
b. The ion current at m/z 15 from [U-13
C] acetate was regarded as arising from the background
production of unlabeled CH4.
c. The percentage of 13
CH4 to total methane calculated from background TIC-corrected 12
CH4 and
actual 13
CH4 values.
d. The ion current at m/z 44 from [U-13
C] acetate was regarded as arising from the background
production of unlabeled CO2.
Calculation of isotope distribution
The production of CH4 and /or CO2 from acetate depends on the metabolic pathways. During
acetoclastic methanogenesis, methane is formed from the methyl group carbon atom in acetate and
carbon dioxide is formed from the carboxyl group carbon atom (6). In synthrophic acetate oxidation
pathway, the carbons of both the groups in acetate are converted to carbon dioxide (7, 8). If
acetoclastic methanogenesis is taking place, 1 mole of 13
CH4 and 1 mole of 12
CO2 will be produced
from 1 mole of 13
CH3COO-. When SAO is taking place, 1 mole of
13CO2, 1 mole of
12CO2 and 4 moles
of H2 will be produced from 1 mole of 13
CH3COO-. Since there is no extra source of H2 available to the
system, only 1 mole of carbon dioxide is reduced by 4 moles of H2 via HM and the other 1 mol of
carbon dioxide is remained unconsumed. Both 13
CO2 and 12
CO2 can be reduced to 13
CH4 and 13
CH4,
respectively, with different proportion according to their rate constants. As 12
CO2 reacts slightly faster
than 13
CO2 during the reduction of carbon dioxide (i.e., 12
k/13
k = 1.065 on average at thermophilic
condition) (9, 10), the production of 12
CH4 is estimated to be slightly higher than 13
CH4 at the given
time point. A schematic representation of calculated mass balance of 13
CH4, 12
CH4, 13
CO2 and 12
CO2 by
assuming 50% methane production from acetoclastic methanogenesis and the remaining 50% from
SAO-HM is shown in Figure S2. In this mass balance calculation, the kinetic isotope effect (12
k/13
k =
1.065) was included. The background TIC was not included in the mass balance calculation since TIC-
corrected MIMS data is used to quantify the contribution of SAO to methane production. Similar mass
balance calculation was reported by Hori et al.(11) but we expanded it by including the kinetic isotope
effect.
44
13CH4(AM)
=50 mM
HMHM
50% AM 50% SAO
13CH3COOH=50 mM
13CH3COOH=100 mM
13CH3COOH=50 mM
12CO2(AM)
=50 mM12CO2(SAO)
=50 mM
H2
=4x50=200 mM
13CO2(SAO)
=50 mM
SAOAM
The sum of all 12CO212CO2(sum)
=50+50
=100 mM
12CH4(HM)
=12CO2(H2eq) x [12CO2(sum)/(12CO2(sum)+
13CO2(SAO)]
=50 x [100/(100+50)]
=33.3mM
13CH4(HM)
=13CO2(H2eq) x [13CO2(SAO)/(12CO2(sum)+
13CO2(SAO)]
=50 x [50/(100+50)]
=16.7 mM
12CO2 left unconsumed
=12CO2(sum)-12CH4(HM,KIE)
=100-35.5
=64.5 mM
12CH4(HM,KIE)
=12CH4(HM) x12k/13k
=33.3 x 1.065
=35.5 mM
13CH4(HM,KIE)
=13CH4(HM) - [12CH4(HM,KIE) -
12CH4(HM)]
=16.7 - [35.5 – 33.3]
=14.5 mM
13CO2 left unconsumed
=13CO2(SAO) - 13CH4(HM,KIE)
=50 – 14.5
=35.5 mM
Final balance13
CH4 =13CH4(AM) +
13CH4(HM,KIE)
= 50 + 14.5 = 64.5 mM12
CH 4=12CH4(HM,KIE)
= 35.5 mM
CH4(total) =12CH4 + 13CH4
= 100 mM13
CH4(atom%)
={13CH4/CH4(total)}x 100
={64.5/100}x 100
= 64.5
Final balance13
CO2 =13CO2 left unconsumed
= 35.5 mM12
CO2=12CO2 left unconsumed
= 64.5 mM
CO2(total) =12CO2 + 13CO2
= 100 mM13
CO2(atom%)
={13CO2/CO2total)}x 100
={35.5/100}x 100
= 35.5
Ratio13CO2(atom%) : 13CH4(atom%)
= 35.5/64.5
= 0.55
45
Figure S2 A schematic representations to determine the mass balance of 12
CH4, 13
CH4, 12
CO2 and 13
CO2. Herein an example of a mass balance is provided when 100mM 13
CH3COOH was degraded by
50% acetoclastic methanogenesis and 50% synthrophic acetate coupled to hydrogenotrophic
methanogenesis. HM is hydrogenotrophic methanogenesis; AM is acetoclastic methanogenesis; SAO is
synthrophic acetate oxidation; atom% of 13
C labeled substrate is {13
C substrate/(13
C substrate + 12
C
substrate)} x 100, for e.g. 13
CO2 (atom%) ={13
CO2/(13
CO2+12
CO2)} x 100; 12
CO2 (H2eq) and 13
CO2
(H2eq) are 1 mol 12
CO2 and 13
CO2 equivalent that can be reduced by the available 4 mol H2 via HM ,
respectively, i.e. only 50 mM of 12
CO2 and 13
CO2 can be reduced by the available 400 mM of H2; KIE
is kinetic isotope effect; 12
k/13
k=1.065 is the ratio of the rate constants of 12
CO2-derived methane to 13
CO2-derived methane; HM,KIE is methane-derived via hydrogenotrophic methanogenesis including
kinetic isotope effect.
3 Cold trap
For measurement of 13
C labeled methane (13
CH4) at m/z 17, the main interference comes from the
fragment ion of water (OH+). The silicone membrane used in our experiment is permeable to low-
molecular mass volatiles, water and gases. A cold trap was used to condense the water vapor before it
reaches the ionization source of the mass spectrometer. A Dewar flask containing dry ice (~ -80°C) was
placed between the membrane probe and mass spectrometer in order to trap water vapor. The cold trap
was very effective in condensing water that otherwise interferes with 13
CH4 measurement. The signal
intensity of water measured at m/z 18 was five-folds reduced when the cold trap was used (Figure S3).
The reduction of the amount of water vapor in the mass spectrometer reduced the pressure inside the
mass spectrometer (data not shown). The provision of high vacuum inside the mass spectromer enables
accurate propagation of ion trajectories which in turn improves the detection limit and the reliability of
measurement (12).
Figure S3 Implementation of the cold trap (operated at ~80°C) caused a significant reduction of the
water vapor content (as monitored by m/z 18), which otherwise permeate through the silicone
membrane. The lines represent the first 95 scans in the absence of cold trap and the last 90 scans in the
presence of cold trap. When cold trap was used, the data points until the attainment of stable MIMS
reading was excluded. That is why a sharp drop between the top left line and the bottom right line was
observed.
0
5
10
15
20
25
30
0 50 100 150 200
Ion
cu
rren
t (n
A)
Scan number
46
4 Effect of sample temperature
The aim of the sample temperature test was to investigate the effect of membrane temperature on the
operating pressure of the MIMS system and performance of the silicone membrane. The temperature of
deionized water as a sample was increased step by step from room temperature to 52.5°C. The choice
of this temperature range was based on the fact that most of biogas production from organic waste is
conducted under either mesophilic (~ 35°C) or thermophilic (~ 52°C) conditions. In our study,
anaerobic digestion was conducted in the incubation chamber held at 52°C. While the MIMS
measurement was conducted, the temperature of the reactor was controlled by a water bath held at
52°C. Constant temperature was maintained during MIMS measurements to avoid process instability
due to temperature fluctuation and to stabilize the MIMS measurements, as temperature affects the
MIMS signals (see also below).
Figure S4 shows changes of pressure inside the mass spectrometer with an increase in temperatures of
water sample from room temperature up to 52.5°C. The pressure hardly changed when temperature
increased from room temperature up to 29°C. However, it showed a linear three-fold increase as the
temperature increased from 29-52.5°C. Moreover, the background intensity of both CO2 and CH4
signals increased by three-fold (data not shown). Higher sample temperature causes the expansion of
membrane micropores, which leads to increase in diffusivity of the analyte through the membrane. As a
consequence of higher analyte flow into the vacuum section of the mass spectrometer, the partial
pressure inside the mass spectrometer increases (13). In our temperature study, the total pressure in the
ionization chamber of the mass spectromer is lower than the recommended safety limit (~ 3.0 x 10-5
mbar). Therefore, it can be concluded that safe operation of a MIMS system is achievable for a typical
sample temperature range (30-52°C).
Figure S4 The pressure of ion source (mbar) of the mass spectrometry as a function of silicone
membrane temperature (°C).
0
1
2
3
4
5
6
7
8
25 29 42 46 50 52.5
Pre
ssu
re (
10
-6 m
ba
r)
Temperature (°C)
47
5 Instrument response time
When calculating response times for polymer membranes, usually 90% or 50% response times are
taken. In our study, we defined the response time as the time required to observe a 90% increase of
signal intensity of the ion. Response time vary with length (L) of the probe tube and inversely with the
radius (r) of the probe (cm). The 90% response time, t90% is given by (14):
t90% = kL2r-1
(S5)
where k is a constant. The value of k is in the order of 1x10-5
and varies with temperature and
properties of membrane (15).
Figure S5 shows the MIMS response time for the measurement of dissolved carbon dioxide and
methane in a fermentation reactor containing substrate-free inoculum and pure water. At time zero the
sample was changed from pure water to substrate-free inoculum and the MIMS was scanned in MID
mode for 4.5 minutes. It can be seen from Figure S5 that the signals of both CO2 and CH4 were
increased to 90% maximum in just 1 min and level off for 4.5 min. Then the membrane probe was
immediately rinsed with deionized water and dipped inside a reactor containing deionized water. The
arrow indicates the change from substrate-free inoculum back to pure water. The signal intensities of
both analytes decreased to 10% in just 1 min and then levels off. Our result was in accordance with
literature (16) where the rise time (10-90% of maximum signal response) is equal to or slightly shorter
than fall time (decrease in signal response from 90-10%). The fast rise and fall time obtained in our
study shows the capability of MIMS technique for online and onsite monitoring of dissolved carbon
dioxide and methane in real time.
Figure S5 MIMS measurement of (a) dissolved methane (m/z 15) and (b) carbon dioxide (m/z 44)
directly inside reactors containing substrate-free inoculum and deionized water. The arrow indicates the
change from substrate-free inoculum back to pure water.
0
500
1000
1500
2000
2500
0 4.5 9
Ion
cu
rren
tm
/z 1
5 (
10
pA
)
Scan time (min)
(a)
0
50
100
150
200
0 4.5 9
Ion
cu
rren
t m
/z4
4 (
nA
)
Scan time (min)
(b)
48
6 Measurement of hydrogen with MIMS
The determination of the concentration of hydrogen at m/z 2 in our reactor was obscured by the high
concentration of dissolved methane (about 1.4 mM) and very low concentration of dissolved hydrogen
ranged from 0.5 to 3 µM in the digester (17). Due to high energetic electron (70 eV) used for the
ionization of the analytes, methane (CH4) molecules fragment to H2+ ion and detected at m/z 2. The
contribution of methane to the signal at m/z 2 was determined by purging methane at different
concentration into deionized water. As shown in Figure 6S, a linear relationship was observed (r2 =
0.995) when m/z 2 signal was plotted versus m/z 15 signal (methane). Due to the high background
signal from methane at m/z 2 and the expected extremely small concentration of hydrogen at steady
anaerobic condition, the ion abundance of hydrogen at m/z 2 could not be determined accurately in our
experiment. However, the hydrogen concentration of anaerobic digester both at steady state organic
loading rate (OLR) and increased OLR was reported by some authors (18-20). In these published
works, the calibration curve for quantifying hydrogen and the interference of methane at m/z 2 were not
reported. The reported hydrogen data might be overestimated due to the interference from fragment ion
of methane. It seems that the concentration of hydrogen in an anaerobic digester can only be
determined accurately by MIMS in a few condition, for instance, at high OLR that leads to high
concentration of hydrogen and with appropriate correction for the interference caused by the fragment
ion of methane.
Figure S6 The fragment ion of methane at m/z 2 divided by m/z 18 (water) and its linear dependence
with the ion intensity of m/z 15 (methane) divided by m/z 18 (water).
7 Measurement uncertainty
The use of ion-current ratio between 13
C and 12
C isotopes of CO2 and CH4 in our study significantly
reduces errors associated with the drift in mass spectrometer and membrane inlet system as well as
slight fluctuations in sample temperature and permeability of water vapor. We have considered two
major sources of uncertainty, which arises from isotope ratio measurement with MIMS and the
calculation of quantifying methanogenesis pathways as follows.
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
0 2 4 6 8
Ion
cu
rren
t m
/z2
-to
-m/z
18
ra
tio
Ion current m/z15-to-m/z18 ratio
49
(i) Uncertainty of MIMS measurement of isotopic distribution of CO2 and CH4:
The MIMS measurement was conducted for 4 minutes in each anaerobic reactor. The maximum
response signal was attained in one minute and an average of data points within the last three minutes
of MIMS measurement was used to assign the signal intensity of the individual measured ion. The
uncertainty associated with taking the average signal intensity was 2.5%, 1.3%, 1.3% and 1.3% for m/z
15 (12
CH4), m/z 17(13
CH4), m/z 44(12
CO2) and m/z 45(13
CO2), respectively.
(ii) Uncertainty of quantifying SAO pathways as a function of MIMS data of 13
CO2 (atom%)/13
CH4
(atom%) using equation 1:
Equation 1 was derived considering the kinetics isotope effect during the reduction of carbon dioxide to
methane. The average value of 12
k/13
k = 1.065 was used to account the kinetic effect (9, 10). Previous
literatures showed that the 12
k/13
k for thermophilic hydrogenotrophic methanogens was in the range of
1.04 to 1.09 (9, 10). The uncertainty associated with the use of 12
k/13
k = 1.065 on average for deriving
equation 1 was 4.9%.
The propagated uncertainty associated with quantification of SAO (%) using equation 1 as a function
of MIMS data of 13
CO2(atom%)/13
CH4(atom%) was calculated from both uncertainty components
listed above and it was estimated to be 5.9%.
8 Microbial community structure
The resulting mass spectra were used to identify and taxonomically attribute individual peptides to
Bacteria and Archaea. At least two peptides with similar identities were required for proper
identification. The proteome analysis showed that a total of 527 proteins were identified with a wide
range of functional categories and representing 18 phyla, 30 classes and 107 genera. The bacterial
proteome constituted 74.5% of the proteins with a relative high diversity whereas the majority of the
remaining proteins affiliated to Archaea (Table S2). We found only a few fungal and plant proteins in
the sample.
50
Table S2 Distribution of protein sequences analyzed in this study
Phyla/Class/Genus/ Percentage
Actinobacteria 2.3
Actinobacteria 2.1
Actinobacteridae 0.2
Aquificae 1.1
Bacteroidetes/Chlorobi group 5.7
Bacteroidetes 4.4
Chlorobi 1.1
Ignavibacteriae 0.2
Chlamydiae/Verrucomicrobia group 0.4
Chloroflexi 0.9
Caldilineae 0.2
Chloroflexi 0.4
Thermomicrobia 0.4
Cyanobacteria 0.2
Deferribacteres 0.2
Fibrobacteres/Acidobacteria group 0.6
Firmicutes 30.9
Bacilli 8
Clostridia 19.7
Erysipelotrichia 0.4
Negativicutes 2.8
Fusobacteria 0.9
Planctomycetes 1.1
Proteobacteria 16.5
Alphaproteobacteria 5.3
Betaproteobacteria 2.8
Gammaproteobacteria 5.7
Deltaproteobacteria 2.5
Epsilonproteobacteria 0.2
Spirochaetes 3.4
Synergistetes 7
Synergistia 7
Thermotogae 0.8
Uncultured bacteria 2.5
51
Euryarchaeota 23.3
Methanobacteria 0.4
Methanomicrobia 23
Methanocella 0.2
Methanocorpusculum 0.8
Methanoculleus 8.7
Methanofollis 1.3
Methanolinea 0.6
Methanoplanus 0.6
Methanosarcina 10.8
Korarchaeota 0.4
Candidatus Korarchaeum 0.4
Thaumarchaeota 0.2
Uncultured Archaea 1.5
52
References
1. Hattori, S. Syntrophic acetate-oxidizing microbes in methanogenic environments.
Microbes Environ. 2008, 23, 118-127.
2. Sasaki, D.; Hori, T.; Haruta, S.; Ueno, Y.; Ishii, M.; Igarashi, Y. Methanogenic pathway
and community structure in a thermophilic anaerobic digestion process of organic solid waste. J.
Biosci. Bioeng. 2011, 111, 41-46.
3. Heyer, R.; Kohrs, F.; Benndorf, D.; Rapp, E.; Kausmann, R.; Heiermann, M.; Klocke, M.;
Reichl, U. Metaproteome analysis of the microbial communities in agricultural biogas plants. New
Biotechnol. 2013.
4. Kılıç, N. K.; Stensballe, A.; Otzen, D. E.; Dönmez, G. Proteomic changes in response to
chromium (VI) toxicity in Pseudomonas aeruginosa. Bioresour. Technol. 2010, 101, 2134-2140.
5. Bastidas-Oyanedel, J.-R.; Mohd-Zaki, Z.; Pratt, S.; Steyer, J.-P.; Batstone, D. J.
Development of membrane inlet mass spectrometry for examination of fermentation processes. Talanta
2010, 83, 482-492.
6. Thauer, R. K.; Möller-Zinkhan, D.; Spormann, A. M. Biochemistry of acetate catabolism
in anaerobic chemotrophic bacteria. Annu. Rev. Microbiol. 1989, 43, 43-67.
7. Zinder, S. H.; Koch, M. Non-aceticlastic methanogenesis from acetate: acetate oxidation
by a thermophilic syntrophic coculture. Arch. Microbiol. 1984, 138, 263-272.
8. Schnürer, A.; Houwen, F. P.; Svensson, B. H. Mesophilic syntrophic acetate oxidation
during methane formation by a triculture at high ammonium concentration. Arch. Microbiol. 1994, 162,
70-74.
9. Conrad, R. Quantification of methanogenic pathways using stable carbon isotopic
signatures: a review and a proposal. Org. Geochem. 2005, 36, 739-752.
10. Pohlman, J.; Kaneko, M.; Heuer, V.; Coffin, R.; Whiticar, M. Methane sources and
production in the northern Cascadia margin gas hydrate system. Earth Planet. Sc. Lett. 2009, 287, 504-
512.
11. Hori, T.; Sasaki, D.; Haruta, S.; Shigematsu, T.; Ueno, Y.; Ishii, M.; Igarashi, Y.
Detection of active, potentially acetate-oxidizing syntrophs in an anaerobic digester by flux
measurement and formyltetrahydrofolate synthetase (FTHFS) expression profiling. Microbiol. 2011,
157, 1980-1989.
12. Tarkiainen, V.; Kotiaho, T.; Mattila, I.; Virkajarvi, L.; Aristidou, A.; Ketola, R. A. On-
line monitoring of continuous beer fermentation process using automatic membrane inlet mass
spectrometric system. Talanta 2005, 65, 1254-1263.
13. Johnson, R. C.; Cooks, R. G.; Allen, T. M.; Cisper, M. E.; Hemberger, P. H. Membrane
introduction mass spectrometry: Trends and applications. Mass Spectrom. Rev. 2000, 19, 1-37.
14. Lloyd, D.; Bohátka, S.; Szilágyi, J. Quadrupole mass-spectrometery in the monitoring and
control of fermentations. Biosensors 1985, 1, 179-212.
15. Cowie, G.; Lloyd, D. Membrane inlet ion trap mass spectrometry for the direct
measurement of dissolved gases in ecological samples. J. Microbiol. Meth. 1999, 35, 1-12.
16. Ketola, R. A.; Ojala, M.; Sorsa, H.; Kotiaho, T.; Kostiainen, R. K. Development of a
membrane inlet mass spectrometric method for analysis of air samples. Anal. Chim. Acta 1997, 349,
359-365.
53
17. Kuroda, K.; Gaiger Silveira, R.; Nishio, N.; Sunahara, H.; Nagai, S. Measurement of
dissolved hydrogen in an anaerobic digestion process by a membrane-covered electrode. J. Ferment.
Bioeng. 1991, 71, 418-423.
18. Whitmore, T.; Lloyd, D.; Jones, G.; Williams, T. Hydrogen-dependent control of the
continuous anaerobic digestion process. Appl. Microbiol. Biotechnol. 1987, 26, 383-388.
19. Scott, R. I.; Williams, T. N.; Whitmore, T. N.; Lloyd, D. Direct measurement of
methanogenesis in anaerobic digestors by membrane inlet mass spectrometry. Eur. J. Appl. Microbiol.
1983, 18, 236-241.
20. Zhang, Y.; Zhang, F.; Chen, M.; Chu, P.-N.; Ding, J.; Zeng, R. J. Hydrogen
supersaturation in extreme-thermophilic (70° C) mixed culture fermentation. Appl. Energ. 2013, 109,
213-219.
54
Chapter 5: Paper II-Acetate oxidizing microbial communities during
acid accumulation in anaerobic digestion
In preparation for peer-reviewed journal.
55
Acetate oxidizing communities during acid accumulation in anaerobic digestion
Freya Mosbæk1, Henrik Kjeldal
1, Daniel Girma Mulat
2, Mads Albertsen
1, Alastair James Ward
2, Anders
Feilberg2, Jeppe Lund Nielsen
1
1Department of Chemistry and Bioscience, Aalborg University, Fredrik Bajers Vej 7H, 9220 Aalborg East,
Denmark 2Department of Engineering, Aarhus University, Hangøvej 2, DK-8200 Aarhus N, Denmark
#Address correspondence to Jeppe Lund Nielsen, [email protected]
Abstract
The microorganisms involved in the biological recovery of an acetic acid accumulated anaerobic batch
reactor were investigated using 100 mM 13
C labeled acetate combined with proteomic identifications
(Protein-SIP). The isotope composition of the produced CH4 and CO2 in the anaerobic batch reactors
was monitored using membrane inlet quadrupole mass spectrometry (MIMS) while changes in the
microbial community during the incubations were investigated by amplicon sequencing of the 16S
rRNA genes. Higher proportion of 13
C labeled CO2 and CH4 was detected immediately following the
incubation of all the reactors with [U-13
C]acetate, indicating high degradation rate of acetate to
methane and carbon dioxide. Protein-SIP revealed that Methanosarcina and Methanoculleus were
actively involved in acetate consumption. Furthermore, five subspecies of Clostridia were involved in
degradation of 13
C labeled acetate. 13
C labeled peptides were identified by a combination of homology
search in the Uniprot database and in metagenomes of two full-scale biogas plants. The metagenome
search increased the identification of 13
C labeled peptides. All 13
C labeled bacteria contained the fhs
gene coding for formyltetrahydrofolate synthetase, a key enzyme for reductive acetogenesis, indicating
that these are possible synthrophic acetate oxidizing bacteria (SAOB) that can facilitate acetate
consumption via syntrophic acetate oxidation coupled with hydrogenotrophic methanogenesis (SAO-
HM) occurring in anaerobic digestion. To our knowledge, this is the first study applying protein-SIP
for analysis of complex biogas samples, which is a promising method for identifying key
microorganisms involved in specific pathways.
56
Introduction
Anaerobic digestion (AD) of organic materials to biogas has several advantages, since the methane-rich
biogas is a source of renewable energy that potentially can replace fossil fuels. In addition, the process
has beneficial potential in removal of pathogens as well as odor and pollution reductions from
agricultural, industrial and municipal wastes. The AD of organic matter to biogas involves coordinated
activity of a complex microbial association for the process to remain stable. Failure to maintain the
balance between these groups can lead to reactor inhibitions and breakdown (1). Accumulation of
volatile fatty acids (VFAs) may cause the pH to decrease and result in reduced process performance or
under the worst circumstances can result in reactor failure (2–4). Acetate is one of the important VFAs
seen to increase during these inhibitory periods. Several studies have investigated the inhibitory effect
of VFAs accumulation on reactor performance and on the microbial community (5–9).
It is well-established that methane production from acetate is carried out by either acetoclastic
methanogens or by syntrophic acetate oxidizing bacteria (SAOB) coupled to hydrogenotrophic
methanogens. The former has been well studied, as it was believed to be the main pathway for acetate
consumption to methane production in biogas plants. Less is known of the SAOB and only a few
species have been isolated and studied to date (10).
Microbial community compositions studied by molecular techniques are hampered by the reduced link
between phylogenetic information and the specific functions carried out by these microorganisms.
Thus, it is often impossible to couple the major metabolisms such as hydrolysis, acidogenesis, and
acetogenesis to the organisms involved. Stable or radioactive isotope-labeled substrates are frequently
applied to investigate the functions and activities of the microbes in AD. Previous studies of microbial
communities in AD have applied methods such as MAR-FISH, radiolabelling, stable isotope probing
(SIP) and FISH coupled with nanoscale secondary ion mass spectrometry (nanoSIMS) (11–15). The
methanogenic pathway is restricted to a distinct phylogenetic group of Archaea and was recently
studied in a batch AD using membrane inlet quadrupole mass spectrometry (MIMS) by tracing the
incorporation of 13
C into the produced CO2 and CH4 in real time during the degradation of [2-13
C]acetate (16). The results showed that syntrophic acetate oxidation coupled to hydrogenotrophic
methanogenesis (SAO-HM) played a key role in the conversation of acetate to methane (16). Likewise,
in another study using radio isotope labeled substrate, [2-14
C]acetate in batch incubations inoculated
with manure and food waste, it was found that acetate oxidation was the dominant methanogenic
pathway in the absence of the aceticlastic Methanosaetaceae (11).
The application of DNA-SIP have recently been applied in several studies investigating microbial
interactions in AD (15, 17, 18), and in sludge batch reactors fed with 100 mM 13
C-labeled acetate at
low and high ammonia concentrations to simulate ammonia inhibitions (15). Amplicon sequencing of
ribosomal intergenic spacer regions and 16S rRNA gene sequences revealed a high incorporation of
heavy isotopes in Methanosarcina from the archaeal population and in Clostridia from the bacterial
population (15). Another study identified acetate-utilizing bacterial (Synergistes group 4) and archaeal
(Methanosaeta) populations by MAR-FISH and RNA-SIP in anaerobic sludge batch reactors with
acetate concentrations between 2.5-10 mM (12). DNA- and RNA-SIP are powerful methods for the
identification of microbial species involved in specific substrate uptake.
57
SIP can also be applied with proteins, but this has to our knowledge not been done on very complex
samples such as those from AD. Labelling of proteins has a great advantage as the proteins reflect the
actual activity of the microbes. Protein-SIP can be used to determine both the metabolic activity and
the identity of active key microorganisms (19). Protein-SIP has the potential of linking key metabolic
functions while also extracting the phylogenetic information of active microorganisms in highly
complex microbial environments such as the AD. In this context, we applied protein-SIP aiming to
identify the microorganisms involved in the biological recovery of acid accumulation in batch AD.
Protein-SIP, a culture-independent method, was applied for the identification of key players and
enzymes involved in consumption of 13
C labeled acetate, [U-13
C]acetate, in batch anaerobic reactors
(Figure 1). Inoculum was obtained from a full-scale biogas digester working with a mixture of pig and
cattle manure, maize silage and deep litter manure. 13
C-labeled acetate at high and low concentrations
was incubated under thermophilic condition. The aim of this study was to identify the key
microorganisms involved in degradation of acetate to methane. It was also the aim to identify the effect
of different acetate concentrations on the microbial community structure. To our knowledge, this is the
first study applying protein-SIP in an anaerobic digester.
58
Figure 1. Time resolved analysis of proteins in anaerobic batch reactors using protein-SIP. Proteins are
extracted and subjected to tryptic digestion followed by liquid chromatography-tandem mass
spectrometry (LC-MS/MS) analysis. Two metagenomes from biogas AD serves as the underlying
template in the identification of peptides in the subsequent protein-SIP analysis. Peptides are identified
by matching against the metagenomes (merged with a subspace of the NCBI database) using OpenMS.
59
From the incorporation of 13
C results the change in relative isotope abundance (RIA) as well as the
ratio of labeled to unlabeled peptide, the labeling ratio (LR), were calculated whereas functional
information was obtained from the identification of peptides.
Materials and Methods
Sources of inoculum
Inoculum was obtained from a commercial full-scale biogas digester at research centre Foulum,
Denmark. The digester works with a mixture of pig and cattle manure, maize silage and deep litter
manure. It runs under thermophilic condition ca. 52°C. The total solid (TS), volatile solid (VS), pH
value and total ammonia nitrogen (TAN) of the inoculum were 50.2 g/L, 40.2 g/L, 7.64 and 1.54 g/L,
respectively. The inoculum was pre-incubated under anaerobic condition at 52°C for two weeks prior
to the main tracer experiment to reduce the background contribution of carbon dioxide and methane
from the original substrates. A digestate sample was collected at Lynggård biogas plant. Lynggård is
operated at 52°C and runs with swine manure, and grass and maize silage.
Operation of anaerobic digestion
Glass serum bottles (500-mL) were used for preparing the anaerobic incubation assay. Aliquots (195
mL) of inoculum were transferred into the 500-mL serum bottles, which were then sealed with butyl
rubber stoppers and aluminium crimps. Five treatments were prepared with 13
C labeled and unlabeled
acetate; detailed information of the reactor setup is summarized in Table 1. Appropriate substrate was
added once at the beginning of the experiment for the high acetate concentration incubation, whereas
the appropriate substrate was added once daily for the low acetate concentration incubation. All
treatments were run in triplicate whereas blank reactors were run in duplicate under static incubation
condition for 9 days at 52°C.
Table 1 Description of batch reactors setup.
Description of reactors
Substrate Feeding rate
195 mL Inoculum+ 5mL water Once at the beginning
195 mL Inoculum + 5mL concentrated CH3COONa
to give a final of 4mM acetate
Once daily
195 mL Inoculum + 5mL concentrated CH3COONa
to give a final of 100mM acetate
Once at the beginning
195 mL Inoculum + 5mL concentrated [U-13
C]acetate
to give a final of 4mM acetate
Once daily
195 mL Inoculum + 5mL concentrated [U-13
C]
acetate to give a final of 100mM acetate
Once at the beginning
60
MIMS measurement
MIMS was used to monitor the incorporation of 13
C into the produced methane and carbon dioxide
during the degradation of 13
C fully labeled acetate as described before (16). Briefly, the MIMS system
consisted of a quadrupole mass spectrometer (QMS) connected to a silicone membrane probe. The
membrane probe was directly submerged into a fermentation broth to provide an interface for the
selective and direct introduction of dissolved gaseous analytes (via pervaporation mechanism) into the
vacuum section of the QMS. Characteristic ions of each compound were monitored in multiple ion
detection (MID) mode by the QMS: m/z 17 for 13
C methane (13
CH4), m/z 15 for 12
C methane (12
CH4),
m/z 45 for 13
C carbon dioxide (13
CO2), m/z 44 for 12
C carbon dioxide (12
CO2) and m/z 18 for water
(H2O). The MIMS data are finally reported in terms of atom percent as follows:
13
X (atom%) ={13
X/(13
X+12
X)} (1),
where 13
X represents 13
CO2 or 13
CH4 and 12
X represents 12
CO2 or 12
CH4.
Basic analytical methods
The volume of a produced biogas was measured using an acidified water displacement method at room
temperature and atmospheric pressure. Headspace biogas was collected using a gas tight syringe with a
needle through a septum and transferred into a 20 mL headspace vial. The compositions of CH4 and
CO2 in the biogas samples were analyzed using Agilent technologies 7890A gas chromatograph
equipped with a thermal conductivity detector and an Agilent technology GC sampler 80. Methane and
carbon dioxide were isolated using a CTR I stainless steel column. It is essentially a column within a
column which was packed with different materials: the outer column (1.8 m x 0.64 cm) was packed
with an activated molecular sieve and the inner column (1.8 m x 0.32 cm) was packed with a porous
polymer mixture. The carrier gas was helium at 30 mL/min. The temperatures of injection port, column
oven and detector were set at 110, 40 and 150 °C, respectively.
Liquid samples for VFA analysis were collected periodically. The liquid samples (1.000 g) were first
acidified with 4 mL of 0.3 M oxalic acid containing the internal standard dimethylpropanoic acid, then
centrifuged at 4,500 rpm for 12 min and filtered through a 0.45m GHP membrane. Finally an aliquot of
the supernatant solution was transferred into a vial. The concentrations of volatile fatty acids were
determined by Agilent Technologies 7890A gas chromatograph equipped with flame ionization
detector (FID). A polar phase capillary column, HP-INNOWax (30m x 0.25 mm x 0.25 μm), was used
for separation. Helium was used as a carrier gas at 1.8 ml/min flow rate. The analyses were performed
using a temperature programme: 5 min at 100°C, a linear gradient from 100-120°C at the rate of 10
°C/min, 5 min at 120°C, a linear gradient from 120°C to the final temperature of 220°C at the rate of
30°C/min and final hold at
220°C for 3 min. The temperatures of injection port and detector were set at
285°C and 300°C, respectively.
Protein-SIP and amplicon analysis
Liquid samples for protein-SIP and amplicon analysis were collected from all reactors periodically and
stored at -20°C until analysis. Samples were collected at different time points at 8, 24, 32, 48, 96, 144,
and 192 hrs after starting the experiment.
61
Protein extraction, precipitation and SDS-PAGE
One biological replicate of time points 8, 24, 48 and 192 hrs of each condition was used for Protein-SIP
analysis. Samples were mixed 1:1 (V/V) with B-PER® reagent (Thermo Scientific) in impact resistant
micro packaging vials (Thermo Scientific) containing 0.3 g of glass beads (106 µm diameter, Sigma-
Aldrich) and 0.3 g of ceramic beads (1.4 mm diameter, MoBio). Samples were homogenized using a
Precellys 24 bead-based homogenizer (Stretton Scientific, Stretton, UK) running 3 cycles of 6000 rpm
for 20 sec, keeping a 10 sec break in between. Three cycles of fast freeze-thaw were carried out, by
alternatingly submerging samples in liquid nitrogen and hot water (95°C). The step of homogenization
and freeze-thaw was repeated for all samples. Cell debris was removed by centrifugation at 14,500 g
for 10 min at 4°C.
Samples were mixed with ice cold 100 % acetone in the ratio 1:5 (V/V) and vortexed thoroughly. After
1 hr incubation at -20°C, proteins were pelleted at 14,500 g for 10 min at 4°C. The supernatant was
discarded and the pellet was dried down in a centrifugal concentrator (Centrivap, Labconco). The pellet
was resuspended in water and the protein content was estimated using the Qubit® protein assay kit
(Invitrogen). Following the determination of the protein content, proteins were pelleted with 100 %
acetone, as previously described. The pellet was resuspended in SDS-PAGE sample buffer
supplemented with dithiothreitol to a final concentration of 40 mM denatured by boiling at 95°C for 10
min. Samples were cooled to room temperature and loaded onto a pre-cast 4-15% gradient SDS-gel
(BioRad), followed by separation for 5 min at 160V.
In-gel digestion
In-gel digestion of samples was performed as previously described (20). Briefly, each gel lane was cut
out and then further excised. Gel pieces were washed, reduced and alkylated prior to being digested
with trypsin. Digested peptides were extracted, dried down and resuspended in a solution of 0.1%
(V/V) Trifluoroacetic acid (TFA) and 2% Acetonitrile (V/V).
Desalting
Peptides were desalted using the StageTip protocol originally published as described elsewhere (21). In
brief, two C18 disks (3 M Empore C18 extraction disk) were stacked on top of each other in a gel
loader tip (Eppendorf, VWR). R3 material (Poros, Applied Biosystem) was loaded on top of the C18
disks and the gel loader tip was washed using 80% (V/V) acetonitrile in 0.1% (V/V) TFA and
subsequently equilibrated with 0.1% (V/V) TFA. Samples were loaded, washed in 0.1% (V/V) TFA
and eluted using 80% (V/V) acetonitrile in 0.1% (V/V) TFA.
62
LC-MS/MS
Tryptic peptides were analyzed by an automated liquid chromatograph-electrospray ionization –tandem
mass spectrometer (LC-ESI-MS/MS) consisting of an UltiMate 3000 RSLCnano system (Thermo
Scientific) coupled to a Q Exactive mass spectrometer (Thermo Scientific) via a Nanospray Flex ion
source (Thermo Sientific). The analytical conditions were as previously described (22), with minor
modification of the chromatographic condition. Analytes were eluted during a 120 min linear gradient,
ranging from 12-40% (V/V) of solvent B (0.1% (V/V) formic acid (FA), 0.005% (V/V)
heptafluorbutyric acid (HFBA), 90% (V/V) acetonitrile).
Protein analysis
A six-frame translation and prediction of open reading frames (ORFs) in the in-house constructed
metagenome of the anaerobic reactor of Foulum was carried out in MaxQuant (v. 1.5.1.2). A two-
search strategy (23) was then used for creating a sub search space of the NCBI database: An initial
survey search was performed creating the search space that was merged with the six-frame translation
of metagenome used in the subsequent main search.
The initial survey search, was performed using Proteome Discoverer v. 1.4 (Thermo Scientific) and
Mascot v. 2.3 (Matrix Science). For the survey search, raw files from all reactors were searched against
all the prokaryotic entries of NCBInr (21st of August 2014). Trypsin was selected as the digestion
enzyme. A precursor mass tolerance of 20 ppm and a fragment mass tolerance of 0.05 Da was used,
and two missed cleavages were allowed. Oxidation of methionine was set as a dynamic modification
and carboxymethylation of cysteine as a fixed modification. Search reports from all reactors were
loaded into a single multiconsensus report in Proteome discoverer. Low confidence Petide Spectral
Matches (PSMs) were allowed and delta Cn was set to 0.5. The list of weakly filtered protein
identifications was exported as a fasta file. Prior to the main search contaminants and reverse entries as
decoys were added to the fasta file.
For the main search used OMSSA (Geer et al., 2004) in an OpenMS pipeline (Sturm et al., 2008).
Fixed and static modifications were set as in the survey search. The precursor mass tolerance was set to
5 ppm and the fragment mass tolerance was set to 0.5 Da. False discovery rate (FDR) was set to 1%.
Raw files were analyzed using OpenMS (24). Data was matched against a merged database comprised
of the six-frame translation of the ORFs of the metagenome and the pre-search results from the NCBI
database.
RIA and protein LR were, as described elsewhere (25), determined using the opensource software
OpenMS and the MetaproSIP tool (http://openms.de/metaprosip).
DNA extraction
DNA extraction was conducted in triplicate from the liquid samples collected at 8, 32 and 192 hr of
incubation periods. The samples were centrifuged for 10 min 14,000 g and the supernatant was
discarded. Phosphate buffer (978 µL) and MT buffer (122 µL) from FastDNA™ Spin Kit for Soil was
added and mixed. Proteinase K, SDS (1 %) and CaCl2 (1.2 mM) was added and samples were
incubated at 37°C for 1 hr. Sample was transferred to FastDNA™ Spin Lysing Matrix tube E and
homogenized in bead beater for 5 cycles at 6000 rpm for 40 sec followed by 3 cycles of freeze/thaw in
63
liquid nitrogen and 95°C water. Samples were centrifuged and pellet was mixed with 250 µL PPS
reagent and protocol from FastDNA™ Spin Kit for Soil was followed hereafter.
Amplicon PCR
Amplicon libraries were prepared for the 16S rRNA V4 region as described elsewhere (26) using a
one-step PCR followed by Agencourt AMPure XP system for purification. Master mix was prepared as
follows: 10x buffer Platinum® High Fidelity, 400 µM dNTP, 1.5 mM MgSO4, 2 mU Platinum® Taq
DNA polymerase High Fidelity, 400 nM 515f and 806r barcoded primers (26), DEPC H2O and 5 ng/µL
template DNA. PCR cycle: 94°C for 3 min, 35 cycles of 94°C for 45 sec, 50°C for 60 sec, and 72°C for
90 sec, followed by 72°C for 10 min. DNA concentration were measured using Quant-iT™ High-
Sensitivity DNA Assay Kit on TECAN Infinite M1000 plate reader. PCR product was visualized on
Agilent 2200 TapeStation prior to multiplexing. Sequencing was performed on an Illumina Miseq.
Sequence analysis
Forward and reverse reads were assembled using FLASH (27). Unique reads observed less than 2 times
were discarded to reduce sequencing noise. QIIME (28) was used for clustering (uclust) at a 97 %
identity level, picking representative set of sequences, and for taxonomic assignment with RDP
classifier, against Greengenes database (29). Each sequence library was sub-sampled to 10,000
sequences. R was used for the remaining analysis, using phyloseq, ggplot2 and an in-house script
(https://github.com/MadsAlbertsen/ampvis). The phylogenetic analysis does not take into account 16S
copy number and primer specificity, thus this analysis is considered semi-quantitative.
Metagenome preparation
Two metagenomes were prepared from samples collected at the Foulum and Lynggård biogas plants.
DNA was extracted following a CTAB and enzyme based method as described elsewhere (30–32). The
sequencing and analysis were as described elsewhere (33).
64
Results
Degradation of acetate and methane production
The degradation of acetate in the blank reactor (inoculum only) and in the reactors fed with low (4 m
M) and high (100 mM) concentrations of [U-13
C]acetate is shown in Figure 2. [U-13
C]acetate was
added once at the beginning of the incubation in the reactor of high acetate concentration where as it
was added once daily for nine days in the reactor of low acetate concentration.
The degradation of high concentration of acetate started immediately and followed a linear trend in the
first 120 hrs until it reached a low concentration (18 mM). Almost 83 % of [U-13
C]acetate was
consumed in this period. After 120 hrs, the rate of degradation of [U-13
C]acetate started to decrease and
finally reached a stationary phase during the last 68 hrs of the incubation where the concentration of
acetate was very low (~1.6 mM). On the other hand in the low acetate concentration reactor, the
residual acetate concentration measured before the feeding event during the first 120 hrs was
characterized with relatively large variation among triplicates. After 120 hrs, the variation among
replicates was small and it remained constant at about 2 mM. The concentration of acetate in the blank
reactor was about 2 mM for the first 48 hrs and later on it was below the detection limit.
Figure 2. Temporal change of the residual acetate in the blank reactors (BR; ▲); reactors fed with low
concentration (4mM) of [U-13
C]acetate (LLR; ■ ); and reactors fed with high concentration (100mM)
of [U-13
C]acetate (HLR; ●). The lines represent mean values (n = 3) and error bars denote the standard
deviation. We used 0 to represent the residual acetate concentrations below the detection limit in the
blank reactors at 72-216 hours.
MIMS was used to monitor the temporal change in the amount of the produced 13
CH4, 12
CH4, 12
CO2
and 13
CO2 during the degradation of low and high concentrations of [U-13
C]acetate and the results are
presented as atom percent of
13CO2 and
13CH4 (Figure 3). In the low acetate reactor, the proportion of
the produced 13
CO2 to total carbon dioxide increased almost linearly from 6 atom% at 24 hrs up to 32
0
20
40
60
80
100
120
0
1
2
3
4
5
6
0 24 48 72 96 120 144 168 192 216 240
Ace
tate
(m
M)
in H
LR
Ace
tate
(m
M)
in L
LR
an
d B
R
Time (hour)
65
atom% at 216 hrs whereas the proportion of the produced 13
CH4 to total methane increased gradually
from 45 atom% at 24 h up to 76 atom%. In the high acetate reactor, the production of 13
CO2 increased
from 20 atom% at 24 h to about 45 atom% at 120 h and later on stabilized at this value whereas the
production of 13
CH4 reached the peak value (80 atom%) at 24 h and remained almost stable for 120 hrs
before it started to reduce later on. Almost 83% of the acetate was degraded during the time the atom
percent of 13
CH4 remained constant. Such a high level of 13
C labeled CO2 and CH4 at both low and high
concentrations of
13C labeled acetate incubations show that high amounts of
13CO2 and
13CH4 were
produced from the degradation of 13
C labeled acetate.
Figure 3. Temporal change of the atom percent of 13
CO2 and 13
CH4 in the reactors fed with low
concentration of [U-13
C]-acetate (LLR; ■ 13
CH4; □ 13
CO2); and in the reactors fed with high
concentration of [U-13
C]acetate (HLR; ● 13
CH4; ○ 13
CO2). The lines represent mean values (n = 3) and
error bars denote the standard deviation.
Metagenome analysis
In total 130,190,434 reads from the Foulum sample and 79,999,986 reads from the Lynggård sample
were sequenced. When assembled this amounted to a size of 165,123,257 bp for the Foulum
metagenome, while the Lynggård metagenome was a bit smaller, namely 82,793,162 bp. Both of the
metagenomes represent approximately 98 % Bacteria and 2 % Archaea. The coverage of the two
metagenomes, scaffold length and phylogeny are plotted in Figure 4. The metagenomes were used as a
reference database for the identified and 13
C labeled peptides, Figure 5.
0
10
20
30
40
50
40
45
50
55
60
65
70
75
80
85
0 24 48 72 96 120 144 168 192 216 240
13C
O2 (
ato
m%
)
13C
H4 (
ato
m%
)
Time (hour)
66
Figure 4. The scaffold coverage from two metagenomes are plotted. The dot sizes indicate the scaffold
length. Coloring is according to phylogeny.
67
Figure 5. The scaffold coverage from two metagenomes are plotted. The dot sizes indicate the scaffold
length. Coloring is according to identified and 13
C-labeled peptides observed in the protein-SIP
analysis. Labeling was seen in 6 clusters of scaffolds belonging to 5 subspecies of Clostridia and
Methanoculleus.
Protein-SIP analysis
Protein-SIP analysis of the batch reactors samples at different time points showed that 13
C was
incorporated into peptides for the reactor fed with 100 mM [U-13
C] acetate starting from 48 hrs (Table
2, Figure 6). No 13
C labeling was detected in the reactor fed with 4 mM [U-13
C]acetate, control (fed
with unlabeled acetate) and blank reactors as well as in the controls (data not shown) at any time point.
In the reactor fed with high concentration of acetate a total of five peptides incorporated 13
C at 48 hrs
(Table 2) after the start of the incubation. From these 13
C labeled peptides, two of them assigned to the
domain of bacteria and the other two to the domain of archaea. The last peptide could not be assigned
68
to any specific kingdom. Following 192 hrs, a total of 56 peptides could be identified that showed
incorporation of 13
C (Figure 6 and Table 2). These peptides were dominated by bacterial species
(predominantly Clostridia). Three labeled peptides originated from the domain of archaea (two from
the genus of Methanoculleus and one peptide from Methanosarcina Barkeri).
Table 2 Peptides for which an increase in isotopic incorporation of 13
C was observed in the reactor fed
with 100 mM [U-13
C]acetate.
Peptide Sequence1 Description2 Exp. m/z3 Theo. m/z4 Charge5 RIA 16 RIA 27 LR8 Taxonomy9
48 hrs
AGDDAAGLSISEK Flagellin, subunit
protein B 617.3017 617.3015 2 0.6 90.3 91.3 Clostridia1 [class OPB54]
SVAVNLAGIQGALASGK
Chain B, Methyl-
Coenzyme M
Reductase
778.4442 778.4438 2 0.6 57.6 5.8 Methanosarcina Barkeri
GPNEPGGLSFGHLSDIIQTSR
methyl coenzyme
M reductase
subunit alpha
728.0337 728.0344 3 1.2 51.1 7.4 Archaea
LIGHGPFILDQYK
500.9459 500.9452 3 0.6 25.1 0.8 Clostridia2 [genus Gelria]
LTPEEFVSTFIPADLTWM
(Oxidation)R 757.0406 757.0452 3 1.1 14.0 60.6 Unassigned
192 hrs
FITVGEKYPEGLTAPR
593.3216 593.3210 3 1.2 24.0 18.7 Clostridia1 [class OPB54]
AVTGPLPPLVWASR
732.4232 732.4221 2 0.6 20.7 21.6 Clostridia1 [class OPB54]
LLIALQTSDK
551.3309 551.3293 2 0.7 34.3 21.4 Clostridia2 [genus Gelria]
FVAIEHVSADAAR
462.5779 462.5772 3 0.7 93.8 35.0 Clostridia2 [genus Gelria]
SVAVNLAGIQGALASGK
Chain B Methyl-
Coenzyme M
Reductase
778.4450 778.4438 2 0.6 58.1 39.2 Methanosarcina Barkeri
GFVSNPYTGNYM
(Oxidation)PHR 585.9351 585.9351 3 1.2 21.0 20.7 Clostridia1 [class OPB54]
AGELAFSASK
490.7562 490.7560 2 0.7 25.4 16.8 Clostridia1 [class OPB54]
GFVSNPYTGNYM
(Oxidation)PHR 585.9357 585.9351 3 1.2 20.7 13.4 Clostridia1 [class OPB54]
TLDEFFQIAK
606.3202 606.3190 2 0.6 22.3 38.3 Clostridia1 [class OPB54]
IDEVWLAAQR
600.8230 600.8222 2 0.6 25.9 19.6 Clostridia1 [class OPB54]
YNVEVEFKPVPR
492.9338 492.9330 3 0.6 20.3 12.9 Clostridia1 [class OPB54]
RGEEIGGTIR
544.2969 544.2964 2 0.7 26.6 10.4 Clostridia1 [class OPB54]
VDELLELGR
ABC transporter
substrate-binding
protein
522.2914 522.2902 2 0.7 31.7 26.1 Clostridia1 [class OPB54]
VAIGVEDLGHTSLAER
556.2979 556.2968 3 1.2 26.8 15.0 Clostridia3
RGEEIGGTIR
544.2968 544.2964 2 0.7 25.8 0.9 Clostridia1 [class OPB54]
AIISVDANTQSHGVVIR
593.9960 593.9952 3 1.3 27.2 17.6 Clostridia3
ITITDINDVAHHQFK
438.7331 438.7323 4 1.2 24.2 18.4 Clostridia1 [class OPB54]
AIGINANPAFPDAGVYNDR
987.9882 987.9871 2 1.2 26.3 17.6 Clostridia1 [class OPB54]
HFTNSIRPIR
414.2352 414.2352 3 0.6 28.9 13.9 Clostridia2 [genus Gelria]
IDSDLSKYDVYLQSAAR
648.6618 648.6618 3 1.2 24.8 22.0 Clostridia1 [class OPB54]
DFPLYGAGDRTEDNLIK
641.9862 641.9864 3 1.2 25.4 13.6 Clostridia1 [class OPB54]
69
VVEAAIAAGK
simple sugar
transport system
substrate-binding
protein
464.7768 464.7767 2 0.7 30.9 12.6 Clostridia3
DDNWWGNAVFGQPKPK
620.3004 620.3006 3 1.2 13.6 11.4 Clostridia1 [class OPB54]
DLFVQAGLPTPNELQNEGR
1049.5332 1049.5318 2 1.2 21.7 22.9 Clostridia1 [class OPB54]
AIGINANPAFPDAGVYNDR
658.9934 658.9938 3 1.2 15.0 7.1 Clostridia1 [class OPB54]
ITITDINDVAHHQFK
584.6417 584.6407 3 1.2 22.8 17.4 Clostridia1 [class OPB54]
AAELGVTLR
signal
transduction
histidine kinase-
like protein
465.2752 465.2744 2 0.7 26.8 6.2 Clostridia1 [class OPB54]
IFTVDQISFIPK
704.3993 704.3978 2 0.6 21.4 9.5 Clostridia4
QIVGELFQEDLAALGIK
922.5129 922.5118 2 1.2 27.8 9.3 Clostridia1 [class OPB54]
IAIVFATGGLGDK
631.3625 631.3612 2 0.6 30.5 8.1 Clostridia2 [genus Gelria]
ILDLLDSAPDLATAK
778.4337 778.4325 2 1.2 31.9 14.9 unassigned
QAADEAQLILAR
649.8583 649.8568 2 0.6 36.0 9.0 Clostridia2 [genus Gelria]
VLELALKDSVR
414.9192 414.9187 3 0.6 23.9 17.7 Clostridia5
AILPSPYGAFTR
646.8548 646.8535 2 0.6 32.2 10.7 Clostridia1 [class OPB54]
GVIDPETFILNYDQYIEK
1079.0426 1079.0412 2 1.2 25.0 13.9 Clostridia1 [class OPB54]
LLDEAGYTVDPATGIR
845.9382 845.9360 2 1.2 32.2 7.0 Clostridia2 [genus Gelria]
EYEISEDGTEVTFYLR
975.9533 975.9520 2 1.3 20.3 14.7 Clostridia1 [class OPB54]
FQVGFEEGVK
570.2911 570.2902 2 0.6 30.4 9.9 Clostridia1 [class OPB54]
NLTFAEVGFR
577.3048 577.3037 2 0.7 17.3 8.8 Clostridia2 [genus Gelria]
WTTLNIKPVR
409.9120 409.9118 3 0.6 24.6 6.5 Clostridia2 [genus Gelria]
ELLAEAGIEPGEISIR
848.9621 848.9594 2 1.2 27.5 6.2 Clostridia1 [class OPB54]
VYYALDEPQAINALR
868.4566 868.4543 2 1.2 24.5 7.8 Clostridia1 [class OPB54]
SGAQVLLSR
465.7724 465.7720 2 0.6 25.5 5.9 Clostridia1 [class OPB54]
REPLADDVLR
putative TetR
family
transcriptional
regulator
592.3258 592.3251 2 0.7 30.1 5.9 Clostridia2 [genus Gelria]
ELDLDIVGNKDAVISK
576.9853 576.9841 3 1.2 51.7 2.8 Methanoculleus
TIAVNLGGIEGALK
678.3993 678.3983 2 0.6 49.6 2.3 Methanoculleus
DNPESIFVPLPIVIDPLVEER
797.7637 797.7651 3 1.2 28.4 12.2 Clostridia1 [class OPB54]
IEVTVEEGLPVAK
692.3909 692.3901 2 0.6 26.9 13.9 Clostridia1 [class OPB54]
ALAFAVNPEIIVER
771.4387 771.4379 2 0.6 26.6 7.9 Clostridia1 [class OPB54]
LIGHGPFILDQYK
500.9458 500.9452 3 0.6 26.2 6.1 Clostridia2 [genus Gelria]
IDQELILVR
549.8302 549.8295 2 0.7 25.1 11.0 Clostridia1 [class OPB54]
LELLINENR
557.3177 557.3168 2 0.7 24.5 2.4 Clostridia1 [class OPB54]
EIALSLDLSPR
RNA polymerase
sigma-70 factor
expansion family
1
607.3438 607.3430 2 0.7 24.4 15.4 Bacteriodetes
TLDLSNYFIPGVPAIK
874.4877 874.4851 2 1.5 23.7 9.1 Clostridia1 [class OPB54]
NDNYYEFDEEGNRLPYLNR
807.6958 807.6963 3 1.2 23.1 12.2 Clostridia1 [class OPB54]
ALELSLEDSPR
615.3230 615.3222 2 0.6 20.7 5.0 Clostridia4
1Peptide sequence (with identified modification),
2description of the identified protein,
3experimentally
determined mass-to-charge ratio of the observed sequence, 4theoretically determined mass-to-charge
ratio of the observed sequence, 5charge of the tryptic peptide,
6relative isotope abundance of the
70
naturally occurring isotopic cluster of the peptide, 7relative isotope abundance of isotopic cluster of the
peptide showing incorporation of 13
C, 8labeling Ratio,
9Highest possible taxonomic classification.
Figure 6. A) RIA as a function of LR for peptides showing incorporation of 13
C at 192 hrs in the reactor
fed with 100 mM of [U-13
C]acetate (colors represent highest possible taxonomic rank). Functionally
annotated proteins are indicated in the figure. The peptide, SVAVNLAGIQGALASKGK, which
showed a high degree of 13
C incorporation was identified belonging to Methanosarcina Barkeri B)
Time resolved analysis of the incorporation of 13
C in the peptide SVAVNLAGIQGALASKGK
belonging to the methanogenic archaea Methanosarcina Barkeri. Incorporation of 13
C in the peptide
71
was evident after 48 hours (the arrows indicate the 13
C labeled isotopologue). The peptide was
identified as a Methyl-Coenzyme M Reductase, which catalyzes anaerobic oxidation of methane.
Besides the Methyl-Coenzyme M Reductase, proteins with functions related to transport of substrate
and sugars, signal transduction and translation and a transcriptional repressor were also identified
(Figure 6 and Table 2).
Phylogeny
The microbial community compositions were evaluated by amplicon sequencing of triplicate reactors
run with low and high concentrations of acetate. Samples at three time points were chosen for amplicon
sequencing (8, 32 and 192 hrs). In total 1,804,008 reads were sequenced of which 705,001 assembled
sequences passed filtering. At least 29,428 sequences per sample passed filtering and were used for
sub-sampling. From the archaeal population the most abundant genera were affiliated with
Methanobacterium, Methanosarcina, Methanobrevibacter, and Methanoculleus. However,
Methanomassiliicoccus and Methanothermobacter were also detected. The 5 most abundant classes of
bacteria belonged to Clostridia, Bacteroidia, Bacilli, Thermotogae, and Anaerolineae.
In the reactor of low acetate concentration 40 % of total OTUs were shared between the 3 time points.
While in the reactor of high acetate concentration, 38 % of total OTUs were shared. The differences in
the microbial communities were seen to occur during time as observed by PCA analysis Figure 7.
Among abundant genera (> 0.1%) several genera from the bacterial population were seen to either
decrease or increase from 8 to 192 hrs.
72
Figure 7. PCA taken for the replicates at 3 time points. Approximately 63 % of the total variance is
explained in this analysis. PC1 describes time during which the microbial communities in the samples
change.
Discussion
The aim of this study was to identify the key microorganisms involved in recovery of acetate
accumulation in AD batch reactors using protein-SIP and metagenomics. Furthermore, we intended to
identify the effect of low and high acetate concentrations on the microbial community structure. In this
study protein-SIP was successfully applied to investigate the microorganisms involved in acetate
consumption in batch AD. The study was composed of a combined metabolomic, metagenomic and
metaproteomic approach. Acetate degradation products were investigated through an isotopic
evaluation of CO2 and CH4. Phylogenetic identification of the AD batch reactors was examined by
amplicon sequencing. While the specific detection of heavy labelling was performed with protein-SIP
combined with a search against two biogas metagenomes.
73
The production of highly 13
C labeled CO2 and CH4 immediately after the start of incubation in both the
low and high concentrations of [U-13
C]acetate showed that CO2 and CH4 are the main products of the
degradation of acetate. The relatively large variation of acetate concentration among triplicates for the
low acetate reactors are probably due to adaption of the microbial community. Almost all of the acetate
was degraded to CO2 and CH4 within 120 hrs in the high concentration reactor.
Since both the methyl and carboxyl groups are 13
C labeled in the [U-13
C]acetate, it is difficult to
estimate the proportion of methane produced through SAO-HM using the values of atom percent of 13
CO2 and 13
CH4. The produced 13
CO2 is not specific to one of the pathways as it can be produced
either through acetoclastic methanogenesis or SAO during the degradation of [U-13
C]acetate. In our
previous study (16), we used methyl 13
C labeled acetate, [2-13
C]acetate, which allowed us to estimate
the SAO-HM contribution to methane since 13
CO2 can be produced from the degradation of [2-13
C]acetate through SAO-HM alone. Nevertheless, the measured atom percent of 13
CH4 and CO2
during the degradation of [U-13
C]acetate can be qualitatively interpreted as follows. The atom% of 13
CO2 was generally lower than the 13
CH4 in both the low and high concentrations of [U-13
C]acetate due
to the high background pool of unlabeled CO2 in the system. The produced 13
CH4 in the former reached
a maximum of 75 atom% and 80 atom% in the latter, indicating the production of unlabeled CH4. One
possibility is that H2 and unlabeled acetate were produced by the fermentation of the background
biomass in the inoculum and subsequently used by the hydrogenotrophs and acetoclastic methanogens
for generating unlabeled CH4. The results from the blank reactor (only inoculum) showed that the
background unlabeled acetate from the inoculum was lower than 2 mM. Therefore, the contribution of
the inoculum to the production of unlabeled acetate and H2 in the [U-13
C]acetate reactors is likely very
low. Another possibility is via the reduction of the background pool of unlabeled CO2 by
hydrogenotrophs in syntrophy with the electron provided by SAOB in the form of H2. This assumption
is in line with our previous findings where SAO-HM played a key role for the production of methane
during the degradation of high concentration of [2-13
C]acetate (16). All the experimental conditions in
this study and the previous study (16) was identical except the inoculum was obtained at different time
points from the same full-scale biogas plant running with a mixture of pig and cattle manure, maize
silage and deep litter manure under the same operating conditions. Although the microbial community
from the same plant can change over time, it remained relatively stable during the time of this and the
previous studies (16) (data not shown). The dominant microbial communities are almost similar in both
studies (see the discussion below), indicating SAO-HM played significant role in the reactor fed with
the high concentration of [U-13
C]acetate.
The microbial community composition was investigated during the incubation by amplicon sequencing
of the 16S rRNA gene in both low and high acetate reactors. In general the microbial community
observed in both batch reactors are in accordance with previous studies of AD communities (32, 34–
36). The majority of the identified cells was affiliated to Clostridia that are capable of taking part in
several processes in the biogas plant such as hydrolysis of cellulosic plant biomass and acetate
oxidation (32, 37, 38). The bacterial classes of Bacteroidia, Bacilli, Thermotogae, Anaerolineae,
Synergistia as well as several Proteobacteria were also highly abundant. This composition is in
accordance with our metagenome constructed from the Foulum and Lynggård biogas plants. Only few
of the identified bacterial classes observed in AD have been affiliated with functional roles in the
ecosystem so far. Species from Clostridia have been observed to take part in hydrolysis, acidogenesis
and acetogenesis. This correlates well with the high abundance of Clostridia detected in our reactors.
74
Species from class Bacteroidia, which were similarly highly abundant and have previously been
described as hydrolyzers of various polymers (32). Cultured species from Anaerolineae, within the
Chloroflexi, has been shown to have significantly stimulated growth in the presence of
hydrogenotrophic methanogens (39) . Compared to the three pathways described above,
methanogenesis are well studied, both for heterogenotrophic and acetoclastic methanogens.
Methanobacterium, Methanosarcina, Methanobrevibacter, and Methanoculleus were the 4 most
abundant genera of Archaea detected in our reactors. These methanogens were commonly found in AD
in previous studies (30, 31, 40–42).
The abundance of Archaea (1-2%) compared to Bacteria (98-99%) was relatively low. Several
previous studies of the microbial diversity in AD were based on separate analysis for Archaea and
Bacteria thus the abundance of the two were not directly comparable (35, 43–45). However, in our
study we applied a universal primer set targeting both kingdoms, and it is therefore possible that the
true abundance of Archaea is higher. The approach used in this study do not take into account 16S
rRNA gene copy number, and it has previously been shown that Archaea in general has a lower copy
number than Bacteria (46).
There was no statistically significant difference between the communities at low and high acetate
concentrations. However, we did observe minor differences in relative abundance of some of the
microorganisms at the two different acetate concentrations. Thus, the concentration of acetate had a
minor effect on the microbial communities during the incubations. Time had the major impact on the
microbial communities as verified by PCA. As we suspected from the acetate degradation study, the
microbial community require time to adjust to new growth conditions. While the abundance of
Methanobrevibacter and Methanoculleus were constant during the incubation, the abundance of
Methanobacterium decreased and the abundance of Methanosarcina increased. The increase of
Methanosarcina is interesting since this organism is capable of both acetoclastic and hydrogenotrophic
pathways. Since the residual acetate concentration in both reactors was above 1.5 mM over the 9 days
of incubation, such a high acetate concentration possibly favors the growth of Methanosarcina.
Members of Methanosarcina are favored at an acetate concentration higher than 1 mM (Hori et al.,
2006; Karakashev et al., 2005). The relative abundance of several microorganisms remained largely
unchanged with time, but a significant increase in Bacillales specifically the genera Ureibacillus were
observed, and might have benefitted from the presence of acetate amendments, although it was not
found in the labeled fraction of proteins.
Protein-SIP was employed for the identification of active microorganisms and determining the level of
activity. 13
C labeled peptides were detected only in the reactors fed with high concentration of acetate.
Sufficient incorporation of the 13
C for detection of labeled peptides of the microorganisms was
observed 48 hrs after starting the incubation experiment (5 peptides) and further increased to 56
peptides after 192 hrs. This indicates that the microbial community requires an adaptation period to
adjust to a changing environment such as the high concentration of acetate. The number of 13
C labeled
peptides detected increased by using the two metagenomes as a reference database. Still, the number of 13
C labeled peptides detected in our reactors is considered low and can be explained by the very dense
nature of the samples and that protein extractions are often contaminated by humic substances, which
can disturb the MS measurement (47). If humic substances affected the detection of the 13
C labeled
peptides, we would suspect the unlabeled peptides to be affected as well. However, we detected
between 1000-2100 unlabeled peptides in our reactors, which is much larger than those labeled with
75
13C. There is also the possibility that the incubation for a total of 9 days may have been too short for the
slow growing methanogens to adjust and synthesize new proteins. Furthermore, the microbial
communities in the AD are very complex and hence, difficult to analyze using MS. Consequently, it
would be challenging to detect an increase in 13
C-incorporation (48).
Peptides from 5 subspecies of Clostridia and one Bacteroidetes as well as Methanosarcina and
Methanoculleus incorporated 13
C from the 100 mM 13
C labeled acetate. Many of the peptides came
from hypothetical proteins but none of them could be involved in SAO pathway. The gene encoding
the enzyme formyltetrahydrofolate synthetase (fhs) is an ecological biomarker for reductive
acetogenesis(49, 50). Thus, we searched for the fhs gene in the metagenomes, which is present in all of
the Clostridia we detected with the 13
C labelling approach. Therefore, we hypothesize that these
phylotypes are responsible for synthrophic acetate oxidation. Formyltetrahydrofolate synthetase also
catalyzes the formation of acetate from H2 and CO2 (49, 50), but has previously been associated with
syntrophic acetate oxidation in an anaerobic digester (Hori et al., 2011).
Most of the peptides had a RIA (proportion of the peptide 13
C labeled) of around 13-36 %, however,
four peptides had higher RIA values of 49.6 % and 51.7 % (Methanoculleus), 58.1 %
(Methanosarcina), and 93.8 % (genus Gelria in the class Clostridia), respectively. The LR (how much
of a peptide population is labeled) of these four peptides were 2.3 % and 2.8 % (Methanoculleus), and
39.2 % (Methanosarcina) and 35.0 % (Gelria). These data clearly confirm that these microorganisms
are capable of acetate degradation.
We detected unlabeled peptides from several methanogens both acetoclastic and hydrogenotrophic.
Among the methanogens, Methanosarcina have the highest 13
C labelling with a RIA of 58.1% and an
LR of 39.2 %. The labeled peptide came from the methyl coenzyme M reductase β subunit, which is a
common intermediate reaction of all metabolic pathways leading to methane formation (51). Thus,
methyl coenzyme M reductase take part in both methanogenic pathways (52). The highest 13
C-labelling
of Methanosarcina is in accordance with the measured acetate consumption rate and the observed
increase in the relative abundance of Methanosarcina during the course of incubation as shown by the
amplicon sequencing analysis. This methanogen participates in acid recovery in anaerobic digestions
by transformation of acetate in accordance with previous studies (53). Methanosarcina was also highly
abundant in the reactor fed with 100 mM [2-13
C]acetate presented in our previous study and the results
of the isotope analysis demonstrated the key role of HM-SAO in the degradation of acetate (16).
Adjusting the parameters of the AD to fit Methanosarcina can possibly improve the biogas production.
There are several advantages of having Methanosarcina in the reactor, it is both acetoclastic and
heterogenotrophic, due to this it is more tolerant to several inhibitions such as high ammonium levels,
low retention times, and high organic loading rates (54). Methanosaeta, which is strictly acetoclastic
methanogens were not detected with amplicon sequencing. This could point in the direction of HM-
SAO taking place, this pathway is likewise dominant at thermophilic temperatures and increased levels
of acetate (54). However, since Methanosarcina is capable of both methanogenesis pathways it is not
possible to define the exact pathway in this study.
76
Methanoculleus were the other methanogen found with labeled peptides. This particular methanogen is
hydrogenotrophic and grow on H2 and CO2 (55). Applying [2-13
C]acetate would have allowed to detect
the exact pathway occurring by using MIMS. Furthermore, the use of primers targeting the fhs gene
would allow us to identify the specific SAOB present in our reactors. Since Methanoculleus
incorporated 13
C into their peptides while growing on 100 mM [2-13
C]acetate, they likely carried out
SAO-HM pathway in syntrophy with SAOB.
Conclusion
In this study we showed that protein-SIP is a method that can detect the active microorganisms that
incorporated 13
C into their proteins in complex samples from AD batch reactors. We conclude that the
combined use of a metagenome is highly recommended to improve the identification of the 13
C labeled
proteins. Thus, construction of a metagenome is highly relevant for this type of study. MIMS was a
valuable method for tracing the incorporation of 13
C into the produced CO2 and CH4 from the
degradation of 13
C labeled acetate. Peptides from Clostridia, Bacteroidetes, Methanosarcina, and
Methanoculleus were labeled with 13
C and therefore confirmed that these microorganisms were
involved in the degradation of acetate to methane.The 13
C labeled Clostridia are possibly oxidizing
acetate as part of a synthrophy since they contain the fhs gene coding for formyltetrahydrofolate
synthetase. This gene is a key enzyme in reductive acetogenesis and the presence hereof strongly
indicates that these cells are possible synthrophic acetate oxidizing bacteria (SAOB) that can facilitate
acetate consumption via syntrophic acetate oxidation coupled with hydrogenotrophic methanogenesis
(SAO-HM). Since Methanosarcina is a mixotrophic methanogens, its exact role as acetoclastic or
hydrogenotrophic methanogenesis was not verified. Methanoculleus are heterogenotrophic and thus
likely involved in SAO-HM pathway in the reactors fed with high concentration of acetate. Since 13
C
labeled peptides was not detected in the low acetate concentration, it was not possible to identify the
active microorganisms metabolizing acetate in this reactor. Nevertheless, the results from amplicon
sequencing indicated that the acetate concentration only had a minor effect on the microbial
community.
Acknowledgements
This study was funded by the Danish Strategic Research Council (Grant No. 10-093944).
77
References
1. Demirel B, Yenig�n O. 2002. Two-phase anaerobic digestion processes: a review. J. Chem.
Technol. Biotechnol. 77:743–755.
2. Chen Y, Cheng JJ, Creamer KS. 2008. Inhibition of anaerobic digestion process: a review.
Bioresour. Technol. 99:4044–64.
3. Krakat N, Schmidt S, Scherer P. 2011. Potential impact of process parameters upon the
bacterial diversity in the mesophilic anaerobic digestion of beet silage. Bioresour. Technol.
102:5692–701.
4. Angelidaki I, Ahring BK. 1993. Thermophilic anaerobic digestion of livestock waste: the effect
of ammonia. Appl. Microbiol. Biotechnol. 38.
5. Labatut RA, Angenent LT, Scott NR. 2014. Conventional mesophilic vs. thermophilic
anaerobic digestion: a trade-off between performance and stability? Water Res. 53:249–58.
6. Fotidis IA, Karakashev D, Angelidaki I. 2013. Bioaugmentation with an acetate-oxidising
consortium as a tool to tackle ammonia inhibition of anaerobic digestion. Bioresour. Technol.
146:57–62.
7. Lü F, Hao L, Guan D, Qi Y, Shao L, He P. 2013. Synergetic stress of acids and ammonium on
the shift in the methanogenic pathways during thermophilic anaerobic digestion of organics.
Water Res. 47:2297–306.
8. Palatsi J, Viñas M, Guivernau M, Fernandez B, Flotats X. 2011. Anaerobic digestion of
slaughterhouse waste: main process limitations and microbial community interactions.
Bioresour. Technol. 102:2219–27.
9. Rajagopal R, Massé DI, Singh G. 2013. A critical review on inhibition of anaerobic digestion
process by excess ammonia. Bioresour. Technol. 143:632–41.
10. Müller B, Sun L, Schnürer A. 2013. First insights into the syntrophic acetate-oxidizing
bacteria--a genetic study. Microbiologyopen 2:35–53.
11. Karakashev D, Batstone DJ, Trably E, Angelidaki I. 2006. Acetate oxidation is the dominant
methanogenic pathway from acetate in the absence of Methanosaetaceae. Appl. Environ.
Microbiol. 72:5138–41.
12. Ito T, Yoshiguchi K, Ariesyady HD, Okabe S. 2011. Identification of a novel acetate-utilizing
bacterium belonging to Synergistes group 4 in anaerobic digester sludge. ISME J. 5:1844–56.
13. Ito T, Yoshiguchi K, Ariesyady HD, Okabe S. 2012. Identification and quantification of key
microbial trophic groups of methanogenic glucose degradation in an anaerobic digester sludge.
Bioresour. Technol. 123:599–607.
14. Limam RD, Chouari R, Mazéas L, Wu T-D, Li T, Grossin-Debattista J, Guerquin-Kern J-
L, Saidi M, Landoulsi A, Sghir A, Bouchez T. 2014. Members of the uncultured bacterial
candidate division WWE1 are implicated in anaerobic digestion of cellulose. Microbiologyopen
3:157–67.
15. Hao L, Lü F, Mazéas L, Desmond-Le Quéméner E, Madigou C, Guenne A, Shao L,
Bouchez T, He P. 2014. Stable isotope probing of acetate fed anaerobic batch incubations
shows a partial resistance of acetoclastic methanogenesis catalyzed by Methanosarcina to sudden
increase of ammonia level. Water Res. 69C:90–99.
16. Mulat DG, Ward AJ, Adamsen APS, Voigt NV, Nielsen J, Feilberg A. 2014. Quantifying
contribution of synthrophic acetate oxidation to methane production in thermophilic anaerobic
reactors by membrane inlet mass spectrometry. Environ. Sci. Technol. 140117091552006.
78
17. Chapleur O, Bize A, Serain T, Mazéas L, Bouchez T. 2014. Co-inoculating ruminal content
neither provides active hydrolytic microbes nor improves methanization of 13
C-cellulose in batch
digesters. FEMS Microbiol. Ecol. 87:616–29.
18. Werner JJ, Garcia ML, Perkins SD, Yarasheski KE, Smith SR, Muegge BD, Stadermann
FJ, DeRito CM, Floss C, Madsen EL, Gordon JI, Angenent LT. 2014. Microbial community
dynamics and stability during an ammonia-induced shift to syntrophic acetate oxidation. Appl.
Environ. Microbiol. 80:3375–83.
19. Jehmlich N, Schmidt F, Taubert M, Seifert J, Bastida F, von Bergen M, Richnow H-H,
Vogt C. 2010. Protein-based stable isotope probing. Nat. Protoc. 5:1957–66.
20. Shevchenko A, Tomas H, Havlis J, Olsen J V, Mann M. 2006. In-gel digestion for mass
spectrometric characterization of proteins and proteomes. Nat. Protoc. 1:2856–60.
21. Rappsilber J, Mann M, Ishihama Y. 2007. Protocol for micro-purification, enrichment, pre-
fractionation and storage of peptides for proteomics using StageTips. Nat. Protoc. 2:1896–906.
22. Kjeldal H, Pell L, Pommerening-Röser A, Nielsen JL. 2014. Influence of p-cresol on the
proteome of the autotrophic nitrifying bacterium Nitrosomonas eutropha C91. Arch. Microbiol.
196:497–511.
23. Seifert J, Herbst F-A, Halkjaer Nielsen P, Planes FJ, Jehmlich N, Ferrer M, von Bergen M.
2013. Bioinformatic progress and applications in metaproteogenomics for bridging the gap
between genomic sequences and metabolic functions in microbial communities. Proteomics
13:2786–2804.
24. Sachsenberg T, Herbst F-A, Taubert M, Kermer R, Jehmlich N, von Bergen M, Seifert J,
Kohlbacher O. 2014. MetaProSIP: Automated Inference of Stable Isotope Incorporation Rates
in Proteins for Functional Metaproteomics. J. Proteome Res.
25. Kleindienst S, Herbst F-A, Stagars M, von Netzer F, von Bergen M, Seifert J, Peplies J,
Amann R, Musat F, Lueders T, Knittel K. 2014. Diverse sulfate-reducing bacteria of the
Desulfosarcina/Desulfococcus clade are the key alkane degraders at marine seeps. ISME J.
8:2029–44.
26. Caporaso JG, Lauber CL, Walters W a, Berg-Lyons D, Huntley J, Fierer N, Owens SM,
Betley J, Fraser L, Bauer M, Gormley N, Gilbert J a, Smith G, Knight R. 2012. Ultra-high-
throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. ISME J.
6:1621–4.
27. Magoč T, Salzberg SL. 2011. FLASH: fast length adjustment of short reads to improve genome
assemblies. Bioinformatics 27:2957–63.
28. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, Fierer
N, Peña AG, Goodrich JK, Gordon JI, Huttley GA, Kelley ST, Knights D, Koenig JE, Ley
RE, Lozupone CA, McDonald D, Muegge BD, Pirrung M, Reeder J, Sevinsky JR,
Turnbaugh PJ, Walters WA, Widmann J, Yatsunenko T, Zaneveld J, Knight R. 2010.
QIIME allows analysis of high-throughput community sequencing data. Nat. Methods 7:335–6.
29. DeSantis TZ, Hugenholtz P, Larsen N, Rojas M, Brodie EL, Keller K, Huber T, Dalevi D,
Hu P, Andersen GL. 2006. Greengenes, a chimera-checked 16S rRNA gene database and
workbench compatible with ARB. Appl. Environ. Microbiol. 72:5069–72.
30. Klocke M, Mähnert P, Mundt K, Souidi K, Linke B. 2007. Microbial community analysis of
a biogas-producing completely stirred tank reactor fed continuously with fodder beet silage as
mono-substrate. Syst. Appl. Microbiol. 30:139–51.
79
31. Blume F, Bergmann I, Nettmann E, Schelle H, Rehde G, Mundt K, Klocke M. 2010.
Methanogenic population dynamics during semi-continuous biogas fermentation and
acidification by overloading. J. Appl. Microbiol. 109:441–50.
32. Wirth R, Kovács E, Maróti G, Bagi Z, Rákhely G, Kovács KL. 2012. Characterization of a
biogas-producing microbial community by short-read next generation DNA sequencing.
Biotechnol. Biofuels 5:41.
33. Albertsen M, Hugenholtz P, Skarshewski A, Nielsen KL, Tyson GW, Nielsen PH. 2013.
Genome sequences of rare, uncultured bacteria obtained by differential coverage binning of
multiple metagenomes. Nat. Biotechnol. 31:533–8.
34. Sundberg C, Al-Soud WA, Larsson M, Alm E, Yekta SS, Svensson BH, Sørensen SJ,
Karlsson A. 2013. 454 pyrosequencing analyses of bacterial and archaeal richness in 21 full-
scale biogas digesters. FEMS Microbiol. Ecol. 85:612–26.
35. Lee S-H, Kang H-J, Lee YH, Lee TJ, Han K, Choi Y, Park H-D. 2012. Monitoring bacterial
community structure and variability in time scale in full-scale anaerobic digesters. J. Environ.
Monit. 14:1893–905.
36. St-Pierre B, Wright A-DG. 2014. Comparative metagenomic analysis of bacterial populations
in three full-scale mesophilic anaerobic manure digesters. Appl. Microbiol. Biotechnol.
98:2709–17.
37. Schlüter A, Bekel T, Diaz NN, Dondrup M, Eichenlaub R, Gartemann K-H, Krahn I,
Krause L, Krömeke H, Kruse O, Mussgnug JH, Neuweger H, Niehaus K, Pühler A, Runte
KJ, Szczepanowski R, Tauch A, Tilker A, Viehöver P, Goesmann A. 2008. The metagenome
of a biogas-producing microbial community of a production-scale biogas plant fermenter
analysed by the 454-pyrosequencing technology. J. Biotechnol. 136:77–90.
38. Hattori S. 2008. Syntrophic Acetate-Oxidizing Microbes in Methanogenic Environments.
Microbes Environ. 23:118–127.
39. Yamada T, Sekiguchi Y, Hanada S, Imachi H, Ohashi A, Harada H, Kamagata Y. 2006.
Anaerolinea thermolimosa sp. nov., Levilinea saccharolytica gen. nov., sp. nov. and Leptolinea
tardivitalis gen. nov., sp. nov., novel filamentous anaerobes, and description of the new classes
Anaerolineae classis nov. and Caldilineae classis nov. in the. Int. J. Syst. Evol. Microbiol.
56:1331–40.
40. Ziganshina EE, Bagmanova AR, Khilyas I V, Ziganshin AM. 2014. Assessment of a biogas-
generating microbial community in a pilot-scale anaerobic reactor. J. Biosci. Bioeng. 117:730–6.
41. Hagen LH, Vivekanand V, Linjordet R, Pope PB, Eijsink VGH, Horn SJ. 2014. Microbial
community structure and dynamics during co-digestion of whey permeate and cow manure in
continuous stirred tank reactor systems. Bioresour. Technol.
42. Tuan NN, Chang Y-C, Yu C-P, Huang S-L. 2014. Multiple approaches to characterize the
microbial community in a thermophilic anaerobic digester running on swine manure: a case
study. Microbiol. Res. 169:717–24.
43. Rodríguez E, Lopes A, Fdz-Polanco M, Stams AJM, García-Encina P a. 2012. Molecular
analysis of the biomass of a fluidized bed reactor treating synthetic vinasse at anaerobic and
micro-aerobic conditions. Appl. Microbiol. Biotechnol. 93:2181–91.
44. Patil SS, Kumar MS, Ball AS. 2010. Microbial community dynamics in anaerobic bioreactors
and algal tanks treating piggery wastewater. Appl. Microbiol. Biotechnol. 87:353–63.
80
45. Kim J, Shin SG, Han G, O’Flaherty V, Lee C, Hwang S. 2011. Common key acidogen
populations in anaerobic reactors treating different wastewaters: molecular identification and
quantitative monitoring. Water Res. 45:2539–49.
46. Lee ZM-P, Bussema C, Schmidt TM. 2009. rrnDB: documenting the number of rRNA and
tRNA genes in bacteria and archaea. Nucleic Acids Res. 37:D489–93.
47. Heyer R, Kohrs F, Benndorf D, Rapp E, Kausmann R, Heiermann M, Klocke M, Reichl U.
2013. Metaproteome analysis of the microbial communities in agricultural biogas plants. N.
Biotechnol. 30:614–22.
48. Doherty MK, Beynon RJ. 2006. Protein turnover on the scale of the proteome. Expert Rev.
Proteomics 3:97–110.
49. Xu K, Liu H, Du G, Chen J. 2009. Real-time PCR assays targeting formyltetrahydrofolate
synthetase gene to enumerate acetogens in natural and engineered environments. Anaerobe
15:204–13.
50. Hori T, Sasaki D, Haruta S, Shigematsu T, Ueno Y, Ishii M, Igarashi Y. 2011. Detection of
active, potentially acetate-oxidizing syntrophs in an anaerobic digester by flux measurement and
formyltetrahydrofolate synthetase (FTHFS) expression profiling. Microbiology 157:1980–9.
51. Grabarse W, Mahlert F, Duin EC, Goubeaud M, Shima S, Thauer RK, Lamzin V, Ermler
U. 2001. On the mechanism of biological methane formation: structural evidence for
conformational changes in methyl-coenzyme M reductase upon substrate binding. J. Mol. Biol.
309:315–30.
52. Rademacher A, Zakrzewski M, Schlüter A, Schönberg M, Szczepanowski R, Goesmann A,
Pühler A, Klocke M. 2012. Characterization of microbial biofilms in a thermophilic biogas
system by high-throughput metagenome sequencing. FEMS Microbiol. Ecol. 79:785–99.
53. Petersen S. 1991. Acetate oxidation in a thermophilic anaerobic sewage-sludge digestor: the
importance of non-aceticlastic methanogenesis from acetate. FEMS Microbiol. Lett. 86:149–
152.
54. De Vrieze J, Hennebel T, Boon N, Verstraete W. 2012. Methanosarcina: the rediscovered
methanogen for heavy duty biomethanation. Bioresour. Technol. 112:1–9.
55. Mikucki JA, Liu Y, Delwiche M, Colwell FS, Boone DR. 2003. Isolation of a Methanogen
from Deep Marine Sediments That Contain Methane Hydrates, and Description of
Methanoculleus submarinus sp. nov. Appl. Environ. Microbiol. 69:3311–3316.
81
Chapter 6: Paper III- Effect of exogenous hydrogen addition on process
performance, methanogenesis and homo-acetogenesis pathways during an in situ
biogas upgrading
In preparation for peer-reviewed journal.
82
Effect of exogenous hydrogen addition on process performance, methanogenesis and homo-
acetogenesis pathways during an in situ biogas upgrading under thermophilic anaerobic
digestion
Daniel Girma Mulat
1, Freya Mosbæk
2, Alastair James Ward
1, Daniela Polag
3, Markus Greule
3, Frank
Keppler3, Jeppe Lund Nielsen
2, Anders Feilberg
1#
1Department of Engineering, Aarhus University, Hangøvej 2, 8200 Aarhus N, Denmark
2Department of Chemistry and Bioscience, Aalborg University, Fredrik Bajers Vej 7H, 9220
Aalborg East, Denmark
3Institute of Earth Sciences, University of Heidelberg, Im Neuenheimer Feld 234-236, 69120
Heidelberg, Germany
#Corresponding author: Anders Feilberg, [email protected]
ABSTRACT
In this study the possibility of exogenous H2 gas addition for the in situ biogas upgrading (increase
the biogas CH4 content) via conversion of the CO2 in the biogas into methane was investigated in
batch incubation fed with maize leaf as a substrate. Inoculum for incubating the H2 reactors and
control reactors (without H2 addition) was obtained from a full-scale biogas plant. Since H2 is a key
intermediate that regulates several steps of anaerobic digestion processes, the aim of this study was
to study the effects of H2 concentration under steady state condition (without H2 addition) and high
partial pressure (with H2 addition) on the process performance, methanogenesis and homo-
acetogenesis. The results showed that all the added H2 was almost completely utilized and the
methane content of the biogas reached up to 90% with concomitant decrease in CO2 content.
Volatile fatty acid (VFA) degradation rate decreased in the H2 reactors during the first few days of
the incubation and finally accumulated to a large amount whereas it was consumed in about 10 days
in the control reactors. After flushing the headspace of the H2 reactors with helium, H2
concentration was reduced in the headspace and the accumulated acetate was consumed completely.
Despite the dominating role of hydrogenotrophic methanogenesis (HM) in converting CO2 and H2
into methane in the H2 reactors, the observed carbon isotope fractionation between CO2 and CH4
was lower in the H2 reactors than the control reactors. Since HM leads to higher isotope
fractionation under a low H2 concentration environment, the observed lower isotope fractionation is
possibly explained by the differential reversibility concept, indicating exogenous H2 addition may
have led to high H2 concentration within micro-aggregates of methanogens. In such high H2
concentration environments, the H2-consuming bacteria with low H2 affinity, such as the homo-
acetogens, might be stimulated. This underlines that H2 addition rate (or gas retention time) should
be adjusted to match the H2-uptake rate and growth of hydrogenotrophic methanogens in order to
avoid VFA accumulation.
Key words: biogas upgrading, methanogenesis, homo-acetogenesis, stable isotope
83
INTRODUCTION
Methane is produced by a complex community consisting of hydrolytic, fermentative, acetogenic
and methanogenic microorganisms, with CH4 production by methanogenic archaea being the
terminal process(Conrad, 2005). Volatile fatty acids (VFA) such as acetate, propionate and butyrate
are key intermediates in anaerobic digestion and the accumulation of the VFA is usually associated
with digester instability (Ahring et al., 1995; Scholten & Conrad, 2000). The oxidation of VFA
other than acetate to H2 (or formate), acetate and CO2 is endergonic under standard conditions
(Equations 1 and 2; Table 1) and is thermodynamically feasible only when the H2 partial pressure
(or formate concentration) is kept low with H2-consuming methanogens according to equation 4
(Table 1) (Dolfing et al., 2008; Schmidt & Ahring, 1993). Thermodynamic estimation, for example,
predicted that the H2 partial pressure as low as 10 Pa and 100 Pa are necessary for oxidation of
propionate and butyrate, respectively. Such low H2 partial pressure is achieved by syntrophic
transfer of H2 from H2-producing bacteria to H2-consuming methanogens (Schmidt & Ahring,
1993).
Table 1 Standard Gibbs free-energy changes (∆G´) at 25 and 55 °C for reactions involved in
anaerobic digestion at pH 7a
Equation Reaction
∆G°´
(kJ/mol)
∆G´55
(kJ/mol)
1 CH3CH2CH2COO- + 2H2O → 2CH3COO
- + 2H2 + H
+ 48.1 37.9
2 CH3CH2COO- + 3H2O → CH3COO
- + HCO3
- + 3H2 + H
+ 76.1 62.3
3 CH3COO- + H2O → CH4 + HCO3
- -31.0 -34.7
4 4H2 + HCO3- + H
+ → CH4 + 3H2O -135.6 -122.5
4HCOO- + H2O + H
+ → CH4 + 3HCO3
- -130.0 -118.9
5 HCO3- + 2H2 + 0.5 H
+ → 0.5CH3COO
- + 2H2O -55
a Data obtained from literatures (Schmidt & Ahring, 1993; Thiele & Zeikus, 1988)
The importance of H2 in methanogenesis has been already well studied in several natural
environments. Methane production was stimulated in marine sediments when incubated with 70%
H2 (Oremland, 1975). Another study demonstrated that the addition of H2 to lake sediment
stimulated the reduction of CO2 to methane. Methane production increased significantly when more
H2 was supplemented into the sediment incubation (Winfrey et al., 1977). Recently exogenous
addition of H2 into biogas digester for an in situ (Luo & Angelidaki, 2013; Luo et al., 2012) or ex
situ biogas upgrading to higher CH4 content (Luo & Angelidaki, 2012) has been demonstrated as a
good strategy to store surplus electricity from renewable energy sources in the form of storable
energy form (i.e. CH4). In situ represents the conversion of the CO2 in the biogas to methane in
existing digesters working with other substrates whereas ex situ refers to biogas upgrading in a
separate digester. H2 can be produced via electrolysis of water using surplus wind energy or other
renewable energy source. Whereas storage of H2 is problematic due to safety issues and the
requirement of new infrastructures, upgraded methane can be easily stored in existing natural gas
grids (Luo et al., 2012).
84
Addition of H2 for the conversion of CO2 into methane is carried out by hydrogenotrophic
methanogens via hydrogenotrophic methanogenesis (HM) and this reaction is thermodynamically
feasible according to Equation 4 (Table 1). However, the addition of H2 might rapidly increase H2
partial pressure in the reactors to the level that thermodynamically inhibits VFA oxidation as
discussed above. In addition to HM, homo-acetogens might be stimulated and lead to the production
of acetate from CO2 and H2 through the homo-acetogenesis pathway. A better understanding of the
influence of H2 concentration on process performance, methanogenesis and homo-acetogenesis is of
paramount significance to implement the necessary strategy needed for optimizing in situ biogas
upgrading through the addition of H2.
In this study the possibility of in situ biogas upgrading to higher CH4 content with H2 addition was
investigated in batch incubation fed with unlabeled and U-13
C labeled maize leaf as a substrate and
with the inoculum obtained from full-scale biogas plant under thermophilic condition. Control
reactors without H2 addition were also incubated in parallel with the H2 pressurized reactors.
Moreover, it was aimed to investigate the effects of H2 addition on overall process performance,
methanogenesis and homo-acetogenesis. The stable carbon isotope signatures of CO2 and CH4 at
natural abundance were measured to identify methanogenic pathways. The degradation of VFA,
biogas composition and methane production rate were monitored to evaluate process performance.
U-13
C labeled maize leaf was used to trace the 13
C incorporation into CO2 and CH4 over the course
of incubation by stable isotope analysis and to identify the key microorganisms involved in the
process of anaerobic digestion of U-13
C labeled maize leaf to methane by protein-SIP and DNA-
SIP.
MATERIALS AND METHOD
Sources of inoculum
Inoculum was obtained from a commercial full-scale biogas digester at research centre Foulum,
Denmark. The digester works with a mixture of pig and cattle manure, maize silage and deep litter
manure. It runs under thermophilic condition ca. 52°C. The total solid (TS), volatile solid (VS), pH
value and total ammonia nitrogen (TAN) of the inoculum were 69.8 g/L, 57.8 g/L, 7.59 and 1.46
g/L, respectively. The inoculum was pre-incubated under anaerobic condition at 52°C for two
weeks prior to the main experiment to reduce the residual biodegradable organic material.
Main batch incubation reactor setup
Inoculum (75 g) was transferred into 100-mL serum bottles followed by the addition of the
following substrates and then sealed with butyl rubber stopper and aluminum crimp (Table 2): i)
Reactors fed with 13
C labeled maize and with H2 addition (LMH); 1.33 g (on a wet mass basis) of
U-13
C labeled maize leaf (97 > % 13
C; VS 42.12% of a wet mass basis; ground and sieved < 1.0
mm) was added once at the beginning and the reactors was pressurized with H2 periodically.. The U-13
C-labeled maize leaf was obtained from Isolife (Wageningen, The Netherlands). The amount of
H2 supplied into the reactors headspace differs from time to time depending on the expected
stoichiometric amount of the produced CO2 (stoichiometric mol ratio of H2:CO2 is 4:1). The
volumes of the pure H2 supplied were 44, 36, 52, 52, 72, 80, 80, 80, 40 and 100 mL on days 0, 1, 2,
3, 4, 6, 8, 10, 12 and 15, respectively. (ii) Reactors fed with unlabeled maize and with H2 addition
(UMH); 1.33 g (on a wet mass basis) of unlabeled maize leaf (VS was 42.33% of a wet mass basis;
ground and sieved < 1.0 mm) was added once at the beginning and the same amount of H2 was
supplied into the headspace as the 13
C labeled maize reactors. (iii) Control reactors fed with 13
C
labeled maize and without H2 addition (LM); 1.33 g (on a wet mass basis) of U-13
C-labeled maize
85
leaf was added once at the beginning and without H2 addition , (iv) Control reactors fed with
unlabeled maize and without H2 addition (UM); 1.33 g (on a wet mass basis) of unlabeled maize
leaf was added once at the beginning and without H2addition, and (v) Blank reactors (BR); instead
of maize substrate, deionized water (1.33 mL) was added once at the beginning. The reactors were
incubated in triplicate in a shaking incubator at 100 rpm for 24 days under thermophilic condition
(Ca. 52°C).
Table 2 Reactor setup for the main batch incubation experiment
Substrate composition
Reactors
description
Reactors
name
Inoculum
(g) H2
U-13
C
labeled
maize (g)
Unlabeled
maize (g)
water
(mL)
H2 reactors LMH 75 Periodically
pressurized
1.33
UMH 75 Periodically
pressurized
1.33
Control
reactors
LM 75 1.33
UM 75 1.33
Blank
reactors
BR 75 1.33
Supplementary incubation to investigate the effect of H2 concentration on acetate degradation
Since acetate accumulated in the H2 reactors from day 10 of the incubation, digestate was collected
from the LMH on day 13 and transferred into 6 glass serum bottles (12-mL) followed by the
addition of the following substrates and finally sealed with butyl rubber stoppers and screw caps
(Table 3): i) Acetoclastic methanogenesis (AM) uninhibited reactors, LMH-U; 1 mL of
concentrated unlabeled sodium bicarbonate solution was added to give 20 mmol/L total bicarbonate
solution and the headspace was flushed with H2-CO2 mixture (80-20 V/V) for 5 min, (iii) AM
inhibited reactors, LMH-I; 1 mL of concentrated sodium bicarbonate solution was added to give 20
mmol/L total bicarbonate solution and the headspace was flushed with H2-CO2 mixture (80-20 V/V)
for 5 min and finally fluoromethane (1.3%) as specific inhibitor to AM was added into the
headspace (iii) 1 mL water was added instead of substrate and the headspace was flushed with
helium. The reactors were flushed with H2-CO2 mixture and helium periodically. Each treatment
was run in duplicate and incubated in a shaking incubator at 100 rpm for 7 days under thermophilic
condition (Ca. 52°C).
86
Table 3 Reactor setup for the supplementary incubation experiment to investigate the effect of H2
concentration on acetate degradation
Substrate composition
Reactors
description
Reactors
name
Inoculum
(g)
H2-CO2
mixture (80-
20 V/V) Helium
CH3F
(%)
Concentrated
NaHCO3a
Water
(mL)
AM
uninhibited
LMH-U 4 Periodically
flushed for 5
min
Added once
AM
inhibited
LMH-I 4 Periodically
flushed for 5
min
1.3 Added once
Blank
reactors
LMH-B 4 Periodically
flushed for
5 min
1
a1 mL of concentrated unlabeled sodium bicarbonate was added once at the beginning of the
incubation to give a total concentration of 20 mM
Basic analytical methods
The volume of a produced biogas was measured using an acidified water displacement method at
room temperature and atmospheric pressure. Headspace biogas was collected using a gas tight
syringe with a needle through a septum and transferred into a 12 mL evacuated vial. The
composition of biogas (CH4, CO2 and H2) was analyzed using micro-gas chromatograph (µ-GC,
Agilent 3000) equipped with a thermal conductivity detector (TCD) as described before (Ward et
al., 2011). In brief, the µ-GC was equipped with two parallel GC columns containing different
coatings (MolSieve 5Å PLOT, 10 m x 0.32 mm x 12 µm and PLOT Q, 10 m x 0.32 mm x10 µm)
for the separation of the components of biogas. Argon was used as a carrier gas with a column
pressure of 80 psi.
pH was measured immediately after sampling of a solution and adjusted with 5M HCl to maintain a
pH in the range of 7.0 -7.8. Liquid samples for volatile fatty acid (VFA) analysis were collected
periodically. The liquid samples (1.000 g) were first acidified with 4 mL of acidifying solution (pH
< 2; containing mixtures of 0.3 M oxalic acid and 2.1 mM dimethylpropanoic acid as internal
standard). Then it was centrifuged at 14,000 g for 12 min at room temperature and clear solution
was filtered through a 0.45m Acrodisc® syringe filter (Sigma-Aldrich, Denmark) and finally the
supernatant solution was transferred into analytical vials. Aliquot (1.0 µL) of the solution was
injected into the GC system using an autosampler. The concentrations of volatile fatty acids were
determined by Agilent Technologies 7890A gas chromatograph equipped with flame ionization
detector (FID). A polar phase capillary column, HP-INNOWax (30m x 0.25 mm x 0.25 μm), was
used for separation. Helium was used as a carrier gas at 1.8 ml/min flow rate. The analyses were
performed using a temperature programme: 5 min at 100°C, a linear gradient from 100-120°C at the
rate of 10 °C/min, 5 min at 120°C, a linear gradient from 120°C to the final temperature of 220°C at
87
the rate of 30°C/min and final hold at
220°C for 3 min. The temperatures of injector and detector
were set at 285°C and 300°C, respectively.
TS and VS of maize leaf and inoculum were analyzed according to the standard methods (APHA,
2005).
Protein-SIP and DNA-SIP analysis
Liquid samples for protein-SIP and DNA-SIP analysis were collected from all reactors periodically
and stored at -20°C until analysis. The samples were immediately centrifuged at 14,000 g for 30
min at room temperature. The solid fraction and supernatant solution were separated and stored in
freezer (-20 °C) until further analysis.
Stable isotope analysis
Biogas samples were periodically collected from the headspace of the reactors and stored in a gas
tight evacuated vial (12 mL) until further analysis. The stable carbon isotope of gas samples was
analyzed using a gas chromatography combustion isotope ratio mass spectrometery (GC-C-IRMS)
system consisting of a gas chromatograph (HP 6890 Series, Agilent Technology, Santa Clara, USA)
coupled with IRMS (Finnigan MAT 253, Thermofinnigan, Bremen, Germany) via a combustion
interface. For the GC separation of CH4 and CO2, gas samples (30 µL) were injected with
autosampler into the GC instrument equipped with a CP-Porabond Q column
(50 m × 0.32 mm × 0.5 μm, Varian, USA). The column temperature was kept constant at 40 ºC and
flow rate of Helium was 2 mL/min.
The stable carbon isotope data was reported in delta notation (δ13
C) in parts per thousand (‰) unit
versus the Vienna Pee Dee Belemnite (V-PDB):
δ13
C = [(Ra)sample/(Ra)standard - 1]*103 (‰)
where Ra is the 13
C/12
C ratio (Whiticar, 1999).
The apparent fractionation factor (αmc) between CO2 and CH4 was calculated according to the
following equation:
αmc = (δ13
C-CO2 + 1000)/(δ13
C-CH4 + 1000)
RESULTS AND DISCUSSION
Reactor performance and H2 consumption rate
13C labeled and unlabeled maize leaf were added once at the beginning of the main batch
experiment and incubated with inoculum obtained from a full-scale biogas plant under thermophilic
condition. The 13
C labeled (LMH) and unlabeled maize leaf reactors (UMH) were periodically
pressurized with H2 to investigate the conversion of the CO2 in the biogas and the added H2 into
methane through HM. Control reactors that received only 13
C labeled maize leaf (LM) and
unlabeled maize leaf (UM) were incubated with the same inoculum but without H2 addition. Blank
reactors (BR) which received only inoculum were incubated as well.
Accumulated and daily methane production as well as VFA degradation is shown in Figure 1. The
methane production in all the maize fed reactors was characterized with a typical batch incubation
with an immediate methane production followed by exponential phase until day 13 and a stationary
88
phase afterwards (Figure 1A). The cumulative methane production in BR was relatively much lower
than those fed with maize. The daily methane production profile showed that methane production
increased until day 6 and after that decreased before finally stopping almost completely on day 15
(Figure 2B). The rapid onset of hydrolytic and acidogenic activity led to the transient accumulation
of VFA, reaching about 2300-2800 mg/L of acetate and about 200 mg/L of propionate on day 1 in
all maize fed reactors (Figure 2). The acetate concentration in LMH and UMH are slightly higher
than their corresponding controls (LM and UM) on day 1. The acetate concentration started to
decrease slightly until day 3 in all reactors with and without H2 addition and later on decreased
significantly in all reactors, indicating the activity of acetate-consuming microorganisms in this
period. On day 6, the concentration of acetate reached very low levels in UM and LM whereas it
remained accumulated in UMH and LMH at about the concentration of 1300 mg/L from day 10
until end of the incubation (day 24). Propionate degradation in the LM and UM started when acetate
reached very low concentrations on day 6 whereas it remained accumulated in the LMH and UMH
from day 6 until the end of incubation. Acetate and propionate concentration gradually increased in
the BR from day 6 until day 24, probably due to the degradation of a residual biodegradable organic
material in the inoculum.
Figure 1 Temporal change of (A) accumulated methane production; (B) daily methane production;
(C) acetate degradation; and (D) propionate degradation. UMH (▲); LMH(■ ); LM (□); UM (△);
and BR (○).
0
50
100
150
200
250
300
350
0 4 8 12 16 20 24
Acc
um
ula
ted
CH
4(m
L)
Incubation time (days)
0
10
20
30
40
50
60
70
0 4 8 12 16 20 24
CH
4(m
L)
Incubation time (days)
(A) (B)
0
500
1000
1500
2000
2500
3000
0 4 8 12 16 20 24
Ace
tate
(m
g/L
)
Incubation time (days)
(C)
0
200
400
600
800
1000
1200
0 4 8 12 16 20 24
Pro
pio
na
te (
mg/L
)
Incubation time (days)
(D)
89
The production of methane and biogas composition (CH4, CO2 and H2) was monitored over the 24
incubation days (Figures 1 and 2). Different amounts of H2 were added periodically into LMH and
UMH depending on the expected CO2 production rate in stoichiometric proportion (stoichiometric
mol ratio of H2:CO2 is 4:1). Accumulated methane until the first 8 days was very similar in LMH,
UMH LM and UM and then started to differ until the end of incubation with more methane
produced in the H2 eactors than the controls (Figure1). During the first 10 days except day 4, all the
added H2 (Figure 2) was almost completely utilized in the LMH and UMH with concomitant
increase of CH4 content in the biogas (Figure 2). After day 10, H2 consumption reduced
significantly and accumulated to a large extent until the end of the experiment. The biogas CH4
content in the MH and LMH increased from 40% to peak value of 90% in six days and remained
stable at 90% until day 10. After day 10, it decreased sharply to almost 4% on day 15. During the
first 10 days, a decrease in CO2 content was accompanied with an increase in CH4 content while the
significant depletion in CO2 content (less than 1% of CO2 was detected) after day 10 was correlated
with a sharp decrease in CH4 content. The accumulation of H2, the depletion of CO2 and the sharp
decrease in CH4 content in the LMH and UMH after day 10 indicates that CH4 production was
influenced by the unavailability of CO2 in the system at this incubation period. This also showed
that CO2 was converted to CH4 with electrons from H2. It also demonstrated that after day 10,
excess amount of H2 was supplied in compared to the stoichiometric amount of CO2 that was
produced.
Figure 2 Temporal change of (A) CH4 content; (B) CO2 content; and (C) H2 content in the biogas.
UMH (▲); LMH(■ ); LM (□); UM (△); and BR (○).
0
20
40
60
80
100
0 4 8 12 16 20 24
CH
4(%
)
Incubation time (days)
0
20
40
60
80
100
0 4 8 12 16 20 24
CO
2(%
)
Incubation time (days)
0
20
40
60
80
100
0 4 8 12 16 20 24
H2
(%)
Incubation time (days)
(A) (B)
(C)
90
H2 concentration in the control and blank reactors were generally below 0.1% except in few
instances it reached 1.4%. CO2 content in the control and blank reactors decreased from 60% to
26% in the first few days and later on slightly increased to maximum values of 92 and 58% in the
blank and control reactors, respectively. CH4 content increased from about 40% to 61% to the peak
values of 61 and 73% in the blank and control reactors, respectively and later on decreased
gradually.
Carbon isotope fractionation and methanogenic pathway
The stable carbon isotope (δ13
C) of the produced biogas (CH4 and CO2) was monitored for 24 days
(Figure 3). The δ13
C-CH4 and δ13
C-CO2 values in the UMH after day 10 and in the UM and BR
after day 15 were not reported since the measurement was influenced by the low concentration of
CH4 and CO2 injected into the IRMS system. The δ13
C-CH4 enriched over time in all unlabeled
reactors during the course of the incubation. The enrichment was highest in the UMH and lowest in
the BR. Enriched δ13
C-CH4 values are characteristics to an increase in the contribution of AM to the
degradation of acetate into methane (Conrad, 2005). The δ13
C-CO2 in the UMH and UM increased
slightly during the first few days and stabilized afterwards. The δ13
C-CO2 in the BR decreased
during the first 6 days and stabilized at the same value as the UMH and UM.
Figure 3 Temporal profile of the stable carbon isotope (δ13
C) of the produced (A) CH4; and (B)
CO2; as well as (C) the fractionation factor, αmc, between CO2 and CH4. UMH (▲); LMH(■ ); LM
(□); UM (△); and BR (○).
1.03
1.04
1.05
1.06
1.07
1.08
1.09
1.10
1.11
0 4 8 12 16 20 24
αm
c
Incubation time (days)
1000
11000
21000
31000
41000
51000
61000
71000
81000
-80
-70
-60
-50
-40
-30
-20
-10
0 4 8 12 16 20 24
δ1
3C
-CH
4 (‰
) in
LM
H
& L
M
δ1
3C
-CH
4 (‰
) in
UM
,
UM
H &
BR
Incubation time (days)
0
5000
10000
15000
20000
25000
30000
35000
40000
45000
5
10
15
20
25
30
35
0 4 8 12 16 20 24
δ1
3C
-CO
2(‰
) in
LM
H
& L
M
δ1
3C
-CO
2(‰
) in
UM
,
UM
H &
BR
Incubation time (days)
(A) (B)
(C)
91
Highly enriched δ13
C-CH4 and δ13
C-CO2 in the reactors fed with U-13
C labeled maize (UMH and
UM) compared to the reactors fed with unlabeled maize (UMH and UM) showed that a significant
amount of the carbon flows to the production of CO2 and CH4 during the anaerobic digestion of
maize as a substrate (Figure 3). The δ13
C-CH4 in the LMH and LM enriched during the first 6 days
of incubation and reached the most enriched value on day 6 and later on depleted until the end of the
experiment. This is in line with the observed daily methane production profile where methane
production increased within 6 days and reached the peak value on day 6 and later on dropped. The
δ13
C-CH4 on days 6 and 10 in the LMH was higher than LM, which corresponds with the observed
higher methane production in the LMH between days 10 and 13 compared to the LM. The δ13
C-
CO2 in the LM increased until day 6, indicating the production of 13
CO2 during the several steps of
the anaerobic digestion of U-13
C labeled maize within 6 days. The δ13
C-CO2 in the LM after day 6
decreased significantly, which corresponds with hydrolysis, acidogenesis, minimal conversion of 13
C labeled acetate to 13
CO2, and active hydrogenotrophic methanogens after day 6. The δ13
C-CO2
in the LMH after day 1 significantly reduced, indicating either i) an immediate consumption of the
produced 13
CO2 with H2 addition, ii) reduced production rate of 13
CO2 during the digestion of 13
C
labeled maize or iii) reduced rate of conversion of 13
C labeled acetate to 13
CO2.
The apparent carbon fractionation factor (αmc) between CO2 and CH4 is commonly employed to
identify which methanogenic pathway dominates the methane production. According to literature
(Conrad, 2005; Qu et al., 2009), αmc > 1.065 and αmc < 1.055 indicate the predominance of HM and
AM pathways, respectively, while the αmc value between 1.055 and 1.065 are characteristics to the
presence of both methanogenesis pathways. The fractionation factors in the BR throughout the
whole incubation period were always higher than 1.065 (Figure 3), indicating most of the methane
production was contributed by HM. In the UM, methane production started with the dominance of
HM and later on both pathways contributed to methane production. This trend is in line with the
acetate degradation rate where acetate accumulated in first few days of the incubation and
consumed afterwards. The αmc in the UMH for the first 3 days was the same as the UM and
significantly decreased in the former afterwards. The observed αmc < 1.055 in the UMH after day 3
cannot be represented as the dominance of AM since acetate was accumulated and methane
production was mainly derived from CO2 reduction during this period. Instead it is explained
according to a recently introduced new concept known as “differential reversibility”.
A ΔG-dependent differential reversibility effect has been proposed to explain the extent of variation
of αmc under different H2 concentration environment (Hattori et al., 2012). According to this
concept, the extent of reversibility in multistep enzymatic steps of the H2/CO2 pathway is expected
to impact the extent at which fractionations are expressed from each enzymatic step. The H2
concentration is proposed as a primary controlling factor of the ΔG values and hence affect the
value of αmc (Valentine et al., 2004). This assumption is well supported in microbial communities
obtained from groundwater in a deep aquifer. The results showed that under high H2 concentration,
αmc was lower compared to the one at low H2 concentration, because of reduced reversibility in the
multiple enzymatic processes in CH4 production under high H2 concentration (Hattori et al., 2012).
In our study, the observed αmc < 1.055 and the dominance of methane production from the reduction
of CO2 in the UMH after day 3 is possibly explained by the differential reversibility concept,
indicating exogenous H2 addition may have led to high H2 concentration within micro-aggregates of
methanogens. Since the measurement of H2 in the bulk liquid or in the headspace of a reactor does
not necessarily represent the energetically relevant H2 concentration within the micro-aggregates of
methanogens environment, the αmc value can be used to indirectly indicate whether H2 addition lead
to higher H2 concentration in the micro-environment or not (Penning et al., 2005).
92
Acetate and propionate degradation under different H2 concentration
Since acetate was accumulated from day 10 until day 24 in the H2 reactors, separate batch
incubation was conducted to test the effect of H2 partial pressure on acetate degradation. Digestate
was collected on day 13 from the LMH and used as inoculum to incubate additional reactors (LMH-
U, LMH-I and LMH-B). Both LMH-U and LMH-I were flushed with H2-CO2 mixture (80/20 V/V)
for 5 min while 1.3% of CH3F was added into the latter to inhibit acetoclastic methanogenesis (AM)
and the former did not receive CH3F. Both LMH-U and LMH-I were also supplemented with
bicarbonate whereas the blank reactors (LMH-B) did not receive any substrate. The LMH-U and
LMH-I simulated high H2 partial pressure and presence of stoichiometric amount of H2 and CO2
where as LMH-B simulated the low H2 partial pressure (or steady state H2 concentration).
The VFA degradation and carbon isotope signatures of CH4 and CO2 profiles are shown in Figure 4.
Acetate concentration reduced slightly during the first five days of the incubation in the AM
uninhibited reactors (LMH-U and LMH-B) and later on dropped considerably in both reactors with
higher degradation rate in the latter. The difference in acetate degradation in the H2 reactor (LMH-
U) and the blank reactors (LMH-B) is possibly due to H2 partial pressure needing to be reduced to a
low level before acetate degradation started in the former. Since acetate was consumed in both
LMH-U and LMH-B, the depletion of bicarbonate in the H2 reactor did not directly influence the
degradation of acetate. The observed higher acetate degradation rate in the LMH-B in 7 days but
not in the reactors pressurized with H2 (LMH) during the 14 days of incubation is a strong evidence
for the inhibition of acetate degradation under high H2 partial pressure and the inhibition was
reversible when H2 partial pressure was kept low. Acetate concentration hardly changed in the AM
inhibited reactors (LMH-I) over 7 incubation days. This indicates that syntrophic acetate oxidation
(SAO) hardly participated in the degradation of acetate in the H2 reactors. In all the reactors, the
concentration of propionate changed only slightly over 7 incubation days.
93
Figure 4 Supplementary incubation experiment showing the temporal variation in (A) acetate
degradation; (B) propionate degradation; (C) δ13
C-CH4 ; and (D) δ13
C-CO2. LMH-U (●); LMH-I
(■); LMH-B (▲).
Stable isotope signatures of CO2 and CH4 in the LMH-U, LMH-I and LMH-B were monitored to
investigate methanogenic pathway. The isotope measurement of the LMH-I sample on day 7 was
not available. The slight change in δ13
CH4 during the first 5 days correlates with the observed slow
acetate degradation rate within this time period. The δ13
CH4 increased significantly in the LMH-B
from day 5 to day 7, indicating high conversion rate of the 13
C labeled acetate to 13
C labeled
methane (13
CH4). This is in line with the observed high acetate degradation rate as described above.
Despite significant amount of acetate degradation in the LMH-U from day 5 to 7, the δ13
CH4 hardly
changed during these days. The carbon dioxide pool in the LMH-UI and LMH-B were enriched
with 13
C which comes from the inoculum but it was highly diluted with the unlabeled CO2 in the
former due to the addition of unlabeled bicarbonate and unlabeled CO2. This was evident from the
δ13
CO2 values in the LMH-B and LMH-U where it was always higher in the former than the latter.
Therefore, the reason for the observed constant δ13
CH4 in the LMH-UI is possibly due to the
methane produced from the degradation of 13
C labeled acetate and the reduction of the unlabeled
pool of CO2 was almost equal. The increase in δ13
CO2 in all the reactors during the first 5 days may
be influenced by the distribution of CO2 between the headspace and liquid phase as well as the slow
degradation rate of 13
C labeled acetate observed in this period. The enrichment in δ13
CO2 after day
5 in the LMH-B indicates the 13
CO2 was produced during the degradation of 13
C labeled acetate
either through AM or SAO.
0
200
400
600
800
1000
1200
1400
0 1 2 3 4 5 6 7
Ace
tate
(mg
/L)
Incubation time (days)
400
500
600
700
800
900
1000
0 1 2 3 4 5 6 7
Pro
pio
nate
(mg/L
)
Incubation time (days)
9900
11900
13900
15900
17900
19900
0 1 2 3 4 5 6 7
δ1
3C
-CH
4 (‰
)
Incubation time (days)
1900
3900
5900
7900
9900
11900
13900
0 1 2 3 4 5 6 7
δ1
3C
-CO
2(‰
)
Incubation time (days)
(A) (B)
(C) (D)
94
Effects of H2 partial pressure on acetate degradation, methanogenesis and homo-acetogenesis
In this study the possibility of exogenous H2 gas addition for the in situ biogas upgrading to higher
CH4 content was investigated in batch incubation fed with maize leaf as a substrate. Since H2 is a
key intermediate that regulates several steps of anaerobic digestion processes, the effects of H2
concentration under steady state condition (without H2 addition) and high partial pressure (with H2
addition) on the process performance, methanogenesis, homo-acetogenesis and microbial
community structure were studied. The results of the present study showed that methane was
produced by the reduction of CO2 with the added H2, reaching a CH4 content of 90%. This result is
in agreement with previous studies that demonstrated exogenous addition of H2 led to the reduction
of CO2 to methane (Luo et al., 2012). In a lab scale continuous stirred tank reactor (CSTR) fed with
manure, H2 addition led to a reduction in CO2 content from 38% to 15% and a corresponding
increase in CH4 content as well an increase in methane production rate by 22% (Luo et al., 2012).
Another important aspect in this study was the influence of CO2 reduction on the pH of the H2
reactors. Our results showed that pH was between 7.0 and 8.0 without H2 addition and increased to
higher than 8.0 with H2 addition (data not shown). Previous study showed that the addition of H2
resulted in increase of pH (from 8.0 to 8.3) as a result of the consumption of bicarbonate, which
subsequently caused slight inhibition of methanogenesis (Luo et al., 2012). In our experiment, the
pH was corrected by adding buffer solution when it became higher than 8.0 in the H2 reactors.
Recently it was demonstrated that the co-digestion of manure with acidic whey helped to keep the
pH below 8.0 in a H2 reactor without inhibiting the process (Luo & Angelidaki, 2013). Therefore,
pH should be monitored and regulated continuously to optimum range for stably operating an in situ
biogas upgrading process.
An unanticipated disadvantage of H2 addition during an in situ biogas upgrading is that H2
concentration may rapidly increase in the micro-aggregate of the methanogens environment to the
level that thermodynamically inhibits VFA oxidation. The oxidation of VFA is endergonic under
standard conditions and is thermodynamically feasible only when the H2 partial pressure is kept
very low. The partial pressure of H2 should be maintained as low as 10 Pa and 100 Pa to allow the
oxidation of propionate and butyrate, respectively (Dolfing et al., 2008; Schmidt & Ahring, 1993).
In our study, the degradation of acetate and propionate was reduced with the H2 addition in
compared to without H2 addition. We even observed a complete inhibition of acetate and propionate
degradation at the late stage of the incubation experiment. The observed lower isotope fractionation
factor (αmc) in the H2 reactors also indicated that the H2 concentration in the micro-aggregate
methanogens environment possibly increased due to the supply of H2. Since CO2 significantly
depleted and H2 accumulated in the system at this stage, further supplementary experiment was
conducted to investigate whether CO2 depletion or high H2 concentration inhibited the process. The
results demonstrated that acetate degradation was inhibited under high H2 concentration and the
inhibitory effect was reversible when the H2 concentration was kept very low. CO2 depletion was
not inhibitory as bicarbonate addition in the system did not improve VFA degradation compared to
the reactors without bicarbonate addition. Propionate was slightly degraded within 7 days of
incubation, probably due to propionate degraders being the slowest growing among the known
acetogenic bacteria (Nielsen et al., 2007). Acetate remained unconsumed in an up-flow anaerobic
fixed bed reactor fed continuously with mixture of H2-CO2 and supplemented with acetate at a final
concentration of 200 mg/L (Lee et al., 2012), which agrees with our observation.
95
Addition of H2 could change the microbial structure, promoting the growth of H2-consuming
microbes. Terminal restriction fragment length polymorphism (T-RFLP) analysis of 16S rRNA
genes was used in a previous study (Leybo et al., 2006) to monitor the changes in the composition
of methanogenic community in enrichment cultures under high and low H2 concentrations. H2
concentration influenced the methanogenic community with the dominance of the mixotrophic
Methanosarcina under high H2 concentration and more diverse community in the low H2
concentration culture. Moreover, members of Methanosarcina could alter their metabolic pathway
from acetoclastic to hydrogenotrophic under specific conditions (Qu et al., 2009), that could also be
an alternative reason for the dominance of Methanosarcina at such high H2 concentration. In our H2
reactors, members of Methanosarcina might dominate the methanogenic community and change
their metabolic activity to dominating role of HM. Growth of pure cultures of Methanosarcina sp.
strain 227 and Methanosarcina mazei on mixtures of H2-CO2 and acetate showed that CO2 was
rapidly reduced to CH4 in the presence of H2 with low acetate consumption rate until H2 was
exhausted (Ferguson & Mah, 1983). Our observation of faster acetate degradation rate in reactors
without H2 addition (LMH-B) compared to those with addition of mixtures of H2-CO2 (LMH-U) is
in accord with the findings of the previous study (Ferguson & Mah, 1983). Another study with pure
culture of Methanosarcina thermophila TM-1 showed that acetate consumption was inhibited with
even slight increase in the concentration of H2. The H2 partial pressure showed to give initial, 50%,
and 100% inhibition of M. thermophila was 250, 1010, and 5070 Pa, respectively (Ahring et al.,
1991).
The higher acetate concentration in the H2 reactors could also be the result of inhibition of SAO
pathway under the high H2 partial pressure. In our study, it was shown that SAO played less
importance in degradation of acetate in the AM inhibited reactors under high H2 partial pressure.
Moreover, the possible participation of homo-acetogens in the production of acetate from CO2 and
H2 may explain the higher acetate concentration in the H2 reactors. We have developed a rapid and
accurate GC/MS method for determining the isotope enrichment of underivatized acetate and
concentration of underivatized VFA in liquid samples from reactors with and without H2 addition.
The isotope enrichment data and high concentration of acetate in the H2 reactors showed that under
high H2 partial pressure, carbon dioxide was reduced to acetate via homo-acetogenesis pathway
(Mulat & Feilberg, 2005). Another study showed that H2 addition greatly stimulated acetate
production through homo-acetogenesis in an anaerobic mesophilic digester operating with cattle
manure as substrate (Boone, 1982), which is in agreement with our findings. Under low H2
concentration, the H2-consuming bacteria with low H2 affinity like the homo-acetogenes could not
compete with those with high affinity such as the sulfate reducers and hydrogenotrophic
methanogens. For example, members of homo-acetogenic Acetobacterium have much higher
threshold concentration of H2 (52–95 Pa) (Cord-Ruwisch et al., 1988). Homo-acetogens might be
stimulated with the exogenous H2 addition (Hao et al., 2013). Nevertheless, further data is needed
for the identification of functionally active microbes. Since samples from the U-13
C labeled maize
leaf reactors with and without H2 addition have been under investigation by protein-SIP and DNA-
SIP, the molecular data that will be obtained later will increase our understanding of the identity of
the microorganisms involved in the specific metabolic processes.
96
Implementation of in situ biogas upgrading process
H2 addition should be adjusted to the production rate of CO2 in stoichiometric amount (H2:CO2 mol
ratio is 4:1) to upgrade biogas to higher methane content in situ. H2 addition should be also adjusted
to the capacity of the system to convert the CO2 in the biogas into methane considering the H2-
uptake rate and the growth of hydrogenotrophic methanogens. As demonstrated in the present and
previous study (Luo et al., 2012), H2 addition in the presence of stoichiometric amount of the
produced CO2 led to higher methane content up to 90% with slight reduction in the degradation of
acetate and other VFA. The results of the H2 reactors from day 10 indicates that VFA accumulated
in the system when the available CO2 was very low and excess amount of H2 was supplemented
into the system. This underlines the importance of regulating the influent H2 flow rate continuously
to avoid accumulation of VFA during in situ biogas upgrading. Low supply of H2 gas would lead to
higher CO2 content in the effluent biogas while excess H2 would lead to accumulation of VFA and
process instability that in turn reduces the overall methane yield. It is also important to regulate H2
influent rate and gas residence time for avoiding the conversion of H2 and CO2 into acetate by
homo-acetogens that could increase the concentration of VFA and may ultimately lead to process
instability.
CONCLUSION
A mixed culture obtained from a full-scale biogas plant was incubated under thermophilic condition
with maize leaf as a substrate and with exogenous H2 addition into the reactor headspace. The data
show that almost all the added H2 was consumed and methane content of the biogas reached up to
90% with concomitant decrease in CO2 content. However, the degradation of acetate and other VFA
reduced during the first few days of the incubation in the H2 reactors and accumulated afterwards
whereas it was consumed almost completely in 10 days in the control reactors without H2 addition.
Flushing the headspace of the H2 reactors with helium reduced the H2 concentrations in the
headspace and led to the degradation of acetate to almost completion. Despite the dominating role
of HM in converting CO2 and H2 into methane in the H2 reactors, the observed carbon isotope
fractionation between CO2 and CH4 was lower in the H2 reactors than the control reactors. Since
higher isotope fractionation under low H2 concentration is characteristic of the dominance of HM,
the observed lower isotope fractionation is possibly explained by the differential reversibility
concept, indicating exogenous H2 addition may have led to high H2 concentration within micro-
aggregates of methanogens. In such high H2 concentration environments, the H2-consuming bacteria
with low H2 affinity like the homo-acetogens might be stimulated and compete with
hydrogenotrophic methanogens for H2. This underlines the importance of regulating the H2 addition
rate (or gas retention time) to the H2-uptake rate and growth of hydrogenotrophic methanogens in
order to avoid VFA accumulation during an in situ biogas upgrading process.
ACKNOWLEDGEMENTS
This research was financially supported by the Danish Strategic Research Council (Grant 10-
093944). We would like to thank Sabine Lindholst for the great help with micro-GC instrument.
97
References
Ahring, B.K., Sandberg, M., Angelidaki, I. 1995. Volatile fatty acids as indicators of process
imbalance in anaerobic digestors. Applied Microbiology and Biotechnology, 43(3), 559-565.
Ahring, B.K., Westermann, P., Mah, R.A. 1991. Hydrogen inhibition of acetate metabolism and
kinetics of hydrogen consumption by Methanosarcina thermophila TM-1. Archives of
microbiology, 157(1), 38-42.
APHA. 2005. Standard methods for the examination of water and wastewater. in: American Public
Health Association, Vol. 21. Washington DC.
Boone, D.R. 1982. Terminal reactions in the anaerobic digestion of animal waste. Applied and
environmental microbiology, 43(1), 57-64.
Conrad, R. 2005. Quantification of methanogenic pathways using stable carbon isotopic signatures:
a review and a proposal. Org. Geochem., 36(5), 739-752.
Cord-Ruwisch, R., Seitz, H.-J., Conrad, R. 1988. The capacity of hydrogenotrophic anaerobic
bacteria to compete for traces of hydrogen depends on the redox potential of the terminal
electron acceptor. Archives of Microbiology, 149(4), 350-357.
Dolfing, J., Jiang, B., Henstra, A.M., Stams, A.J., Plugge, C.M. 2008. Syntrophic growth on
formate: a new microbial niche in anoxic environments. Applied and environmental
microbiology, 74(19), 6126-6131.
Ferguson, T.J., Mah, R.A. 1983. Effect of H2-CO2 on methanogenesis from acetate or methanol in
Methanosarcina spp. Applied and environmental microbiology, 46(2), 348-355.
Hao, L., Lü, F., Li, L., Shao, L., He, P. 2013. Response of anaerobes to methyl fluoride, 2-
bromoethanesulfonate and hydrogen during acetate degradation. Journal of Environmental
Sciences, 25(5), 857-864.
Hattori, S., Nashimoto, H., Kimura, H., Koba, K., Yamada, K., Shimizu, M., Watanabe, H., Yoh,
M., Yoshida, N. 2012. Hydrogen and carbon isotope fractionation by thermophilic
hydrogenotrophic methanogens from a deep aquifer under coculture with fermenters.
Geochemical Journal, 46(3), 193-200.
Lee, J.C., Kim, J.H., Chang, W.S., Pak, D. 2012. Biological conversion of CO2 to CH4 using
hydrogenotrophic methanogen in a fixed bed reactor. Journal of Chemical Technology and
Biotechnology, 87(6), 844-847.
Leybo, A., Netrusov, A., Conrad, R. 2006. Effect of hydrogen concentration on the community
structure of hydrogenotrophic methanogens studied by T-RELP analysis of 16S rRNA gene
amplicons. Microbiology, 75(6), 683-688.
Luo, G., Angelidaki, I. 2013. Co-digestion of manure and whey for in situ biogas upgrading by the
addition of H(2): process performance and microbial insights. Appl Microbiol Biotechnol,
97(3), 1373-81.
Luo, G., Angelidaki, I. 2012. Integrated biogas upgrading and hydrogen utilization in an anaerobic
reactor containing enriched hydrogenotrophic methanogenic culture. Biotechnology and
bioengineering.
Luo, G., Johansson, S., Boe, K., Xie, L., Zhou, Q., Angelidaki, I. 2012. Simultaneous hydrogen
utilization and in situ biogas upgrading in an anaerobic reactor. Biotechnology and
bioengineering.
Mulat, D.G., Feilberg, A. 2005. GC/MS method for determining carbon isotope enrichment and
concentration of underivatized short-chain fatty acids by direct aqueous solution injection of
biogas digester samples Under review in Talanta, X(X), P.X.
Nielsen, H.B., Uellendahl, H., Ahring, B.K. 2007. Regulation and optimization of the biogas
process: Propionate as a key parameter. Biomass and Bioenergy, 31(11–12), 820-830.
98
Oremland, R.S. 1975. Methane production in shallow-water, tropical marine sediments. Appl
Microbiol, 30(4), 602-8.
Penning, H., Plugge, C.M., Galand, P.E., Conrad, R. 2005. Variation of carbon isotope fractionation
in hydrogenotrophic methanogenic microbial cultures and environmental samples at
different energy status. Global Change Biology, 11(12), 2103-2113.
Qu, X., Mazéas, L., Vavilin, V.A., Epissard, J., Lemunier, M., Mouchel, J.M., He, P.j., Bouchez, T.
2009. Combined monitoring of changes in δ13CH4 and archaeal community structure
during mesophilic methanization of municipal solid waste. FEMS microbiology ecology,
68(2), 236-245.
Schmidt, J.E., Ahring, B.K. 1993. Effects of hydrogen and formate on the degradation of propionate
and butyrate in thermophilic granules from an upflow anaerobic sludge blanket reactor.
Applied and environmental microbiology, 59(8), 2546-2551.
Scholten, J.C., Conrad, R. 2000. Energetics of syntrophic propionate oxidation in defined batch and
chemostat cocultures. Applied and environmental microbiology, 66(7), 2934-2942.
Thiele, J.H., Zeikus, J.G. 1988. Control of interspecies electron flow during anaerobic digestion:
significance of formate transfer versus hydrogen transfer during syntrophic methanogenesis
in flocs. Applied and environmental microbiology, 54(1), 20-29.
Valentine, D.L., Chidthaisong, A., Rice, A., Reeburgh, W.S., Tyler, S.C. 2004. Carbon and
hydrogen isotope fractionation by moderately thermophilic methanogens. Geochimica et
Cosmochimica Acta, 68(7), 1571-1590.
Ward, A.J., Bruni, E., Lykkegaard, M.K., Feilberg, A., Adamsen, A.P., Jensen, A.P., Poulsen, A.K.
2011. Real time monitoring of a biogas digester with gas chromatography, near-infrared
spectroscopy, and membrane-inlet mass spectrometry. Bioresour. Technol. , 102(5), 4098-
4103.
Whiticar, M.J. 1999. Carbon and hydrogen isotope systematics of bacterial formation and oxidation
of methane. Chemical Geology, 161(1–3), 291-314.
Winfrey, M., Nelson, D., Klevickis, S., Zeikus, J. 1977. Association of hydrogen metabolism with
methanogenesis in Lake Mendota sediments. Applied and environmental microbiology,
33(2), 312-318.
99
Chapter 7: Paper IV- GC/MS method for determining carbon isotope
enrichment and concentration of underivatized short-chain fatty acids by direct
aqueous solution injection of biogas digester samples
Submitted to Talanta. Under review.
100
GC/MS method for determining carbon isotope enrichment and concentration of
underivatized short-chain fatty acids by direct aqueous solution injection of biogas digester
samples
Daniel Girma Mulat1, Anders Feilberg
1*
1Department of Engineering, Aarhus University, Hangøvej 2, DK-8200 Aarhus N, Denmark
*Corresponding author:
Anders Feilberg; phone: +45 30896099; e-mail: [email protected]
ABSTRACT
In anaerobic digestion of organic matter, several metabolic pathways are involved during the
simultaneous production and consumption of short-chain fatty acids (SCFA) in general and acetate
in particular. Understanding the role of each pathway requires both the determination of the
concentration and isotope enrichment of intermediates in conjunction with isotope labeled
substrates. The objective of this study was to establish a rapid and simple GC/MS method for
determining the isotope enrichment of acetate and concentration of underivatized short-chain fatty
acids (SCFA) in biogas digester samples by direct liquid injection of acidified aqueous samples.
Sample preparation involves only acidification, centrifugation and filtration of the aqueous solution
followed by direct injection of the aqueous supernatant solution onto a polar column. With the
sample preparation and GC/MS conditions employed, well-resolved and sharp peaks of
underivatized SCFA were obtained in a reasonably short time. Good recovery (96.6-102.3%) as
well as low detection (4-7 µmmol/L) and quantification limits (14-22 µmmol/L) were obtained for
all the 6 SCFA studied. Good linearity was achieved for both concentration and isotope enrichment
measurement with regression coefficients higher than 0.9978 and 0.9996, respectively. The method
has a good intra and inter-day precision with a relative standard deviation (RSD) below 6 and 6.5 %
for determining the tracer-to-tracee ratio (TTR) of both [2-13
C]acetate and [U-13
C]acetate,
respectively. It has also a good intra and inter-day precision with a RSD below 6% and 5% for
determining the concentration of standard solution and biogas digester samples, respectively.
Acidification of biogas digester samples with oxalic acid provided the low pH required for the
protonation of SCFA and thus, allows the extraction of SCFA from the complex sample matrix.
Moreover, oxalic acid was the source of formic acid which was produced in the injector set at high
temperature. The produced formic acid prevented the adsorption of SCFA in the column, thereby
eliminating peak tailing and ghost peaks. The optimized and validated GC/MS method was then
employed to determine the 13
C isotope enrichment of acetate and concentration of SCFA in
anaerobic digester samples. The concentration of acetate and the TTR of [2-13
C]acetate were higher
for hydrogen-added biogas digesters compared to the control digesters in which hydrogen was not
supplied. These results demonstrated that under high hydrogen partial pressure, carbon dioxide was
reduced to acetate via homo-acetogenesis pathway.
Key words: gas chromatography/mass spectrometry (GC/MS); direct liquid injection; short-chain
fatty acids (SCFA); stable isotope; homo-acetogenesis; biogas digester
101
INTRODUCTION
The conversion of organic compounds from waste materials to methane is an important strategy for
sustainable energy supply. The production of methane involves a consortium of several
microorganisms, which undertakes major biological processes such as hydrolysis/acidogenesis,
acetogenesis and methanogenesis steps. Fermentative bacteria hydrolyze polymers to simpler
monomers and oligomers, which afterward are fermented to alcohols, short-chain fatty acids
(SCFA), carbon dioxide and hydrogen. Then acetogenic bacteria convert the SCFA other than
acetate to acetate, carbon dioxide and hydrogen. Finally, methanogens produce methane from the
direct cleavage of acetate through acetoclastic methanogenesis pathway as well as the reduction of
carbon dioxide with hydrogen via hydrogenotrophic methanogenesis pathway [1, 2].
SCFA including acetate are key intermediates in an anaerobic digestion of organic matters to
methane [2-6]. In particular to acetate, fermentative, acetogenic and homo-acetogenic bacteria are
the main contributors to its production. Fermentative and acetogenic bacteria are involved in the
production of H2, CO2, acetate and other products [3]. On the other hand, homo-acetogenic bacteria
utilize H2 and CO2 for the production of acetate via homo-acetogenesis pathway according to
equation 1. The feasibility of the homo-acetogenesis pathway is highly sensitive to the amount of
hydrogen. In the presence of H2-consuming microorganisms (e.g. hydrogenotrophic methanogens),
the amount of hydrogen is kept very low. In such an environment, homo-acetogenic bacteria are
outcompeted by the hydrogenotrophic methanogens due to the latter having a lower hydrogen
threshold than the former [7]. However, homo-acetogenic bacteria has been shown to be stimulated
in a high hydrogen concentration environment such as a biohydrogen producing digester [8] as well
as actual land fill [9]. Understanding the role of different metabolic pathways involved in the
production and consumption of acetate is paramount important for the efficient operation and
optimization of biogas process. This is because of an imbalance in the production and consumption
of acetate could lead to biogas process inhibition [4-6, 10]. In this regard, an analytical technique
that helps us to quantify the concentration of acetate and its isotope species from a biogas digester
sample is highly needed.
2CO2 + 4H2 → CH3COOH + 2H2O (1)
Radio isotopes are commonly used as tracers to study the fate of acetate in anaerobic digester [7,
11]. Radio isotope tracer is not a method of choice due to the strict safety requirement in handling
radio isotopes [3]. An alternative method is the use of stable isotope pairing and the measurement of
isotopically enriched products with hyphenated mass spectrometry techniques (GC/MS and
HPLC/MS). GC/MS and HPLC/MS have been successfully employed for determining isotope
enrichment of SCFA from plasma and urine samples in the field of clinical chemistry [12]. The
separation of individual SCFA from a complex biological sample is provided by either HPLC or
GC, whereas the tracer/tracee ratio (TTR) which is proportional to the ion-current ratio of the
labelled/unlabelled species is determined by the mass spectrometer.
GC/MS is a widely employed technique for determining the isotope enrichment of SCFA [12-14].
Previous GC/MS method for SCFA measurement involves different sample preparation procedures,
such as, extraction with organic solvents [13], purge and trap technique [15], steam distillation [13],
ultrafiltration [13] and solid-phase microextraction [16] as the most common cleanup pretreatments
followed by derivatization step [12, 13]. Derivatization of SCFA with different derivatizing
102
reagents is usually performed for increasing the volatility of the analytes. Although these sample
preparation procedures were demonstrated to be a good cleanup strategy, they are relatively time-
consuming and labor-intensive. They may also reduce analyte recovery and affect the accuracy and
repeatability of the method due to the requirement of multiple sample preparation procedures [13,
17, 18]. Moreover, derivatization of the analytes potentially increases natural isotope background
due to the introduction of additional isotopomers from the derivatizing reagents which challenges
accurate determination of low isotope enrichment assay [12].
An alternative method to derivatization of SCFA is a direct injection of aqueous sample into a
GC/MS system after a minor pretreatment of the liquid sample. GC method for determining the
concentration of fermentation products including SCFA from the direct injection of aqueous
solution into GC was shown to be rapid, accurate and reliable [4, 13, 18-20]. However, direct
aqueous solution injection could lead to peak tailing and ghosting due to contamination of the GC
column. Biogas digester liquid samples are complex in nature containing several organic and
inorganic compounds and solid particles. Therefore, an appropriate choice of glass liner, cleaning
solvent and sample preparation strategy should be employed to avoid column contamination which
ultimately provides good baseline separation of individual SCFA and peak shape.
To date there has not been any report on the application of GC/MS method for determining the
isotope enrichment of underivatized SCFA from the direct injection of aqueous biogas digester
sample into GC/MS system. To our knowledge this is the first time GC/MS was applied for
determining the isotope enrichment of acetate in a biogas digester samples without involving any
derivatization step, though GC/MS method was employed in plasma samples in one study [14]. In
this study we report a very simple, accurate, reproducible and rapid GC/MS method for determining
both the isotope enrichment of acetate and concentration of underivatized SCFA in biogas digester
sample by direct aqueous sample injection into the GC system. A water resistant polar column was
used for separating underivatized SCFA from the complex biogas digester matrix. In this study,
oxalic acid was used for acidification of biogas digester samples and other purposes such as to
improve peak sharpness, reduce peak tailing and to clean unwanted residues coming from the
complex sample matrix. As an example of application of this method, this paper also presents that a
stable isotope tracer experiment in combination with tracer-to-tracee ratio (TTR) determination by
the GC/MS method can be used to verify the activity of homo-acetogenic bacteria in a biogas
digester process.
EXPERIMENTAL
Reagents
A short-chain fatty acids (SCFA) mixture containing acetic acid (C2), propionic acid (C3),
isobutyric acid (i-C4), n-butyric acid (C4), isovaleric acid (i-C5) and n-valeric acid (C5) was obtained
from Sigma-Aldrich (Denmark). The concentration of all the analytes was 10 mmol/L in deionized
water. Sodium acetate, acetic acid, propionic acid, butyric acid, tert-pentanoic acid (t-C5), oxalic
acid as well as sodium salts of [2-13
C]acetate (99 atom%) and [U-13
C]acetate (99 atom%) were
purchased from Sigma-Aldrich (Denmark). Fluoromethane as specific inhibitor to acetoclastic
methanogens was obtained from Sigma-Aldrich (Denmark).
103
Instrumentation and analytical condition
The GC/MS analysis was performed using a CP-3800 gas chromatograph (Varian Inc.) coupled to a
Saturn 2000 ion trap mass spectrometer (Varian Inc.). The gas chromatograph was equipped with an
electron impact ion source, split-splitless injector and an autosampler. A high polarity capillary
column with a cross-linked and bonded polyethylene glycol (PEG) phase (HP-INNOWax, 30m x
0.25 mm i.d coated with 0.25 μm film thickness, Agilent Technologies Inc., Denmark) was used for
separation of the SCFA. GC/MS conditions for determining both the isotope enrichment and
concentration were as follows: positive electron impact mode; injector temperature, 285 °C; helium
constant flow, 1 ml/min; initial column temperature was 100 °C and hold for 1 min, then increased
at the rate of 10 °C/min to 120 °C and hold for 5 min, and finally increased at the rate of 30 °C/min
to 220 °C and kept at 220 °C for 3 min; solvent delay time, 3 min; total run time, 14.3 min. The
volume injected was 1µL in 1:10 split mode. All the samples were analyzed in full scan mode with
a mass range of m/z 40-300. Finally, the following extracted ions were used for determining the
concentration and isotope ratio of SCFA: m/z 60 for unlabeled C2, C4, i-C5 and C5; m/z 61 for [1-13
C]acetate/[2-13
C]acetate; m/z 62 for [U-13
C]acetate; m/z 57 for t-C5; m/z 73 for C3 and i-C4. The
total concentration of acetate was determined from sum of ion currents of m/z 60, 61 and 62. Data
acquisition and analysis was done with Saturn GC/MS workstation software (Varian Inc.)
Operation of biogas digester
Inoculum was obtained from batch anaerobic digesters fed with 13
C fully labeled maize leaf
(IsoLife BV,Wageningen, the Netherlands) under thermophilic conditions. Effluent was collected
and aliquots of the effluent (10 mL) was immediately transferred into 6 glass serum bottles (20-
mL), which were then sealed with butyl rubber stopper and aluminium crimp. Then, three
treatments were prepared: i) Control reactor, RC; only 1 mL water was added instead of substrate
and the headspace was flushed with helium (ii) Uninhibited reactor, RUI; 1 mL of concentrated
sodium bicarbonate solution was added to give 20 mmol/L total bicarbonate solution and the
headspace was flushed with H2-CO2 mixture (80-20 V/V) for 5 min, (iii) Acetoclastic
methanogenesis-inhibited reactor, RI; 1 mL of concentrated sodium bicarbonate solution was added
to give 20 mmol/L total bicarbonate solution, the headspace was flushed with H2-CO2 mixture (80-
20 V/V) for 5 min and fluoromethane (1.3%) as specific inhibitor to acetoclastic methanogens was
added into the headspace. The above mentioned substrates were added periodically and each
treatment was run in duplicate. The reactors were incubated in a shaking incubator at 100 rpm for
15 days under thermophilic conditions (52°C). Biogas digester liquid samples were collected
periodically for determining the isotope composition of acetate and concentration of SCFA. The
liquid sample was subjected to sample preparation procedure as describe below prior to the GC/MS
analysis.
Sample preparation
Liquid sample (1.000 g) was transferred to a 10 mL plastic centrifuge tube and acidified with 4 mL
of acidifying solution (pH < 2; containing mixtures of 0.3 mol/L oxalic acid and 2.1 mmol/L tert-
pentanoic acid as internal standard). Then it was centrifuged at 14,000 g for 12 min at room
temperature and the clear solution was filtered through a 0.45m Acrodisc® syringe filters (Sigma-
Aldrich, Denmark) and finally the supernatant solution was transferred into analytical vials. Aliquot
104
(1.0 µL) of the solution was injected into the GC/MS system using autosampler. The processed
sample can be stored in the refrigerator (~ 4 °C) for a week and the freezer (~ -20 °C) for a month
prior to the GC/MS analysis (studied as a part of this work; data not shown).
Method validation
Recovery study
The recovery of the method was estimated by comparing the measured concentration of SCFA from
non-spiked and spiked biogas digester liquid sample. A liquid sample from a biogas digester was
spiked with a mixture of SCFA solution containing 5 mmol/L of each SCFA. Both spiked and non-
spiked samples were prepared in triplicate and were further subjected to the sample preparation
procedure described above prior to the GC/MS analysis. The peak areas of the SCFA from the
spiked and non-spiked sample were used to calculate the concentration of each SCFA.
Linearity range for isotope ratio measurement
The linearity range for the isotope ratio measurement was tested from a series of standard solutions
containing mixtures of 13
C labeled and unlabeled acetate. The isotope enrichment of the standard
solutions were expressed in terms of tracer-to-tracee ratio (TTR, %) and it was prepared in the
range from 0 to 100%. Working solutions of [U-13
C]acetate (5 mmol/L) and unlabeled acetate (5
mmol/L) were used for preparing the isotope enrichment standard solutions. Series of the standard
solutions were prepared by increasing the labeled [U-13
C]acetate ratio relative to the unlabelled
acetate and followed by dilution of all the standard solutions to the same concentration (2.5
mmol/L). The standard solutions were subjected to the sample preparation procedure described
above prior to the GC/MS analysis. Similarly standard solutions containing mixtures of [2-13
C]acetate and unlabeled acetate were prepared separately in the range of 0 to 100% TTR.
Individual calibration curves were plotted for both [U-13
C]acetate and [2-13
C]acetate standard
solutions from the measured TTR against the theoretical TTR.
Linearity range for quantifying concentration
Using a stock solution containing mixtures of SCFA, five calibration standards ranging from 0.5 to
10 mmol/L were prepared. For acetate, the linearity was investigated in the concentration range of
0.5-20 mmol/L due to the high concentration of acetate in the biogas digester studied herein. Four
mL of 2.1 mmol/L tert-pentanoic acid as the internal standard was added to each calibration
standards followed by sample preparation and GC/MS measurement as described above. Individual
calibration curves were plotted from the concentration ratio of each SCFA to internal standard
against their peak area ratio.
Limits of detection (LOD) and quantification (LOQ)
The LOD and LOQ for determining the concentration of SCFA were calculated from 3 and 10 times
signal-to-noise of a diluted standard solution, respectively.
Precision
The intra and inter-day precision of the method, expressed as percentage relative standard deviation
(% RSD), was obtained from the analysis of a sample injected 6 times in a single day and over 3
days. The precision of the method for determining the concentration of SCFA was studied from
105
biogas digester sample and standard solution. On the other hand, biogas digester samples spiked
with standard solutions of [2-13
C]acetate and [U-13
C]acetate were used to study the precision of the
method for determining the isotope enrichment of acetate.
Isotope calculation
Isotope enrichment of acetate was calculated from the ions intensities of M0, M0+1 and M0+2, which
were monitored at m/z 60, 61 and 62, respectively. The ion intensity ratio of labeled-to-unlabeled
species measured by the GC/MS was presented as tracer-to-tracee ratio (TTR). In the study
presented herein, the Rosenblatt approach [21] was used for calculating the TTR of acetate, which
is a widely used approach in the tracer studies of SCFA. According to Rosenblatt et al. [21],
several corrections need to be included in the calculation of TTR. The following three corrections
were considered in our TTR calculation. First, the contribution of labeled molecule from the
background sample was corrected for the sample data. Second, the abundance in the spectrum of
labeled molecule is different from those in the unlabeled molecule, and this proportionality “skew”
was corrected. The “skew” abundance distribution correction can be approximated by the value
[1/(1+nA)] , where A is the natural abundance of the element used as the label and n is the number
of these atoms labeled. The “skew” correction factors are 0.9890 and 0.9783 for the [2-13
C]acetate
and [U-13
C]acetate, respectively (A=0.01108). Third, the isotope overlapping of neighboring peaks
was corrected. Each compound studied (Table 1) not only contributed to its major peak, but also
contributed to the mass isotopomer distribution of the neighboring peaks (M0, M0+1 and M0+2)
significantly. Taking all the corrections stated above, equations 2 and 3 were used for calculating
the TTR of both [2-13
C]acetate and [U-13
C]acetate, respectively.
TTR(M0+1)=[(M0+1/M0)sample – (M0+1/M0)nat] x [1/(1+A)] – TTR(M0+2) x contribution(M0+2) (2)
TTR(M0+2)=[(M0+2/M0)sample – (M0+2/M0)nat] x [1/(1+2A)] – TTR(M0+1) x contribution(M0+1) (3)
where A corresponds to the carbon natural abundance, i.e. 0.01108
(M0+1/M0)nat and (M0+2/M0)nat are the ratio of the integrated peak areas of m/z61-to-m/z60 and m/z62-
to-m/z60 of acetate at natural abundance, respectively.
(M0+1/M0)sample and (M0+2/M0)sample are the ratio of the integrated peak areas of m/z61-to-m/z60 and
m/z62-to-m/z60 of 13
C enriched acetate samples, respectively.
Equations 2 and 3 are transformed to equations 4 and 5 with two unknown variables for solving the
TTR(M0+1) and TTR(M0+2), respectively. Using equations 4 and 5, the solutions to TTR(M0+1) and
TTR(M0+2) were given in equations 6 and 7, respectively.
TTR(M0+1)=X1-TTR(M0+2)Y1 (4)
TTR(M0+2)=X2-TTR(M0+1)Y2 (5)
TTR(M0+1)=(X1-X2Y1)/(1-Y2Y1) (6)
TTR(M0+2)=X2-[ (X1Y2 -X2Y1Y2)/(1-Y1Y2)] (7)
where X1=[(M0+1/M0)sample – (M0+1/M0)nat] x [1/(1+A)]
106
X2=[(M0+2/M0)sample – (M0+2/M0)nat] x [1/(1+2A)]
Y1= contribution(M0+2)
Y2=contribution(M0+1)
Table 1 Mass isotopomer distribution measured by GC/MS for unlabeled acetate, [2-13
C]acetate
and [U-13
C]acetate
Ions
m/z
Unlabeled acetate
%
[2-13
C]acetate
%
[U-13
C]acetate
%
60 100 9.2 5.9
61 15.7 100 5.2
62 2.4 15.1 100
As given in Table 1, the Y1 and Y2 are 5.2 and 15.1%, respectively. The measured TTR values
exceeded the theoretical values because the instrument was not specifically set for tuning acetate
and the low resolution of the quadrupole ion trap. In addition to the corrections stated above, one
need to use a linear regression curve plotted from the measured TTR against the theoretical TTR for
calculating the true TTR value of 13
C labeled acetate (see Figure 4 and the discussion in the section
“method validation”).
RESULTS AND DISCUSSION
Chromatographic separation of SCFA
Figure 1 shows a typical total ion chromatogram of an aqueous standard solution containing the
seven SCFA including the internal standard (C2, C3, i-C4, t-C5, C4, i-C5 and C5, listed according to
increasing retention time). For straight-chain SCFA, analytes with low molecular weight eluted later
than those with high molecular weight. On the other hand, the branched-chain isomer eluted earlier
than the corresponding straight-chain isomer. The identification of the analytes was based on mass
spectra of individual analyte. The peak observed at 5.34 min retention time appeared in all the
standard solution and biogas digester samples. The mass spectrum analysis showed that the source
of the peak was formic acid formed from the oxalic acid used for acidification of the samples. It
should be noted that formic acid elutes later than acetic acid on very polar columns due to higher
polarity. Since formic acid was not injected into the system, it is likely generated from oxalic acid
in the injector set at high temperature (285 °C).
107
Figure 1 The total ion chromatogram of SCFA (C2–C5) and t-C5 as internal standard obtained from
aqueous standard solution (* is formic acid produced from the oxalic acid used for acidification of
the sample)
Oxalic acid was used for several purposes. One was for the acidification of the biogas digester
samples. Acidification of samples is an important strategy to ensure protonation of SCFA before
their separation on a polar column. Previous study showed that protonated SCFA were better
extracted from complex matrix than ionized SCFA [13]. This is most likely due to ionized SCFA
has higher interaction with the constituents of the complex matrix than the protonated SCFA.
Theoretical consideration shows that pH less than 3 of the matrix provides a complete protonation
of the SCFA due to the studied SCFA have a pKa of roughly 4.8. In the present study, the added
oxalic acid brought the samples matrix to about pH 2. Another purpose was to reduce column
contamination due to its volatility nature opposed to the nonvolatile inorganic acids such as
hydrochloric and phosphoric acids employed in previous studies [4, 19]. Thirdly it was aimed to
improve peak sharpness and reduce peak tailing due to the liberated formic acid in the injector. In
previous study, formic acid was added to saturate the polar column that reduces adsorption of SCFA
on the column and ultimately provides the above mentioned benefits [22]. Moreover, the liberated
formic acid from oxalic acid was used for the purpose of cleaning unwanted residues of SCFA from
the polar column. Previous studies showed that after the injection of formic acid, ghost peaks did
not appear in a chromatogram [13, 20]. Our results also showed that ghost peaks were not observed
even after several injections of underivatized analytes into GC/MS system.
108
With the employed chromatographic condition, a good baseline separation of individual SCFA and
peak shape was obtained for a biogas digester sample with a reasonable short run time of less than
14 min (Figure 2). As shown in Figure 2, acetate and propionate were the main constituents of the
sample. The peaks shape was sharp and well resolved even if the sample matrix was very complex
as expected from organic waste treating biogas digester. These results showed that a good
separation of individual SCFA could be achieved without employing the time-consuming
derivatization and conventional liquid extraction techniques reported in literature [12]. It took only
less than half an hour for biogas digester sample preparation and a single GC/MS run, showing the
high throughput of the employed method.
Figure 2 The total ion chromatogram of SCFA obtained from biogas digester sample (* is formic
acid produced from the oxalic acid used for acidification of the sample)
The main challenge to direct injection of aqueous sample into GC/MS system is a possible
contamination of the column which would shorten the life column and could lead to peak
broadening and tailing [13]. In the present study, a glass liner packed with glass wool was inserted
into the injection port, which was demonstrated to be good for protecting a column from
contamination by non-volatile compounds [4, 13, 18]. We did not observe any ghost peaks after
running samples for more than 1000 injections. In our laboratory, SCFA from biogas digester
samples have been routinely quantified with GC method. We use the same column and similar GC
condition as the GC/MS method presented herein. Our experience with the GC method showed that
109
the life time of the polar column is around 4-5 years for the GC that runs almost all the time for the
day to day measurement of SCFA in a biogas digester samples. In the worst case, when the peak
shape starts to broaden and tailing occurs, the gas liner is most probably contaminated and need to
be changed in a matter of few minutes. Moreover, glass liner contamination was reduced due to the
use of oxalic acid.
Method validation
Linearity
The linearity of the isotope enrichment measurement was tested by analyzing series of calibration
solutions containing mixture of [U-13
C]acetate and unlabeled acetate or mixture of [2-13
C]acetate
and unlabeled acetate in the range from 0 to 100 TTR%. The mass spectrum was acquired with full
scan mode and the ions (m/z 60, 61 and 62) that represent isotope species of acetate (M0, M0+1 and
M0+2, respectively) were extracted and integrated for quantifying the isotope enrichment of 13
C
enriched acetate. As shown in Figure 3, the selected ion chromatograms of isotope species of
acetate were very clean which improved the signal to noise ratio and the accuracy of the method.
110
Figure 3 Selected ion chromatograms of acetate and tert-pentanoic acid used as internal standard
for biogas digester samples: (A) Unlabeled acetate, m/z 60; (B) [2-13
C]acetate, m/z 61; and (C) [U-13
C]acetate, m/z 62
The measured TTR was plotted individually against the theoretical TTR of [2-13
C]acetate and [U-13
C]acetate. Figure 4 shows that the measured TTR closely matched the theoretical TTR (R² =
0.9997 and 0.9996 for [2-13
C]acetate and [U-13
C]acetate, respectively). Since the slope of the
calibration curves were different from 1, the true TTR values were calculated using the linear
calibration equations as demonstrated in previous studies [23, 24]. The use of calibration curves
avoids problems inherent to direct determination of isotope ratio. These problems include the
unavoidable instrument drift over time, changes in isotope ratio depending on the concentration of
analyte, non-linear response of the mass spectrometer with higher amount of tracer, day-to-day
variation in tuning conditions within the instrument [23]. Therefore, the calibration curves should
be prepared in time series with the biogas digester samples being analyzed [24]. The accuracies of
isotope enrichment determination for [2-13
C]acetate and [U-13
C]acetate in the entire range of 0-100
111
TTR% were 90-101 and 89-107%, respectively. These results demonstrated the good analytical
accuracy of the employed isotope enrichment measuring method.
Figure 4 Relationship between the theoretical and measured tracer-to-tracee ratio of (A) [2-13
C]acetate-to-unlabeled acetate (y = 0.8605x – 0.8948, R² = 0.9997) in the range of 0-100 TTR%
and (B) [U-13
C]acetate-to-unlabeled acetate (y = 0.8889x + 0.1753, R² = 0.9996) in the range of 0-
100 TTR%
The calibration curves for determining the concentration of unlabeled SCFA (C3, i-C4, C4, i-C5 and
C5) were individually plotted from the GC/MS measurement at concentration levels from 0.5 to 10
mmol/L containing known amount of 2.1 mmol/L tert-pentanoic acid (t-C5) as internal standard
(Table 2). Whereas for acetate, the linearity was tested in the concentration ranged from 0.5 to 20
mmol/L. Correlations between the GC/MS responses and the amount of analytes were verified by
plotting signal intensity ratio of individual SCFA to internal standard against the ratio of their
concentrations. Good linearity was achieved in all cases with regression coefficients higher than
0.9978. The linearity range was wide enough to cover the amount of SCFA determined in the
biogas digester samples (Figure 5).
0
20
40
60
80
100
0 20 40 60 80 100Mea
sure
d t
race
r-to
-tra
cee
rati
o
(%)
Theoretical tracer-to-tracee ratio
(%)
0
20
40
60
80
100
0 20 40 60 80 100Mea
sure
d t
race
r-to
-tra
cee
rati
o
(%)
Theoretical tracer-to-tracee ratio
(%)
(A) (B)
112
Table 2 Calibration equation (y = mx + b)a, LOD and LOQ of the GC/MS method for determining
the concentration of SCFA
SCFA
Linearity
range
(mmol/L)b Slope m Intercept b R²
LOD
(µmol/L)
LOQ
(µmol/L)
C2 0.5-20 0.1335 0.0245 0.9995 4.27 14.22
C3 0.5-10 0.1835 -0.0015 0.9995 5.99 19.96
i-C4 0.5-10 0.3646 0.0013 0.9997 6.59 21.97
C4 0.5-10 0.6503 0.0037 0.9985 5.89 19.64
i-C5 0.5-10 0.7318 0.0099 0.9978 4.51 15.04
C5 0.5-10 0.7151 0.0034 0.9998 4.36 14.52
a y is the ratio of the concentration of SCFA to internal standard whereas x is the ratio of peak area
of the SCFA to internal standard b
the linearity range represent the amount of standard solutions of SCFA before acidification of the
sample
Figure 5 Concentration of total acetate in samples from biogas digesters (RC □; RI ■; RUI ▲). The
lines represent mean values (n=2) and error bars denote data range
Limits of detection (LOD) and quantification (LOQ)
The limits of detection (LOD) and quantification (LOQ) were 4-7 and 14-22 µmol/L, respectively
(Table 2). The low LOD and LOQ showed that the developed method is sensitive enough to
identify and quantify even lower concentration of SCFA in a biogas digester sample than those
detected in our study.
0
1
2
3
4
5
6
7
8
9
0 2 4 6 8 10 12 14 16
Ace
tate
(m
mol/
L)
Incubation time (days)
113
Recovery
The recovery of the method was investigated by spiking a mixture of acetic acid, propionic and
butyric acid to a biogas digester sample. The spiked concentration of each SCFA was 5 mmol/L.
Since acetic acid, propionic acid and butyric acid are the major SCFA produced in a biogas digester
[4, 10], the recovery test involved only those three acids. As given in Table 3, the recovery for C2,
C3 and C4 was 98.7, 96.6 and 102.3%, respectively. The results showed that the sample preparation
procedure leads to a very good quantitative extraction of the individual SCFA.
Table 3 Recovery of C2, C3 and C4 after spiking biogas digester sample with 5 mmol/L of each
SCFA
Sample description C2 (mmol/L) C3 (mmol/L) C4 (mmol/L)
Original biogas digester sample 5.23±0.17 2.15±0.09 0.65±0.02
Standard solution added 5 5 5
Measured value (spiked with 5 mmol/L SCFA) 10.10±0.06 6.91±0.24 5.78±0.07
Recovery (%) 98.7 96.6 102.3
Precision
The intra and inter-day precision of the method was studied from a sample injected 6 times in one
day and over 3 different days, respectively. As given in Table 4, the intra and inter-day variability
for determining the concentration of SCFA in standard solutions was below 6%. The RSD was
below 5% for both the intra and inter-day precision of the biogas digester samples, showing the
excellent precision of the employed method for quantifying the concentration of SCFA in complex
biogas digester sample.
TABLE 4 Precision of the method for the measurement of concentrations of SCFA
Precision C2 C3 i-C4 C4 i-C5 C5
Intra-day of standard
solution
Concentration (mmol/L) (n=6) 2.76 2.63 2.53 2.49 2.48 2.55
RSD (%) 3.61 3.70 5.72 2.17 5.29 5.89
Intra-day of samplea
Concentration (mmol/L) (n=6) 5.24 2.18 NDb 0.61 ND ND
RSD (%) 3.80 3.94
2.08
Inter-day of standard
solution
Concentration (mmol/L) (n=18) 2.52 2.60 2.44 2.46 2.36 2.38
RSD (%) 4.15 3.22 5.62 2.42 5.66 4.22
Inter-day of sample
Concentration (mmol/L) (n=18) 5.16 2.06 ND 0.63 ND ND
RSD (%) 4.17 2.15 1.64 a biogas digester sample
b Not detected
The method was also investigated for its precision in measuring the isotope enrichment of acetate
from a biogas digester sample spiked with [2-13
C]acetate and [U-13
C]acetate. As deduced from RSD
in Table 5, the within day and day-to-day variations were below 6% for [2-13
C]acetate and 6.5% for
[U-13
C]acetate. Therefore, the method has a good precision for simultaneously quantifying the
isotope enrichment of acetate and concentration of SCFA.
114
TABLE 5 Precision of the method for the measurement of isotope enrichment of 13
C labeled
acetate
Tracer-to-Tracee ratio
Day to Day
Sample No. of
injections Day 1 Day 2 Day 3 Mean SD RSD (%)
[2-13
C]acetate 1 0.167 0.181 0.176 0.175 0.007 4.103
2 0.162 0.163 0.158 0.161 0.003 1.715
3 0.167 0.183 0.175 0.175 0.008 4.349
4 0.158 0.175 0.165 0.166 0.008 5.123
5 0.150 0.163 0.155 0.156 0.007 4.305
6 0.165 0.169 0.157 0.163 0.006 3.816
Mean 0.161 0.172 0.164 0.166
SD 0.007 0.008 0.009 0.008
RSD (%) 4.163 4.934 5.636 4.532
[U-13
C]acetate 1 0.145 0.142 0.153 0.146 0.006 3.919
2 0.134 0.144 0.146 0.141 0.006 4.593
3 0.137 0.144 0.149 0.143 0.006 4.140
4 0.141 0.152 0.141 0.144 0.006 4.462
5 0.133 0.145 0.148 0.142 0.008 5.616
6 0.146 0.161 0.166 0.158 0.010 6.410
Mean 0.139 0.148 0.150 0.146 0.006 3.982
SD 0.006 0.008 0.009 0.006
RSD (%) 4.075 5.096 5.662 4.251
Method application
After the optimization and validation of the GC/MS method as described above, it was applied to
study metabolic pathways in anaerobic digestion. The concentration of SCFA and isotope
enrichment of acetate in liquid samples obtained from biogas digesters were determined by GC/MS.
The aim of this study was to investigate if biogas digester under high hydrogen partial pressure
leads to the reduction of CO2 to acetate via homo-acetogenesis pathway (equation 1). The
concentration of SCFA and TTR values of 13
C enriched acetate (M0+1 and M0+2) were compared
between hydrogen-added reactors and control reactors (RC) in which hydrogen was not supplied.
Fluoromethane (CH3F) was added in the H2-added reactors (RI) to inhibit acetoclastic
methanogenesis (AM) whereas the other H2-added reactors were not supplied with CH3F and hence,
named uninhibited reactors (RUI). The TTR values of the 13
C enriched acetate in the RUI throughout
the incubation is influenced by both acetoclastic and homo-acetogenesis pathways (if any) and only
by homo-acetogenesis pathway in the RI. Comparison of the concentration and isotopic
enrichment of acetate between RI and RUI during the course of incubation would give a clear
115
indication about the activity of homo-acetogenesis pathway under high hydrogen partial pressure
(see the discussion below).
The total concentration of acetate in RC decreased continuously from the initial 5.2 mmol/L and
finally reached a minimum value of 0.3 mmol/L on day 15 (Figure 5). The concentration of acetate
in RUI followed similar temporal trend as in RC but the concentration of acetate was higher in the
former throughout the incubation. However, the concentration of acetate in the RI increased slightly
and finally reached 7.8 mmol/L on the last day. In addition to acetate, propionate was detected in all
the reactors at some points of the incubation. The concentration of propionate in RI was almost
constant (~1.0 mmol/L) throughout the incubation period. In both RUI and RC, the concentration of
propionate was below 1.0 mmol/L for most of the time and below the detection limit in few cases
(data not shown). The observed low concentrations of propionate indicate that the turnover rate of
propionate may be very fast.
The continuous reduction in the concentration of acetate in RC and RUI showed that acetate was
likely consumed by acetate-utilizing microorganisms. Concentration of acetate was higher in the
RUI compared to the control reactor RC, showing either acetate degradation rate became lower or
more acetate was produced in the former via homo-acetogenesis according to equation 1. Since
acetate concentration increased in RI over time, the acetate-consuming acetoclastic methanogens
was inhibited by the added CH3F. Moreover, the increase in acetate concentration in RI could be
attributed to the stimulation of homo-acetogenic bacteria under the high hydrogen partial pressure.
However, there is also a possibility of acetate production by the fermentative and acetogenic
bacteria by utilizing the background biomass that limits the use of acetate concentration alone to
prove the activity of homo-acetogenic bacteria. Therefore, further information from the isotopic
enrichment of acetate is required to verify whether homo-acetogenic bacteria were stimulated under
high hydrogen partial pressure.
Acetate M0+1 and M0+2 were 89 and 97 TTR% in the inoculum which was used for incubating all the
reactors and therefore, for all the reactors the data presented at day 0 in Figure 6 was the one
determined in the original inoculum. The observed high TTR values of acetate M0+1 and M0+2
showed that the inoculum was the source of 13
C enriched fermentation products (13
CH3COOH,
CH313
COOH, 13
CH313
COOH, and/or 13
CO2). This is expected since the starting inoculum was
obtained from a batch digester treating 13
C fully labeled maize leaf. As shown in Figure 6, the TTR
of acetate M0+1 and M0+2 decreased in all the reactors overtime. The decrease in the TTR of acetate
M0+1 was the highest in the RC and lowest in the RI. It decreased to the minimum 7, 24 and 58
TTR% on day 15 in RC, RUI and RI, respectively. Similar temporal trend was observed in all the
reactors for the acetate M0+2 as the M0+1. Acetate M0+2 decreased up to the final 20 and 29 TTR% in
the RUI and RC, respectively. It only decreased up to 54 TTR% in the RI during the whole
incubation period.
116
Figure 6 Isotope enrichment of acetate in samples from biogas digesters. The unbroken lines show
the TTR of acetate M0+2 while the dashed lines are the TTR of acetate M0+1 (RC □; RI ■; RUI ▲).
The lines represent mean values (n=2) and error bars denote data range
If homo-acetogenic bacteria are stimulated, fermentation of 13
CO2 in the presence of hydrogen as a
reductant leads to the incorporation of 13
C into the acetate. Since the headspace of RI and RUI was
flushed with unlabeled carbon dioxide (CO2) and they were supplemented with unlabeled
bicarbonate, the 13
CO2 originated from the inoculum would be diluted significantly. Therefore, in
the presence of homo-acetogenic bacteria both unlabeled and 13
C labeled acetate will be produced
simultaneously with the former product being in higher proportion in the RI. In the RUI, however,
still the proportion of 13
CO2 was higher than CO2 due to the production of the former through
acetoclastic methanogenesis (AM). The continuous decrease in the TTR of acetate M0+1 and M0+2 in
the AM-inhibited reactor (RI) as well as the continuous increase in acetate concentration in this
reactor over time showed that higher proportion of unlabeled acetate was produced via homo-
acetogenesis pathway compared to the production of 13
C enriched acetate. Both the molecular ion
intensity of unlabeled acetate (m/z 60) and the 13
C enriched acetate (13
CH3COOH and/or
CH313
COOH; m/z 61) increased over time with the former was in higher proportion than the latter
(data not shown). This is in agreement with the presence of higher proportion of CO2 than 13
CO2 in
the RI. Whereas the molecular ion intensity of 13
CH313
COOH (m/z 62) as well as the fragment ion of
CH313
COOH and/or 13
CH313
COOH (i.e. 13
COOH at m/z 46) did not change over time in the RI (data
not shown). The increase in the ion intensity of m/z 61 but not m/z 46 showed that the 13
CO2 was
incorporated at the methyl position of the acetate by homo-acetogenic bacteria, which is in
agreement with the findings of Wolin et al. [11]. Acetate utilization by acetoclastic methanogens in
both RUI and RC was apparent from the continuous decrease in the TTR of acetate M0+1 and M0+2
over the course of the incubation in both reactors compared to RI. 13
CH3COOH was produced by
homo-acetogenic bacteria in the RUI as the TTR of acetate M0+1 and acetate concentration were
slightly higher in the RUI compared to RC. These findings demonstrated that it is important to
simultaneously monitor the concentration of acetate and its isotope enrichment in order to prove the
activity of homo-acetogenesis pathway in a biogas process.
0
20
40
60
80
100
0
20
40
60
80
100
0 2 4 6 8 10 12 14 16
Ace
tate
Mo+
2 (
TT
R%
)
Ace
tate
Mo+
1 (
TT
R%
)
Incubation time (days)
117
CONCLUSION
A simple, rapid and accurate GC/MS method was established for determining the isotope
enrichment of underivatized acetate and concentration of underivatized short-chain fatty acids
(SCFA) in samples from biogas digesters. Acidification of biogas digester samples with oxalic acid
provided the low pH required for protonation of SCFA and improved extraction of SCFA from the
complex sample matrix. Oxalic acid was converted to formic acid in the injector set at high
temperature and the latter proved to overcome problems normally associated with direct injection of
aqueous samples into GC/MS such as peak tailing, ghost peak and column contamination. The
applicability of the method was evaluated in samples obtained from anaerobic biogas reactor with
the aim of identifying the role of high hydrogen partial pressure for the reduction of CO2 to acetate
via homo-acetogenic pathway. The results of the concentration of SCFA and isotope enrichment of
acetate under high hydrogen partial pressure showed that carbon dioxide was reduced to acetate via
homo-acetogenesis pathway. Unlike the conventional time consuming and labor-intensive sample
preparation procedure, herein we presented a method involving only a minor sample preparation
steps without derivatization of the analytes in a reasonably short time. Moreover, the method could
be further optimized for measuring the isotope enrichment of other SCFA such as propionate and
butyrate. Since SCFA are key intermediates of anaerobic digestion, understanding the role of
different metabolic pathways involved in the production and consumption SCFA under different
biogas operation conditions would contribute to improving process control and optimization
strategies.
Acknowledgments
This research was financially supported by the Danish Strategic Research Council (Grant No. 10-
093944).
References
[1] D. Sasaki, T. Hori, S. Haruta, Y. Ueno, M. Ishii, Y. Igarashi, Journal of Bioscience and
Bioengineering, 111 (2011) 41-46.
[2] A.O. Wagner, C. Reitschuler, P. Illmer, Biochemical Engineering Journal, 90 (2014) 154-161.
[3] D.G. Mulat, A.J. Ward, A.P.S. Adamsen, N.V. Voigt, J.L. Nielsen, A. Feilberg, Environmental
science & technology, 48 (2014) 2505-2511.
[4] V. Diamantis, P. Melidis, A. Aivasidis, Analytica Chimica Acta, 573–574 (2006) 189-194.
[5] K. Boe, I. Angelidaki, Water Science & Technology, 66 (2012) 2496-2503.
[6] K. Boe, D.J. Batstone, I. Angelidaki, Biotechnology and bioengineering, 96 (2007) 712-721.
[7] T.M. Hoehler, D.B. Albert, M.J. Alperin, C.S. Martens, Limnol. Oceanogr, 44 (1999) 662-667.
[8] V. Siriwongrungson, R.J. Zeng, I. Angelidaki, Water research, 41 (2007) 4204-4210.
[9] A.C. Chen, A. Ohashi, H. Harada, Environmental technology, 24 (2003) 435-443.
[10] P. Weiland, Applied Microbiology and Biotechnology, 85 (2010) 849-860.
[11] M.J. Wolin, T.L. Miller, M.D. Collins, P.A. Lawson, Applied and environmental
microbiology, 69 (2003) 6321-6326.
[12] K. Tomcik, R.A. Ibarra, S. Sadhukhan, Y. Han, G.P. Tochtrop, G.-F. Zhang, Analytical
Biochemistry, 410 (2011) 110-117.
[13] A. Tangerman, F.M. Nagengast, Analytical biochemistry, 236 (1996) 1-8.
118
[14] C.L. Kien, D.H. Chang, R.D. Murray, A. Ailabouni, J. Kepner, Biomedical & environmental
mass spectrometry, 19 (1990) 554-558.
[15] R. Marsili, N. Miller, G. Kilmer, R. Simmons, Journal of chromatographic science, 32 (1994)
165-171.
[16] R.B. Batista, C.C. Grimm, M.A. Godshall, Journal of chromatographic science, 40 (2002) 127-
132.
[17] M.-H. Yang, Y.-M. Choong, Food Chemistry, 75 (2001) 101-108.
[18] M.-H. Yang, Y.-M. Choong, Food Chemistry, 75 (2001) 101-108.
[19] P.F. Pind, I. Angelidaki, B.K. Ahring, Biotechnology and Bioengineering, 82 (2003) 54-61.
[20] G. Zhao, J.-f. Liu, M. Nyman, J.Å. Jönsson, Journal of Chromatography B, 846 (2007) 202-
208.
[21] J. Rosenblatt, D. Chinkes, M. Wolfe, R.R. Wolfe, American Journal of Physiology-
Endocrinology And Metabolism, 263 (1992) E584-E596.
[22] A. Tangerman, A. van Schaik, M.T. Meuwese-Arends, J.H. van Tongeren, Clinica chimica
acta; international journal of clinical chemistry, 133 (1983) 341-348.
[23] E. Pouteau, I. Meirim, S. Métairon, L.B. Fay, Journal of mass spectrometry, 36 (2001) 798-
805.
[24] J.M. Hernandez-Perez, E. Cabre, L. Fluvia, A. Motos, C. Pastor, A. Corominas, M.A. Gassull,
Clinical chemistry, 48 (2002) 906-912.
119
Chapter 8: Paper V- Changing feeding regimes to demonstrate flexible biogas
production: effects on process performance, microbial community structure
and methanogenesis pathways
In submission for Bioresource Technology Journal.
120
Changing feeding regimes to demonstrate flexible biogas production: effects on process
performance, microbial community structure and methanogenesis pathways
Daniel Girma Mulat
a, H. Fabian Jacobi
b,c, Anders Feilberg
a, Anders Peter S. Adamsen
a, Marcell
Nikolauszd
aDepartment of Engineering, Aarhus University, Hangøvej 2, DK-8200 Aarhus N, Denmark
bDepartment of Biochemical Conversion, Deutsches Biomasseforschungszentrum (DBFZ), Torgauer Str.
116, 04347 Leipzig, Germany cLandesbetrieb Hessisches Landeslabor, Schloß Str. 26, 36251 Bad Hersfeld, Germany
dDepartment of Environmental Microbiology, Helmholtz Centre for Environmental Research-
UFZ, Permoser Str. 15, 04318 Leipzig, Germany
ABSTRACT
Distillers dried grains with solubles (DDGS) treating reactors were operated under different
feeding intervals to demonstrate flexible biogas production and its effect on process performance,
microbial community structure and methanogenesis pathway. Due to the temporal variation in
biogas production in the less frequently compared to the more frequently fed reactor, biogas can
be produced flexibly to meet the goal of producing more biogas during high energy demand
periods. Moreover, methane yield was significantly higher by 14 and 27% in the less frequently
fed reactor during the steady state and stress conditions, respectively. The composition of the
bacterial community varied between reactors whereas methanogenic community remained stable,
with the dominance of Methanosarcina followed by Methanobacterium. Our observation of the
dominating methanogens was supported by stable isotope analysis of the produced biogas,
showing similar contribution of both acetoclastic and hydrogenotrophic methanogenesis to the
total methane produced during each feeding event.
Key words: flexible biogas production; demand-driven electricity; feeding regimes; stillage,
isotope
121
1. INTRODUCTION
The share of renewable energy sources for the generation of electricity is increasing in Germany
and worldwide. A transition of electricity supply based on renewable energies requires an
integration of electricity from biomass, particularly from biogas, to balance the supply of
electricity generated from fluctuating sources such as solar and wind (Hahn et al., 2014; Szarka et
al., 2013). Biogas plants are traditionally operated with a continuous and constant substrate feed
to achieve nearly the same amount of biogas and electricity generation throughout the day with a
market value equivalent to base loads market prices. Since energy prices vary throughout the day,
biogas plants could be operated on demand when the market price for electricity is high. The
option to generate electricity on demand could possibly maximize profit (Hochloff & Braun,
2014).
Biogas offers the advantage of high availability and predictability allowing the generation of
electricity on demand. The concept of demand-driven electricity generation from biogas could be
realized with different strategies such as storage of biogas and intermediate products, biogas
upgrading to biomethane and subsequently injection into the natural gas grid as well as flexible
biogas production (Hahn et al., 2014; Szarka et al., 2013). Flexible biogas production could be
attained through an adapted feeding regime to control the gas production on demand. In such way
the need for gas storage could be minimized to reduce investment costs (Szarka et al., 2013).
An efficient mineralization of organic materials to biogas is dependent on mutual and syntrophic
interactions of functionally distinct microorganisms. Environmental perturbation and differences
in reactor operating conditions such as pH, feeding interval, organic loading rate (OLR),
temperature, hydraulic retention time (HRT) and feed composition have been reported to affect
the stability of the process as well as the activity of microorganisms in general and the
methanogens in particular (Franke-Whittle et al., 2014; Schmidt et al., 2014; Ziganshin et al.,
2013). Shorter feeding interval is usually preferred, because in common practice it is believed to
help process stability. Whereas longer feeding interval is believed to favor conditions under
which the balance between volatile fatty acid (VFA) producing and consuming microbes
mismatch. This VFA imbalance is due to the slower growth rate and metabolic activity of
methanogens in compared to acid-producing bacteria (acidogenic and acetogenic bacteria). In
such condition, VFA could accumulate to the level that may have a direct toxic effect and could
also lower the pH to suboptimal values that may further reduce the activity of methanogens
(Franke-Whittle et al., 2014; Weiland, 2010). These implies that knowledge on biogas process
and microbial community structure up on a change in feeding regime is highly required to realize
the transition to flexible biogas production.
Since the concept of flexible biogas production is still under discussion, the effects of varied
feeding regimes have not been well documented. Lv and his colleagues compared the effects of
feeding a digester once and twice per day with the same daily organic loading on process
performance and methanogenesis pathways of maize silage reactors (Lv et al., 2014). The reactor
fed once per day showed a transient accumulation of VFA after feeding event and subsequent
utilization of the VFA afterwards. Short-term stable isotope analysis of gases demonstrated a
temporal variation in methanogenesis pathways which correlated with the change in
concentration of VFA (Lv et al., 2014). On the other hand, the reactor that was fed twice per day
did not show significant changes in concentration of VFA and gas production between feeding
122
events. Another study of acetate-fed reactors demonstrated that an hourly fed reactor was
dominated by the strictly acetoclastic Methanosaeta whereas the daily fed reactor was dominated
by the versatile Methanosarcina capable of utilizing all three methanogenic pathways (Conklin et
al., 2006). The Methanosarcina dominated reactor was more tolerant to environmental
perturbation (organic overloading) compared to the Methanosaeta dominated reactor. Another
study based on synthetic feed compared the effect of two feeding patterns (daily and every 2 days
feeding) on bacterial community dynamics (De Vrieze et al., 2013). The average biogas
production was similar in both reactors despite higher variation in daily biogas production for the
reactor fed every 2 days. However, the reactor fed every 2 days had higher tolerance to organic
shock load of 8 gCOD L-1
and high total ammonia nitrogen (TAN) levels up to 8000 mgNH4+-N
L-1
which was in part due to the higher degree of bacterial dynamics in the every 2 days fed
reactor. Despite these previous studies, our understanding of the microbial community structure
and biogas process performance for reactors fed at different feeding intervals is still limited.
Moreover, the findings from the reactors fed with acetate (Conklin et al., 2006) and synthetic
feed materials (De Vrieze et al., 2013) cannot be easily translated to typical biogas plants
working with agricultural and organic industrial wastes. It was recently reported that biogas can
be produced flexibly by applying discontinuous feeding of agricultural feed stocks such as maize
silage, sugar beet silage and cattle slurry in continuous stirred tank reactors (CSTR) (Mauky et
al., 2014). However, the effect of discontinuous feeding on methanogenic pathways and
microbial communities were still unclear.
In this study, two laboratory-scale CSTRs were fed with distillers dried grains with solubles
(DDGS) under identical conditions except different feeding regimes (every 2 hours, once a day,
and every 2 days). The 2 hours feeding interval mimics the traditional full-scale biogas plants
operating on a semi-continuous substrate feeding mode to achieve nearly the same amount of
biogas and electricity generation throughout the day. The other feeding intervals (once a day and
every 2 days) assume flexible biogas production with the aim of high biogas production for few
hours after feeding and lower biogas production before the next feeding period. DDGS is a
byproduct of bio-ethanol production plants. It is produced from a fermentation waste product
known as stillage after passing through centrifugation and drying stages. Due to the reduction in
water content, DDGS can be easily transported to a place where it is used as animal feed. Since
DDGS drying accounts for approximately 30% of energy consumption of a bioethanol production
plant (Eskicioglu et al., 2011), the use of stillage prior to drying for biogas production has
environmental and economic benefits, especially where both plants are located in close proximity
to each other (Moestedt et al., 2013; Wilkie et al., 2000).
The aim of this study was to investigate flexible biogas production by changing feeding regimes
and the assessment of its effect on process performance, microbial community structure and
methanogenesis pathway of DDGS-fed reactors under mesophilic conditions. Bacteria and
archaea community structures were monitored by T-RFLP analysis of 16S rRNA and mcrA
genes, respectively. Carbon isotope signatures of methane and carbon dioxide were analyzed to
quantify the methanogenic pathways. Any correlation between methanogenic pathways studied
by isotope analysis and methanogenic community structure studied by molecular methods were
further assessed. Moreover, reactors fed at different feeding regimes were exposed to stress
condition by increasing OLR for short period to evaluate the tolerance level to environmental
perturbation.
123
2. MATERIALS AND METHOD
2.1 Substrate
All biogas reactors were fed with distillers dried grains with solubles (DDGS) and supplemented
with FerroSorp® DG (HeGo, Biotec, Germany) and trace element solution (TES). DDGS was
obtained from an industrial scale bio-ethanol plant (CropEnergies AG; Zeitz) with wheat as a
main raw material. The FerroSorp is a powdery substance containing iron hydroxide that was
added to the reactors to precipitate hydrogen sulfide. The TES contains (all are in g/L):
Ni(II).6H2O, 2.13; Co(II).6H2O, 0.531; NaMoO4.2H2O, 0.332; (NH4)6H2W12O40·xH2O, 0.423.
2.2 Reactor setup
Two laboratory-scale continuous stirred tank reactors (CSTR) were operated under mesophilic
conditions (38 ºC) with DDGS as the only substrate. The total volume of each reactor was 15 L
with a working volume of 10 L. Before the start of this study, the two reactors were operated
under identical conditions and then the contents of the reactors were mixed together to setup
equivalent starting conditions for this study. Three different feeding regimes were studied (once
daily, every 2 days and every 2 hours). DDGS was fed manually in daily and every 2 days
feeding regimes whereas a peristaltic pump was used for every 2 hours feeding The whole
experiment was divided into 4 phases: (1) reactors A and B were fed once daily (phase I, day 1-
29); (2) reactor A was fed once daily and B was fed every 2 hours (phase II, day 30-63); (3)
reactor A was fed once every 2 days and B was fed every 2 hours (phase III, day 64-107); (4)
reactor A was fed once every 2 days and B was fed every 2 hours but with higher OLR compared
to phase III (phase IV, day 108-118) (see Table 1). During phases I-III, the daily organic load of
both reactors under every feeding regimes was kept constant at 4 gVS L-1
d-1
. To achieve this
organic loading rate, the daily substrate was prepared by mixing 66.34 g DDGS, 2 mL TES and
2.56 g FerroSorp® DG and finally water was added to the mixture to make a total of 380 mL
solution. In phase IV, the OLR was increased continuously from 4 to 11 gVS L-1
d-1
at the same
rate for both reactors A and B. Hydraulic retention time (HRT) was kept constant at 26 days in
phases I-III. In phase IV, the HRT was 26 days for OLR ranged 5-7 gVS L-1
d-1
whereas HRT was
changed to 14 and 10 days at OLR of 9 and 11 gVS L-1
d-1
, respectively. Due to the difficulty to
pump the feed at OLR of 9 and 11 gVS L-1
d-1
in a total feed volume of 380 mL (higher total solid
compared to lower OLR feed), the feed was further diluted with water to a total daily feed
volume of 700 and 1000 mL and as a consequence, the HRT was shortened to 14 and 10 days,
respectively.
2.3 Basic process parameters
The volume of biogas was automatically recorded by a gas meter (TG 0.5, Dr.-Ing. Ritter
Apparatebau Gmbh & Co KG, Bochum, Germany). The volume of the biogas was corrected to
standard temperature (273.15 K) and pressure (101.325 kPa). The biogas vented from the gas
meter was collected in a gas bag and once the gas bag was full, all the biogas was transferred into
AwiFlex gas analyzer (Awite Bioenergie GmbH, Germany) for measuring the composition of the
biogas. The AwiFlex gas analyzer was equipped with optical infrared for determining the
concentration of CH4 and CO2 as well as electrochemical sensors for determining the
concentration of O2, H2, and H2S (detection limit 5,000 ppm).
124
Liquid effluent was periodically collected from all reactors and its pH was measured
immediately. It was then centrifuged at 10,000 rpm and 10 ºC for 12 min and the supernatant
solution was filtered through a sieve with mesh size of 1 mm. The supernatant solution was used
for further analysis of volatile fatty acids (VFA), volatile organic acids (VOA) and total inorganic
carbonate (TIC) as well as total ammonia nitrogen (TAN) expressed in gNH4+-N L
-1 (i.e. TAN is
defined as a sum of free ammonia nitrogen (FAN) plus ammonium nitrogen). For VFA analysis,
aliquot (5 mL) of the supernatant solution was transferred into a 20 mL vial followed by the
addition of 1 mL internal standard (2-ethylbutyric acid, 180 mg L-1
) and 1 mL phosphoric acid
(50% V/V). The vial was immediately sealed with butyl rubber stopper and aluminum crimp. The
concentration of VFA was determined from the headspace by a 5890 series II gas chromatograph
(Hewlett Packard, Palo Alto, USA) equipped with an HS40 automatic headspace sampler (Perkin
Elmer, Waltham, USA). HP-FFAP column (30m x 0.32 mm x 0.25 μm; Agilent Technology,
Germany) was used for chromatographic separation. The GC conditions were as described earlier
(Ziganshin et al., 2011). For TAN analysis, aliquot (125 µL) of the solution was diluted to 1:2000
with distilled water and the TAN was determined with a standard Nessler method using a
benchtop spectrometer (Hach-Lange DR 3900, Loveland, USA). Aliquot (10 mL) of the
supernatant solution was used for determining VOA and TIC using Titration Excellence T 90
titrator (Mettler-Toledo, Switzerland ) by titrating with 0.025-0.1 M H2SO4 in a pH range from
4.5 to 3.5 (Ziganshin et al., 2011). Total solid (TS) and volatile solid (VS) was analyzed
according to standard method (APHA, 2005) The VS reduction (VSR) was calculated based on
the gVS of the feed and effluent according to the following equation: VSR= [1- (gVS of
effluent)/gVS of feed)] x 100% .
2.4 Stable isotope analysis
Biogas samples were periodically collected from the headspace of the reactor in triplicate and
stored in a gas tight evacuated vial (20 mL) until further analysis. The stable isotope analysis of 13
C/12
C gas samples was performed using a gas chromatography combustion isotope ratio mass
spectrometery (GC-C-IRMS) system consisting of a gas chromatograph (HP 6890 Series, Agilent
Technology, Santa Clara, USA) coupled with IRMS (Finnigan MAT 253, Thermofinnigan,
Bremen, Germany) via a combustion interface. For the GC separation of CH4 and CO2, gas
samples (30 µL) were injected manually into the GC instrument equipped with a CP-
Porabond Q column (50 m × 0.32 mm × 0.5 μm, Varian, USA). The column temperature was
kept constant at 40 ºC and flow rate of Helium was 2 mL/min for carbon.
The analysis of the δ13
C of biomass (DDGS) was performed with an elemental analyzer (EA)-
IRMS system consisting of an EA (Euro EA, HEKAtech GmbH, Wegberg, Germany) and an
IRMS ((Finnigan MAT 253, Thermofinnigan, Bremen, Germany), coupled via an interface
(ConFlo III; Thermo Fisher Scientific).
The stable carbon isotope data was reported in delta notation (δ13
C) in parts per thousand (‰)
unit versus the Vienna Pee Dee Belemnite (V-PDB) standard: δ13
C= [(Ra)sample/(Ra)standard -
1]*103 (‰), where Ra is the
13C/
12C (Whiticar, 1999).
125
The relationship among the 13
C isotopic signature of total CH4 (δ13
CH4), acetate-derived CH4
(δma) and CO2-derived CH4 (δmc) can be described using the following mass balance equation
(Conrad, 2005):
δ13
CH4 = fmc* δmc + (1- fmc)* δma (1)
where fmc is the fraction of CH4 produced from the reduction of CO2 via hydrogenotrophic
methanogenesis.
To estimate the fmc, the values of δma and δmc need to be known beforehand. Therefore, additional
batch experiment was conducted using inoculum obtained from both CSTRs (A and B) in phase
III on day 72 (see supporting information, SI). Fluoromethane (CH3F) was added into batch
bottles to inhibit acetoclastic methanogenesis. Controls without inhibitor were incubated in
parallel to CH3F-added incubations. Any methane production in the CH3F-added (acetoclastic
inhibited) batch incubation is, therefore, due to hydrogenotrophic methanogenesis, which was
used to estimate the value of δmc. The δ13
C–CH4 in the CH3F-added incubation was on average -
68.5‰ (data not reported), and thus, the δmc was equal to -68.5‰. Then, the δma was estimated
using equation 2:
δma = δac + εma (2)
where εma is the isotopic enrichment by acetoclastic methanogenesis. In pure culture when
substrate is not limiting, the isotope fractionation factor for the acetoclastic methanogenesis alone
can reach up to 1.027 (equivalently isotope enrichment factor (εma) of -27‰) for the genus
Methanosarcina (Goevert & Conrad, 2009), which was the dominant acetoclastic methanogens in
all of our DDGS reactors (Figure 4). Due to the fact that the εma value is affected by the
concentration of acetate (Goevert & Conrad, 2009) and since the concentration of acetate in all of
the DDGS reactors were generally low (less than 365 mg L-1
in our DDGS reactor versus 1200
mg L-1
of acetate used in the previous study (Goevert & Conrad, 2009)), we assumed a low
fractionation factor (εma = -10‰). Since the carbon isotope fractionation between biomass (δ13
DDGS = -25.5‰) and fermentatively produced acetate is negligible (Conrad, 2005) as well as the
fact that the δ13
of biomass may be used as a proxy for mass balance calculations (Conrad et al.,
2014), we assumed the same value for 13
C isotopic signature of acetate (δac) as the biomass (i.e,
δac = -25.7‰). Using this assumption and equation 2, the value of δma was estimated to be
-35.7‰. This value is more realistic since δ13
C–CH4 value of -40‰ was observed in the control
batch incubation with DDGS as a monosubstrate (data not reported). We also considered a high
isotope fractionation factor (for instance, εma = -21‰) and the calculated δma value was -46.7‰ in
this case. Since both acetoclastic and hydrogenotrophic methanogens were represented in the
control batch incubation, the δ13
C value of the produced CH4 should be more negative than the
one produced exclusively through AM (i.e more negative than -46.7‰). However, the observed
δ13
C–CH4 value of -40‰ in the control incubation showed that our assumption of high
fractionation factor (εma = -21‰) must be wrong.
To summarize, the fractions of hydrogenotrophic methanogenesis to methane production in the
CSTRs were estimated using equation 1, the measured values of δ13
CH4 in the CSTRs, δmc = -
68.5‰ and δma = -35.7‰.
126
2.5 DNA extraction and purification
Samples for microbial community analyses were taken right before substrate feeding during all
experimental phases (phases I-IV). During each phase, liquid samples were collected at two
different sample points (around two weeks when the new phase began and right before starting
the next phase) (days 16 and 29 in phase I; days 37 and 59 in phase II; days 81 and 107 in phase
III; and day 115 in phase IV). Samples were collected into sterile Eppendorf tubes and stored in a
freezer (-80 ºC) until further analysis. DNA was isolated in duplicate from 400 mg of the sample
with a Nucleospin® Soil Kit (MACHEREY-NAGEL Gmbh & Co KG, Düren, Germany)
according to the manufacturer’s instructions and its quality was checked with 1.5% agarose gel
electrophoresis and quantified with a NanoDrop® ND-1000 UV/Vis spectral photometer
(PeqLab, Germany).
2.6 Polymerase chain reaction
Bacterial 16S rRNA gene fragments were amplified by PCR using universal bacterial primers
27F (5′-AGAGTTTGGATCMTGGCTCAG-3′) and 1492R (5′-
TACGGYTACCTTGTTACGACTT-3′) as described earlier (Ziganshin et al., 2011) whereas
mcrA genes were amplified by using the mcrA/mrtA specific forward primer mlas and the
reverse primer mcrA-rev as described earlier (Nikolausz et al., 2013). For terminal restriction
fragment length polymorphism (T-RFLP) analysis, the same primers were used for PCR
amplification except the reverse primer was 5’-labeled with phosphoramidite fluorochrome 6-
carboxyfluorescein. PCR products were checked in 1.5% gel electrophoresis. Then the PCR
products were purified using a SureCleanPlus kit (Bioline, Germany) and quantified using
NanoDrop® ND-1000 UV/Vis spectral photometer (PeqLab, Germany).
2.7 T-RFLP analysis based on 16S rRNA and mcrA genes
T-RFLP analysis was performed as described before (Nikolausz et al., 2013). Briefly, bacterial
16S rRNA genes were subjected to restriction enzyme digestion with either the restriction
endonucleases HaeIII or MspI (New England Biolabs, Schwalbach, Germany)
whereas mcrA genes were digested with BstNI (New England Biolabs, Schwalbach, Germany).
After the purification of the terminal restriction fragments (T-RFs), it was resuspended in HiDi
formamide containing GeneScan-500 ROX standard (Applied Biosystems, Weiterstadt,
Germany) and MapMarker 1000 standard (Bioventures Inc., Murfreesboro, TN, USA) for mcrA
and bacterial 16S rRNA genes analyses, respectively. Fluorescently labeled T-RFs were
separated using capillary electrophoresis with an automatic sequencer ABI PRISM 3130xl
Genetic Analyzer (Applied Biosystems, Weiterstadt, Germany). T-RFLP data were retrieved by
comparison with the internal standards using GeneMapper V3.7 software (Applied Biosystems,
Weiterstadt, Germany). T-RFs in the size ranges of 50-500 and 50-1000 bp were considered for
subsequent analysis of the archaeal mcrA and bacterial 16S rRNA genes, respectively. For
statistical analysis, T-RFLP data sets were reduced by removing low abundance T-RFs below
1%. Relative T-RF abundances were calculated by dividing individual T-RF peak area to the total
peak areas. Linking taxonomic information to the major mcrA T-RFs were done by using the
sequence and T-RFLP database from previous studies of our group (Lv et al., 2014; Nikolausz et
al., 2013).
127
2.8 Statistical analysis
Multivariate statistical analysis were performed by using the “vegan” package of R version 3.0.1
(Oksanen, 2011) applying Bray-Curtis dissimilarity indices, which includes relative abundance
information in addition to presence and absence of T-RFs. The multidimensional dissimilarity
matrix was reduced to two dimensions showing the dissimilarities between community structures
of samples as distances on the plot. The major process parameters correlating with community
composition were fitted using the “envfit” algorithm. Significance of single process parameters
on the non-metric multidimensional scaling ordination (nMDS) results were tested using a
Monte-Carlo test with 1000 permutations and only significant parameters are shown (p < 0. 05).
3. RESULTS AND DISCUSSION
3.1 Process performance
The effect of change in feeding regimes (feeding every 2 hours, once per day and every second
days) on biogas performance was studied for DDGS treating laboratory-scale CSTRs. Basic
process parameters were monitored over 128 days during four experimental phases (Figure 1).
During phase I, both reactors A and B were operated under identical conditions (fed once per day
and OLR was 4 gVS L-1
d-1
) to test the homogeneity of the two reactors before proceeding to the
next phase where different feeding regimes were employed. During phases II-IV, both reactors
(A and B) were operated under identical conditions except the feeding intervals were different.
The reactor characteristics and the steady-state data are summarized in Table 1. The steady-state
data represent the average value of the process parameters within phase’s monitored over the last
two weeks. Since both A and B were operated under identical conditions with daily feeding
during phase I, process performance was almost identical (Table 1). This result demonstrated that
the process performance of both reactors was almost identical before employing any change in
feeding intervals in subsequent experimental phases.
128
Data (row no)
23-36
This is Final Table
Phase 1
10
20
30
40
50
60
0
10
20
30
40
50
60
70
80
0 10 20 30 40 50 60 70 80 90 100 110
Meth
ane
(LN
d-1
)
Bio
gas
(LN
d-1
)
A-Biogas B-Biogas A-Methane B-Methane
0
200
400
600
800
1000
0
2
4
6
8
10
12
0 10 20 30 40 50 60 70 80 90 100 110
SM
P, S
BP
(mL
gV
S-1
)
OL
R (
gV
S L
-1d
-1)
A/B-OLR A-SBP B-SBP A-SMP B-SMP
7.0
7.5
8.0
8.5
9.0
0
1
2
3
4
5
0 10 20 30 40 50 60 70 80 90 100 110
pH
TA
N (
gN
H4
+-N
.L-1
)
Days
A-TAN B-TAN A-pH B-pH
0
20
40
60
80
100
50
55
60
65
70
0 10 20 30 40 50 60 70 80 90 100 110
Hydro
gen (
ppm
)
Meth
ane (%
)
A-Methane B-Methane A-Hydrogen B-Hydrogen
(a)
Phase I Phase II Phase IIIPhase
IV
129
Figure 1 Long-term process parameters for DDGS treating reactors monitored over phase I-IV
(a) organic loading rate (OLR), specific methane production (SMP), specific biogas production
(SBP), biogas and methane production, methane content, concentration of hydrogen, total
ammonia nitrogen (TAN) and pH; (a) total volatile fatty acids (VFA), acetate, propionate,
butyrate, volatile organic acids (VOA), volatile organic acids to total inorganic carbon ratio
(VOA/TIC ratio). The value for butyrate is the sum of the concentrations of n-butyrate and iso-
butyrate. Reactors A and B were fed once daily in phase 1; reactor A was fed daily and B was fed
every 2 hours in phase II; reactor A was fed once every 2 days and B was fed every 2 hours in
phase III; reactor A was fed once every 2 days and B was fed every 2 hours in phase IV but with
higher OLR compared to phase III.
Phase 3
Phase 4
Phase 1
(1-29 days)
Phase 2
(30-63 days)
Phase 3
(64-107 days)
Phase 4
0
0.25
0.5
0
0.5
1
1.5
2
0 10 20 30 40 50 60 70 80 90 100 110
VO
A/T
IC r
ati
o (
gV
OA
gC
aC
O3
-1)
VO
A(g
L-1
)
Days
A-VOA B-VOA A-VOA/TIC ratio B-VOA/TIC ratio
0
20
40
60
80
100
0
20
40
60
80
100
120
140
0 10 20 30 40 50 60 70 80 90 100 110
Ace
tate
(m
g L
-1)
VF
A (
mg
HA
C e
q.L
-1) A-VFA B-VFA A-Acetate B-Acetate
0
1
2
3
4
0
2.5
5
7.5
10
12.5
15
0 10 20 30 40 50 60 70 80 90 100 110
Bu
tyra
te (
mg
L-1
)
Pro
pio
na
te (
mg
L-1
) A-Propionate B-Propionate A-Butyrate B-Butyrate
(b)Phase I Phase II Phase III
Phase
IV
130
Table 1 DDGS reactors operating conditions and process parameters
Reactors
OLR
(gVS
L-1
d-
1)
Feeding
mode Total
CH4
(LN
d-1
)b
Total
biogas
(LN d-
1)
b
SMP
(mLN
gVS-
1)
b
SBP
(mLN
gVS-
1)
b
CH4
(%)
H2
(ppm)
pH
VOA
(g L-
1)
VOA/TIC
ratio
(gVOA
gCaCO3-
1)
TAN
(gNH4+-
N L-1
)
Acetate
(mg L-
1)
VFA
(mgHAC
eq. L-1
)
VSR
(%)
Phase
I A 4 Daily 16 28 402 710 57 59 7.82 1.3 0.1 3.5 58 67 62
(1-29
days) B 4 Daily 16 28 400 706 57 68 7.86 1.3 0.1 3.6 52 61 63
Phase
II A 4 Daily 16 29 414 724 57 61 7.81 1.1 0.1 3.4 39 55 61
(30-
63
days) B 4
Every 2
h 14 25 363 622 58 32 7.78 1.1 0.1 3.3 43 58 59
Phase
III A 4
Every 2
d 16 29 402 724 58 41 7.78 1.1 0.1 3.1 43 52 62
(64-
107
days) B 4
Every 2
h 14 24 354 613 58 9 7.71 1.1 0.1 2.6 28 35 59
Phase
IV A 5-11
Every 2
d 29 51 387 689 60 75 7.76 1.2 0.1 3.3 54 62 68
(108-
118
days) B 5-11
Every 2
h 23 40 304 526 57 20 7.67 12 0.1 2.8 45 59 67 aThe process parameters for phases I-III are the average values of the last two weeks of each phase. Process parameters in phase IV are
the average values of 10 days reactors operation under stress condition (when OLR was increased from 5 to 11 gVS L-1
d-1
). OLR, organic
loading rate; VS, volatile solid; VSR, volatile solid reduction; SMP, specific methane production; SBP, specific biogas production; VOA,
volatile organic acids; TIC, total inorganic carbon; TAN, total ammonia nitrogen which is defined as a sum of free ammonia nitrogen
(FAN) plus ammonium nitrogen; VFA, volatile fatty acids; HAc eq., acetic acid equivalent of all VFA bData was corrected to standard temperature (273.15 K) and pressure (101.325 kPa)
131
Since the feeding interval was not changed for reactor A during phase II, the daily fed reactor A
performed almost the same as in phase I (average biogas and methane production were 29 and 16
LN d-1
, respectively; SMP and SBP were 414 and 724 mLN gVS-1
) whereas the every 2 h fed
reactor B showed a reduced performance (biogas production, CH4 production, SMP and SBP
were 25 LN d-1
, 14 LN d
-1, 363 mLN gVS
-1and 622 mLN gVS
-1, respectively). Process efficiency
significantly deteriorated in the reactor B when feeding was changed from daily to every 2 hours
(14 and 16% of less CH4 and biogas production, respectively, p <0.05). The volatile solid
reduction (VSR) was 61 and 59% for reactors A and B, respectively, showing more
biodegradable organic matter remained unutilized in the reactor B compared to the reactor A. The
observed higher biogas and CH4 production in reactor A could partly be due to the better VS
utilization. Interestingly, there was no appreciable difference between reactors A and B regarding
other process parameters (pH 7.8, VOA 1.1 g L-1
, VOA/TIC 0.1 gVOA gCaCO3-1
, TAN 3.3-3.4
gNH4+-N L
-1). The concentration of residual VFA was similar in both reactors (acetate 40-43 mg
L-1
and total VFA 58-60 mgHACeq. L-1
). The average daily biogas composition remained almost
the same in both reactors (CH4 57-58%, CO2 43-42% and H2 32-61 ppm). However, the biogas
composition and the concentration of VFA varied a lot between two feeding events for the daily
fed reactor whereas it remained constant for the every 2 h fed reactor (see the discussion below).
In phase III, feeding interval was changed from once per day to once every 2 days in reactor A
whereas the feeding interval of reactor B was kept the same at every 2 h interval. Process
efficiency was almost the same in the daily (phase II) and every 2 days (phase III) fed reactors
(Table1). On the other hand, SBP, SMP, CH4 and biogas production was still higher by 16-18,
13-14 and 16-18 % (p <0.05) in the reactors fed at longer intervals (daily and every 2 days)
compared to the every 2 hours fed reactor. The concentrations of TAN, H2, acetate and total VFA
were significantly higher in the every 2 days fed reactor than in the every 2 h fed reactor (p
<0.05). The higher TAN level in the every 2 days fed reactors implies that the protein fraction of
the DDGS was better hydrolyzed and fermented in subsequent steps. However, free ammonia has
not reached an inhibitory level. Despite the significant difference between the concentrations of
total VFA and acetate in both reactors, the residual acetate and VFA concentrations were low
(acetate 29-40 mg L-1
and total VFA 30-53 mgHACeq. L-1
). This shows, irrespective of the
feeding intervals, the VFA production rate by fermentative bacteria and subsequent production of
acetate by acetogenic bacterial were well balanced with the consumption rate of the acetate by
methanogens and syntrophic acetate oxidizing bacteria (SAOB). The pH of all reactors under all
phases remained stable and ranged from 7.7 to7.8, which is optimal for the biogas process (De
Vrieze et al., 2012; Weiland, 2010).
In phase IV, short-term stress test was conducted by gradually increasing the OLR from 4 to 11
gVS L-1
d-1
within ten days (Figure 1). Biogas production increased significantly in both reactors
with an increase in OLR. However, the SBP and SMP reduced in both reactors with the OLR
increase in phase IV (stress condition). Compared to phase III (steady state), SBP and SMP
reduced by only 4.8 and 3.6% in the every 2 days fed reactor , respectively whereas it was even
much lower in the every 2 hours fed reactor (lower by about 14%). Despite the reduced biogas
and methane yield in both reactors, the every 2 days fed reactor still performed much better than
the every 2 h fed reactor under the stress condition. SBP, SMP, methane and biogas production
was higher by 31, 27, 27 and 31% on average in the every 2 days fed reactor compared to the
every 2 hours fed reactor. Nevertheless, both reactors maintained a stable biogas process under
stress condition as demonstrated with optimum pH range and low residual VFA, VOA and
132
VOA/TIC ratio. VOA/TIC values below ca. 0.25 imply that the process is stable without
indication of stress (Munk & Lebuhn, 2014).
In addition to the long-term process parameters presented above, short-term process parameters
were also monitored at certain time points for the periods of 48 hours in both reactors in phase III
(Figure 2). Biogas production was highly dynamic for the every 2 days fed reactor whereby
highest biogas production rate (4.6 LN h-1
) was observed within one hour after feeding and the
lowest value was 0.5 LN h-1
right before the next feeding event (Figure 2A). However, hourly
biogas production rate was almost constant for the every 2 hours fed reactor. In the every 2 days
fed reactor, almost 50% of the total biogas produced within 48 hours was generated within the
first 12 hours whereas only 25% of the biogas was generated during the same time in the every 2
hours fed reactor. If an electricity demand is assumed for the first 12 hours after the feeding of
reactor A and no demand between 13 and 48 hours, then 50% and 75% of the biogas generated
by reactors A and B, respectively, need to be stored intermittently during the second period. This
simplified calculation demonstrates that dynamic feeding of the biogas process could allow either
a higher flexibility of the electricity production or a lower demand for storage capacity and thus
save extra investment. More rigorous calculations and different scenarios need to be considered
for practical implementations (Mauky et al., 2014).
133
0
1
2
3
4
5
0 4 8 12 16 20 24 28 32 36 40 44 48
Pro
du
ctio
n (
LN h
-1) A-Biogas B-Biogas
A-CH4 B-CH4
(a)
0
20
40
60
0 4 8 12 16 20 24 28 32 36 40 44 48
Acc
um
ula
ted
bio
gas
(LN)
A-Biogas B-Biogas
0
25
50
75
100
125
150
40
45
50
55
60
65
70
0 4 8 12 16 20 24 28 32 36 40 44 48
H2 (
pp
m)
CH
4 (
%)
Hours
A-CH4 B-CH4
A-H2 B-H2
134
(b)
4
9
14
-60
-55
-50
-45
-40
0 4 8 12 16 20 24 28 32 36 40 44 48
δ1
3C
-CO
2 (‰)
δ1
3C
-CH
4 (‰)
A-δ13C-CH4 B-δ13C-CH4
A-δ13C-CO2 B-δ13C-CO2
1.050
1.055
1.060
1.065
1.070
1.075
1.080
0.2
0.4
0.6
0.8
0 4 8 12 16 20 24 28 32 36 40 44 48
f mc
A-fmc B-fmc
A-αmc B-αmc
0
100
200
300
400
500
0
100
200
300
400
500
600
700
800
0 4 8 12 16 20 24 28 32 36 40 44 48
Ace
tate
(m
g L
-1)
VF
A (
mg L
-1)
A-VFA B-VFA
A-Acetate B-Acetate
0
2
4
6
8
10
12
14
0
50
100
150
200
250
300
350
400
0 4 8 12 16 20 24 28 32 36 40 44 48
Bu
tyra
te (
mg L
-1)
Pro
pio
nate
(m
g L
-1)
Hours
A-Propionate B-Propionate
A-Butyrate B-Butyrate
135
Figure 2 Short-term process parameters and carbon isotope signatures for the every 2 days
(reactor A) and 2 hours (reactor B) fed DDGS reactors during phase III (a) biogas production,
methane production, accumulated biogas production, CH4 content and concentration of H2; (b) 13
C isotope signature of CH4 and CO2, fraction of methane produced through hydrogenotrophic
methanogenesis (fmc) and concentrations of total VFA, acetate, propionate and butyrate.
For reactor A (every 2 days fed reactor), the CH4 content was 60% right before feeding and then
reduced to a minimum of 47% within 4 h and finally returned to 60% and remained at 60%
afterwards (Figure 2A). The quality of the biogas in terms of combustibility was lower with in 4
h after feeding because of the lower methane content in a biogas. However, in a full-scale plant
biogas is collected for several hours in a storage tank. This allows biogas having different
composition to get mixed in the storage tank before it will be used for electricity generation in
CHP unit (Mauky et al., 2014). Therefore, the quality of the biogas that reaches the CHP unit is
better than the biogas generated within 4 h after feeding. For reactor A, hydrogen was always
higher than 18 ppm and the maximum value (124 ppm) was recorded at 1 h after feeding. The
concentration of total VFA changed over wider ranges of 35-664 mgHACeq. L-1
as well as
acetate, propionate and butyrate ranged 30-343, 2-278 and 0.2-11.5 mg L-1
, respectively (Figure
2B). Due to the simultaneous production and consumption of VFA, the amounts of individual
VFA produced by hydrolysis and fermentation steps are even higher than the one reported here.
On the other hand, biogas process was less dynamic in every 2 h fed reactor with biogas
production rate (0.8-1.6 LN h-1
), almost constant CH4 composition (59%), narrow range for total
and individual VFA (total VFA 21-56 mgHACeq. L-1
, acetate 15-41 mg L
-1, propionate 2.2-5.6
mg L-1
and butyrate < 1.8 mg L-1
) and lower hydrogen concentration (7-40 ppm). In both
reactors, the average methane content in 48 h was similar (58-60%, Table 1). Moreover, the daily
fed reactors investigated in phases I-II were also more dynamic with regards to biogas and VFA
production compared to the every 2 hours fed reactor (data not shown).
3.2 Stable isotope signatures and methanogenic pathways
The carbon isotopic signature (δ13
C) of CO2 and CH4 were monitored over long time (118 days).
The long term isotope signatures of CO2 and CH4 were obtained from the gas samples collected
right before feeding (Figure 3). δ13
C-CH4 values can be used to differentiate between acetoclastic
methanogenesis (AM) and hydrogenotrophic methanogenesis (HM) pathways, with more
depleted values being indicative for the dominance of HM (Conrad, 2005). As described in the
experimental section, the isotope signature of CH4 was used to estimate the fraction of methane
produced through HM (fmc). Interestingly, for all feeding regimes, the δ13
C-CH4 remained almost
constant (-48.2 to -51.8 ‰) throughout all experimental phases (Figure 3A). The same is true for
δ13
C-CO2, which remained in a narrow range of 7.6 to 11.4 ‰ for both reactors. The fraction of
methane derived from H2/CO2 (fmc) ranged from 44 to 53 % for both reactors considering the
values measured before feeding, indicating almost equal contribution of both pathways to
methane production irrespective of the different feeding intervals.
136
Figure 3 Long-term carbon isotope signatures of CH4 and CO2 over phases I-IV (a)
13C isotope
signature of CH4 and CO2; (b) fraction of methane produced through hydrogenotrophic
methanogenesis (fmc).
Short-term carbon isotope signatures were also monitored for the periods of 48 hours during
phase III (Figure 2B). In the every 2 days fed reactor ( reactor A), δ13
C-CH4 was -48 ‰ right
before feeding and started to get depleted in 13
C after feeding and reached a minimum of -59.4 ‰
within 2 hours after feeding, showing HM was dominating during this time interval (fmc at 2 h
was 0.72). The subsequent increase in δ13
C-CH4 from the lowest -59.4 ‰ to the highest -46.1 ‰
at 9 h, shows the shift from HM to AM over time (fmc=0.32 at 9 h). Finally, δ13
C-CH4 stabilized
at -49 ‰ and fmc was 0.49 right before the next feeding, showing almost equal contribution of
both pathways to the CH4 being produced at this time point. δ13
C-CO2 was 9.0 ‰ right before
feeding and depleted in 13
C to the lowest value of 6.3 ‰ at 6 h and later on became heavier and
finally reached to the same value as before the feeding. This temporal trend in isotope signatures
were not observed in the every 2 hours fed reactor (reactor B). The δ13
C-CH4 and δ13
C-CO2 of
every 2 hours fed reactor remained at a narrow range of -49 to -52 and 8.2 to 9.4 ‰, respectively,
indicating the less dynamic nature of the process. The fmc for the every 2 h fed reactor ranged
between 0.41 and 0.49, showing almost equal contribution of both pathways to methane
production. To estimate the average CH4 production through HM between two feeding events
(within 48 hours) in the every 2 d fed reactor, we multiplied the amount of CH4 production and
fmc value at each time points within 48 hours. The result showed that on average about 42% of the
CH4 being produced through HM. To summarize, the methane production in 48 hours was almost
equally contributed by both methanogenesis pathways irrespective of the feeding intervals.
4
6
8
10
12
14
-60
-55
-50
-45
-40
-35
0 10 20 30 40 50 60 70 80 90 100 110
δ1
3C
-CO
2(‰
)
δ1
3C
-CH
4(‰
)A-δ13C-CH4 B-δ13C-CH4
A-δ13C-CO2 B-δ13C-CO2
0.0
0.2
0.4
0.6
0.8
0 10 20 30 40 50 60 70 80 90 100 110
f mc
Days
A-fmc B-fmc
(a)
(b)
137
3.3 Microbial community structure
Community members with the relative abundance of 1% or less were excluded from the T-RFLP
analysis. The community structure of bacteria and methanogenic archaea was studied with T-
RFLP profiles of 16S rRNA and mcrA/mrtA functional genes, respectively. The results from
duplicate analysis showed the high reproducibility of the method both for the presence and the
relative abundance of terminal restriction fragments (T-RFs ) (Figure 4). The diversity of the
methanogens in all reactors was not much affected by the different feeding regimes and it can be
described with the dominance of the genus Methanosarcina (69-83%), followed by the genus
Methanobacterium (11-31%). Low abundance of the genus Methanoculleus was detected during
all phases except phases III and IV. Methanosaeta was detected in few occasions with very low
abundance. Moreover, minor contributions of other methanogens were detected by T-RFs not
matching our database clone sequences.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Rela
tive
ab
un
dan
ce
52_53_m Methanosarcina Methanoculleus Methanosaeta
344 407-409_m Methanobacterium
Figure 4 T-RFLP profiles of mcrA/mrtA functional genes of archaeal community structure.
Sample name consists of four variables: the capital letters (A and B) refer to reactors A and B, the
number next to the letters indicate the phase number, number at the far right side represent the
sampling time points and the small letters (a and b) refer to duplicate samples.
138
The resulting bacterial T-RFLP profiles of all reactors were characterized by a large number of
distinct peaks, a more diverse and varied bacterial groups (supporting information, Figure S1).
Figures S1A and B show the T-RFs pattern determined by the restriction endonuclease enzymes,
HaeIII and MspI, respectively. A total of 37 and 43 different T-RFs were observed using HaeIII
and MspI enzymes, respectively. In general, the HaeIII data set yielded similar pattern as MspI
data set.
Further insight into the relationship between the observed bacterial community patterns and the
reactor parameters (pH, TAN, total VFA, acetate, propionate, butyrate, SBP, VSR, CH4 content
and production, CO2 content, H2, biogas production, VOA and VOA/TIC ratio) was investigated
more extensively by nMDS (Figure 5). The bacterial community structure was investigated from
a total of 7 samples collected at two different sampling points during phases I-III and one
sampling point during phase IV. The community structure was similar in both HaeIII and MspI
data set but the correlation with the reactor parameters was not that strong in case of MspI.
Except phase I, there is a clear segregation between the bacterial community structures of both
reactors fed at different feeding regimes. Those samples (in phase I) from similar feeding pattern
clustered together as shown in the plot. The clear segregation between the two reactors during
phases II –IV and the clustering of both reactors during phase I showed that the bacterial
community structure was clearly affected by the change in feeding regimes. Moreover, samples
from the same feeding regime over time were more similar to each other in community structure
than those samples from the same reactor fed at different feeding regimes. A statistical
comparison between bacterial community structure and reactor parameters suggested that the
observed T-RFLP patterns (generated using HaeIII enzyme) was best explained by the difference
in the pH value and concentrations of TAN and H2 and pH (statistically highly significant, p <
0.01) as well as by the SBP and concentrations of total VFA and acetate to some extent
(statistically significant, p < 0.05). On the other hand, the pH value and concentrations of H2,
TAN, total VFA and acetate were relatively the main reactor parameters (p < 0.05) to explain the
observed bacterial community patterns (generated using MspI enzyme).
139
NMDS1
NM
DS
2(a)
140
Figure 5 Non-metric multidimensional scaling (nMDS) ordination of bacterial community
TRFLP data for 14 samples collected from two DDGS fed reactors (reactors A and B) over
phases I-IV (a) with HaeIII and (b) MspI restriction enzyme digestion. Statistically highly
significant (P < 0.01) and significant (P < 0.05) reactor parameters correlating with the
community structures are indicated by blue and magenta solid arrows, respectively. The direction
of the arrows show the correspondence to the community structures and the length of the arrow
indicates the strength of the correlation with the ordination axis. Samples from reactors A and B
are indicated in open circles and triangles with digester name displayed. Sample name consists of
three variables: the capital letters (A and B) refer to reactors A and B, the number next to the
letters indicate the phase number and number at the far right side represent the sampling time
points.
NMDS1
NM
DS
2(b)
141
3.4 Effects of change in feeding intervals on methanogenic pathways and microbial
communities
The higher diversity of the bacteria compared to the methanogens community was not surprising
as they are involved in several steps of biomass degradation from hydrolysis up to acetogenesis
whereas the archaea are responsible in the methanogenesis pathways. Moreover, the bacterial
community varied a lot under the different feeding regimes (Figure 5), while methanogens were
only indirectly impacted. Our results were in agreement with a previous study which
demonstrated that different bacterial community structures were observed under different feeding
regimes (De Vrieze et al., 2013). Further research on the identity of the bacterial communities
will provide a better insight into the role of each functional bacterial community.
Unlike bacterial community structure, methanogenic community structure remained stable under
all feeding intervals. The most predominant methanogens in all the DDGS reactors were
affiliated to the genus Methanosarcina followed by the genus Methanobacterium. Members of
the genus Methanobacterium (strictly hydrogenotrophic methanogens) has been widely detected
in biogas digesters treating agricultural waste products (Nettmann et al., 2010; Ziganshin et al.,
2011). Methanosarcina is a multipotent methanogenic group with members capable of utilizing
acetate, H2/CO2, methanol, and methylamines as well as metabolic possibility for acetate
oxidation to CO2 and H2. Whereas the strict acetoclastic methanogens (Methanosaeta) were
detected in only few occasions with very low abundance. This showed that both
hydrogenotrophic and acetoclastic methanogens are represented in both reactors. Carbon isotope
results were in agreement with the observed methanogenic community whereby on average both
pathways almost equally contributed to the daily methane production. Such a high methane
production through HM and Methanosarcina dominating community were previously reported in
biogas digesters where acetate was converted to methane through an alternative pathway in
addition to the well-known AM (Mulat et al., 2014; Nettmann et al., 2010). This alternative
pathway involves a syntrophic relationship between hydrogenotrophic methanogens and
syntrophic acetate-oxidizing bacteria (SAOB) where the latter oxidizes the acetate to CO2 and H2
and the former reduces the CO2 to methane. Since syntrophic acetate oxidation is endergonic
under standard condition, hydrogenotrophic methanogens need to maintain a low H2 partial
pressure in order for the reaction to be energetically favorable.
The genus Methanosarcina has higher threshold for acetate in compared to the genus
Methanosaeta, which gives a competitive advantage for Methanosarcina to dominate in the
presence of high acetate concentration (Karakashev et al., 2005; Mulat et al., 2014). Within two
feeding events, accumulated transient acetate concentration in the reactors fed at longer feeding
interval was very dynamic (50-343 mg/L) compared to every 2 hours fed reactor. Despite the
difference in concentration of acetate between the reactors, all the reactors were dominated with
Methanosarcina. Therefore, the dominance of the Methanosarcina in both reactors cannot be
explained by the concentration of acetate alone. Instead, the high abundance of Methanosarcina
could be better explained in terms of the concentration of TAN. It is well known that
hydrogenotrophic methanogens and Methanosarcina is more tolerant to higher ammonia level.
Whereas members of the genus Methanosaetaare sensitive to ammonia (Karakashev et al., 2005)
and may no longer be detected at TAN concentration exceeding 2.5 gNH4+-N L
-1 (De Vrieze et
al., 2012; Nettmann et al., 2010). The abundance of Methanosarcina and Methanobacterium was
most likely due to their tolerance to high TAN concentration in our DDGS reactors (2.6-3.6
142
gNH4+-N.L
-1). Our findings are in accordance with previous findings where Methanosarcina
dominated DDGS reactors operated at high OLR of 5 gVS L-1
d-1
with TAN concentration of
2.94 gNH4+-N L
-1 whereas Methanosaeta dominated at low OLR of 2 gVS L
-1d
-1 with TAN
concentration of 1.82 gNH4+-N L
-1 (Ziganshin et al., 2011).
Interestingly, methanogenic pathways changed very much within two feeding events in the less
frequently (once per day and every 2 days) fed reactors compared to the every 2 hours fed
reactors despite the existence of identical methanogenic community in both reactors. As shown in
Figure 2, the maximum concentration of H2 and the highest contribution of HM to CH4
production was observed at 2 h after feeding for the every 2 days fed reactor whereas the highest
contribution of AM to CH4 production and maximum concentration of acetate were observed
around 9 hours after feeding. This temporal variation in methanogenesis pathway was influenced
by the availability of fermentation products (acetate, H2/CO2) to methanogens, which is in
agreement with maize silage fed reactors in previous study (Lv et al., 2014). On the other hand,
temporal variation in methanogenesis was minimal for the every 2 hours fed reactors and both
pathways contributed almost equally over the course of the day. As a result of the short feeding
interval in the every 2 hours fed reactor, perhaps the methanogens were not limited by any one of
the fermentation products (H2, CO2 and acetate).
3.5 Effect of change in feeding intervals on process performance
Despite the difference in feeding intervals, both reactors were in stable condition without any
accumulation of VFA. The less frequently fed reactor performed much better with regards to
methane and biogas production under stress condition (increased OLR). Our results are in
agreement with previous study where a less frequently fed reactor appeared to have higher degree
of functional stability and even more tolerance to an organic shock load (De Vrieze et al., 2013).
However, to our knowledge this is the first report that methane yield was significantly higher by
about 14% in the less frequently fed reactor compared to the more frequently fed one.
Two possible explanations can be provided for the significantly higher methane yield in the less
frequently fed reactor. Since there is no accumulation of VFA in any reactors, the difference in
methane yield is governed by the first two stages of anaerobic digestion (hydrolysis and
acidogenesis steps). One is the difference in average time the freshly added substrate spends in a
reactor before effluent was taken out from the reactor. For daily, every 2 days and every 2 hours
fed reactors, effluents were taken out once per day, in two days and in 2 hours intervals. This
means for substrate containing slowly and easily degradable fractions, there is a high probability
of removing the slowly degradable fraction of the substrate before degradation completes in the
every 2 hours fed reactor. Whereas the slowly degradable fraction spends more time before it is
taken out of the daily and every 2 days reactors, which can give at least considerable time for
microorganisms to solubilize the substrate. This is a possible explanation based on the fact that
the chemical composition of particulate substrates such as DDGS, maize silage and other solid
agricultural waste materials is generally heterogeneous and contains rapidly as well as slowly
degradable fractions (Brule et al., 2014; Schofield et al., 1994; Vavilin et al., 2008). Nevertheless,
further work is needed to determine the rate of hydrolysis of DDGS in batch reactors. The second
explanation is related to the difference in an environment that is conducive to bacteria. Our
results demonstrate that there was a change in bacterial community, TAN concentration and VSR
upon different feeding regimes. Moreover, the less frequently fed reactors showed a highly
143
dynamic environment associated with a transient accumulation of fermentation products (VFA,
H2/CO2) and temporal variation in pH compared to the less dynamic every 2 hours fed reactor.
This dynamic environment was also confirmed by the short-term isotope measurement, showing
temporal variation in methanogenic pathways. This dynamic environment could provide more
functional niches for the hydrolyzing and acidogenesis bacteria to grow and solubilize the DDGS
efficiently. Different pH, redox and concentration of fermentation products have been shown to
affect the growth and activity of fermenting microorganism (Vavilin et al., 2008). Higher
concentration of TAN, which is a product of protein fermentation and an increase in the reduction
of volatile solid in these reactors, also indicate the improved degradation of DDGS in less
frequently fed reactors. Nevertheless, the observed increased biogas yield for less frequently fed
DDGS reactors cannot be generalized to reactors run with other feedstocks. Since different
substrates have different proportions of slowly and easily degradable fractions, further research is
warranted to improve our understanding of process efficiency of reactors fed with different
feedstock and feeding intervals.
3.6 Implication for reactor operation
In conventional biogas plants, small portions of substrate are fed in short time intervals for
generating constant production of biogas and methane. To demonstrate the principle of flexible
biogas production aimed at a production of a higher amount of biogas on demand, reactors were
fed once daily and every two days (less frequently fed) with higher substrate loading at once. The
results show that by feeding less frequently more biogas can be produced for hours after the
feeding event to meet a high energy demand and feeding can be stopped during a low energy
demand for up to 2 days. The higher biogas production on demand in a less frequently fed reactor
compared to the continuously fed reactors would reduce the amount of biogas that is needed to be
stored on site, which in turn reduces the size of gas storage and saves extra investments.
Another key finding of this study is the improved process efficiency and high tolerance of
substrate overloading in less frequently fed reactors. This demonstrates that for even a
conventional biogas plant which runs on a continuous feeding basis, less frequently feeding can
be employed once in a while to improve functional stability. An additional advantage from a
practical standpoint is that less frequently feeding provides ease of plant operation and
mechanical stability. Due to the possibility of manual feeding and few feeding events, it allows a
plant operator to closely monitor and service the pump used to feed a substrate and take out
effluent, which reduces pump maintenance cost. Nevertheless further research is needed to
demonstrate the benefit of less frequently feeding regime in full-scale DDGS reactors.
4. CONCLUSION
Biogas production using DDGS as a monosubstrate was relatively constant in the more frequently
fed CSTR (every 2 hours) whereas it followed a temporal trend in the less frequently fed CSTR
(once per day and every 2 days) which allows for the flexibility to produce more biogas at times
of high energy demand. Interestingly, methane yield was significantly higher by about 14%
during steady state and by 27% under stress condition in the less frequently fed CSTR compared
to the more frequently fed CSTR. The bacterial community structure varied between CSTRs fed
under different feeding regimes whereas methanogens remained stable.
144
ACKNOWLEDGEMENTS
This research was financially supported by the Danish Strategic Research Council (Grant 10-
093944).
References
APHA. 2005. Standard methods for the examination of water and wastewater. in: American
Public Health Association, Vol. 21. Washington DC.
Brule, M., Oechsner, H., Jungbluth, T. 2014. Exponential model describing methane production
kinetics in batch anaerobic digestion: a tool for evaluation of biochemical methane
potential assays. Bioprocess Biosyst Eng, 37(9), 1759-70.
Conklin, A., Stensel, H.D., Ferguson, J. 2006. Growth kinetics and competition between
Methanosarcina and Methanosaeta in mesophilic anaerobic digestion. Water Environ Res,
78(5), 486-96.
Conrad, R. 2005. Quantification of methanogenic pathways using stable carbon isotopic
signatures: a review and a proposal. Organic Geochemistry, 36(5), 739-752.
Conrad, R., Claus, P., Chidthaisong, A., Lu, Y., Fernandez Scavino, A., Liu, Y., Angel, R.,
Galand, P., Casper, P., Guerin, F. 2014. Stable carbon isotope biogeochemistry of
propionate and acetate in methanogenic soils and lake sediments. Organic Geochemistry.
De Vrieze, J., Hennebel, T., Boon, N., Verstraete, W. 2012. Methanosarchina: The rediscovered
methanogen for heavy duty biomethanation. Bioresour. Technol. , 112, 1-9.
De Vrieze, J., Verstraete, W., Boon, N. 2013. Repeated pulse feeding induces functional stability
in anaerobic digestion. Microb Biotechnol, 6(4), 414-24.
Eskicioglu, C., Kennedy, K.J., Marin, J., Strehler, B. 2011. Anaerobic digestion of whole stillage
from dry-grind corn ethanol plant under mesophilic and thermophilic conditions.
Bioresour Technol, 102(2), 1079-86.
Franke-Whittle, I.H., Walter, A., Ebner, C., Insam, H. 2014. Investigation into the effect of high
concentrations of volatile fatty acids in anaerobic digestion on methanogenic
communities. Waste Management(0).
Goevert, D., Conrad, R. 2009. Effect of substrate concentration on carbon isotope fractionation
during acetoclastic methanogenesis by Methanosarcina barkeri and M. acetivorans and in
rice field soil. Applied and environmental microbiology, 75(9), 2605-2612.
Hahn, H., Krautkremer, B., Hartmann, K., Wachendorf, M. 2014. Review of concepts for a
demand-driven biogas supply for flexible power generation. Renewable and Sustainable
Energy Reviews, 29(0), 383-393.
Hochloff, P., Braun, M. 2014. Optimizing biogas plants with excess power unit and storage
capacity in electricity and control reserve markets. Biomass and Bioenergy, 65(0), 125-
135.
Karakashev, D., Batstone, D.J., Angelidaki, I. 2005. Influence of environmental conditions on
methanogenic compositions in anaerobic biogas reactors. Appl Environ Microbiol, 71(1),
331-8.
Lv, Z., Leite, A.F., Harms, H., Richnow, H.H., Liebetrau, J., Nikolausz, M. 2014. Influences of
the substrate feeding regime on methanogenic activity in biogas reactors approached by
molecular and stable isotope methods. Anaerobe, 29, 91-9.
145
Mauky, E., Jacobi, H.F., Liebetrau, J., Nelles, M. 2014. Flexible biogas production for demand-
driven energy supply - Feeding strategies and types of substrates. Bioresour Technol.
Moestedt, J., Påledal, S., Schnürer, A., Nordell, E. 2013. Biogas Production from Thin Stillage on
an Industrial Scale—Experience and Optimisation. Energies, 6(11), 5642-5655.
Mulat, D.G., Ward, A.J., Adamsen, A.P.S., Voigt, N.V., Nielsen, J.L., Feilberg, A. 2014.
Quantifying contribution of synthrophic acetate oxidation to methane production in
thermophilic anaerobic reactors by membrane inlet mass spectrometry. Environmental
science & technology, 48(4), 2505-2511.
Munk, B., Lebuhn, M. 2014. Process diagnosis using methanogenic< i> Archaea</i> in maize-
fed, trace element depleted fermenters. Anaerobe.
Nettmann, E., Bergmann, I., Pramschufer, S., Mundt, K., Plogsties, V., Herrmann, C., Klocke, M.
2010. Polyphasic analyses of methanogenic archaeal communities in agricultural biogas
plants. Appl Environ Microbiol, 76(8), 2540-8.
Nikolausz, M., Walter, R.F.H., Sträuber, H., Liebetrau, J., Schmidt, T., Kleinsteuber, S.,
Bratfisch, F., Günther, U., Richnow, H.H. 2013. Evaluation of stable isotope
fingerprinting techniques for the assessment of the predominant methanogenic pathways
in anaerobic digesters. Applied Microbiology and Biotechnology, 97(5), 2251-2262.
Oksanen, J. 2011. Multivariate analysis of ecological communities in R: vegan tutorial. 2.0-6 ed.
Schmidt, T., Ziganshin, A.M., Nikolausz, M., Scholwin, F., Nelles, M., Kleinsteuber, S., Pröter,
J. 2014. Effects of the reduction of the hydraulic retention time to 1.5 days at constant
organic loading in CSTR, ASBR, and fixed-bed reactors – Performance and
methanogenic community composition. Biomass and Bioenergy, 69(0), 241-248.
Schofield, P., Pitt, R., Pell, A. 1994. Kinetics of fiber digestion from in vitro gas production.
Journal of animal science, 72(11), 2980-2991.
Szarka, N., Scholwin, F., Trommler, M., Fabian Jacobi, H., Eichhorn, M., Ortwein, A., Thrän, D.
2013. A novel role for bioenergy: A flexible, demand-oriented power supply. Energy,
61(0), 18-26.
Vavilin, V.A., Fernandez, B., Palatsi, J., Flotats, X. 2008. Hydrolysis kinetics in anaerobic
degradation of particulate organic material: An overview. Waste Management, 28(6), 939-
951.
Weiland, P. 2010. Biogas production: current state and perspectives. Applied Microbiology and
Biotechnology, 85(4), 849-860.
Whiticar, M.J. 1999. Carbon and hydrogen isotope systematics of bacterial formation and
oxidation of methane. Chemical Geology, 161(1–3), 291-314.
Wilkie, A.C., Riedesel, K.J., Owens, J.M. 2000. Stillage characterization and anaerobic treatment
of ethanol stillage from conventional and cellulosic feedstocks. Biomass and Bioenergy,
19(2), 63-102.
Ziganshin, A., Schmidt, T., Scholwin, F., Il’inskaya, O., Harms, H., Kleinsteuber, S. 2011.
Bacteria and archaea involved in anaerobic digestion of distillers grains with solubles.
Applied Microbiology and Biotechnology, 89(6), 2039-2052.
Ziganshin, A.M., Liebetrau, J., Proter, J., Kleinsteuber, S. 2013. Microbial community structure
and dynamics during anaerobic digestion of various agricultural waste materials. Appl
Microbiol Biotechnol, 97(11), 5161-74.
146
Chapter 9: Paper VI- Stable isotope techniques as a tool for process
monitoring of biogas reactors operating under different condition
In preparation for peer-reviewed journal.
147
Stable isotope techniques as a tool for process monitoring of biogas reactors operating
under different conditions
Daniel Girma Mulat
a, Anders Feilberg
a, H. Fabian Jacobi
b,c, Marcell Nikolausz
d
aDepartment of Engineering, Aarhus University, Hangøvej 2, DK-8200 Aarhus N, Denmark
bDepartment of Biochemical Conversion, Deutsches Biomasseforschungszentrum (DBFZ), Torgauer Str.
116, 04347 Leipzig, Germany cLandesbetrieb Hessisches Landeslabor, Schloß Str. 26, 36251 Bad HerRh2eld, Germany
dDepartment of Environmental Microbiology, Helmholtz Centre for Environmental Research-
UFZ, Permoser Str. 15, 04318 Leipzig, Germany
ABSTRACT
CSTRs were fed with dried distillers grains with solubles (DDGS) under different operating
conditions (feeding interval and organic loading rate) to compare the suitability of stable isotope
composition of biogas and basic process parameters as a process monitoring tool. Longer feeding
interval led to a dynamic process, as depicted in the short term changes of biogas production rate,
biogas composition, hydrogen concentration, δ13
C-CH4, δD-CH4, total VFA, acetate and
propionate. The stable isotope composition of biogas and all the basic process parameters except
the total and the individual VFA responded to an increase in OLR. The short-term changes in
isotope composition of biogas were correlated with substrate availability and hence reflect the
changes in activity of microorganisms and relative proportion of methanogenic pathways. The
results demonstrated that a process monitoring tool based on stable isotope measurement of
biogas in conjunction with biogas production, biogas composition and hydrogen concentration
would indicate the actual state and performance of the process as well as process imbalance at an
early stage.
Key words: feeding interval, biogas, stable isotope, process monitoring, stillage
148
1. INTRODUCTION
Conversion of organic waste to biogas under anaerobic digestion has gained considerable
importance for its potential as waste management strategy, renewable source of energy and
reduction of greenhouse gas emissions. However, the microorganisms involved in the
degradation of organic material to biogas are sensitive to operating and environmental conditions
such as changes in organic loading rate (OLR), substrate composition, feeding interval, pH,
temperature, ammonia (Karakashev et al., 2005). Therefore, an appropriate process monitoring
and control strategy should be employed to achieve process stability and optimum biogas
production.
Several parameters such as pH, biogas production, biogas composition, total volatile fatty acids
(VFA) concentration, individual VFA concentration and hydrogen concentration have been
proposed as a tool for monitoring biogas process (Ahring et al., 1995). Individual and total VFA
concentrations are the most commonly monitored parameters for indicating process balance
between acid producing and consuming microorganisms. Butyrate and isobutyrate were
suggested to be particularly good indicators for predicting process instability following
perturbation in a mixture of cattle and swine manure treating continuous stirred tank reactor
(CSTR) (Ahring et al., 1995). In another study, isoforms of butyrate and valerate were reported
to be the best indicators of process instability in swine manure treating CSTR (Hill & Bolte,
1989). Since acetate is one of the key precursor of methane, acetate concentration higher than 13
mM suggested to indicate process imbalance (Hill et al., 1987) whereas acetate concentration as
high as 38 mM was detected in stably operating biogas digester (Franke-Whittle et al., 2014).
Others suggested propionate (Boe et al., 2008) and acetate to propionate ratio (Marchaim &
Krause, 1993) as an important parameter for indicating process imbalance. These all findings
demonstrated that using one or combinations of individual VFA as a tool for process monitoring
would seem to depend on the specific digester conditions. Moreover, several studies have
underlined that monitoring the relative changes of individual and total VFA levels over time is
more important than their absolute concentrations to indicate process imbalance (Angelidaki et
al., 1993; Franke-Whittle et al., 2014; Ward et al., 2011).
pH measurement can indicate process imbalance in a digester with low buffer capacity. In such
system with a low buffer capacity, a drop in pH correlated with an accumulation of VFA.
However, in highly buffered systems, pH changes can be small, even when the process is
extremely stressed (Angelidaki & Ahring, 1994), suggesting pH is less important to indicate
process imbalance under those conditions. Hydrogen concentration in the liquid and gas phases
has also been suggested to indicate the process imbalance, since its higher concentration is
associated with accumulation of VFA (Björnsson et al., 2001). Propionate and butyrate
degradation by obligate syntrophic acetogenic bacteria is dependent on hydrogen concentration
and would be inhibited at elevated hydrogen concentration (Björnsson et al., 2001).
Biogas production is routinely monitored with a gas sensor at full-scale biogas plants to indicate
overall process performance. However, it can poorly indicate process imbalance since a decrease
in biogas process is not necessarily a consequence of process inhibition and it could be rather a
result of low reactor loading. The change in composition of biogas with a decrease in methane
content and an increase in carbon dioxide content is assumed to be associated with a process
disturbance. Since the concentration of CO2 in the headspace is also a function of pH, any change
149
in pH can results in a change of ratio of CO2 and methane, therefore, it is not necessarily a direct
consequence of inhibited or imbalanced biogas production.
Measurement of stable isotope composition of methane, CO2 and acetate has been extensively
used for quantifying metabolic pathways in environmental research fields (Conrad, 2005). This is
based on the isotope measurement of stable isotope labeled or unlabeled substances at natural
abundance. The former can be measured with low resolution mass spectrometer (MS) such as
quadrupole and ion trap MS. Incubation of 13
C labeled acetate under anaerobic condition and the
subsequent monitoring of the incorporation of the 13
C into the produced CH4 and CO2 with gas
chromatography−mass spectrometry (GC/MS), is widely employed for identifying specific
degradation kinetics and methanogenic pathways (Sasaki et al., 2011). More recently, we
employed an on-line method based on membrane inlet quadrupole mass spectrometry (MIMS)
for measuring the isotope composition of dissolve CH4 and CO2 in conjunction with 13
C labeled
acetate (Mulat et al., 2014). It is demonstrated that MIMS can be used to quantify temporal
variation in methanogenic pathways at higher time resolution. Despite the importance of labeled
substrate for quantifying methanogenic pathways in a laboratory condition, it is impractical for
monitoring biogas process in full-scale biogas plants or even in larger-scale continuous
laboratory experiments.
Measurement of stable isotope ratio of biogas at natural abundance is a promising tool for process
monitoring (Lv et al., 2014b; Nikolausz et al., 2013; Polag et al., 2014). The isotope signature of
the produced biogas in anaerobic digesters can be used to identify methanogenic pathways
because of different methanogenic pathways results in a distinct variation in isotope abundances
of the produced biogas (Conrad, 2005). Since this variation in the isotopic ratio between the
heavy and light stable carbon isotopes differs by only few percent, a highly precise analytical
techniques such as continuous-flow gas chromatography-combustion isotope ratio mass
spectrometry (GC/C/IRMS) (Meier-Augenstein, 1999) and tunable diode laser absorption
spectrometry (TDLAS) are required (Keppler et al., 2010). Gas phase samples are measured
offline with the former technique whereas the latter enables on-line measurement of the carbon
isotope ratios of CH4 and CO2 at higher time resolution (Keppler et al., 2010). More recently,
TDLAS was applied for real-time monitoring of stable carbon isotopes of methane (δ13
CH4) in a
pilot-scale biogas digester fed with maize silage and the results suggested that the δ13
CH4
responded earlier than other basic process parameters to process perturbation with high organic
loading rate (Polag et al., 2014). In another study, the stable isotope signatures of biogas in less
frequently fed reactor followed a temporal trend which was highly correlated with the availability
of methane precursors and the change in the activity of methanogens (Lv et al., 2014b). Stable
isotope signature of the produced biogas was proved as an early warning tool to ammonia
inhibition during co-digestion of dry chicken waste and maize silage (Lv et al., 2014a).
Biogas plants are traditionally operated with a continuous and constant substrate feed to achieve
nearly the same amount of biogas and electricity generation throughout the day with a market
value equivalent to base loads market prices. Recently, the importance of flexible biogas
production to balance the supply of electricity generated from fluctuating sources such as solar
and wind has been emphasized (Hahn et al., 2014; Szarka et al., 2013). Flexible biogas
production can be attained by feeding substrate at different intervals in order to produce more
biogas during high energy demand periods (Lv et al., 2014b; Mauky et al., 2014). Since change in
feeding interval could lead to process imbalance, an appropriate process monitoring and control
150
tools are needed for achieving process stability and optimum biogas production. In this regard,
the responses of basic process parameters and isotope composition of produced biogas as tools
for indicating actual state of the process and imbalance at an early stage under different operating
conditions (different feeding intervals and OLR) have not been well documented.
This paper presents the application of both carbon and hydrogen isotope signatures for
monitoring the actual state of biogas reactors under different feeding regimes and stress condition
(continuous increase in OLR). In this study, laboratory-scale continuous stirred tank reactors
(CSTRs) were fed with distillers dried grains with solubles (DDGS) under different feeding
regimes (every 2 hours, once a day, and every 2 days). Then the CSTRs were exposed to a short-
term stress condition by continuously increasing OLR. DDGS is a byproduct of bioethanol
industry and its use for biogas production has gained importance given the environmental and
economic benefits (Wilkie et al., 2000). The objective this study was to investigate whether short-
term changes in carbon and hydrogen isotope composition of produced biogas could be used to
indicate the state of biogas process and stress condition at an early stage. In parallel to isotope
analysis, other process parameters were measured to suggest which combination of parameters
could be used as a process monitoring tool.
2. MATERIALS AND METHOD
2.1 Substrate
All biogas reactors were fed with distillers dried grains with solubles (DDGS) and supplemented
with FerroSorp® DG (HeGo, Biotec, Germany) and trace element solution (TES). DDGS was
obtained from an industrial scale bio-ethanol plant (CropEnergies AG; Zeitz) with wheat as a
main raw material. The FerroSorp is a powdery substance containing iron hydroxide that was
added to the reactors to precipitate hydrogen sulfide. The TES contains (all are in g/L):
Ni(II).6H2O, 2.13; Co(II).6H2O, 0.531; NaMoO4.2H2O, 0.332; (NH4)6H2W12O40·xH2O, 0.423.
2.2 Reactor setup
Two laboratory-scale continuous stirred tank reactors (CSTRs) were operated under mesophilic
conditions (38 ºC) with DDGS as the only substrate. The total volume of each reactor was 15 L
with a working volume of 10 L. Before the start of this study, the two reactors were operated
under identical conditions and then the contents of the reactors were mixed together to setup
equivalent starting conditions for this study. Three different feeding regimes were studied (once
daily, every 2 days and every 2 hours). DDGS was fed manually in daily and every 2 days
feeding regimes whereas peristaltic pump was used for every 2 hours feeding The whole
experiment was divided into 4 phases: reactors A and B were fed once daily (phase I, day 1-29);
reactor A was fed once daily and B was fed every 2 hours (phase II, day 30-63); reactor A was
fed once every 2 days and B was fed every 2 hours (phase III, day 64-107); reactor A was fed
once every 2 days and B was fed every 2 hours but with higher OLR compared to phase III
(phase IV, day 108-118) (Table 1). Daily, every 2 d and every 2 h fed reactors were re-named as
Rd1, Rd2 and Rh2. Moreover, the Rd2 and Rd1 were collectively named as longer feeding
interval (LF). During phases I-III, the daily organic load of both reactors under every feeding
regime was kept constant at 4 gVS L-1
d-1
. To achieve this organic loading rate, the daily substrate
151
was prepared by mixing 66.34 g DDGS, 2 mL TES and 2.56 g FerroSorp® DG and finally water
was added to the mixture to make a total of 380 mL solution. In phase IV, the OLR was increased
continuously from 4 to 11 gVS L-1
d-1
at the same rate for both reactors A and B. Hydraulic
retention time (HRT) was kept constant at 26 days in phases I-III. In phase IV, the HRT was 26
days for OLR ranged 5-7 gVS L-1
d-1
whereas HRT was reduced to 14 and 10 days at OLR of 9
and 11 gVS L-1
d-1
, respectively. Reduction of the HRT is not just simulating real case scenarios
of increased OLR, but it was a technical necessity to dilute the substrate in our experiments to
avoid difficulties in pump feeding.
Table 1 Operating condition of CSTRs fed with DDGS
Phase
No.a
Days Original
reactor
name
New
reactor
nameb
OLR
(gVs
L-1
d-1
)
Feeding
Mode
HRT
T
(°C)
CH4
(%)c
SMP
(mLN
gVS-1
)d
SBP
(mLN
gVS-
1)
d
1 1-29 A Rd1 4 Daily 26 38 57 402 710
B Rd1 4 Daily 26 38 57 400 706
2 30-63 A Rd1 4 Daily 26 38 57 414 724
B Rh2 4 Every 2 h 26 38 58 363 622
3 64-107 A Rd2 4 Every 2 d 26 38 58 402 724
B Rh2 4 Every 2 h 26 38 58 357 613
4
108-
118 A
Rd2
5-11 Every 2 d 10-26 38 60
387 689
B Rh2 5-11 Every 2 h 10-26 38 57 304 526 aIn phase I, the operating condition was identical for both reactors. During phases II-III, both
reactors were fed at the same OLR (4 gVs L-1
d-1
) but differ in feeding interval. In phase IV, a
short stress test was conducted by continuously increasing OLR to 11 gVs L-1
d-1
. OLR, organic
loading rate; HRT, hydraulic retention time; SMP, specific methane production; SBP, specific
biogas production bReactor A was fed once per day during phases I-II and the feeding interval was changed to once
every 2 days during phases III-IV. Reactor B was fed once per day in phase I and he feeding was
changed to once every 2 hours during phases II-IV. Rd1, Rd2 and Rh2 represent reactors fed
once per day, once every 2 days and once every 2 hours, respectively cThe average daily methane content
dData was corrected to standard temperature (273.15 K) and pressure (101.325 kPa)
2.3 Acetoclastic methanogenesis inhibition batch experiment with CH3F
In addition to the lab-scale CSTRs, batch experiment was prepared using inoculum sourced from
the CSTRs (every 2 d and 2 h fed) on day 72 (phase III). In an anoxic glove-box, aliquot of the
inoculum (30 mL) was transferred into a 50 mL serum bottles and then DDGS as the only
substrate was added. It was also supplemented with Ferrosorb and TES and finally sealed with
rubber stopper and aluminum crimp. Three treatments were prepared for each inoculum source:
(i) 1.3% CH3F-added incubation; (ii) 5% CH3F-added incubation; and (iii) control incubation -
without CH3F. In all treatments, the OLR was the same as their corresponding lab-scale CSTRs.
Fluoromethane (CH3F) was added into the bottles of (i) and (ii) incubations in order to inhibit
acetoclastic methanogenesis (Table 2). Acetoclastic inhibited treatments allowed us to
152
specifically determine the isotopic signature of the CH4 produced through hydrogenotrophic
methanogenesis (δmc, see section 2.5). Fluoromethane treatments and controls were incubated in
duplicate and quadruplicate at 38°C.
2.4 Basic process parameters
Basic process parameters such as biogas production and composition, total and individual fatty
acids were monitored as described before (our paper). Briefly, the volume of biogas was
automatically recorded by a gas meter (TG 0.5, Dr.-Ing. Ritter Apparatebau Gmbh & Co KG,
Bochum, Germany). The volume of the biogas was corrected to standard temperature (273.15 K)
and pressure (101.325 kPa). An AwiFlex gas analyzer equipped with optical infrared and
electrochemical sensors was used for determining the contents of CH4 and CO2 in the biogas as
well as for determining the concentrations of O2, H2, and H2S (detection limit 5,000 ppm),
respectively.
Liquid effluent was periodically collected from all reactors and its pH was measured
immediately. It was then centrifuged at 10,000 rpm and 10 ºC for 12 min and the supernatant
solution was filtered through a sieve with mesh size of 1mm. The supernatant solution was used
for further analysis of volatile fatty acids (VFA) and total ammonia nitrogen (TAN) expressed in
gNH4+-N L
-1 (i.e. TAN is defined as a sum of free ammonia nitrogen (FAN) plus ammonium
nitrogen). An aliquot (5 mL) of the supernatant solution was transferred into a 20 mL vial
followed by the addition of 1 mL internal standard (2-ethylbutyric acid, 180 mg/L) and 1 mL
phosphoric acid (50% V/V). The vial was immediately sealed with butyl rubber stopper and
aluminum crimp. The concentration of VFA was determined from the headspace by a 5890 series
II gas chromatograph (Hewlett Packard, Palo Alto, USA) equipped with an HS40 automatic
headspace sampler (Perkin Elmer, Waltham, USA). HP-FFAP column (30m x 0.32 mm x 0.25
μm; Agilent Technology, Germany) was used for chromatographic separation. The GC
conditions were as described before (Ziganshin et al., 2011). For TAN analysis, an aliquot (125
µL) of the solution was diluted to 1:2000 with distilled water and the TAN was determined with a
standard Nessler method using a benchtop spectrometer (Hach-Lange DR 3900, Loveland, USA).
Total solid (TS) and volatile solid (VS) were analyzed according to standard method (APHA,
1995).
2.5 Stable isotope analysis
Biogas samples were periodically collected from the headspace of the reactor in triplicate and
stored in a gas tight evacuated vial (20 mL) until further analysis. The stable carbon isotope
analysis of gas samples was performed using a gas chromatography combustion isotope ratio
mass spectrometery (GC-C-IRMS) system consisting of a gas chromatograph (HP 6890 Series,
Agilent Technology, Santa Clara, USA) coupled with IRMS (Finnigan MAT 253,
Thermofinnigan, Bremen, Germany) via a combustion interface. For the stable hydrogen isotope
analysis of gas samples, the same GC and IRMS systems were used as for carbon but a pyrolysis
unit was used instead of combustion interface. For the GC separation of CH4 and CO2, gas
samples (30 µL) were injected manually into the GC instrument equipped with a CP-
Porabond Q column (50 m × 0.32 mm × 0.5 μm, Varian, USA). The column temperature was
kept constant at 40 ºC and flow rate of Helium was 2 mL/min for carbon and 1.6 mL/min for
hydrogen.
153
The analysis of the stable carbon isotope of biomass (DDGS) was performed with an elemental
analyzer (EA)-IRMS system consisting of an EA (Euro EA, HEKAtech GmbH, Wegberg,
Germany) and an IRMS ((Finnigan MAT 253, Thermofinnigan, Bremen, Germany), coupled via
an interface (ConFlo III; Thermo Fisher Scientific). In addition, the hydrogen isotope
composition of H2O in a tap water, which was used for preparing the daily substrate, was
analyzed as described before (Gehre et al., 2004).
The stable carbon and hydrogen isotope data was reported in delta notation (δ13
C and δD) in parts
per thousand (‰) unit versus the Vienna Pee Dee Belemnite (V-PDB) and Vienna Standard
Mean Ocean Water (V-SMOW), respectively:
δx= [(Ra)sample/(Ra)standard - 1]*103 (‰) (1)
where δx is the δ13
C or δD; Ra is the 13
C/12
C or D/H ratios (Whiticar, 1999).
The apparent fractionation factor (αmc) between CO2 and CH4 was calculated according to the
following equation:
αmc = (δ13
C-CO2 + 1000)/(δ13
C-CH4 + 1000) (2)
The relationship among the 13
C isotopic signature of total CH4 (δ13
C-CH4), acetate-derived CH4
(δma) and CO2-derived CH4 (δmc) can be described using the following mass balance equation
(Conrad, 2005):
δ13
C-CH4 = fmc* δmc + (1- fmc)* δma (3)
where fmc is the fraction of CH4 produced from the reduction of CO2 through hydrogenotrophic
methanogenesis (HM). To estimate the fmc, the values of δma and δmc need to be known
beforehand. The δma was estimated as follows:
δma = δac + εma (4)
where εma is the isotopic enrichment by acetoclastic methanogenesis. We assumed δac = -25.5‰
and εma=-10‰ and then the value of δma was estimated to be -35.5‰ using equation 2. The δ13
C–
CH4 in the acetoclastic inhibited batch incubation (5% CH3F added reactor) was used to determine
the isotopic signature of the CH4 produced through hydrogenotrophic methanogenesis (δmc).
δ13
C–CH4 in this reactor was on average -68.5‰, and thus, the δmc was assumed equal to -68.5‰.
To summarize, the fmc in CSTRs was estimated using equation 3, the measured δ13
C-CH4 in the
CSTRs, δma=-35.7‰ and δmc=-68.5 (detail discussion and calculation was provide in the text,
section 3.2).
2.6 RNA extraction and purification
Liquid samples for microbial community analyses were collected at different time points in 48
hours interval (right before substrate feeding (0 h) and 2, 7, 24 and 48 h after feeding reactor A)
during phase III (days 105-107). Samples were collected into a sterile Eppendorf tubes and stored
in a freezer (-80 ºC) until further analysis. Total RNA was isolated in duplicate using the RNeasy
Mini Kit (Qiagen, Hilden, Germany) according to manufacturer’s protocol including
modifications to use mechanical cell lysis as described earlier (Lv et al., 2014b; Nikolausz et al.,
2013). Remaining DNA contamination was eliminated from the isolated RNA by the TURBO
DNA-free kit (Ambion, California, USA) according to the manufacturer’s protocol.
154
2.7 Reverse transcription and polymerase chain reaction (PCR)
The purified RNA was used for preparing cDNA with mcrA-rev primer (Steinberg & Regan,
2008) using Revert Aid H Minus M-MuLV RT, (Fermentas Thermo Fisher Scientific, Inc.,
Waltham, MA, USA) as described earlier (Lv et al., 2014b; Nikolausz et al., 2013). mcrA genes
were amplified by using the mcrA/mrtA specific forward primer mlas and the reverse primer
mcrA-rev (Lv et al., 2014b; Nikolausz et al., 2013). For T-RFLP analysis, the same primers were
used for PCR amplification except the reverse primer was 5’-labeled with FAM
(phosphoramidite fluorochrome 6-carboxyfluorescein). PCR products were checked in 1.5% gel
electrophoresis. Then the PCR products were purified using a SureCleanPlus kit (Bioline,
Germany) and quantified using NanoDrop® ND-1000 UV/Vis spectral photometer (PeqLab,
Germany).
2.8 T-RFLP analysis
T-RFLP analysis was performed as described before (Lv et al., 2014b; Nikolausz et al., 2013).
Briefly, mcrA genes were digested with BstNI (New England Biolabs, Schwalbach, Germany).
After the purification of the terminal restriction fragments (T-RFs) by ethanol precipitation,
pellets were resuspended in HiDi formamide containing GeneScan-500 ROX standard (Applied
Biosystems, Weiterstadt, Germany) followed by separation of fluorescently labeled T-RFs using
capillary electrophoresis with an automatic sequencer ABI PRISM 3130xl Genetic Analyzer
(Applied Biosystems, Weiterstadt, Germany). T-RFLP data were retrieved by comparison of
samples with the internal standards using GeneMapper V3.7 software (Applied Biosystems,
Weiterstadt, Germany). T-RFs in the size range of 50-500 were considered for subsequent
analysis of the archaeal mcrA genes. For statistical analysis, T-RFLP data sets were reduced by
removing low abundance T-RFs below 1%. Relative T-RF abundances were calculated by
dividing individual T-RF peak area to the total peak areas. Linking taxonomic information to the
major mcrA T-RFs were done by using the sequence and T-RFLP database from previous studies
of our group (Lv et al., 2014b; Nikolausz et al., 2013).
3. RESULTS and DISCUSSION
3.1 Response of basic process parameters under different feeding intervals
Basic process parameters such as biogas production rate, biogas composition, pH, hydrogen
concentration as well as individual and total volatile fatty acids (VFA) concentration were
monitored over short-time interval under three different feeding intervals (once per day, every 2
hours and every 2 days). The response of these parameters subjected to feeding 1/12 portion of
the daily substrate once every 2 h (Rh2), all the daily substrate at once per day (Rd1) and twice
the amount of the daily substrate at once every 2 days (Rd2) are shown in Figures 1-3.
155
7.75
7.80
7.85
7.90
7.95
0
1
2
3
4
0 4 8 12 16 20 24
pH
Bio
gas
(LN h
-1)
A-Biogas B-BiogasA-pH B-pH
(a) Phase I
Days 16-17
0
50
100
150
50
52
54
56
58
60
62
0 4 8 12 16 20 24
H2 (
pp
m)
CH
4 (
%)
A-CH4 B-CH4A-H2 B-H2
0
20
40
60
80
100
120
140
45
65
85
105
125
145
0 4 8 12 16 20 24
Ace
tate
(mg L
-1)
VF
A (
HA
Ceq
. m
gL
-1) A-VFA B-VFA
A-Acetate B-Acetate
0.0
0.5
1.0
1.5
0
5
10
15
20
25
0 4 8 12 16 20 24
Isob
uty
rate
,
Isovale
rate
(mg L
-1)
Pro
pio
nate
(m
g L
-1)
Hours
A-Propionate B-PropionateA-Isobutyrate B-IsobutyrateA-Isovalerate B-Isovalerate
156
Figure 1 Short-term basic process parameters and stable isotope signatures of two CSTRs (A and
B) with identical operating condition. Both CSTRs were fed once per day during phase I and they
were collectively referred as “Rd1” in the text. (a) basic process parameters (biogas production
rate; pH; CH4 content; H2 concentration; total VFA concentration in unit of mgHAc eq. L-1
, mg
of acetic acid equivalent per liter; acetate; propionate; isobutyrate, and isovalerate); (b) isotope
signatures (δ13
C-CH4; δ13
C-CO2; fmc, fraction of methane derived from the reduction of CO2;and
αmc, fractionation factor between CO2 and CH4). 0 h represents the period just before feeding
event.
6
8
10
12
-60
-55
-50
-45
0 4 8 12 16 20 24
δ1
3C
-CO
2 (‰)
δ1
3C
-CH
4 (‰)
A-δ13C-CH4 B-δ13C-CH4 A-δ13C-CO2 B-δ13C-CO2
1.055
1.060
1.065
1.070
1.075
1.080
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 4 8 12 16 20 24
αm
c
f mc
Hours
A-fmc B-fmcA-αmc B-αmc
(b) Phase I
Days 16-17
157
7.75
7.80
7.85
7.90
7.95
0
1
2
3
4
0 4 8 12 16 20 24 28 32 36 40 44 48
pH
Bio
gas
(LN h
-1)
Rd1-Biogas Rh2-BiogasRd1-pH Rh2-pH
(a) Phase II Days 37-38
Phase II
Days 59-60
0.0
50.0
100.0
150.0
50
55
60
65
0 4 8 12 16 20 24 28 32 36 40 44 48
H2 (
pp
m)
CH
4 (
%)
Rd1-CH4 Rh2-CH4
0
20
40
60
80
100
0
25
50
75
100
125
0 4 8 12 16 20 24 28 32 36 40 44 48
Ace
tate
(m
g L
-1)
VF
A (
HA
Ceq
. m
gL
-1) Rd1-VFA Rh2-VFA
Rd1-Acetate Rh2-Acetate
0.0
0.5
1.0
0.0
2.5
5.0
7.5
10.0
0 4 8 12 16 20 24 28 32 36 40 44 48
Iso
bu
tyra
te (
mg L
-1)
Pro
pio
na
te (
mg L
-1) Rd1-Propionate Rh2-Propionate
Rd1-Isobutyrate Rh2-Isobutyrate
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0 4 8 12 16 20 24 28 32 36 40 44 48
Con
c. (
mg
L-1
)
Hours
Rd1-Butyrate Rh2-ButyrateRd1-Isovalerate Rh2-Isovalerate
158
Figure 2 Short-term basic process parameters and stable isotope signatures of two CSTRs (Rd1
and Rh2) operated under different feeding intervals. Rd1 and Rh2 refer to CSTRs fed once per
day and once every 2 h, respectively, during phase-II; (a) basic process parameters (biogas
production rate; pH; CH4 content; H2 concentration; total VFA concentration in unit of mgHAc
eq. L-1
, mg of acetic acid equivalent per liter; acetate; propionate; isobutyrate, butyrate; and
isovalerate); (b) isotope signatures (δ13
C-CH4; δ13
C-CO2; fmc, fraction of methane derived from
the reduction of CO2; αmc, fractionation factor between CO2 and CH4; and δD-CH4. The δD-CH4
values for the first 24 h (days 37-28) were not available. 0 and 24 h represent the period just
before feeding event. The arrow represent the period when the second feeding event was
conducted. The data at [0 h, 24 h) and [24 h, 48] were collected on days 37-38 and 59-60,
respectively, and they were plotted on the same graph for simplicity.
4
6
8
10
12
-60
-50
-40
0 4 8 12 16 20 24 28 32 36 40 44 48
δ1
3C
-CO
2 (‰)
δ1
3C
-CH
4 (‰)
Rd1-δ13C-CH4 Rh2-δ13C-CH4 Rd1-δ13C-CO2 Rh2-δ13C-CO2
(b)
1.055
1.060
1.065
1.070
1.075
1.080
0.2
0.3
0.4
0.5
0.6
0.7
0 4 8 12 16 20 24 28 32 36 40 44 48α
mc
f mc
Rd1-fmc Rh2-fmc
Rd1-αmc Rh2-αmc
-370
-365
-360
-355
-350
0 4 8 12 16 20 24 28 32 36 40 44 48
δD
-CH
4 (‰)
Hours
Rd1-δD-CH4 Rh2-δD-CH4
Phase II Days 37-38
Phase II
Days 59-60
159
7.50
7.65
7.80
7.95
8.10
0
2
4
6
0 8 16 24 32 40 48 56 64 72 80 88 96
pH
Bio
gas
(LN h
-1)
Rd2-Biogas Rh2-Biogas
Rd2-pH Rh2-pH
0
25
50
75
100
125
150
175
40
50
60
70
0 8 16 24 32 40 48 56 64 72 80 88 96H
2 (
pp
m)
CH
4 (
%)
Rd2-CH4 Rh2-CH4Rd2-H2 Rh2-H2
(a)
0
50
100
150
200
250
300
350
0
100
200
300
400
500
600
700
800
0 8 16 24 32 40 48 56 64 72 80 88 96
Ace
tate
(m
g L
-1)
VF
A (
HA
Ceq
.mg L
-1) Rd2-VFA Rh2-VFA
Rd2-Acetate Rh2-Acetate
0
2
4
6
8
10
12
0
50
100
150
200
250
300
350
0 8 16 24 32 40 48 56 64 72 80 88 96
Isob
uty
rate
(m
g L
-1)
Pro
pio
nate
(m
g L
-1) Rd2-Propionate Rh2-Propionate
Rd2-Isobutyrate Rh2-Isobutyrate
0
2
4
6
0 8 16 24 32 40 48 56 64 72 80 88 96
Con
c. (
mg L
-1)
Hours
Rd2-Butyrate Rh2-Butyrate
Rd2-Isovalerate Rh2-Isovalerate
Phase III Days 79-81
Phase III
Days 105-107
160
Figure 3 Short-term basic process parameters and stable isotope signatures of two CSTRs (Rd2
and Rh2) operated under different feeding intervals. Rd2 and Rh2 refer to CSTRs fed once every
2 d and once every 2 h, respectively, during phase-III; (a) basic process parameters (biogas
production rate; pH; CH4 content; H2 concentration; total VFA concentration in unit of mgHAc
eq. L-1
, mg of acetic acid equivalent per liter; acetate; propionate; isobutyrate, butyrate; and
isovalerate); (b) isotope signatures (δ13
C-CH4; δ13
C-CO2; fmc, fraction of methane derived from
the reduction of CO2; αmc, fractionation factor between CO2 and CH4; and δD-CH4. 0 and 24 h
represent the period just before feeding event. The arrow represent the period when the second
feeding event was conducted. The data at [0 h, 48 h) and [48 h, 96] were collected on days 79-81
and 105-107, respectively, and they were plotted on the same graph for simplicity.
(b)
1.055
1.060
1.065
1.070
1.075
1.080
0.2
0.4
0.6
0.8
0 8 16 24 32 40 48 56 64 72 80 88 96α
mc
f mc
Rd2-fmc Rh2-fmc
Rd2-αmc R2h-αmc
-370
-365
-360
-355
-350
-345
0 8 16 24 32 40 48 56 64 72 80 88 96
δD
-CH
4 (‰)
Hours
Rd2-δD-CH4 Rh2-δD-CH4
4
6
8
10
12
-60
-55
-50
-45
-40
0 8 16 24 32 40 48 56 64 72 80 88 96
δ1
3C
-CO
2 (‰)
δ1
3C
-CH
4 (‰)
Rd2-δ13C-CH4 Rh2-δ13C-CH4
Rd2-δ13C-CO2 Rh2-δ13C-CO2
Phase III Days 79-81
Phase III
Days 105-107
161
7.50
7.65
7.80
7.95
8.10
0
5
10
0 8 16 24 32 40 48 56 64 72
pH
Bio
gas
(LN h
-1)
Rd2-Biogas Rh2-BiogasRd2-pH Rh2-pH
0
50
100
150
200
250
300
350
400
20
40
60
80
0 8 16 24 32 40 48 56 64 72
H2 (
pp
m)
CH
4 (
%)
Rd2-CH4 Rh2-CH4Rd2-H2 Rh2-H2
(a)
0
25
50
75
100
125
150
0
50
100
150
200
250
300
0 8 16 24 32 40 48 56 64 72
Ace
tate
(m
g L
-1)
VF
A (
HA
Ceq
.mg L
-1) Rd2-VFA Rh2-VFA
Rd2-Acetate Rh2-Acetate
0
2
4
6
8
10
12
14
0
20
40
60
80
100
0 8 16 24 32 40 48 56 64 72
Isob
uty
rate
(m
g L
-1)
Pro
pio
nate
(m
g L
-1) Rd2-Propionate Rh2-Propionate
Rd2-Isobutyrate Rh2-Isobutyrate
0.0
0.2
0.4
0.6
0.8
0 8 16 24 32 40 48 56 64 72
Con
c. (
mg L
-1)
Hours
Rd2-Butyrate Rh2-Butyrate
Rd2-Isovalerate Rh2-Isovalerate
Phase IV Days 113-116
162
Figure 4 Short-term basic process parameters and stable isotope signatures of two CSTRs (Rd2
and Rh2) under stress condition (continuous increase in OLR). Rd2 and Rh2 refer to CSTRs fed
once every 2 d and once every 2 h, respectively, during phase-IV; (a) basic process parameters
(biogas production rate; pH; CH4 content; H2 concentration; total VFA concentration in unit of
mgHAc eq. L-1, mg of acetic acid equivalent per liter; acetate; propionate; isobutyrate, butyrate;
and isovalerate); (b) Isotope signatures (δ13
C-CH4; δ13
C-CO2; fmc, fraction of methane derived
from the reduction of CO2; αmc, fractionation factor between CO2 and CH4; and δD-CH4. 0 and 24
h represent the period just before feeding event. The arrow represent the period when the second
feeding event was conducted.
(b)
1.050
1.060
1.070
1.080
0.3
0.5
0.6
0.8
0.9
0 8 16 24 32 40 48 56 64 72α
mc
f mc
Rd2-fmc Rh2-fmcRd2-αmc R2h-αmc
-385
-380
-375
-370
-365
-360
-355
-350
-345
0 8 16 24 32 40 48 56 64 72
δD
-CH
4 (‰)
Hours
Rd2-δD-CH4 Rh2-δD-CH4
1
3
5
7
9
11
-65
-60
-55
-50
-45
-40
0 8 16 24 32 40 48 56 64 72
δ1
3C
-CO
2 (‰)
δ1
3C
-CH
4 (‰)
Rd2-δ13C-CH4 Rh2-δ13C-CH4 Rd2-δ13C-CO2 Rh2-δ13C-CO2
Phase IV Days 113-116
163
Biogas production was highly dynamic in the daily and every 2 days fed CSTRs referred as
longer feeding interval (LF) further in the manuscript. In the LF, biogas production reached a
peak level almost 1 h after feeding and then started to decrease and finally reached the same level
as before feeding. Since Rd2 received twice of the amount of the daily substrate, the peak of
biogas production was higher in the Rd2 than Rd1 (4.4 LN h-1
in Rd2 vs 3.0-3.5 LN h-1
in the Rd1).
On the other hand, the biogas production rate was almost constant (about 1 LN h-1
on average) in
the reactor in the Rh2.
Unlike the Rh2, the methane content in the LF also changed shortly after feeding event. In the
Rd2, the CH4 content was 60-61% right before feeding and then dropped to a minimum value of
44-47 % at 4 h and finally increased until it reached the same value as before feeding. The CH4
content in Rd1 followed the same temporal trend as the Rd2. The only difference was the
minimum CH4 content, which was 53% at 4 h in Rd1. Methane content in the Rh2 remained
relatively stable at about 58%. Similar to methane content, the concentration of H2 in the LF was
influenced by feeding event. It increased shortly after feeding event and returned to the initial
value before the next feeding event. It increased almost by 3 and 1.5-2 folds within 1 to 2 h after
feeding event in the Rd2 and Rd1, respectively. However, H2 fluctuated slightly after feeding in
the Rh2 and it did not show any clear trend for the measurements carried out at different days for
the reactor operated under identical condition. Generally the H2 concentration was lower in the
Rh2 than in the LF.
Following feeding in case of LF, the individual and total VFA concentrations fluctuated whereas
it remained stable in the Rh2. Total VFA and acetate concentrations fluctuated by about two
folds in the Rd1 with the maximum value was observed at bout 6-8 h after feeding event. In the
Rd1, the total VFA and acetate concentration ranged 29-128 mgHACeq. L-1
and 29-114 mg L-1
,
respectively. Propionate also fluctuated by two folds from the baseline value (value right before
feeding event), which ranged 4-14 and 0.6-4 mg/L in the Rd1 during phases I and II, respectively.
When feeding interval was changed from daily to once every 2 days, the concentrations of
acetate, propionate and total VFA varied over wider range. The total VFA in the Rd2 ranged 50-
664 mgHACeq. L-1
and 52-376 mgHACeq. L-1
during phase III on days 79-81 and 105-107,
respectively whereas acetate concentration ranged 36-343 mg L-1
and 44-284
mg L
-1, respectively.
Propionate began to accumulate after feeding and ranged 2-277 and 5-83 mg L-1
. Whereas
feeding 1/12 portion of the daily substrate once every 2 h (Rh2) resulted in almost a stable
concentration of individual and total VFA before and after feeding event. The concentration of
acetate, propionate and total VFA was generally less than 42 mg L-1
, 6 mg L-1
and 63 mgHACeq.
L-1
in the Rh2 in particular during phase III.
Under all feeding regimes, the concentrations of valerate, isovalerate, butyrate and isobutyrate
were relatively small. The concentration of butyrate did not show any clear trend as propionate,
acetate and total VFA. Iso-butyrate did not show any temporal trend either except in one instance,
which ranged between 0.13 and 9 mg L-1
in the Rd2 during phase III on days 79-81. Valerate was
detected only in few occasions, in particular, during the first two weeks after changing a new
feeding regime. The trend of isovalerate fluctuation was similar as the total VFA, acetate and
propionate during only phase III on days 79-81, which ranged 1-4.4 mg L-1
in the Rd2. Generally
isovalerate fluctuated between 0.3 and 1.3 mg L-1
in the Rd1 whereas it remained stable at 1.20-
1.39 mg L-1
in the Rh2.
164
3.2 Isotope signatures under different feeding intervals
In parallel to measuring the basic process parameters, the stable carbon isotope (δ13
C) of the
produced biogas (CH4 and CO2) as well as the stable hydrogen isotope (δD) of CH4 was
monitored in short-time interval. The δ13
C of CH4 and CO2 were used to estimate the contribution
of hydrogenotrophic methanogenesis (HM) and acetoclastic methanogenesis (AM) to the
produced methane. In addition, inoculum sourced from Rd2 and R h2 were individually
incubated in batch bottles with and without CH3F. CH3F was added to specifically inhibit
acetoclastic methanogenesis (AM) and the δ13
C-CH4 in the inhibited incubation was used to
determine the isotopic signature of CH4 produced via hydrogenotrophic methanogenesis alone
(δmc).
The δ13
C-CH4 showed temporal variation following feeding event in the LF whereas it remained
relatively stable in the Rh2 (Figures 1-3). On the other hand, the δ13
C-CO2 after substrate feeding
varied in a narrow range under all feeding regimes (Figures 1-3). The δ13
C-CH4 in the Rd1
became more depleted within 2 h and later on became heavier until about 12 h and finally
returned to the same value as before feeding around 22-24 h. Data between 12 to 22 h was not
available but the trend clearly showed that the δ13
C-CH4 became lighter soon after 12 h before
finally returned to the initial value of -50 ‰. The most depleted δ13
C-CH4 in Rd1 was around -56
‰ at 2 h and the least depleted δ13
C-CH4 was around -46 ‰ at about 12 h. On the other hand, the
δ13
C-CO2 in Rd1 varied in a narrow range of 8 to 10 ‰ on average. The δ13
C-CH4 in Rd2
followed similar temporal trend as Rd1. The most and least depleted δ13
C-CH4 in Rd2 was -59 ‰
at 2 h and -46‰ at 9-11 h, respectively. The δ13
C-CO2 in Rd2 became slightly lighter over time
after feeding event from about 10 ‰ at 0 h to about 6 ‰ at 6 h and then became heavier and
finally stabilized to almost the same value as before feeding. However, the δ13
C-CH4 and δ13
C-
CO2 in Rh2 was relatively stable over time and ranged -49 to -52 ‰ and 9 to 10 ‰, respectively.
In comparison to the Rh2, the δD-CH4 in the LF showed a higher temporal variation after feeding
event. It was depleted to -368 ‰ at 2 h in the Rd1 from the initial value of -345 ‰ and later on
gradually returned to the initial value (Figure 1-3). In Rd2, it became depleted to -369 ‰ at 3-4 h
from the initial value of -348 ‰ and later on became heavier and finally returned to the initial
value. However, the temporal fluctuation of δD-CH4 in the Rh2 was relatively small in the narrow
range of -361 to -351 ‰.
The apparent carbon fractionation factor (αmc) between CO2 and CH4 is commonly employed to
identify which methanogenic pathway dominates the methane production. According to
literatures (Conrad, 2005; Qu et al., 2009), αmc > 1.065 and αmc < 1.055 indicate the
predominance of hydrogenotrophic methanogenesis (HM) and acetoclastic methanogenesis (AM)
pathways, respectively, whereas the αmc value between 1.055 and 1.065 are characteristics to the
presence of both methanogenesis pathways. The fractionation factors throughout the whole
experiment in the AM inhibited batch incubations were always higher than 1.066 (Table 2),
indicating methane was mainly produced through HM. The accumulated acetate in the CH3F
added incubation also supported that AM was inhibited by CH3F. However, the fractionation
factors were always lower in the 1.3% CH3F-added incubation than the 5% CH3F-added
incubation. This implies that 5% CH3F was more effective in inhibiting acetoclastic methanogens
under the studied incubation. Previous study (Hao et al., 2014) showed that the appropriate dose
of CH3F to specifically inhibit acetoclastic methanogens differs under different environmental
165
conditions. Our results demonstrated that the higher dose (5% CH3F) should be optimum to
estimate the value of δmc under this experimental condition (see the discussion below). The αmc
values in the non-inhibited batch incubations at 3 h were lower than 1.065 and ranged between
1.055 and 1.065 at 8 and 11 h (Table 2). This showed that methanogenesis began with a
dominance of HM but later on both pathways contributed to the production of methane. With
regards to the CSTR experiment, the αmc values in the R2h ranged between 1.055 and 1.065
throughout, indicating a mixture of contribution of both pathways. However, the αmc was higher
than 1.065 in the LF within 4 h and later on ranged between 1.055 and 1.065, showing methane
production was dominated by HM with in the first 4 h and both pathways contributed afterwards.
166
Table 2 Process parameters and carbon isotope signatures of batch incubations
Feed
Sample
Name
Sampling
time (h)
Inoculum
sourcec
DDGS
(gVS L-1
)
CH3F
(%)
CH4
(mLN)d CO2 (mLN)
d
Acetate
(mg L-1
)
Propionate
(mg L-1
)
δ13
C-CH4
(‰)f
δ13
C-CO2
(‰) αmc
Control-1a 3 Rd2 4
6.1 9.07
-63.93 8.04 1.077
8
17.9 17.46
-44.99 7.13 1.055
11
NM NMc 152.00 241.49 -50.78 7.18 1.061
Control-2a 3 Rd2 4
6.7 9.75
-63.05 7.89 1.076
8
19.1 17.79 269.84 181.54 -44.41 8.00 1.055
AM-
inhibitedb 3 Rd2 4 1.3 6.6 12.07
-66.80 7.80 1.080
8
12.9 16.52
-59.90 7.92 1.072
11
NM NM 790.58 406.29 -62.00 6.86 1.073
AM-
inhibited 3 Rd2 4 1.5 7.56 13.18
-68.98 8.56 1.083
8
7.62 12.94
-65.93 8.47 1.080
11
NM NM 1269.04 580.12 -67.43 6.76 1.080
Control-1 3 Rh2 4
9.12 12.24
-60.14 9.06 1.074
8
12.30 10.67
-41.35 9.17 1.053
11
0.48 0.42 133.93 BDLe -51.43 8.55 1.063
Control-2 3 Rh2 4
11.61 14.93
-60.57 7.66 1.073
8
12.30 10.76 255.84 82.42 -42.05 7.30 1.052
AM-
inhibited 3 Rh2 4 1.3 8.03 13.47
-66.09 8.14 1.079
8
7.52 9.62
-59.95 8.91 1.073
11
NM NM 635.05 175.18 -62.67 8.54 1.076
AM-
inhibited 3 Rh2
1.5 7.76 13.80
-68.01 8.70 1.082
8
5.95 9.87
-64.80 7.36 1.077
11 NM NM 914.55 257.23 -67.33 8.19 1.081
167
aThe experimental condition of control-1 and control-2 was identical but the latter was incubated
only for 8 h and then terminated for determining the concentration of individual VFA whereas
the former was operated for 11 h bAM-inhibited refers to those reactors where CH3F was added in order to inhibit acetoclastic
methanogenesis (CH3F) cNM refers to not measured
dData was corrected to standard temperature (273.15 K) and pressure (101.325 kPa)
eBDL refers to below detection limit
f Since the
batch incubation was prepare in a sealed bottles (closed system), the δ
13C-CH4 was
corrected for the dilution from the residual methane in the bottle. The data at 3 and 8 h are the
corrected δ13
C-CH4 whereas the data at 11 h was not corrected since the gas volume was not
measured
To estimate the contribution of each pathway to methane production, prior knowledge of the
isotopic signatures of methane produced through HM alone (δmc) and AM alone (δma) is required.
The 5% CH3F-added (AM-inhibited) batch incubation was used to estimate the value of δmc. The
δ13
C–CH4 in the 5% CH3F-added incubation was on average -68.5‰, and thus, the δmc was
assumed equal to -68.5‰. In pure culture when substrate is not limiting, the isotope fractionation
factor for the AM alone can reach up to 1.027 (equivalently isotope enrichment factor (εma) of -
27‰) for the genus Methanosarcina (Goevert & Conrad, 2009), which was the dominant
acetoclastic methanogens in all of our DDGS reactors (Figure 5). In substrate limited
environment, fractionation factor is generally lower (Goevert & Conrad, 2009). The acetate
concentration in the AM-inhibited batch incubation reached as high as 1269 mg L
-1 (Table 2)
whereas it was generally less than 365 mg L-1
in the CSTRs. This showed that acetate was
limiting in the CSTRs most of the time except a short period after feeding and hence, we assumed
a low average fractionation factor (εma = -10‰). This assumption was further supported by the
fact that the measured δ13
C–CH4 in control batch incubation at 8 h was -41‰ (Table 2) and the
δ13
C of the biomass (DDGS) was -25.5‰, indicating that the fractionation factor should be lower
than 1.015 to get δ13
C–CH4 = -41‰. Since the carbon isotope fractionation between biomass (δ13
DDGS = -25.5‰) and fermentatively produced acetate is negligible (Conrad, 2005) as well as the
fact that the δ13
of biomass may be used as a proxy for mass balance calculations (Conrad et al.,
2014), we assumed the 13
C isotopic signature of acetate (δac) is the same as the δ13
C of the
biomass (i.e, δac = -25.5‰). Using this assumption and equation 4, the value of δma was estimated
to be -35.5‰. Finally, the fractions of HM to methane production (fmc) in the CSTRs were
estimated using equation 3 (Figures 1-4).
The fmc values showed that hydrogenotrophic methanogenesis dominated (52-65%) the first 4 h
after feeding of Rd1 and later on dropped to about 32-49% of the total methane production. A
similar temporal trend of fmc was observed in the Rd2. It contributed to about 58-72% in the first
4 h and to about 32-49% afterwards. The fmc in the Rh2 fluctuated in a narrow range of 41-52%,
indicating both pathways almost equally contributed to the total methane produced during each
feeding event. We also calculated the contribution of each pathway to the total methane produced
during each feeding event (within 48 h in Rd2 and 24 h in Rd1). This value was estimated by
multiplying the volume of methane production with the fmc value at each time points. The results
showed that on average about 42-46% of the total CH4 was produced through HM in the LF,
showing that both pathways almost equally contributed to the average methane production during
each feeding event.
168
3.3 Basic process parameters in response to continuous increase of OLR
In addition to the responses of basic process parameters under different feeding intervals, their
responses were also investigated during the continuous increase of OLR (stress condition)
(Figure 4). OLR was continuously increased from the steady state condition (4 gVS L-1
d-1
) up to
11 gVS L-1
d-1
in 10 days at the same rate in both the Rd2 and Rh2. The basic process parameters
and isotope signatures within the first 48 h and the last 24 h represent data at OLR of 7 and 9 gVS
L-1
d-1
, respectively (Figure 4). In addition to these parameters, the biogas and methane yield was
measured, as expressed as specific biogas production (SBP) and specific methane production
(SMP), respectively. Although SBP and SMP are good parameters to indicate process
performance, it takes at least a day to analyze the volatile solids (VS) of the feed and difficult to
accurately determine the VS level of different input of substrates, which makes this parameter
unsuitable for a routine monitoring of biogas performance at full-scale biogas plants (Ahring et
al., 1995). Therefore, we did not consider SMP and SBP as process monitoring tools and instead
used them to indicate the process performance of the DDGS CSTRs during continuous increase
of OLR. In comparison to the steady state condition, the SBP and SMP of both Rd2 and Rh2
decreased during the continuous increase of OLR (Table 1), which reflected incomplete substrate
conversion compared to the steady state conditions. However, Rd2 was less affected by the
overfeeding stress than Rh2.
Immediately after feeding of the Rd2, biogas production increased to a peak level of 8.6 LNh-1
in
2 h and later on dropped to 1 LNh-1
before the next feeding event. However, the biogas production
remained stable (2 LNh-1
on average) in the Rh2 throughout the experiment. The peak level of 8.6
LNh-1
in the Rd2 and the average 2 LNh-1
in the Rh2 were almost two folds higher than their
corresponding values during the steady state condition.
The CH4 content in the Rd2 was 62% right before feeding and then reduced to a minimum of 33
% in 3 h and subsequently increased to 65% and finally returned to the initial value of 62%.
Hydrogen also fluctuated significantly in this reactor. It increased immediately after feeding from
the initial value of 30 ppm to a peak level of 360 ppm in 3 h and later on dropped to the same
value as before feeding. In comparison to the values at steady state, the peak value of H2
increased more than two folds. In Rd2, the CH4 content changed between 44 to 60% and 33 to
65% during the steady state and the stress conditions, respectively. On the other hand, the
composition of CH4 in the Rh2 during stress condition changed in a narrow range (55-60%)
compared to the steady state value (~ 58%). Despite an increase in OLR, H2 concentration was
relatively stable and less than 30 ppm in the Rh2.
The total and individual VFA (acetate and propionate) in the R2d during an increase in OLR
followed similar temporal trend as during the steady state condition. Despite an increase in OLR,
the peak values of acetate, propionate and total VFA in the R2d were even lower than the values
observed during the steady state condition. Although they did not show temporal variation
following feeding event, the absolute levels of acetate, propionate and total VFA in the R2h
increased by about 2 folds during the increase in OLR compared to the steady state condition.
In both Rd2 and Rh2, the concentration of isobutyrate was generally much lower than acetate and
propionate (less than 5 mg L-1
). Isobutyrate in the R2d followed similar temporal trend as acetate
and propionate. The concentrations of butyrate, valerate and iso-valerate were below detection
169
limits in both Rd2 and Rh2. The pH in both reactors increased immediately after feeding event
and afterwards dropped to almost the same value as before feeding. The change in pH was in a
narrow range (less than 0.2 units), which is the same range as the steady state condition. In
general, the pH was slightly higher in the R2d compared to one in the Rh2.
3.4 Isotope signatures in response to continuous increase in OLR
The δ13
C-CH4 in the R2d during continuous increase in OLR followed similar temporal trend as
the steady state condition (Figure 4). It became depleted from the initial value of -50 to -63 ‰ at
2 h and later on became enriched until it reached the least depleted value of -48‰ at 22 h before
it finally returned to the initial value. The same temporal trend was observed in the second
feeding event as well. At 2 h after feeding event, δ13
C-CH4 was depleted by 13 ‰ and 11‰ under
stress and steady state conditions, respectively. On the other hand, the δ13
C-CO2 became slightly
lighter over time form the initial value of about 9 ‰ and reached the most depleted value of
about 2 ‰ around 8 h. After feeding event, the δ13
C-CO2 was depleted by larger extent under
stress condition than the steady state condition (by about 6 ‰ in the former versus about 2-3‰ in
the latter). The variation in both δ13
C-CH4 and δ13
C-CO2 over time in the Rh2 was moderate and
ranged -49.5 to -54.2 ‰ and 6.7 to 8.5 ‰, respectively. These variations in δ13
C-CH4 and δ13
C-
CO2 during stress condition were almost twice in magnitude in comparison to the steady state
condition.
The temporal variation in δD-CH4 in the Rd2 under stress and steady state conditions was similar
except that the difference in magnitude of δD-CH4 depletion after feeding event. It was -355 ‰ at
0 h and was depleted to -381 ‰ at 3 h and finally returned to almost the same value as before
feeding. The most depleted δD-CH4 was about -383‰ during the second feeding event (at OLR 9
gVS d-1
L-1
; results after 48 h) whereas it was about -369 ‰ under steady state condition. After
feeding event, the δD-CH4 depleted by about 20 and 26 ‰ under steady state and stress
conditions, respectively. Moreover, the most depleted value under stress condition was lower by
about 14 ‰ compared to the steady state condition. On the other hand, the temporal variation of
δ13
D-CH4 in the R2h was smaller and ranged between -361 and -365 ‰. The most depleted δ13
D-
CH4 in the Rh2 was about -361 ‰ under steady state condition but it was -365 ‰ under stress
condition.
The αmc and fmc in the Rd2 during continuous increase in OLR showed similar temporal variation
as the steady state condition but the maximum and minimum values between the two conditions
differed slightly. The αmc and fmc in the Rh2 during continuous increase in OLR slightly
fluctuated overtime compared to the constant αmc values observed under the steady state
condition. Despite the variation in absolute values of the fmc and αmc (as a consequence of the
variation in δ13
-CH4 and δ13
C-CO2), the overall results showed that there was no inhibition of
methanogenesis under the stress condition and the proportion of CH4 produced through HM and
AM between feeding events was almost the same as the steady state condition.
170
3.5 Dynamics of methanogenic community and methanogenic pathways
The activity dynamic of the methanogenic community was determined at mcrA transcript level
(mRNA) with T-RFLP analysis of samples collected from the Rh2 and Rd2 at short time interval
(right before feeding and 4 additional time points within 48 h after feeding the Rd2) during phase
III on days 105-107 (Figure 5). In total 8 different sizes of T-RFs (55-57, 62, 78, 87, 93-95, 129,
133-134 and 238-239 bp) were detected across all samples.
Figure 5 Archaeal community structure determined at mcrA transcript level (mRNA) with T-
RFLP analysis. Samples were collected at short time interval (right before feeding Rd2 and R2h
(0 h) as well as 2, 7, 24 and 48 h after feeding Rd2) during phase III (days 105-107). Sample
name consists of two variables: Rd2 and Rh2 refer to the reactors name; the numbers indicate the
sampling time points with reference to a 0 h for the time just before feeding event.
Methanosarcina (55-57 bp); Methanoculleus (93-95 bp); Methanosaeta (129 bp).
The structure of the active methanogenic community in both Rd2 and Rh2 were mainly
dominated by the genus Methanoculleus (93-95 bp, on average 61-65%) and followed by the
genus Methanosarcina (55-57 bp, on average 23-29%). There are two sequences with low
transcript abundance which were not shared by both reactors. Methanosaeta (129 bp) activity was
detected only in the Rd2 whereas activity of a not yet identified methanogen (133-134 bp) was
detected only in the Rh2.
The relative abundance of T-RFs of the mcrA transcripts following feeding event until 48 hours
were compared to investigate the change in activity of microorganisms. In the Rd2, the relative
abundance of the genus Methanosarcina (55-57 bp) transcripts increased by 20 % in 7 h after
feeding event and gradually decreased to original value in 24 h and finally remained stable
between 24 and 48 h. However, the relative abundance of Methanosarcina transcripts stayed
relatively stable in the Rh2 over 48 h. In the Rd2, the relative abundance of Methanoculleus (93-
95 bp) transcripts decreased by 12 % in 2 h to the lowest value of 52% and then increased by
22% to a maximum level of 74%. Similar to the Methanosarcina transcripts, a far less changes of
Methanoculleus activity was observed in the Rh2 over time.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Rd2-0 Rd2-2 Rd2-7 Rd2-24 Rd2-48 Rh2-0 Rh2-2 Rh2-7 Rh2-24 Rh2-48
55-57_m 78 87 93-95_m 129 133-134_m 238-239_m
171
The results showed that diversity of the active methanogens in the two feeding regimes was
almost identical and it can be described with the dominance of the genus Methanoculleus,
followed by the genus Methanosarcina. Members of the genus Methanoculleus (strictly
hydrogenotrophic methanogens) has been widely detected in biogas digesters treating agricultural
waste products (Nettmann et al., 2010; Ziganshin et al., 2011). Methanosarcina is a multipotent
methanogenic group with members capable of utilizing acetate, H2/CO2, methanol, and
methylamines as well as metabolic possibility for acetate oxidation to CO2 and H2. Whereas
Methanosaeta (the strict acetoclastic methanogens) transcripts was detected only in the R2d with
very low abundance.
T-RFLP profiles of mcrA genes was determined at DNA level as well (data not reported). At
DNA level, the methanogenic community was dominated by the genus Methanosarcina, followed
by Methanobacterium. This discrepancy between the mRNA gene transcript and DNA level
profiles were also observed in previous studies (Munk & Lebuhn, 2014; Nikolausz et al., 2013).
The transcript level profiles closely related to the active methanogens while the community
structure observed at DNA level does not necessarily reflect the activity of the methanogens.
Nevertheless, both hydrogenotrophic and acetoclastic methanogens are represented in both
reactors (R2d and R2h). Our observation of the dominating active methanogens was supported
with the stable isotope analysis of the produced biogas, suggesting both methanogenic pathways
almost equally contributed to the total methane produced during each feeding event under all
feeding regimes but a bigger dynamics observed in R2d by both approaches.
The changes in isotopic composition of the produced biogas, methanogens activity and basic
process parameters following feeding of LF can be explained in terms of time-dependent
activation of several steps of anaerobic process. In the LF, feeding event led to the activation of
hydrolysis and acidogenesis pathways, which resulted in the production of CO2 and H2 in large
amounts shortly after feeding event (Polag et al., 2014). Hence, biogas production reached a peak
value with higher CO2 content. Within 2 h after feeding, the δ13
C-CH4 reached the most depleted
value, indicating the dominancy of HM. The higher fmc and αmc values also supported the fact that
HM was dominating the CH4 production pathway in this time interval. Compared to the δ13
C-
CH4, the δ13
C-CO2 depleted slightly. This may imply that CO2 produced from mineralization of
the fresh substrate slightly impacted the isotope composition of the large carbonate pool present
in the system (Lv et al., 2014b). Interestingly, the δD-CH4 reached the most depleted value within
2-4 after feeding event. Previous study showed that the δD-CH4 is possibly influenced by the δD
of environmental H2O, hydrogen concentration, methanogenesis pathway (HM or AM) and δD of
H2 in particular at high H2 concentration (Burke Jr, 1993). The observed highly depleted δD-CH4
in our study was still unclear, but probably influenced by the dominating HM and the increase of
H2 concentration after feeding event. Further research in controlled batch incubations with varied
concentration of hydrogen and inhibition experiment would provide the most direct answer to the
effects of H2 concentration and methanogenesis pathways on the hydrogen isotope composition of
the produced methane (δD-CH4). The increase of the acetate concentration in the LF around 8 h
after feeding coincided with most enriched δ13
C-CH4, indicating AM started to contribute to the
CH4 produced at the later stage. This temporal variation in the substrate availability (i.e methane
precursors such as CO2/H2 and acetate) is most likely influenced the isotope composition of the
produced biogas over time and hence, the proportion of methane produced by HM or AM differs
over time following feeding event. An increase in the activity of Methanosarcina and
172
Methanoculleus in the LF after feeding event also supported the observed temporal variation in
proportion of methanogenesis pathways and changes in substrate availability. Members of the
genus Methanosarcina can switch between different methanogenic pathways and a
predominantly hydrogenotrophic role is assumed after feeding in our system, while many hours
later Methanosarcina probably switched back to mainly acetoclastic pathway.
On the other hand, temporal variation in methanogenesis was minimal in the Rh2 and both
methanogenesis pathways contributed almost equally to the CH4 produced at any time. As a result
of the short feeding interval in the Rh2, perhaps the methanogens were not limited by any of the
methane precursors. This hypothesis was further supported by the fact that the amount of CO2/H2
and acetate as well as the relative abundance of Methanosarcina and Methanoculleus were stayed
relatively stable overtime in the Rh2.
3.6 Comparison of basic process parameters and isotope signatures as a process monitoring
tool
Some important factors of a good process monitoring tool are to indicate the state of the actual
process and an imbalance at an early stage before the process collapses. It is important that the
relative changes in parameter following imbalance should be significantly different from the
baseline fluctuations and measurement uncertainty (Ahring et al., 1995). Therefore, these factors
are considered for subsequent discussion about the suitability of different parameters as a process
monitoring tool.
Under different feeding intervals, individual and total VFA responded differently. Of all VFA,
acetate and propionate as well as total VFA responded to the high amount of substrate loading in
the LF. Whereas these parameters were not influenced by feeding event when smaller portions of
the daily substrate were fed in short time intervals (Rh2). The change in concentrations of
butyrate, isobutyrate, isovalerate and valerate were less obvious and indistinguishable from the
baseline fluctuations in most occasions. Therefore, they are not useful parameters for process
monitoring and control under our experimental conditions. Among VFA, the concentration of
acetate was the highest in the CSTRs but still much lower than the expected yield (compare with
the acetate determined in the acetoclastic inhibited batch experiment, Table 2). This indicates that
acetate utilization processes (acetoclastic methanogenesis and syntrophic acetate oxidation) were
very active following the feeding event in the LF. Moreover, the transient accumulation of
acetate in the LF following feeding and subsequent reduction before the next feeding event is an
indication for the adaptation and stability of the acetate utilizers. Following a short stress test
(continuous increase in OLR), the individual and total VFA in the LF showed temporal variation
as the steady state condition. Despite the continuous increase in OLR, the absolute values of these
parameters were even lower than the values measured under steady state condition. Since the
VFA consuming microorganisms are already adapted to higher loading in the LF, they may grow
fast enough to utilize VFA before it reaches inhibitory level. On the other hand, the concentration
of individual and total VFA increased by two folds under stress condition in the Rh2, showing
that this reactor is less stable under stress condition. Although total VFA, acetate and propionate
are an important parameter to indicate the kinetics of acid producing and acid consuming
microorganisms under different feeding regimes, our results showed that they are less suitable to
indicate process imbalance in the LF during an increase in OLR.
173
Under all feeding regimes, the pH change following feeding event was relatively small and
sometimes not correlated with the concentration of VFA. Moreover, the change was not
significant enough to reflect the process imbalance during the continuous increase in OLR. This
is because of the total ammonia nitrogen (TAN) level was relatively high in these reactors (3.3
and 2.7 gNH4+-N L
-1 in the LF and Rh2, respectively) and the substrate was supplemented with
ferrosorb (iron hydroxide having high pH). In this highly buffered system, pH change is less
important to indicate process imbalance, which in agreement with previous findings (Angelidaki
& Ahring, 1994).
Following feeding event of LF with high amount of substrate at once, biogas production
increased quickly with an increase in H2 concentration and decrease in CH4 content whereas it
was not affected at all in the Rh2 where 1/12 portion of the daily substrate was fed at once in 2 h
interval. Biogas production significantly increased in both Rd2 and Rh2 during the continuous
increase in OLR as well. In this period of time process imbalance was observed as the SMP and
SBP decreased. These results demonstrated that specific biogas production can be used to
indicate the overall performance of the process. However, it does not indicate process imbalance.
CH4 content and H2 concentration also responded to an increase in OLR in the Rd2 whereas they
did not change as such in the Rh2. This demonstrated that biogas composition (CH4, CO2 and H2)
is a good parameter only for LF.
Both carbon and hydrogen stable isotope of CH4 proved to reflect the actual state of
methanogenic pathways under all feeding regimes and indicate process imbalance at an early
stage. The temporal variation in δD-CH4 and δ13
C-CH4 following feeding event in the LF were
correlated with the availability of methane precursors (acetate and H2/CO2) and hence the activity
changes of hydrogenotrophic and acetoclastic methanogens. Process imbalance upon a
continuous increase in OLR was reflected in the δD-CH4 and δ13
C-CH4 values where both
parameters depleted compared to the steady state condition. This depletion was not due to
inhibition of AM, it was rather most likely a consequence of temporal increase in the relative
proportion of CH4 produced through HM as evidenced with an increase in H2/CO2 production
shortly after feeding event. The depletion of δD-CH4 was higher than the δ13
C-CH4, showing the
former is a better warning tool to process imbalance than the latter. The value of δ13
C-CO2 was
less influenced by feeding event under al feeding event except in the LF during continuous
increase in OLR. This demonstrated that δ13
C-CO2 is a good parameter to indicate process
imbalance as a consequence of an increase in OLR. Since the value of δ13
C-CO2 is required to
calculate the fmc and αmc, which gives an insight about the dominant methanogenic pathways,
δ13
C-CO2 should be monitored along with the δ13
C-CH4.
Another important factor which needs to be considered for full-scale biogas plant application is
that sampling and measurements of the intended parameters should be rapid, simple and reliable.
Gas phase measurements are generally easy to perform and more rapid than liquid phase
measurements. Most of full-scale biogas plants are equipped with commercially available gas
174
meter, optical infrared and electrochemical sensors for measuring biogas production, biogas
composition and hydrogen concentration in a gas phase. Therefore, process monitoring based on
hydrogen concentration and biogas production and composition can be employed without high
cost at full-scale biogas plants. In this study, the isotope composition of biogas was measured off-
line with GC/C/IRMS. More recently, an on-line optical system based on TDLAS was prove a
very reliable, precise, rapid and simple method for real time measurement of δ13
C-CH4 in a pilot
scale biogas plant (Polag et al., 2014). Given these benefits, development of a cheaper optical
spectrometry in future would allow the application of stable isotope techniques as a process
monitoring tool in a full-scale biogas plant. Liquid phase parameters such as individual and total
VFA usually involve off-line measurement. Moreover, it is difficult to get representative liquid
samples due to the complex, high-solid and high viscous nature of the fermentation samples.
Several attempts have been under way to develop on-line systems for determining VFA (Ward et
al., 2011) but to date none of them have been routinely applied at full-scale biogas plants.
In summary, the results showed that there is no single parameter that can fulfill all the
requirements of a good process monitoring tool as described above. Therefore, combined process
parameters including short-term changes in stable isotope composition of biogas, biogas
production, biogas composition and hydrogen concentration should be used for indicating the
actual state of the process, process performance and any process imbalance at an early stage. This
is because short-term changes of δ13
C-CH4, δ13
C-CO2 and δD-CH4 reflect the temporal change in
the availability of methane precursors (acetate and CO2/H2) and hence influence the activity of
microorganisms and the relative proportion of CH4 produced through HM and AM. Process
imbalance as a consequence of increase in OLR was better reflected in the δD-CH4 values.
Although biogas production does not reflect process imbalance, it gives an estimation on the
overall process performance. Process imbalance in the LF due to the continuous increase in OLR
was not evident from the dynamics of individual and total VFA concentrations, suggesting these
parameters are less important for indicating process imbalance in the studied DDGS CSTRs.
4. CONCLUSION
The suitability of several parameters as a process monitoring tool was investigated in DDGS
reactors run under different operating conditions (change in feeding interval and continuous
increase in OLR). Longer feeding interval led to a dynamic process, as depicted in the short term
changes of biogas production rate, biogas composition (CH4, CO2 and H2), isotope composition
of methane (δ13
C-CH4 and δD-CH4), total VFA, acetate and propionate. All these parameters
except the total and individual VFA responded to an increase in OLR. Process monitoring tool
based on stable isotope measurement of biogas in conjunction with biogas production, biogas
composition and hydrogen concentration would indicate the actual state and performance of the
process as well as process imbalance at early stage.
ACKNOWLEDGEMENTS
This research was financially supported by the Danish Strategic Research Council (Grant 10-
093944).
175
References
Ahring, B.K., Sandberg, M., Angelidaki, I. 1995. Volatile fatty acids as indicators of process
imbalance in anaerobic digestors. Applied Microbiology and Biotechnology, 43(3), 559-
565.
Angelidaki, I., Ahring, B. 1994. Anaerobic thermophilic digestion of manure at different
ammonia loads: effect of temperature. Water Research, 28(3), 727-731.
Angelidaki, I., Ellegaard, L., Ahring, B.K. 1993. A mathematical model for dynamic simulation
of anaerobic digestion of complex substrates: focusing on ammonia inhibition.
Biotechnology and Bioengineering, 42(2), 159-166.
APHA. 1995. Standard methods for the examination of water and wastewater. 19 ed. American
Public Health Association, Washington, DC, USA.
Björnsson, L., Murto, M., Jantsch, T.G., Mattiasson, B. 2001. Evaluation of new methods for the
monitoring of alkalinity, dissolved hydrogen and the microbial community in anaerobic
digestion. Water Research, 35(12), 2833-2840.
Boe, K., Steyer, J.P., Angelidaki, I. 2008. Monitoring and control of the biogas process based on
propionate concentration using online VFA measurement. in: Water Science and
Technology, Vol. 57, pp. 661-666.
Burke Jr, R.A. 1993. Possible influence of hydrogen concentration on microbial methane stable
hydrogen isotopic composition. Chemosphere, 26(1), 55-67.
Conrad, R. 2005. Quantification of methanogenic pathways using stable carbon isotopic
signatures: a review and a proposal. Organic Geochemistry, 36(5), 739-752.
Conrad, R., Claus, P., Chidthaisong, A., Lu, Y., Fernandez Scavino, A., Liu, Y., Angel, R.,
Galand, P., Casper, P., Guerin, F. 2014. Stable carbon isotope biogeochemistry of
propionate and acetate in methanogenic soils and lake sediments. Organic Geochemistry.
Franke-Whittle, I.H., Walter, A., Ebner, C., Insam, H. 2014. Investigation into the effect of high
concentrations of volatile fatty acids in anaerobic digestion on methanogenic
communities. Waste Management, 34(11), 2080-2089.
Gehre, M., Geilmann, H., Richter, J., Werner, R.A., Brand, W.A. 2004. Continuous flow 2H/1H
and 18O/16O analysis of water samples with dual inlet precision. Rapid Commun Mass
Spectrom, 18(22), 2650-60.
Goevert, D., Conrad, R. 2009. Effect of substrate concentration on carbon isotope fractionation
during acetoclastic methanogenesis by Methanosarcina barkeri and M. acetivorans and in
rice field soil. Applied and environmental microbiology, 75(9), 2605-2612.
Hahn, H., Krautkremer, B., Hartmann, K., Wachendorf, M. 2014. Review of concepts for a
demand-driven biogas supply for flexible power generation. Renewable and Sustainable
Energy Reviews, 29(0), 383-393.
Hao, L., Lü, F., Wu, Q., Shao, L., He, P. 2014. High Concentrations of Methyl Fluoride Affect
the Bacterial Community in a Thermophilic Methanogenic Sludge. PloS one, 9(3),
e92604.
Hill, D., Cobb, S., Bolte, J. 1987. Using volatile fatty acid relationships to predict anaerobic
digester failure. Transactions of the ASAE American Society of Agricultural Engineers.
Hill, D.T., Bolte, J.P. 1989. Digester stress as related to iso-butyric and iso-valeric acids.
Biological Wastes, 28(1), 33-37.
Karakashev, D., Batstone, D.J., Angelidaki, I. 2005. Influence of environmental conditions on
methanogenic compositions in anaerobic biogas reactors. Appl Environ Microbiol, 71(1),
331-8.
176
Keppler, F., Laukenmann, S., Rinne, J., Heuwinkel, H., Greule, M., Whiticar, M., Lelieveld, J.
2010. Measurements of 13C/12C methane from anaerobic digesters: Comparison of
optical spectrometry with continuous-flow isotope ratio mass spectrometry.
Environmental science & technology, 44(13), 5067-5073.
Lv, Z., Hu, M., Harms, H., Richnow, H.H., Liebetrau, J., Nikolausz, M. 2014a. Stable isotope
composition of biogas allows early warning of complete process failure as a result of
ammonia inhibition in anaerobic digesters. Bioresource technology, 167, 251-259.
Lv, Z., Leite, A.F., Harms, H., Richnow, H.H., Liebetrau, J., Nikolausz, M. 2014b. Influences of
the substrate feeding regime on methanogenic activity in biogas reactors approached by
molecular and stable isotope methods. Anaerobe, 29, 91-9.
Marchaim, U., Krause, C. 1993. Propionic to acetic acid ratios in overloaded anaerobic digestion.
Bioresource Technology, 43(3), 195-203.
Mauky, E., Jacobi, H.F., Liebetrau, J., Nelles, M. 2014. Flexible biogas production for demand-
driven energy supply - Feeding strategies and types of substrates. Bioresour Technol.
Meier-Augenstein, W. 1999. Applied gas chromatography coupled to isotope ratio mass
spectrometry. J Chromatogr A, 842(1-2), 351-71.
Mulat, D.G., Ward, A.J., Adamsen, A.P.S., Voigt, N.V., Nielsen, J.L., Feilberg, A. 2014.
Quantifying contribution of synthrophic acetate oxidation to methane production in
thermophilic anaerobic reactors by membrane inlet mass spectrometry. Environmental
science & technology, 48(4), 2505-2511.
Munk, B., Lebuhn, M. 2014. Process diagnosis using methanogenic< i> Archaea</i> in maize-
fed, trace element depleted fermenters. Anaerobe.
Nettmann, E., Bergmann, I., Pramschufer, S., Mundt, K., Plogsties, V., Herrmann, C., Klocke, M.
2010. Polyphasic analyses of methanogenic archaeal communities in agricultural biogas
plants. Appl Environ Microbiol, 76(8), 2540-8.
Nikolausz, M., Walter, R.F.H., Sträuber, H., Liebetrau, J., Schmidt, T., Kleinsteuber, S.,
Bratfisch, F., Günther, U., Richnow, H.H. 2013. Evaluation of stable isotope
fingerprinting techniques for the assessment of the predominant methanogenic pathways
in anaerobic digesters. Applied Microbiology and Biotechnology, 97(5), 2251-2262.
Polag, D., Krapf , L.C., Heuwinkel, H., Laukenmann, S., Lelieveld, J., Keppler, F. 2014. Stable
carbon isotopes of methane for real-time process monitoring in anaerobic digesters.
Engineering in Life Sciences, 14(2), 153-160.
Qu, X., Mazéas, L., Vavilin, V.A., Epissard, J., Lemunier, M., Mouchel, J.-M., He, P.-j.,
Bouchez, T. 2009. Combined monitoring of changes in δ13CH4 and archaeal community
structure during mesophilic methanization of municipal solid waste. FEMS Microbiology
Ecology, 68(2), 236-245.
Sasaki, D., Hori, T., Haruta, S., Ueno, Y., Ishii, M., Igarashi, Y. 2011. Methanogenic pathway
and community structure in a thermophilic anaerobic digestion process of organic solid
waste. J Biosci Bioeng, 111(1), 41-6.
Steinberg, L.M., Regan, J.M. 2008. Phylogenetic comparison of the methanogenic communities
from an acidic, oligotrophic fen and an anaerobic digester treating municipal wastewater
sludge. Applied and environmental microbiology, 74(21), 6663-6671.
Szarka, N., Scholwin, F., Trommler, M., Fabian Jacobi, H., Eichhorn, M., Ortwein, A., Thrän, D.
2013. A novel role for bioenergy: A flexible, demand-oriented power supply. Energy,
61(0), 18-26.
Ward, A.J., Bruni, E., Lykkegaard, M.K., Feilberg, A., Adamsen, A.P.S., Jensen, A.P., Poulsen,
A.K. 2011. Real time monitoring of a biogas digester with gas chromatography, near-
177
infrared spectroscopy, and membrane-inlet mass spectrometry. Bioresource Technology,
102(5), 4098-4103.
Whiticar, M.J. 1999. Carbon and hydrogen isotope systematics of bacterial formation and
oxidation of methane. Chemical Geology, 161(1–3), 291-314.
Wilkie, A.C., Riedesel, K.J., Owens, J.M. 2000. Stillage characterization and anaerobic treatment
of ethanol stillage from conventional and cellulosic feedstocks. Biomass and Bioenergy,
19(2), 63-102.
Ziganshin, A., Schmidt, T., Scholwin, F., Il’inskaya, O., Harms, H., Kleinsteuber, S. 2011.
Bacteria and archaea involved in anaerobic digestion of distillers grains with solubles.
Applied Microbiology and Biotechnology, 89(6), 2039-2052.
178
Chapter 10: General discussion and conclusion
The findings of the works presented in chapters 4-9 are summarized and discussed in line with the
objectives presented in chapter 1. Finally general conclusion and future research perspectives are
provided.
10.1 Developing online MIMS method for quantifying methanogenesis pathway-Paper I
Previously membrane inlet quadrupole mass spectrometry (MIMS) was used only in few studies to
determine the stable nitrogen isotope ratio of N2 for investigating denitrification and nitrogen
fixation in conjunction with 15
N labeled nitrate tracer experiment in aquatic systems (An et al.,
2001; Steingruber et al., 2001) as well as more recently for investigating the N-cycling rates and
fate in aquatic ecosystems with 15
N labeled ammonium tracer (Yin et al., 2014). In our study, for
the first time MIMS was applied to measure carbon isotope ratio of dissolved CO2 and CH4 directly
in a fermentation broth in conjunction with 13
C labeled acetate (paper 1; chapter 4). The objective
was to evaluate whether this unique method can be used to follow the incorporation of 13
C into the
products during AD of [2-13
C]acetate and to quantify the methanogenesis pathways. The inoculum
for incubating [2-13
C]acetate was obtained from a full-scale biogas plant operating with a mixture of
pig and cattle manure, maize silage and deep litter manure.
The MIMS was equipped with a cold trap system (dry ice) to condense the water vapor that
otherwise interferes with measurement of
13CH4. Dissolved carbon dioxide and methane showed a
linear dynamic range in the concentration range of 2.7-27.1 mM (r2
= 0.994) and 0.14-1.4 mM (r2
=
0.998), respectively. The effect of temperature on MIMS operation showed that safe operation of a
MIMS system is achievable for typical anaerobic digestion temperatures (30-52°C). The response
time of the instrument was faster than 1 min for both CH4 and CO2 measurement, indicating MIMS
can be used for online and onsite measurements of dissolved carbon dioxide and methane in real
time.
The results showed that the first 6 days of the [2-13
C]acetate incubation were characterized with a
typical lag phase, slow degradation of acetate and low methane production. After 6 days, the acetate
degradation rate was increased and hence, more methane was produced. The proportion of 13
CO2 to
total carbon dioxide, represented as 13
CO2 (atom%), was hardly increased in the first 6 days but
later on increased from 6% to 26% with concomitant increase of the ratio of 13
CO2(atom%)-to-13
CH4(atom%) from 0.09 to 0.85. The online MIMS measurement of the isotope composition of
CO2 and CH4 showed that the contribution of SAO–HM to the produced CH4 increased
significantly after day 6 and finally reached up to 87% of the total methane production.
Measurement of isotope composition in short time interval provided an insight into the degradation
mechanism of acetate into CH4, where an increase in SAO-HM after 6 days was associated with
significant increase in acetate degradation and hence, methane production was increased.
In previous studies, off-line GC/MS measurement of the stable carbon isotope composition of the
produced unlabeled and 13
C labeled CO2 and CH4 in gas phase (Sasaki et al., 2011) or off-line
radiometric measurement of the produced 14
CH4 and 14
CO2 in gas phase are commonly used
methods to quantify SAO-HM and AM (Karakashev et al., 2006). Radio isotope tracer experiment
is less suitable because of the requirement for strict health and safety regulations of handling radio
isotopes in a lab and waste disposal as well as its high cost associated with training personnel who
works in a radio isotope lab (Pack et al., 2011). Since these techniques are limited to off-line
179
measurements, the sampling interval for isotope analysis is often too long to have high time
resolution data High time resolution measurements are often important in the system where
methanogenesis pathway shifts over the course of acetate degradation. This was demonstrate in our
study where acetate was hardly consumed during the first 6 days of incubation, whereas later on
acetate degradation rate increased significantly with a shift in methanogenesis from AM to SAO-
HM. Therefore, MIMS is a valuable method that allows an online measurement of the isotope ratio
of CO2 and CH4 at high time resolution.
In parallel to the stable isotope measurement, proteome analysis was employed to determine the
microbial community composition in the high concentration of [2-13
C]acetate. The results showed
that the most abundant bacterial communities were the phylum Firmicutes, Proteobacteria, and
Bacteroidetes. The most abundant class in the phylum Firmicutes was Clostridia. Within the
methanogenic community, the mixotrophic Methanosarcina and hydrogenotrophic Methanoculleus
have higher abundance. Methanoculleus and Methanosarcina have been reported as widely
distributed in thermophilic anaerobic reactors, especially those treating manure and agricultural
wastes (Demirel & Scherer, 2008; Kröber et al., 2009). Previous study showed that members of the
class Clostridia with in the phylum Firmicutes exhibit diverse metabolic capabilities such as
homoacetogenesis, SAO and also produce typical fermentation products such as SCFA and alcohols
from cellulosic substrates (Zakrzewski et al., 2012). In those digesters where SAO pathway played
great role in metabolizing acetate, the methanogenic community was dominated by strict
hydrogenotrophic methanogens and the genus Methanosarcina. Our observation of higher
abundance of the order Clostridia showed that these members might play significant role in
metabolizing acetate through SAO in syntrophic association with hydrogenotrophs in our reactor.
The microbial community analysis confirmed the stable isotope analysis results, indicating the key
role of SAO-HM in degrading acetate to methane.
10.2 Role of SAO-HM to methane production from degradation of acetate-Paper II
In paper II (chapter 5), inoculum was obtained from the same full-scale biogas plant as the
experiment in paper 1 was incubated with 13
C fully labeled acetate, [U-13
C]acetate, at low (4 mM)
and high concentrations of acetate (100 mM) to investigate if different concentrations of acetate
have selective effect on methanogenesis pathways and microbial community structure. Moreover,
we aimed to determine the active microorganisms that metabolize acetate and the level of activity in
complex samples from anaerobic biogas reactor. The high concentration of acetate reactor was fed
once with 100 mM [U-13
C]acetate at the beginning of the experiment to simulate high acetate
accumulation whereas the low acetate concentration reactor was fed once daily with 4 mM [U-13
C]acetate. In this study, a combined metabolomic, metagenomic and metaproteomic approach was
applied. Online MIMS was used for tracing the incorporation of 13
C into the produced CO2 and CH4
during the course of acetate degradation. Amplicon sequencing was used to identify the microbial
community composition while protein-SIP combined with a search against two biogas
metagenomes was used to identify the proteins that incorporated 13
C and their phylogenic affiliation
and function.
The production of highly 13
C labeled CO2 and CH4 immediately after the start of incubation in both
the low and high concentrations of [U-13
C]acetate showed that CO2 and CH4 are the main products
from the degradation of acetate. The relatively large difference on the amount of acetate degraded
within the first 120 hrs among triplicates in the low acetate reactors are probably due to adaptation
of the microbial community. Almost all of the acetate was degraded to CO2 and CH4 within 120 hrs
in the high concentration reactor. Since both the methyl and carboxyl groups are 13
C labeled in the
180
[U-13
C]acetate, it is difficult to estimate the proportion of methane produced through SAO-HM
using the values of atom% of 13
CO2 and 13
CH4. The atom% of 13
CO2 was generally lower than the 13
CH4 in both the low and high concentrations of [U-13
C]acetate due to the high background pool of
unlabeled CO2 in the system. The produced 13
CH4 in the former reached a maximum of 75 atom%
and 80 atom% in the latter, indicating the production of unlabeled CH4. The results from the blank
reactor (only inoculum) showed that the background unlabeled acetate from the inoculum was
lower than 2 mM. Therefore, the contribution of the inoculum to the production of unlabeled acetate
and subsequently to unlabeled CH4 in the [U-13
C]acetate reactors is likely very low. Another
possibility for unlabeled CH4 production is via the reduction of the background pool of unlabeled
CO2 by hydrogenotrophs in syntrophy with the electron provided by SAOB in the form of H2. This
assumption is in line with the findings presented in paper I where SAO-HM played a key role for
the production of methane during the degradation of high concentration of [2-13
C]acetate. All the
experimental conditions in both papers I and II was identical except the inoculum was obtained at
different time points from the same full-scale biogas plant running with a mixture of pig and cattle
manure, maize silage and deep litter manure under the same operating conditions. The dominant
microbial communities were almost similar in both studies (see the discussion below), indicating
SAO-HM played significant role in the reactor fed with the high concentration of [U-13
C]acetate.
The stable isotope results were complemented with protein-SIP analysis of the active
microorganisms by following their 13
C incorporation into cellular proteins. Peptides labeled with 13
C were detected only in the reactor with high acetate concentration. In this reactor, five peptides
incorporated 13
C at 48 h and later on the number of peptides that incorporated
13C increased during
the course of incubation and reached 56 peptides at 192 h. Peptides from 5 subspecies of Clostridia
and one Bacteroidetes labeled with 13
C in the bacterial communities while the peptides from the
genera Methanosarcina and Methanoculleus were labeled with 13
C in the methanogenic
communities, indicating effective acetate degradation by these members of microbial communities.
The labeled Clostridia are possibly oxidizing acetate as part of a synthrophy since they contain the
fhs gene coding for the enzyme formyltetrahydrofolate synthetase. This enzyme is considered as a
possible biomarker for SAO (47, 48) and it also catalyzes the formation of acetate from H2 and CO2
via homo-acetogenesis pathway (47, 48). The highest level of 13
C incorporation was observed in the
peptide, methyl coenzyme M reductase (MCR), which take part in both methanogenic pathways
(Friedrich, 2005). Since the peptide of the hydrorogenotrophs Methanoculleus was labeled with 13
C, acetate was first possibly oxidized to CO2 by Clostridia and then subsequently reduced to
methane in syntrophic association with Methanoculleus. Moreover, the highest 13
C labeled protein
belongs to the genus Methanosarcina and the abundance of this methanogens increases during the
course of incubation as detected by amplicon sequencing analysis. This implies that the importance
of Methanosarcina in the degradation of high concentrations of acetate to methane. Since
Methanosarcina is capable of both pathways it was not possible to define the exact pathway in this
study.
Since 13
C labeled peptide was not detected in the low acetate concentration, it was not possible to
identify the active microorganisms metabolizing acetate in this reactor. However, amplicon
sequencing analysis showed that there was no statistically significant difference between the
communities at low and high concentrations of acetate. The abundance of Methanosarcina in all
reactors is possibly due to the effect of high acetate concentration in both reactors. The residual
acetate in both reactors never reached below 1.6 mM. Members of Methanosarcina are favored at
an acetate concentration higher than 1 mM (Hori et al., 2006; Karakashev et al., 2005).
181
The results presented in papers I and II as well as in recent publications (Hattori, 2008; Karakashev
et al., 2006; Krakat et al., 2010; Laukenmann et al., 2010; Sasaki et al., 2011) demonstrated that
syntrophic acetate oxidizing bacteria (SAOB) and hydrogenotrophs play a dominating role in
methane production from acetate. Recently, inoculum from a maize silage reactor was collected and
incubated with 13
C labeled acetate (either 13
CH3CO2Na or CH313
CO2Na). The results suggested that
most of the methane was produced by two step reactions of SAO and HM, demonstrating SAO-HM
can be the major pathway to methane production from acetate (Laukenmann et al., 2010). The key
role of SAO-HM demonstrated in these recent findings contradicts with the widely accepted
anaerobic digestion model (ADM1), which emphasizes that two third of methane is produced by
AM with less importance of the SAO-HM to methane production (Batstone et al., 2006). The
dominating role of AM was observed in anaerobic reactors treating wastewater and sewage sludge
as well as in natural environments such as lake sediments, rice field soil etc (Conrad, 2005; Leclerc
et al., 2004; Liu & Whitman, 2008; Zakrzewski et al., 2012). However, ADM1 need to be modified
to account for the importance of SAO-HM to methane production as observed in typical biogas
plants running with agricultural waste products (e.g. manure and deep litter) and energy crops (e.g.
maize silage).
10.3 Influence of H2 on methanogenesis and homoacetogenesis -Papers III and IV
In paper III (chapter 6), the possibility of exogenous H2 gas addition for the in situ biogas upgrading
to higher CH4 content was investigated in batch incubation fed with maize leaf as a substrate. Since
H2 is a key intermediate that regulates several steps of anaerobic digestion processes, the effects of
H2 concentration under steady state condition (without H2 addition) and high partial pressure (with
H2 addition) on the process performance, methanogenesis and homo-acetogenesis were studied.
Inoculum was obtained from a full-scale biogas plant and incubated with U-13
C labeled and
unlabeled maize leaf individually with and without H2 addition under thermophilic condition. The
purpose of using U-13
C labeled maize leaf in our study was to follow the incorporation of 13
C into
the produced CO2 and CH4 by stable isotope analysis and to identify the key active microorganisms
involved during the anaerobic digestion of U-13
C labeled maize by protein-SIP and DNA-SIP
analysis.
The methane production in all the 13
C labeled and unlabeled maize fed reactors was characterized
with typical batch incubation with an immediate methane production followed by exponential phase
until day 13 and a stationary phase afterwards. Accumulated methane until the first 8 days was very
similar in all reactors and then started to differ until the end of incubation with more methane
produced in the H2 reactors than the controls. During the first 10 days, all the added H2 was almost
completely utilized in the H2 reactors with reduction of CO2 content and an increase of CH4 content
up to 90%. The observed higher CH4 content in our H2 reactor is in agreement with previous studies
that demonstrated exogenous addition of H2 led to the reduction of CO2 to methane (Luo et al.,
2012). In a lab scale continuous stirred tank reactor (CSTR) fed with manure, H2 addition led to a
reduction in CO2 content from 38% to 15% and a corresponding increase in CH4 content as well an
increase in methane production rate by 22% (Luo et al., 2012). However, H2 consumption reduced
significantly after day 10 and accumulated to a large extent until the end of the experiment due to
unavailability of CO2 in the reactor (less than 1% of CO2 was detected). In the control reactors, H2
concentration was generally below 0.1% and CH4 content ranged between 40 to 61%. VFA
degradation rate decreased in the H2 reactors during the first few days of the incubation and finally
accumulated to a large amount whereas it was consumed in about 10 days in the control reactors.
The isotope analysis showed that highly enriched δ13
C-CH4 and δ13
C-CO2 were produced during the
182
anaerobic digestion of U-13
C labeled maize, indicating a significant amount of the carbon flows to
the production of CO2 and CH4 during the anaerobic digestion of maize as a substrate.
The carbon isotope fractionation between CO2 and CH4 (αmc) data showed that methane production
started with the dominance of HM and later on both pathways contributed to methane production in
the unlabeled control reactors. The αmc in the unlabeled H2 reactors for the first 3 days was the same
as the control reactors and significantly decreased afterwards. The observed lower αmc after day 3 in
the unlabeled H2 reactors cannot be represented as the dominance of AM since acetate was
accumulated and methane production was mainly derived from CO2 reduction during this period.
Instead it is explained according to a recently introduced new concept known as “differential
reversibility”. According to this concept, the H2 concentration is proposed as a primary controlling
factor of the ΔG values and hence affect the value of αmc (Valentine et al., 2004). The results of
microbial communities obtained from groundwater in a deep aquifer showed that under high H2
concentration, αmc was lower compared to the one at low H2 concentration, because of reduced
reversibility in the multiple enzymatic processes in CH4 production under high H2 concentration
(Hattori et al., 2012). In our study, the observed lower αmc values in the H2 reactors after day 3 is
possibly explained by the differential reversibility concept, indicating exogenous H2 addition may
have led to high H2 concentration within micro-aggregates of methanogens.
We carried out additional batch incubation to investigate why acetate degradation was inhibited in
the H2 reactors after day 10. Flushing the headspace of the H2 reactors with helium decreased the H2
concentration in the headspace and led to the degradation of acetate to almost completion. The
results demonstrated that acetate degradation was inhibited under high H2 concentration and the
inhibitory effect was reversible when the H2 concentration was kept very low. A previous study
showed that acetate remained unconsumed in an up-flow anaerobic fixed bed reactor fed
continuously with mixture of H2-CO2 and supplemented with acetate at a final concentration of 200
mg L−1
(Lee et al., 2012), which is in agreement with our findings. This underlines that during an in
situ biogas upgrading the H2 concentration may rapidly increase in the micro-aggregate of the
methanogens environment to the level that thermodynamically inhibits VFA oxidation. The
oxidation of VFA is endergonic under standard conditions and is thermodynamically feasible only
when the H2 partial pressure is kept very low (Dolfing et al., 2008; Schmidt & Ahring, 1993). The
observed lower isotope fractionation factor (αmc) in our H2 reactor also indicated that the H2
concentration in the micro-aggregate methanogens possibly increased due to the supply of H2. It is.
Therefore, suggested that H2 addition rate (or gas retention time) should be adjusted to match the
H2-uptake rate and growth of hydrogenotrophic methanogens in order to avoid VFA accumulation.
The higher acetate concentration in the H2 added reactors could also be the result of inhibition of
SAO pathway under the high H2 partial pressure. In the present study it was shown that SAO played
less importance in degradation of acetate in the AM inhibited reactor under high H2 partial pressure.
Moreover, the possible participation of homo-acetogenes in the production of acetate from CO2 and
H2 may explain the higher acetate concentration in the H2 reactors. To prove this assumption, we
carried out additional batch incubation with and without H2 addition (paper IV, chapter 7). In this
study, we have also developed a rapid and accurate GC/MS method for determining the isotope
enrichment of underivatized acetate and concentration of underivatized VFA in liquid samples from
reactors with and without H2 addition. By comparing the concentration and isotope enrichment of
acetate between reactors supplemented with H2 and without H2 helped us to prove the activity of
homo-acetogens. The concentration of acetate and the isotope enrichment of [2-13
C]acetate in
reactors with H2 addition were higher compared to those controls without H2 addition. These results
183
demonstrated that under high hydrogen partial pressure, carbon dioxide was reduced to acetate via
homo-acetogenesis. Another study showed that H2 addition greatly stimulated acetate production
through homo-acetogenesis in an anaerobic mesophilic digester operating with cattle manure as
substrate (Boone, 1982), which is in agreement with our findings.
Addition of H2 could change the microbial structure and promotes the growth of H2-consuming
microbes. Terminal restriction fragment length polymorphism (T-RFLP) analysis of 16S rRNA
genes was used in a previous study (Leybo et al., 2006) to monitor the changes in the composition
of methanogenic community in enrichment cultures under high and low H2 concentrations. H2
concentration had influenced the methanogenic community in which Methanosarcina dominated
under high H2 concentration and more diverse community was observed under low H2
concentration environment. Moreover, members of Methanosarcina could alter their metabolic
pathway from acetoclastic to hydrogenotrophic under specific conditions (Qu et al., 2009), that
could also be an alternative reason for the dominance of Methanosarcina at such high H2
concentration environment. In our H2 reactors, members of Methanosarcina might dominate the
methanogenic community and change their metabolic activity from AM to HM. Growth of pure
cultures of Methanosarcina sp. strain 227 and Methanosarcina mazei on mixtures of H2-CO2 and
acetate showed that CO2 was rapidly reduced to CH4 in the presence of H2 with low acetate
consumption rate until H2 was exhausted (Ferguson & Mah, 1983). Our observation of faster acetate
degradation rate in control reactors compared to those with addition of mixtures of H2-CO2 is in
accord with the findings of the previous study (Ferguson & Mah, 1983). Nevertheless, further data
is needed for the identification of functionally active microbes in the H2 reactors and control
reactors presented in this study. Since samples from the U-13
C labeled maize leaf reactors with and
without H2 addition have been under investigation by protein-SIP and DNA-SIP, the molecular data
that will be obtained later will increase our understanding of the identity of the microorganisms
involved in the specific metabolic processes.
10.4 Factors regulating biogas process and process monitoring tool-papers V and VI
In previous experiments (papers I-IV) most of the isotope analysis was based on measurement of
isotope composition of labeled substances in conjunction with 13
C labeled tracer substrates in a
batch incubation. In this part of the study (papers V and VI; chapters 8 and 9), the main emphasis
was on determination of isotope composition of biogas (CO2 and CH4) at natural abundance in lab-
scale CSTRs fed with distillers dried grains with solubles (DDGS) under different operating
conditions (feeding interval and increase in OLR). Short-term changes in isotope composition of
biogas and process parameters between feeding events were monitored to evaluate to what extent
these parameters reflect the actual state of the biogas process under different feeding intervals
(paper VI). Moreover, archaeal community structure was determined at mcrA transcript level
(mRNA) with T-RFLP analysis. The results showed that longer feeding interval (LF; once per day
and once every second day) led to a dynamic process, as depicted in the short term changes of
biogas production rate, biogas composition, hydrogen concentration, δ13
C-CH4, δD-CH4, total VFA,
acetate and propionate. The relative abundance of the microbial community composition between
two feeding events also changed slightly in the every 2 d fed reactor (Rd2). The changes in the
isotope composition of the produced biogas, methanogens activity and basic process parameters
following feeding of LF can be explained in terms of time-dependent activation of several steps of
anaerobic process. In the LF, feeding led to the activation of hydrolysis and acidogenesis pathways,
which resulted in the production of CO2 and H2 in large amounts shortly after a feeding (Polag et
al., 2014). Hence, biogas production reached a peak value with higher CO2 content at this time
point. The increase in the concentration of H2 after a feeding was concomitant with the depletion of
184
the δ13
C-CH4 value, indicating the dominancy of HM at this time point. Moreover, the calculated
values of the fraction of methane produced through HM (fmc) and the fractionation factor of CO2/H2
(αmc) showed that at this time point the values were higher, supporting the dominating role of HM
to the produced methane. Unlike the δ13
C-CH4, the values of δ13
C-CO2 hardly depleted right after
feeding. This may imply that CO2 produced from mineralization of the fresh substrate slightly
impacted the isotope composition of the large carbonate pool present in the system (Lv et al., 2014).
Interestingly, the δD-CH4 reached the most depleted value within 2-4 h after feeding, which is
possbly influenced by the dominating role of HM and the higher amount of the produced H2.
Previous study showed that the δD-CH4 is possibly influenced by the δD of H2O, H2 concentration,
methanogenesis pathway (HM or AM) and hydrogen isotope exchange between H2 and H2O in
particular at high H2 concentration (Burke Jr, 1993). The increase in acetate concentration in the LF
around 8 h after feeding coincided with the production of highly enriched δ13
C-CH4, indicating that
AM dominated the CH4 production pathway at the later stage. This temporal variation in substrate
availability (CO2/H2 and acetate) most likely influenced the isotope composition of the produced
biogas over time and hence, the proportion of methane produced by HM or AM differs over time
following feeding in the LF. An increase in the activity of Methanosarcina and Methanoculleus in
the LF after feeding also supported the observed temporal variation in the proportion of the
different methanogenesis pathways and the changes in substrate availability. On the other hand, the
carbon and hydrogen isotope composition of biogas in the every 2 h fed reactor (Rh2) hardly
changed following feeding. The values of fmc and αmc indicated that both methanogenesis pathways
contributed almost equally to the CH4 produced at any time. As a result of the short feeding interval
in the Rh2, perhaps the methanogens were not limited by substrate availability. This hypothesis was
further supported by the fact that the amount of CO2/H2 and acetate as well as the relative
abundance of Methanosarcina and Methanoculleus over time remained stable in the Rh2.
After investigating the effect of different feeding intervals on biogas processes, short-term stress
condition was induced in the Rd2 and Rh2 by continuously increasing OLR from the steady state
condition (4 gVS L-1
d-1
) to 11 gVS L-1
d-1
in 10 days. In comparison to the steady state condition,
the specific biogas production (SBP) and specific methane production (SMP) in both CSTRs
decreased during the time of continuous increase in OLR, which indicates process imbalance. But
the level of stress was much higher in the Rh2 than the Rd2. The short-term changes in the basic
process parameters and the stable isotope composition of biogas in the Rd2 under the stress
condition followed the same temporal trend as the steady state condition. The only difference was
the magnitude of the changes was higher in the values of the stable isotope composition of biogas,
composition of biogas, hydrogen concentration and biogas production rate. In comparison to the
steady state condition, the magnitude of depletion of δD-CH4 and δ13
C-CH4 was higher under stress
condition, indicating higher proportion of methane production through HM (not due to inhibition of
AM) as the production of H2/CO2 increased along with the 13
C and D depletion in methane. Despite
the change in magnitude of the stable isotopes, the proportion of methane produced through AM
and HM from each feeding event was identical in both Rd2 and Rh2. On the other hand, the
changes in basic process parameters and isotope signatures after feeding event were moderate in the
Rh2 during the stress condition. In comparison to the steady state condition, the concentrations of
acetate, propionate and total VFA as well as the biogas production rate increased two folds and the
CH4 content fluctuated by 4% under the stress condition. Under stress condition, the changes in the
values of δ13
C-CH4 and δ13
C-CO2 were 2 folds higher than those observed during the steady state
condition. Moreover, δ13
D-CH4 was depleted by 4‰ under stress condition compared to the steady
state condition.
185
Finally, the short-term changes in basic process parameters and stable isotope composition of
biogas were evaluated for their suitability in indicating the actual state of biogas process and
process imbalance at an early stage. It is important that the relative changes in parameter following
imbalance should be significantly different from the baseline fluctuations and measurement
uncertainty (Ahring et al., 1995). The evaluation showed that there is no single parameter that can
fulfill all the requirements of a good process monitoring tool described above. Therefore, combined
process parameters including short-term changes in stable isotope composition of biogas, biogas
production, biogas composition (CH4 content and H2 concentration) should be used for indicating
the actual state of the process, process performance and any process imbalance at an early stage.
The short-term changes of δ13
C-CH4, δ13
C-CO2 and δD-CH4 reflect the temporal change in
substrate availability (acetate and CO2/H2) and hence influence the activity of microorganisms and
the relative proportion of methane produced through HM and AM. Process imbalance as a
consequence of increase in OLR was reflected by the magnitude of the depletion of δ13
C-CH4 and
δ13
C-CO2 values to some extent and the δD-CH4 values much better. Although biogas production
does not indicate process imbalance, it is a good parameter to indicate the overall process
performance. Process imbalance in the LF due to the continuous increase in OLR was not evident
from the dynamics of individual and total VFA concentrations, suggesting these parameters are less
important for indicating process imbalance in the studied DDGS CSTRs.
In addition to the short-term parameters presented above, long-term stable isotope composition of
biogas and basic process parameters were also monitored over 128 days for evaluating the effects of
different feeding intervals and continuous increase of OLR on process performance,
methanogenesis and microbial community composition (paper V). It was also aimed to demonstrate
the principle of flexible biogas production by employing longer feeding intervals. The results of T-
RFLP profiles at DNA level showed that bacterial community composition varied a lot under the
different feeding regimes while the methanogens community remained stable. This result is not
surprising as bacteria are involved in several steps of biomass degradation from hydrolysis up to
acetogenesis whereas the archaea are responsible for methanogenesis pathways alone (De Vrieze et
al., 2013). A statistical comparison between the bacterial community structure and reactor
parameters using non-metric multidimensional scaling (nMDS) suggested that the observed T-
RFLP patterns (generated using HaeIII enzyme) was best explained by the difference in the pH
value and concentrations of TAN and H2 (statistically highly significant, p < 0.01) as well as by the
SBP and concentrations of total VFA and acetate to some extent (statistically significant, p < 0.05).
On the other hand, the pH value and concentrations of H2, TAN, total VFA and acetate were
relatively the main reactor parameters (p < 0.05) to explain the observed bacterial community
patterns (generated using MspI enzyme).
Under all feeding intervals, the T-RFLP profiles at the DNA level showed that the methanogens
community was dominated by the genus Methanosarcina, followed by Methanobacterium. Our
observation of the dominating methanogens was supported by the stable isotope analysis, indicating
on average both pathways almost equally contributed to the methane produced from each feeding
event. Such a high methane production through HM and the dominance of Methanosarcina and
hydrogenotrophs in the methanogens community were previously reported in biogas digesters
where acetate was converted to methane through SAO-HM and AM (Mulat et al., 2014; Nettmann
et al., 2010). The higher abundance of Methanosarcina and Methanobacterium was most likely due
to their tolerance to high TAN concentration in our DDGS reactors (2.6-3.6 gNH4+-N L
-1). These
two groups have higher tolerance to ammonia inhibition whereas members of the genus
Methanosaeta are sensitive to ammonia level (Karakashev et al., 2005) and may no longer be
186
detected at TAN concentration exceeding 2.5 gNH4+-N L
-1 (De Vrieze et al., 2012; Nettmann et al.,
2010).
Interestingly, the methane yield (specific methane production) was significantly different between
CSTRs fed at shorter and longer feeding intervals. In comparison to the Rh2, methane yield was
significantly higher by about 14% in the LF. The biogas production rate hardly changes in the Rh2
after a feeding event whereas it increased immediately in the LF and later on dropped, which could
simulate a fluctuating gas production to suit energy demand: feeding CSTRs at longer intervals
would allow either a higher flexibility of the electricity production or a lower demand for storage
capacity and thus save extra investment. Another key finding of this study was CSTRs fed at longer
interval appeared to have higher degree of tolerance to stress condition (continuous increase in
OLR), which is in accord with previous findings (De Vrieze et al., 2013). This demonstrates that for
even a conventional biogas plant which runs on a continuous feeding basis, less frequently feeding
event can be employed once in a while to improve functional stability.
10.5 General conclusion
The findings of this thesis demonstrated that stable isotope measurements of biogas and key
intermediates in AD have increased our understanding of methanogenesis, homo-acetogenesis, the
influence of hydrogen concentration on biogas processes, the degradation mechanism of acetate to
methane and how changes in biogas operating conditions influence the processes. Based on the
works presented in this thesis, the following conclusions are formulated.
Online MIMS allowed tracing the incorporation of 13
C into the produced CO2 and CH4 in real
time when incubated with 13
C labeled acetate in thermophilic anaerobic reactor. This novel
approach was applied for quantification of the relative contribution of syntrophic acetate
oxidation coupled to hydrogenotrophic methanogenesis pathway (SAO-HM) to methane
production from acetate, which is demonstrated to reach a high degree of contribution during
the time of fast acetate degradation rate to methane. On average, SAO-HM contributed to
almost half of the methane formation from acetate. This study underlines that MIMS due to its
fast response can be used to obtain a better insight into the temporal variation of
methanogenesis pathways.
The combined use of stable isotope measurement with MIMS, protein-SIP and metagenome in
conjunction with 13
C labeled acetate experiment was a novel approach to quantify
methanogenesis pathways, identify active microbial community composition and the level of
their activity in complex biogas digester samples. Peptides from Clostridia, the strict
hydrogenotrophs Methanoculleus and the mixotrophic Methanosarcina were labelled with 13
C
during the degradation of high concentration of 13
C labeled acetate, indicating these
microorganisms were involved in converting the acetate to methane. The production of 13
C
labeled CO2 and CH4 monitored with MIMS confirmed the key role of SAO-HM pathway for
acetate degradation. The 13
C labeled peptide of the Clostridia indicate that part of the acetate
was possibly oxidized by these bacteria in syntrophic association with the hydrogenotrophs.
Exogenous H2 gas addition for an in situ biogas upgrading showed that all the added H2 was
almost completely utilized in the presence of stoichiometric amount of CO2 and the methane
content of the biogas reached up to 90% with concomitant decrease in the CO2 content. Unlike
the control reactors, the degradation of acetate and other VFA decreased in the H2 reactors and
187
finally accumulated. Flushing the headspace of the H2 reactors with helium reduced the H2
concentrations in the headspace and led to the degradation of acetate to almost completion. The
observed lower carbon isotope fractionation between CO2 and CH4 in the H2 reactors is possibly
explained by the differential reversibility concept, indicating exogenous H2 addition may have
led to high H2 concentration within micro-aggregates of methanogens. In such high H2
concentration environment, the H2-consuming bacteria with low H2 affinity like the homo-
acetogens might be stimulated and compete with hydrogenotrophic methanogens for H2. This
underlines the importance of regulating the H2 addition rate (or gas retention time) to the H2-
uptake rate and growth of hydrogenotrophic methanogens in order to avoid VFA accumulation
during an in situ biogas upgrading process.
With an appropriate choice of a glass liner, cleaning solvent (oxalic acid) and minor sample
preparation step, this study presented a very simple, accurate, reproducible and rapid GC/MS
method for determining both the isotope enrichment of acetate and concentration of
underivatized short-chain fatty acids (SCFA) in a biogas digester sample by direct liquid
injection of acidified aqueous samples. As an example of application of this method to a biogas
process, this paper also demonstrated that a stable isotope tracer experiment in combination with
tracer-to-tracee ratio (TTR) determination by the GC/MS method can be used verify that acetate
is produced through the reduction of carbon dioxide under high hydrogen partial pressure,
proving the activity of homo-acetogenic bacteria.
Biogas production rate in the CSTRs fed with DDGS at shorter interval was relatively constant
whereas it followed a temporal trend in the CSTRs fed at longer interval which allows the
flexibility to produce more biogas at times of high energy demand. Interestingly, methane yield
was significantly higher by about 14% in the CSTRs fed at longer interval compared to those
fed at shorter interval. The bacterial community structure varied between CSTRs fed under
different feeding intervals whereas methanogens remained stable.
Longer feeding interval lead to a dynamic process, as depicted in the short-term changes of
biogas production rate, biogas composition (CH4 content, CO2 content and H2 concentration),
isotope composition of methane (δ13
C-CH4 and δD-CH4), total VFA, acetate and propionate. All
these parameters and δ13
C-CO2 except the total and individual VFA responded to an increase in
OLR, indicating these parameters can be used as a warning tool to process imbalance. Process
monitoring tool based on stable isotope measurement of biogas in conjunction with biogas
production rate and biogas composition would indicate the actual state and performance of the
process as well as process imbalance at early stage.
10.6 Perspectives
The advancement of stable isotope techniques and culture-independent molecular biology methods
has increased our understanding about biogas process. Despite this progress, many aspects are still
not completely understood. Better understanding of biogas process could provide us ways to operate
biogas plants at optimum OLR with high reliability of the plant performance, improved productivity
and substrate conversion. Therefore, future research areas that should be investigated to increase
our understanding of biogas process are discussed as follows.
188
Process optimization with the importance of SAO-HM and Methanosarcina
Acetoclastic methanogenesis is considered to be the dominant pathways for methane production in
biogas processes. Therefore, biogas optimization and operation is generally based on maintaining an
appropriate environment for acetoclastic methanogenesis with little importance given to the role of
SAO and HM. Biogas plants are usually operated with sub-optimal OLR given the sensitivity of the
strict acetoclastic methanogens Methanosaeta to the levels of fermentation products such as acetate
and ammonia. Our findings and recent studies showed the importance of SAO coupled to HM as the
key methanogenesis pathways to methane production. In those digesters where SAO-HM played a
key role, the methanogenic community is dominated with the mixotrophic Methanosarcina and
hydrogenotrophs. Members of these methanogenic communities also showed to tolerate high
acetate and ammonia concentrations. Therefore, these recent findings necessitate a re-evaluation of
the previous optimization and operating conditions with consideration of the importance of the SAO
and HM in biogas process. These open new possibilities of looking at the way to increase process
stability and efficiency in commercial biogas digesters by considering suitable conditions for
optimal growth and activity of the SAOB, Methanosarcina and hydrogenotrophs. To date there has
not been comprehensive understanding about the factors regulating the growth and activity of these
communities and this calls on further research, for instance, on ways to improve electron transfer
between syntrophic microorganisms, use of supporting material, bioagumentation of SAOB and
HM enriched mixed culture and effects of change in biogas operating conditions on these
communities etc.
Storing surplus electricity in the form of methane
With an increase in the development of wind and photovoltaic power plants, surplus electricity from
these renewable energy sources during certain periods of the day would be unavoidable. Therefore,
the concept of converting surplus electricity into energy carriers that are easier to store (e.g.
methane) is an interesting research area. The idea tested in this study is the use of hydrolyzer for
converting surplus electricity into H2 and then adding H2 into a biogas digester for converting the
CO2 in the biogas to methane. In this regard, we have showed that hydrogenotrophs are capable of
converting the CO2 to methane using H2 but SCFA accumulated in the digester with H2 addition.
Since the concentration of H2 also regulates other processes such as the degradation of SCFA
including acetate, exogenous addition of H2 in excess of the capacity of the system into the existing
biogas plants could lead to process imbalance. Another challenge that limits our understanding of
the effect of hydrogen addition in biogas process is the difficulty of estimating the hydrogen
concentration in the microenvironment close to the hydrogenotrophs. A hydrogen microsensor was
recently developed in collaboration with Unisense (Aarhus, Denmark) under the HYCON project
for measuring the concentration of dissolved hydrogen in a fermentation broth. Measurement of
dissolved hydrogen level in the liquid phase would give a better understanding about the gas–liquid
mass transfer of H2 during H2 addition in AD. Since the hydrogen concentration measured in the
bulk liquid does not necessarily represent the H2 in the immediate environment of the
microorganisms, this still requires further research. Another research interest is to study the direct
electron transfer from surplus electricity into existing biogas plants through electrochemical
systems. This requires further research, for instance on how direct electron transfer regulates
microbial community structure, and on developing a robust electrode to withstand the harsh
environments in AD etc.
189
MIMS and GC/MS development in future studies
The high precision, rapid sample throughput and high time resolution of MIMS capability makes
this method attractive for future studies of methanogenesis pathways in AD. It can be applied for
studying the inter-conversion between H2 and formate using D2 or D2O, investigating homo-
acetogenesis or HM pathways using H13
CO3- together with GC/MS for determining the
13C isotope
enrichment of acetate. Another future area of study is to couple QMS with multiple membrane
probes that allow real-time measurement of gaseous and volatile compounds from multiple
anaerobic reactors in parallel. This will enable us to study the effects of different treatments in
anaerobic reactors on methanogenesis at the same time with low cost, simplicity and rapid sample
throughput. Moreover, measuring headspace with the membrane placed in front of the ionization
source would be more straightforward and provides fast response time.
The role of homo-acetogenesis pathway in biogas reactors is in general less studied. Since the
developed GC/MS method was rapid and accurate, it can be used for future studies aimed at
investigating the role of homo-acetogenesis in a biogas process. For example, further experiment by
AD of a particular substrate in the inoculum buffered with 13
C labeled bicarbonate and subsequent
tracing the fate of 13
C in acetate and methane by GC/MS and/or IRMS to quantify homo-
acetogenesis and methanogenesis pathways. Moreover, the method could be further optimized for
measuring the isotope enrichment of other SCFA such as propionate and butyrate. Since SCFA are
key intermediates of anaerobic digestion, an understanding of the production and consumption of
SCFA would enable improved process control and optimization strategies.
Effects of feeding managements on process performance
We have shown that feeding at longer intervals improves methane yield and stability of biogas
process in lab-scale CSTRs fed with DDGS. Nevertheless further research is needed to demonstrate
the benefit of less frequent feeding regime in full-scale DDGS reactors. Since different feedstocks
have different chemical composition and structure, further research should be aimed at investigating
the effects of different feeding intervals using a variety of feedstocks. The results would improve
our understanding of how different feedstocks and feeding intervals influences process
performance. Feeding at longer interval also provides the flexibility to produce more biogas during
the times of high energy demand. This shows that further research on feeding management
including change in substrate (depending on the hydrolysis rate of substrates) and different feeding
intervals should be investigated in future to demonstrate the concept of flexible biogas production.
This area of research has a potential since biogas plants can be operated flexibly to balance the
supply of electricity based on fluctuating renewable energy sources.
Process monitoring tool and control strategy
An appropriate process monitoring and control system is needed to operate biogas plants at
optimum condition and detect process imbalances at an early stage and take corrective measures in
due time. Our study demonstrated that measurements of stable isotope composition of biogas (CH4
and CO2) can be used to indicate the actual state of process and as an early warning tool to process
imbalance. In particular, the short-term change in hydrogen isotope of CH4 was very large during
process imbalance, indicating the potential of this parameter to indicate process imbalance at an
early stage. Nevertheless, the reason behind the changes in hydrogen isotope of CH4 is still unclear.
Therefore, further research should be done to increase our understanding of the factors that control
190
the magnitude of hydrogen isotopes in CH4 and hence, dictate the fractionation factor of hydrogen
isotope in CH4. Therefore, future experiments should be done in a controlled batch incubation under
different concentrations of H2, or using tracer experiments (e.g D2O) to investigate how H2
concentration and the hydrogen isotope of water affects hydrogen isotope of the produced CH4.
Moreover, for practical applicability of isotope techniques to be used at a biogas plant, cheaper and
portable optical spectrometers should be developed. CRDS is relatively less expensive and more
compact than TDLAS. Technological advancements in CRDS will likely make it even more
affordable and practical for using at commercial full-scale biogas plants. Optical spectroscopy has
been applied to measure the carbon isotope ratio of methane at a pilot scale biogas plant but it has
not yet been demonstrated to work for the measurement of hydrogen isotope ratio of CH4 at a
biogas plant. Optical spectrometer has already been demonstrated to be suitable for measuring the
carbon and hydrogen isotope ratios simultaneously in a lab environment (Webster & Mahaffy,
2011). Therefore future research should test the practicality of the instrument for measuring
hydrogen isotope of CH4 at a biogas plant.
References
Ahring, B.K., Sandberg, M., Angelidaki, I. 1995. Volatile fatty acids as indicators of process
imbalance in anaerobic digestors. Applied Microbiology and Biotechnology, 43(3), 559-565.
An, S., Gardner, W.S., Kana, T. 2001. Simultaneous measurement of denitrification and nitrogen
fixation using isotope pairing with membrane inlet mass spectrometry analysis. Appl.
Environ. Microbiol., 67(3), 1171-1178.
Batstone, D.J., Kelle, J., Steyer, J. 2006. A review of ADM1 extensions, applications, and analysis:
2002-2005. Water Sci. Technol., 54(4), 1-10.
Boone, D.R. 1982. Terminal reactions in the anaerobic digestion of animal waste. Applied and
environmental microbiology, 43(1), 57-64.
Burke Jr, R.A. 1993. Possible influence of hydrogen concentration on microbial methane stable
hydrogen isotopic composition. Chemosphere, 26(1), 55-67.
Conrad, R. 2005. Quantification of methanogenic pathways using stable carbon isotopic signatures:
a review and a proposal. Organic Geochemistry, 36(5), 739-752.
De Vrieze, J., Hennebel, T., Boon, N., Verstraete, W. 2012. Methanosarchina: The rediscovered
methanogen for heavy duty biomethanation. Bioresour. Technol. , 112, 1-9.
De Vrieze, J., Verstraete, W., Boon, N. 2013. Repeated pulse feeding induces functional stability in
anaerobic digestion. Microb Biotechnol, 6(4), 414-24.
Demirel, B., Scherer, P. 2008. The roles of acetotrophic and hydrogenotrophic methanogens during
anaerobic conversion of biomass to methane: a review. Rev. Environ. Sci. Biotechnol. , 7(2),
173-190.
Dolfing, J., Jiang, B., Henstra, A.M., Stams, A.J., Plugge, C.M. 2008. Syntrophic growth on
formate: a new microbial niche in anoxic environments. Applied and environmental
microbiology, 74(19), 6126-6131.
Ferguson, T.J., Mah, R.A. 1983. Effect of H2-CO2 on methanogenesis from acetate or methanol in
Methanosarcina spp. Applied and environmental microbiology, 46(2), 348-355.
Friedrich, M.W. 2005. Methyl-coenzyme M reductase genes: Unique functional markers for
methanogenic and anaerobic methane-oxidizing Archaea. in: Methods in Enzymology, Vol.
397, pp. 428-442.
191
Hattori, S. 2008. Syntrophic acetate-oxidizing microbes in methanogenic environments. Microbes
Environ., 23(2), 118-127.
Hattori, S., Nashimoto, H., Kimura, H., Koba, K., Yamada, K., Shimizu, M., Watanabe, H., Yoh,
M., Yoshida, N. 2012. Hydrogen and carbon isotope fractionation by thermophilic
hydrogenotrophic methanogens from a deep aquifer under coculture with fermenters.
Geochemical Journal, 46(3), 193-200.
Hori, T., Haruta, S., Ueno, Y., Ishii, M., Igarashi, Y. 2006. Dynamic transition of a methanogenic
population in response to the concentration of volatile fatty acids in a thermophilic
anaerobic digester. Applied and environmental microbiology, 72(2), 1623-1630.
Karakashev, D., Batstone, D.J., Angelidaki, I. 2005. Influence of environmental conditions on
methanogenic compositions in anaerobic biogas reactors. Appl Environ Microbiol, 71(1),
331-8.
Karakashev, D., Batstone, D.J., Trably, E., Angelidaki, I. 2006. Acetate oxidation is the dominant
methanogenic pathway from acetate in the absence of Methanosaetaceae. Appl. Environ.
Microbiol., 72(7), 5138-5141.
Krakat, N., Schmidt, S., Scherer, P. 2010. Mesophilic fermentation of renewable biomass: does
hydraulic retention time regulate methanogen diversity? Applied and environmental
microbiology, 76(18), 6322-6326.
Kröber, M., Bekel, T., Diaz, N.N., Goesmann, A., Jaenicke, S., Krause, L., Miller, D., Runte, K.J.,
Viehöver, P., Pühler, A. 2009. Phylogenetic characterization of a biogas plant microbial
community integrating clone library 16S-rDNA sequences and metagenome sequence data
obtained by 454-pyrosequencing. J. Biotechnol., 142(1), 38-49.
Laukenmann, S., Polag, D., Heuwinkel, H., Greule, M., Gronauer, A., Lelieveld, J., Keppler, F.
2010. Identification of methanogenic pathways in anaerobic digesters using stable carbon
isotopes. Engineering in Life Sciences, 10(6), 509-514.
Leclerc, M., Delgènes, J.P., Godon, J.J. 2004. Diversity of the archaeal community in 44 anaerobic
digesters as determined by single strand conformation polymorphism analysis and 16S
rDNA sequencing. Environmental microbiology, 6(8), 809-819.
Lee, J.C., Kim, J.H., Chang, W.S., Pak, D. 2012. Biological conversion of CO2 to CH4 using
hydrogenotrophic methanogen in a fixed bed reactor. Journal of Chemical Technology and
Biotechnology, 87(6), 844-847.
Leybo, A., Netrusov, A., Conrad, R. 2006. Effect of hydrogen concentration on the community
structure of hydrogenotrophic methanogens studied by T-RELP analysis of 16S rRNA gene
amplicons. Microbiology, 75(6), 683-688.
Liu, Y., Whitman, W.B. 2008. Metabolic, phylogenetic, and ecological diversity of the
methanogenic archaea. Annals of the New York Academy of Sciences, 1125(1), 171-189.
Luo, G., Johansson, S., Boe, K., Xie, L., Zhou, Q., Angelidaki, I. 2012. Simultaneous hydrogen
utilization and in situ biogas upgrading in an anaerobic reactor. Biotechnology and
bioengineering.
Lv, Z., Leite, A.F., Harms, H., Richnow, H.H., Liebetrau, J., Nikolausz, M. 2014. Influences of the
substrate feeding regime on methanogenic activity in biogas reactors approached by
molecular and stable isotope methods. Anaerobe, 29, 91-9.
Mulat, D.G., Ward, A.J., Adamsen, A.P.S., Voigt, N.V., Nielsen, J.L., Feilberg, A. 2014.
Quantifying contribution of synthrophic acetate oxidation to methane production in
thermophilic anaerobic reactors by membrane inlet mass spectrometry. Environmental
science & technology, 48(4), 2505-2511.
192
Nettmann, E., Bergmann, I., Pramschufer, S., Mundt, K., Plogsties, V., Herrmann, C., Klocke, M.
2010. Polyphasic analyses of methanogenic archaeal communities in agricultural biogas
plants. Appl Environ Microbiol, 76(8), 2540-8.
Pack, M.A., Heintz, M.B., Reeburgh, W.S., Trumbore, S.E., Valentine, D.L., Xu, X., Druffel, E.R.
2011. A method for measuring methane oxidation rates using low-levels of 14C-labeled
methane and accelerator mass spectrometry. Limnol. Oceanogr.-Methods, 9, 245-260.
Polag, D., Krapf , L.C., Heuwinkel, H., Laukenmann, S., Lelieveld, J., Keppler, F. 2014. Stable
carbon isotopes of methane for real-time process monitoring in anaerobic digesters.
Engineering in Life Sciences, 14(2), 153-160.
Qu, X., Mazéas, L., Vavilin, V.A., Epissard, J., Lemunier, M., Mouchel, J.M., He, P.j., Bouchez, T.
2009. Combined monitoring of changes in δ13CH4 and archaeal community structure
during mesophilic methanization of municipal solid waste. FEMS microbiology ecology,
68(2), 236-245.
Sasaki, D., Hori, T., Haruta, S., Ueno, Y., Ishii, M., Igarashi, Y. 2011. Methanogenic pathway and
community structure in a thermophilic anaerobic digestion process of organic solid waste. J.
Biosci. Bioeng., 111(1), 41-46.
Schmidt, J.E., Ahring, B.K. 1993. Effects of hydrogen and formate on the degradation of propionate
and butyrate in thermophilic granules from an upflow anaerobic sludge blanket reactor.
Applied and environmental microbiology, 59(8), 2546-2551.
Steingruber, S.M., Friedrich, J., Gächter, R., Wehrli, B. 2001. Measurement of denitrification in
sediments with the 15N isotope pairing technique. Appl. Environ. Microbiol., 67(9), 3771-
3778.
Valentine, D.L., Chidthaisong, A., Rice, A., Reeburgh, W.S., Tyler, S.C. 2004. Carbon and
hydrogen isotope fractionation by moderately thermophilic methanogens. Geochimica et
Cosmochimica Acta, 68(7), 1571-1590.
Webster, C.R., Mahaffy, P.R. 2011. Determining the local abundance of Martian methane and its’
13C/12C and D/H isotopic ratios for comparison with related gas and soil analysis on the
2011 Mars Science Laboratory (MSL) mission. Planetary and Space Science, 59(2–3), 271-
283.
Yin, G., Hou, L., Liu, M., Liu, Z., Gardner, W.S. 2014. A Novel Membrane Inlet Mass
Spectrometer Method to Measure 15NH4+ for Isotope-Enrichment Experiments in Aquatic
Ecosystems. Environmental science & technology, 48(16), 9555-9562.
Zakrzewski, M., Goesmann, A., Jaenicke, S., Jünemann, S., Eikmeyer, F., Szczepanowski, R., Al-
Soud, W.A., Sørensen, S., Pühler, A., Schlüter, A. 2012. Profiling of the metabolically
active community from a production-scale biogas plant by means of high-throughput
metatranscriptome sequencing. Journal of Biotechnology, 158(4), 248-258.