ww.sciencedirect.com

wat e r r e s e a r c h 5 3 ( 2 0 1 4 ) 2 4 9e2 5 8

Available online at w

ScienceDirect

journal homepage: www.elsevier .com/locate/watres

Conventional mesophilic vs. thermophilicanaerobic digestion: A trade-off betweenperformance and stability?

Rodrigo A. Labatut*, Largus T. Angenent, Norman R. Scott

Department of Biological and Environmental Engineering, Cornell University, Ithaca, NY 14853, USA

a r t i c l e i n f o

Article history:

Received 7 October 2013

Received in revised form

6 January 2014

Accepted 18 January 2014

Available online 28 January 2014

Keywords:

Co-digestion

Manure

Temperature

Loading rate

Mixing intensity

LCFA

Adsorption

* Corresponding author. 225 Riley-Robb Hall,E-mail address: labatut@cornell.edu (R.A

0043-1354/$ e see front matter ª 2014 Elsevhttp://dx.doi.org/10.1016/j.watres.2014.01.035

a b s t r a c t

A long-term comparative study using continuously-stirred anaerobic digesters (CSADs)

operated at mesophilic and thermophilic temperatures was conducted to evaluate the

influence of the organic loading rate (OLR) and chemical composition on process perfor-

mance and stability. Cow manure was co-digested with dog food, a model substrate to

simulate a generic, multi-component food-like waste and to produce non-substrate spe-

cific, composition-based results. Cow manure and dog food were mixed at a lower e and an

upper co-digestion ratio to produce a low-fiber, high-strength substrate, and a more

recalcitrant, lower-strength substrate, respectively. Three increasing OLRs were evaluated

by decreasing the CSADs hydraulic retention time (HRT) from 20 to 10 days. At longer HRTs

and lower manure-to-dog food ratio, the thermophilic CSAD was not stable and eventually

failed as a result of long-chain fatty acid (LCFA) accumulation/degradation, which was

triggered by the compounded effects of temperature on reaction rates, mixing intensity,

and physical state of LCFAs. At shorter HRTs and upper manure-to-dog food ratio, the

thermophilic CSAD marginally outperformed the biomethane production rates and sub-

strate stabilization of the mesophilic CSAD. The increased fiber content relative to lipids at

upper manure-to-dog food ratios improved the stability and performance of the thermo-

philic process by decreasing the concentration of LCFAs in solution, likely adsorbed onto

the manure fibers. Overall, results of this study show that stability of the thermophilic

co-digestion process is highly dependent on the influent substrate composition, and

particularly for this study, on the proportion of manure to lipids in the influent stream. In

contrast, mesophilic co-digestion provided a more robust and stable process regardless of

the influent composition, only with marginally lower biomethane production rates (i.e., 7%)

for HRTs as short as 10 days (OLR ¼ 3 g VS/L-d).

ª 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Thermophilic anaerobic digestion (55e60 �C) has the potential

to produce higher biomethane yields, and a more organically-

Cornell University, Ithac. Labatut).

ier Ltd. All rights reserved

stable, pathogen-free effluent compared to conventional

mesophilic digestion (35e40 �C). However, up until now most

commercial-scale anaerobic digesters are operated at meso-

philic temperatures. In addition to the higher energy input,

poor stability and reliability of the thermophilic process are

a, NY 14853, USA. Tel.: þ1 607 339 9429.

.

wat e r r e s e a r c h 5 3 ( 2 0 1 4 ) 2 4 9e2 5 8250

probably the main reasons that have prevented its adoption.

Why thermophilic digestion is unstable is not well under-

stood. Main hypotheses point to a less diverse microbial

community (Leven et al., 2007; Raskin et al., 1994), persistence

of propionate (Speece et al., 2006; Wilson et al., 2008), and

increased intermediate toxicity at thermophilic temperatures.

Temperature and influent substrate may be the most

important parameters determining performance and stability

of the anaerobic digestion process. Together, they influence

the microbial community structure, the biochemical conver-

sion pathways, the kinetics and thermodynamic balance of

the biochemical reactions, and the stoichiometry of the

products formed. Because formation and consumption of

products can occur at different rates, transient accumulation

of potentially inhibitory substances is possible, particularly

with complex substrates. Accumulation of such substances

can slow down or interrupt the digestion process by disrupt-

ing the homeostatic equilibrium of microbes, and/or by

imposing thermodynamic constraints to biochemical re-

actions e inhibitory conditions that usually result in

decreased biomethane production rates and accumulation of

volatile fatty acids. Furthermore, as process temperatures

increase, so does the rate of reactions and the likelihood of

inhibition.

There are several substrates capable of producing inter-

mediate products with the potential of causing inhibition and

instability; however, the most commonly found in commer-

cial digester operations are probably protein- and lipid-based.

Urea- and protein-rich wastes, such as those sourced from

concentrated animal feeding operations (i.e., CAFO’s), food

processing industries, and slaughterhouses, can create high

levels of ammonia in anaerobic digesters. Total ammonia

(NH3 þ NHþ4 ), and particularly its unionized form (NH3), are

inhibitory to methanogens (Koster and Lettinga, 1988;

Angelidaki and Ahring, 1993; Kayhanian, 1994). This is espe-

cially true in thermophilic digesters, where inhibition and

instability have been attributed to the higher concentrations

of NH3 due to its pH interdependency with temperature

(Angelidaki and Ahring, 1994; Chen et al., 2008; Hansen et al.,

1998). Lipid-based, or fats, oil and grease (FOG) wastes, are

usually added as co-substrates in full-scale manure-based co-

digestion systems; mainly sourced from olive oil production,

and food and fish processing industries. Lipid-rich wastes can

be highly inhibitory, particularly of the b-oxidation and

methanogenesis steps (Neves et al., 2009a; Hanaki et al., 1981),

due to the accumulation of long-chain fatty acids (LCFA)

resulting from the hydrolysis of neutral lipids. As with

ammonia, it has also been shown that the inhibitory effect of

LCFA is more pronounced at thermophilic temperatures (Hwu

Table 1 e Summary of the operating conditions evaluated in e

Period (P) Days OLR (g

Mesophilic Thermophilic

Start-up 0e62 0e62 1

I 63e330 63e330 1.5

II 331e430 331e360 2

III 431e498 361e498 2

IV 499e544 499e544 3

and Lettinga, 1997). LCFA accumulation occurs when molec-

ular hydrogen (and/or acetate), a major product of b-oxidation

accumulates to thermodynamically-limiting levels that pre-

vent LCFA (and propionate) to be oxidized. Likewise, higher

accumulation of propionate and resulting inhibition has been

related to increased temperatures (Speece et al., 2006; Wilson

et al., 2008; Kim et al., 2002).

Indeed, thermophilic digesters have been reported to be

more susceptible to inhibition (Angelidaki and Ahring, 1994;

Hwu and Lettinga, 1997; Hansen et al., 1999) and sudden

environmental changes (Biey et al., 2003; Khanal et al., 2010;

Nguyen et al., 2007; VanLier et al., 1996; Zinder, 1986), than

mesophilic digesters. However, two characteristics of ther-

mophilic digestion have been recognized as important ad-

vantages over mesophilic (and psychrophilic) digestion. First,

their capability to produce Class A biosolids, which are

essentially pathogen free streams, with no restrictions on

crop type, harvesting, or site access for land application

(USEPA, 2000); and second, their increased degradation rates,

which can result in increased solids destruction and bio-

methane production rates, and lead to shorter retention times

and/or smaller system footprints.

The aim of this study was to comparatively evaluate per-

formance and stability of mesophilic and thermophilic

anaerobic co-digestion; particularly, the hypotheses that

suggest increased biomethane production and treatment ef-

ficiency of thermophilic digesters, but less stability and reli-

ability compared to its mesophilic counterparts. Specifically,

we evaluated the compounded effects of operating tempera-

ture and substrate chemical composition by co-digesting cow

manure and dog food at two different ratios and three influent

loading rates. Dog food was used to both, simulate the multi-

component properties of a food-like waste, and produce

composition-based results, thereby more suitable for

modeling purposes.

2. Materials and methods

2.1. Experimental design and operating conditions

Cow manure and dry dog food were co-digested at either

mesophilic (37�1 �C) or thermophilic (55�1 �C) temperatures

in two, otherwise identical continuously-stirred anaerobic

digesters (CSAD). Two co-digestion ratios and three organic

loading rates (OLRs) were evaluated over four study periods to

create a set of distinct operating conditions (Table 1). Cow

manure and dog food were mixed at an upper and a lower

co-digestion ratio to create significantly distinct influent

ach CSAD during the four study periods (P).

/L-d) HRT (d) Composition (% VS basis)

Manure Dog food

30 100, 50, 25% Balance

20 25% 75%

15 25% 75%

15 75% 25%

10 75% 25%

wat e r r e s e a r c h 5 3 ( 2 0 1 4 ) 2 4 9e2 5 8 251

substrate compositions while OLRs were increased by

decreasing the CSADs hydraulic retention times (HRTs). The

overall influent volatile solids concentration from themixture

of the two substrates was set to be constant at 30 g VS/L

throughout the study. CSADs were fed manually every

48 � 1 h. Each CSAD had a working volume of 4.5 L and was

continuously stirred by a mechanical mixer (Model 5vb, EMI,

Inc., Clinton, CT) with a 62-mm diameter axial flow impeller

(Lightnin A-310, Rochester, NY). Further details on the CSADs

setup are provided in Labatut (2012) and Usack et al. (2012).

The inoculum used to start-up the anaerobic digesters was

obtained from a mesophilic, farm-based completely-mixed

anaerobic digester (Ridgeline Dairy Farm, Clymer, NY) co-

digesting cow manure and a mixture of high-strength, food-

processing wastes at a 20-day HRT.

2.2. Influent feed

Cow manure was collected at the beginning of the study from

the influent tank of a dairy farm anaerobic digester treating

the daily waste of 600 cows (AA Dairy, Candor, NY). Manure

was homogenized, blended, and initial analyses were per-

formed before storing it in individual 1-L containers at �20 �C.Dry dog food pellets were used to emulate a generic readily-

degradable, high-strength food-like waste. The dog food

from Science Diet (Hill Pet Nutrition, Inc.) was used, because it

contains a highly-reproducible and well-balanced mixture of

carbohydrates, lipids, and proteins. Pellets were ground in an

industrial blender and sieved to produce a feed particle size in

the range of 1e2mm. A complete chemical characterization of

the influent substrate is shown in Table 2.

2.3. Steady-state conditions

A CSAD was considered to be at steady state after the

following criteria were met: (1) the reactor had been contin-

uously operated for three HRT cycles at a fixedHRT; and (2) the

volumetric biogas production rate (L/L-d) has been stable and

had not varied more than 10% during three HRT cycles. Once

at steady-state conditions, a comprehensive set of samples

and measurements was obtained from each digester to

determine final, steady-state performance parameters. Sta-

tistical significance between the mesophilic and thermophilic

CSAD was performed in Microsoft Excel based on the Stu-

dent’s t-test. In addition, a continuous monitoring of selected

Table 2 e Chemical composition of dairy manure, dogfood and resulting influent for the two co-digestionratios; values in g/100 g (VS basis).

Constituent Dairymanure(Mn)

Dogfood(DF)

MnDF2575 MnDF7525

Protein 14.9 34.5 29.6 19.8

Lipids 4.7 30.8 24.3 11.2

Sugars, starch,

pectin

24.4 29.9 28.5 25.8

Hemicellulose 9.6 e 2.4 7.2

Cellulose 32.6 4.8 11.7 25.6

Lignin 13.8 e 3.5 10.4

performance parameters was conducted in both CSAD

throughout the 48-h feeding cycle to investigate degradation

kinetics.

2.4. Analytical methods

Biogas production was monitored at the end of every feeding

cycle using a wet gas meter (Actaris Meterfabriek bv, Delft,

The Netherlands), and reported at STP (0 �C and 1 atm).

Digestate pH was measured with the same frequency using a

single-reference electrode (Thermo Fisher Scientific, Inc.).

Parameters were determined periodically on both CSADs.

Methane, carbon dioxide, and hydrogen sulfide content in the

biogas were measured using a SRI 8610C (SRI Instruments,

Torrance, CA) gas chromatograph (GC) equipped with a ther-

mal conductivity detector (TCD) and a flame photometric de-

tector (FPD), using Helium as a carrier gas in a 1-m Rt-

XLSulfur� packed column and a ramped temperature pro-

gram. Total solids (TS), total volatile solids (VS), total Kjeldahl

nitrogen (TKN), chemical oxygen demand (COD), and total

volatile fatty acids (VFA) were determined according to Stan-

dardMethods, sections 2540B, 2540E, 4500B, 5220C, and 5560C,

respectively (APHA, 1995). Individual VFA were determined

with an HP Agilent GC model 5890 equipped with a flame

ionization detector (FID), using helium as a carrier gas in a

NUKOL� capillary column. A commercially prepared 10-mM

volatile fatty acids (VFA) standard mixture, containing, ace-

tic (C2), propionic (C3), isobutyric (C4), butyric (C4), isovaleric

(C5), valeric (C5), isocaproic (C6), caproic (C6), and enanthic

(C7) acids, was obtained from Sigma Aldrich Co. Long chain

fatty acids (LCFA) were determined using the same GC setup

but following a different temperature program and sampling

preparation adapted from Neves et al. (2009b). The LCFA stock

solution was prepared in our lab using dichloromethane as a

solvent, and a mixture of capric (C10), lauric (C12), myristic

(C14), palmitic (C16), stearic (C18:0), oleic (C18:1), and linoleic

(C18:2) acids as standards, and pentadecanoic acid (C15) as an

internal standard (IS) e all reagents were HPLC grade and

obtained from SigmaeAldrich Co. Total ammonia-N (TAN)

concentration was measured using an ion selective electrode

(Thermo Fisher Scientific, Inc.). Total organic nitrogen was

calculated by subtracting TAN from TKN. Total protein con-

tent was calculated based on the assumption that an average

protein contains 16% organic N. Neutral lipids were deter-

mined according to method of Loehr and Rohlich (1962) for

wastewaters. Hemicellulose, cellulose, and lignin content

were determined according to the neutral detergent fiber

(NDF) and acid detergent fiber and lignin (ADF/ADL) analyses

described by Mertens et al. (2002) and Moller (2009), respec-

tively. Non-lignocellulosic carbohydrates (e.g., sugars, starch)

were obtained using the method of Gaudy (1962).

3. Results and discussion

Both mesophilic and thermophilic CSADs were operated

continuously for a period of over 540 days. Acclimation of

digesters was conducted at an OLR of 1 g VS/L-d (HRT ¼ 30 d)

by first feeding cow manure only and then introducing dog

food in a step-wise manner until the target co-digestion

Table 3 e Average of main effluent parameters measuredduring steady-state conditions at each period.

CSAD Mesophilic Thermophilic

Period I II III IV I II III IV

TS (g/L) 10.6 14.1 21.9 21.3 12.6 F 21.6 20.4

VS (g/L) 8.6 11.7 18.2 17.4 10.6 F 17.7 16.5

wat e r r e s e a r c h 5 3 ( 2 0 1 4 ) 2 4 9e2 5 8252

proportion was reached (i.e., 75% VS basis) (Fig. 1). As expected

during this period, biogas production increased as total VFAs

accumulated and pH decreased in both CSADs (Fig. 1). When

biogas production rate and total VFAs were steady and com-

parable in both reactors (on day 62), feeding rate was

increased to reach the target OLR of 1.5 g VS/L-d (HRT ¼ 20

days) (Table 1).

COD (g/L) 10.9 15.8 22.3 19.3 17.2 F 19.8 21.8

pH (pH units) 7.24 7.27 7.28 7.29 7.37 F 7.72 7.69

TAN (mg/L) 1038 1319 764 786 884 F 791 1008

VFA (mg/L) 47.6 622.9 19.9 129.2 1064.6 F 68.0 184.2

F: digester failure.

3.1. Performance of CSADs at steady-state conditions

3.1.1. Effluent characterizationAs anticipated, TS, VS, and COD concentrations increased in

the effluent as the OLR increased due to hydrolysis rate limi-

tations (Table 3). The same was observed when the content of

slowly degradable lignocellulosic manure in the mesophilic

CSAD was increased from 25% to 75% while maintaining a

constant OLR at 2 g/L-d (HRT ¼ 15 days). The higher pH

observed in the thermophilic digester relative to the meso-

philic has been reported in several studies (Watanabe et al.,

1997; Gavala et al., 2003; Song et al., 2004), and it is predicted

by bioenergetics (see Supporting Information). Also, addi-

tional alkalinity produced due to increased protein degrada-

tion in Periods III and IV would explain the rather large pH

difference between both CSADs observed during those periods

(Table 3). Total ammonia-N (TAN) concentrations were well

below reported inhibitory levels (Chen et al., 2008) in both

digesters during the entire study. However, higher TAN con-

centrations were observed in the thermophilic CSAD during

Period III and especially at IV (Table 3). This was expected,

since production of ammonia is directly related to the degra-

dation of proteinaceous material, which was found to be

higher in the thermophilic CSAD (Labatut, 2012). With the

Fig. 1 e Biogas production rates (STP), total VFA (as acetic acid)

mesophilic CSAD, solid circles e thermophilic CSAD; solid verti

lines denote changes in co-digestion ratio as indicated by the re

callouts denote the start of a period for each CSAD; black arrow

exception of the thermophilic CSAD in Period I, total

measured VFAs at steady-state conditionswere below 700mg/

L for all study periods (Table 3). As discussed in Section 3.2, the

high concentrations of VFAs, particularly acetate and propi-

onate, observed in the thermophilic CSAD by the end of Period

I were attributed to LCFA accumulation/degradation (Table 4).

Similarly, as discussed in Section 3.2, upon increasing the OLR

in Period II, the thermophilic CSAD failed and no data could be

obtained at steady-state conditions.

3.1.2. Biogas production and substrate stabilizationDue to the process instability resulting from LCFAs accumu-

lation and shock loads of acetate and propionate during Period

I, biomethane production rates in the thermophilic CSAD

were ca. 16% lower than those in the mesophilic CSAD

(p < 0.01) (Table 5). Specific biomethane yields (SMY) calcu-

lated at steady-state represented 88% and 72% of the

biochemical methane potential (BMP) for the same co-

digestion mixtures of the mesophilic and thermophilic

, and pH observed during the study; open squares e

cal lines denote changes in OLR/HRT; segmented vertical

ctangular callouts as percent dog food (see Table 1); circular

indicates high-mixing perturbation (see Section 3.2.4).

Table 4 e Concentrations of volatile fatty acids (VFA) andlong-chain fatty acids (LCFA) in the digestate of themesophilic and thermophilic CSAD at steady-stateconditions for each period; concentration is g COD/Lunless otherwise stated; VFA and LCFA correspond to thesum of individual fatty acids; no data obtained at P II forthermophilic CSAD due to process failure.

CSAD Mesophilic Thermophilic

Period I II III IV I II III IV

Acetate 0.04 0.03 0.01 0.02 0.62 F 0.06 0.07

Propionate 0.00 0.00 0.00 0.02 0.59 F 0.01 0.01

Isobutyrate 0.00 0.13 0.00 0.00 0.06 F 0.00 0.00

Butyrate 0.00 0.00 0.00 0.00 0.00 F 0.00 0.00

Isovalerate 0.00 0.38 0.00 0.00 0.10 F 0.00 0.00

Valerate 0.00 0.14 0.01 0.01 0.01 F 0.00 0.03

Isocaproate 0.00 0.06 0.02 0.04 0.00 F 0.02 0.05

Caproate 0.00 0.51 0.00 0.06 0.01 F 0.00 0.06

Heptanoate 0.00 0.00 0.00 0.11 0.00 F 0.00 0.11

Capric acid 0.29 0.00 0.00 0.00 0.04 F 0.00 0.00

Lauric acid 0.40 0.06 0.00 0.00 0.73 F 0.07 0.00

Myristeate 0.44 0.00 0.00 0.00 0.57 F 0.00 0.00

Palmitate 0.55 0.70 0.13 0.17 0.26 F 0.17 0.15

Stearate 0.37 0.31 0.14 0.16 0.12 F 0.14 0.14

Oleate 0.30 0.14 0.04 0.05 0.23 F 0.05 0.03

Linoleate 0.23 0.00 0.04 0.08 0.06 F 0.00 0.06

VFA 0.06 1.27 0.04 0.26 1.39 F 0.08 0.33

LCFA 2.57 1.23 0.35 0.46 2.00 F 0.43 0.38

Acetate/Propionate

(g/g)

16.7 7.3 e 1.3 1.0 F 11.1 4.9

Palmitate/LCFA (g/g) 0.21 0.57 0.37 0.37 0.13 F 0.39 0.40

g COD-LCFA/kg TS 243 87 16 22 159 F 20 19

g COD-LCFA/kg VS 301 105 19 27 189 F 24 23

F: digester failure.

wat e r r e s e a r c h 5 3 ( 2 0 1 4 ) 2 4 9e2 5 8 253

CSADs, respectively (Table 5). Similarly, steady-state SMYs

were 76% and 64% of the theoretical (Buswell) SMY of the

mesophilic and thermophilic CSADs, respectively (Table 5).

Consequently, VS destruction in the thermophilic CSAD was

significantly lower compared to the mesophilic CSAD during

Period I, which was also reflected in the higher total COD of

the thermophilic CSAD (see Table 3). The observed SMY of the

mesophilic digester was only slightly affected by shortening

the HRT from 20 days to 15 days (Period II), representing 72% of

the theoretical SMY. This is not surprising considering that a

small proportion of the influent substrate (i.e., 25%) consisted

of slowly-degradable, high-lignin cow manure.

Table 5 e CSAD performance parameters obtained at steady-s

CSAD Mesophilic

Period I II I

Biomethane production rate (L/L-d) 0.665 0.849 0.5

Biomethane content (% in biogas) 61.7% 61.6% 62

Specific methane yield, SMY (L/g VS added)

CSAD (This study) 444 424 25

BMP (Labatut, 2012) 503 503 31

Theoretical, Buswell 587 587 52

Volatile solids destruction (%) 71.5% 60.9% 39

F: CSAD Failure.

During Periods III and IV biomethane production rates

were 5% and 7%, respectively, higher in the thermophilic

digester (p < 0.01) (Table 5). A considerable decrease in the

biomethane production rate was observed in the mesophilic

CSAD when the co-digestion ratio was changed from 75% to

25% dog food at a constant OLR of 2 g/L-d (Period III) (Table 5).

The slightly higher VS destruction achieved by the thermo-

philic digester during Period III was not statistically significant

(p > 0.05). However, during Period IV, a significantly higher VS

destruction was attained by the thermophilic CSAD relative to

the mesophilic (p < 0.05), as the effects of faster reaction rates

(particularly hydrolysis) become more noticeable at higher

OLRs (shorter HRTs) under thermophilic temperatures.

Regardless of the operating temperature, VS destruction effi-

ciency decreased significantly in both CSADs (p < 0.01) as a

result of the increase of less degradable cow manure.

3.2. Stability of CSADs throughout the study

3.2.1. Ability of mesophilic and thermophilic digesters toreach steady-state conditionsThe mesophilic digester was highly stable and reached

steady-state conditions in each of the four study periods

(Table 6), and biomethane production rates increased with

each increase in OLR, regardless of the co-digestion ratio

(Fig. 1). The thermophilic digester, however, experienced two

major upsets (Fig. 1), which resulted in underperformance

during Period I and precluded it from reaching steady-state

conditions in Period II (Table 6). Key to both upsets were the

influence of temperature on physical properties and the

chemical composition of the influent, specifically the high

lipid content and lower fiber-to-lipid ratio of the co-digestion

mixture during Periods I and II (Table 6). As discussed below,

during both upsets biogas production essentially stopped as a

result of high VFA concentrations and resulting low pH levels

(Fig. 1).

3.2.2. Acetate accumulation and free-energy limitations forpropionate oxidation at thermophilic temperaturesAcetate and propionate accounted for most of the VFA accu-

mulated at the time of the thermophilic CSAD first upset on

day 210 (Fig. 2). Even-carbon LCFAs are primarily degraded by

H2-producing bacteria to acetate and H2 via b-oxidation (Jeris

and McCarty, 1965; McInerney, 1988), while odd-carbon

LCFAs are mainly degraded to acetate and propionate

tate conditions during each period.

Thermophilic

II IV I II III IV

14 0.758 0.561 F 0.541 0.807

.7% 63.6% 59.4% F 61.7% 63.5%

7 253 374 F 271 269

7 317 515 515 310 310

0 520 587 587 520 520

.4% 42.1% 64.7% F 40.9% 45.1%

Table 6e Influent composition data vs. stability of digesters and their ability to reach steady-state conditions at each studyperiod.

Period Organic loading rate (g/L-d) Ratios (VS basis) Stable/steady-state?

Proteins Lipids Carbohydrates Fibera VS Fiber/lipid C/N Mesophilic Thermophilic

I 0.44 0.36 0.43 0.26 1.5 0.72 12.89 Yes/Yes No/Yes

II 0.59 0.49 0.57 0.35 2 0.72 12.89 Yes/Yes No/No

III 0.40 0.22 0.52 0.86 2 3.86 18.27 Yes/Yes Yes/Yes

IV 0.59 0.34 0.77 1.30 3 3.86 18.27 Yes/Yes Yes/Yes

a Neutral detergent fiber (NDF): cellulose, hemicelluloses, and lignin.

wat e r r e s e a r c h 5 3 ( 2 0 1 4 ) 2 4 9e2 5 8254

(Nelson et al., 2008; Gottschalk, 1986). Most of the acetate

observed on day 210 in the thermophilic CSAD came from

LCFA degradation. It was calculated that, 77% of the acetate

was accounted for by LCFAs, on a COD basis e a value that

compares well with the 68.4% of palmitate oxidized to acetate

determined by Jeris and McCarty (1965) using radio-labeled

carbon. Propionate accumulation was likely due to free-

energy limitations imposed by molecular hydrogen, which

partial pressure (PH2) should have reached at least 10�3.2 atm

to prevent propionate oxidation (Fig. 3). The role of bio-

energetics is apparent e thermophilic LCFA- (i.e., palmitate)

oxidizing bacteria have a higher PH2 threshold than their

mesophilic counterparts (10�2.4 atm vs. 10�3 atm, respectively)

(Fig. 3). Thus, LCFAs can be degraded to a greater extent at

thermophilic temperatures, potentially producing additional

H2 and further inhibiting propionate oxidation. This, com-

pounded with faster hydrolysis rates, can make thermophilic

digesters more susceptible to propionate accumulation, and

consequently process upsets. Indeed, accumulation of propi-

onate due to high partial pressures of hydrogen in thermo-

philic digesters has been reportedmany times in the literature

Fig. 2 e Short- (volatile) and long-chain fatty acids and biogas p

thermophilic (below) CSADs before and after the mixing pertur

(right); boxes show calculated g COD-LCFA/kg TS; only VFA dat

(Speece et al., 2006; Kim et al., 2002; McCarty and Smith, 1986).

Further, propionate tends to linger for several HRT cycles in

thermophilic digesters after an initial perturbation (Speece

et al., 2006; McCarty and Mosey, 1991). In this study, it took

approximately three HRT cycles from the initial perturbation

for methanogens to consume H2 and allow propionate

oxidation to re-start in the thermophilic CSAD. This confirms

the ability of organisms to recover from exposure and inhi-

bition from LCFAs, as shown in previous studies (Palatsi et al.,

2010; Nielsen and Ahring, 2006). However, by the end of Period

I acetate and propionate were once again accumulating in the

thermophilic CSAD as a result of LCFA accumulation, despite

that the mesophilic CSAD had almost no VFAs, but greater

accumulation of LCFAs (see Table 4). Moreover, themesophilic

CSAD did not show apparent signs of inhibition as suggested

by steady biomethane production, low VFAs, and high SMY

observed (Table 5), even though LCFA concentrations were

above the threshold of 180e220 g COD-LCFA/kg TS (see Table

4) to prevent process failure at mesophilic temperatures re-

ported by Neves et al. (2009a). It should be noted, however,

that in the aforementioned work LCFAs were extracted from

roduction rate observed in the mesophilic (above) and

bation on day 148 (left) and the OLR increase on day 330

a available on day 148 and no data on day 330.

Fig. 3 e Thermodynamic limits for propionate and

palmitate oxidation, and hydrogenotrophic

methanogenesis at mesophilic (segmented line) and

thermophilic (solid line) temperatures; the limits where all

the reactions are possible at mesophilic and thermophilic

temperatures are enclosed in the lighter and darker boxes,

respectively. The plot was built assuming 1 mM for all fatty

acids, 100 mM for bicarbonate, 0.7 atm for methane, and

0.3 atm for carbon dioxide.

wat e r r e s e a r c h 5 3 ( 2 0 1 4 ) 2 4 9e2 5 8 255

the solid matrix to determine the biomass-associated LCFAs

(adsorbed), whereas in the present study, LCFAs were

extracted from the bulk, i.e., liquid and solid phases, thereby

TS includes all the solid matter in the digestate.

3.2.3. Higher transient accumulation of LCFAs atthermophilic temperaturesWe characterized the thermophilic CSAD with faster lipid

hydrolysis rates through increased LCFA accumulation.

Although both CSADs showed accumulation of LCFAs over

time, the concentration of LCFAs in the thermophilic digester

was 40% higher before the mixing perturbation compared to

itsmesophilic counterpart, i.e., 4.5 vs. 2.8 g COD/L (355 vs. 261 g

COD-LCFA/kg TS), respectively (Fig. 2). This is not surprising

considering that overall digestion rates can be up to 2.25 times

faster at thermophilic temperatures relative to mesophilic

(O’Rourke, 1968). In lipid-containing substrates, degradation

of LCFAs via b-oxidation can be the slowest conversion step

and control the overall kinetics of the digestion process

(Novak and Carlson, 1970; O’Rourke, 1968; Pavlostathis and

Giraldo-Gomez, 1991; Rinzema et al., 1994). Thus, differences

between the rates of hydrolysis of lipids and b-oxidation of

long chain fatty acids can result in a reactant-product imbal-

ance and accumulation of LCFAs over time, resulting in inhi-

bition. Hanaki et al. (1981) reported that LCFA accumulation

could also be accentuated at concentrations greater than 1.4 g

COD/L, because at this level LCFAs were inhibitory for the b-

oxidizing organisms themselves. Furthermore, at a concen-

tration of 2.9 g COD/L the authors observed a lag phase of

nearly 10 days for LCFA degradation, which consequently

extended that ofmethanogenesis to over 20 days. On the other

hand, Angelidaki andAhring (1992), found that concentrations

as low as 0.6 g COD/L of unsaturated oleic acid (C18:1) and

1.5 g COD/L of saturated stearic acid (C:18:0) increased the lag

phase of methanogenesis in batch tests conducted at 55 �C.Koster and Cramer (1987) observed a sharp decrease in the

methanogenic activity at concentrations over 1.3e1.4 g COD/L

for capric (C10:0), myristic, (C14:0) and oleic acids, and of over

0.5 g COD/L for lauric acid (C12:0). Based on above reports and

the work of Neves et al. (2009a), both CSADs in this study were

within the inhibitory range during the period before the

perturbation; however, no apparent decreased biomethane

production rates or differences between the mesophilic and

thermophilic CSADs before the perturbation were observed

(see Figs. 1 and 2). Instead, b-oxidizers seemed to have been

more affected, and particularly at thermophilic temperatures,

as suggested by the greater accumulation of LCFAs observed

in the thermophilic CSAD and higher concentration of acetate

observed in the mesophilic.

3.2.4. The influence of temperature on physical properties andtheir impact on process stabilityIt was apparent that the first upset of the thermophilic CSAD

was triggered by an increase in the mixing intensity per-

formed during days 148e149 of the study (Fig. 1). Although

both CSADs observed an increase in LCFA degradation after

the mixing perturbation, the extent of degradation was more

than two times higher in the thermophilic CSAD compared to

the mesophilic (Fig. 2). It is evident that, such LCFA degrada-

tion was the cause of the sudden and sustained increase in

VFAs observed after the mixing perturbation in the thermo-

philic CSAD (Fig. 1). Mixing intensities were unintentionally

increased in both CSADs from 100 to 1500 RPM in repeated

occasions during days 148e149 (Figs. 1 and 2). It is known that

high mixing intensities can disrupt the syntrophic associa-

tions between H2-producing bacteria and H2-utilizing metha-

nogens and have a detrimental effect on digester stability

(Speece et al., 2006; Hansen et al., 1999; McMahon et al., 2001;

Hoffmann et al., 2008; Vavilin and Angelidaki, 2005; Stroot

et al., 2001). Schmidt and Ahring (1993) observed that after

physically disintegrating granules of a propionate- and n-

butyrate-fed UASB, the PH2 increased to over 10�4 atm,

causing propionate and butyrate degradation rates to

decrease 30 and 20%, respectively. They concluded that, by

disintegrating the granules, syntrophic relationships were

disrupted by the increased distance between H2-producing

and H2-utilizing organisms, which in turn increased the H2

mass transfer resistance. We believe that in this study, high

mixing intensities produced a similar effect for the thermo-

philic digester, affecting H2 transport, increasing PH2, and

halting propionate oxidation.

The fact that the thermophilic CSAD observed higher LCFA

degradation rates compared to the mesophilic can be

explained by the effect of temperature on two main physical

properties, which in turn exacerbated the effect of the

increased mixing intensity. First, the lower water viscosity of

the thermophilic CSAD due to the 18 �C difference with its

mesophilic counterpart resulted in a mixing intensity that

(described by the Reynolds number) was 36% higher in the

former. This translated into a mixing rate that was actually

closer to 2000 RPM during the perturbation, rather than

1500 RPM (See Supporting Information). With high concen-

trations of LCFAs in the digestate, the naturally-occurring lipid

emulsions could have been homogenized by the repeated

wat e r r e s e a r c h 5 3 ( 2 0 1 4 ) 2 4 9e2 5 8256

vigorous mixing. The increased LCFA-to-water interface area

resulting from the dispersed, smaller-diameter emulsions

would favor the contact between substrate and fatty-acid

oxidizing bacteria, and increase degradation rates. Second,

because fatty acids have different melting points depending

on their degree of saturation and chain length, their physical

state changes with temperature. At thermophilic tempera-

tures, LCFAs were in a more liquid form e thereby, a higher

extent of emulsification should have been achieved through

mixing, especially at 36% higher mixing rate, which could

have further increased LCFA availability for b-oxidizers and

therefore their rate of degradation. For example, palmitic acid

(C16), which constituted 42% of the total LCFA pool of the

thermophilic digester before the mixing perturbation, has a

melting point of 62.9 �C, suggesting that palmitic acid should

have been in a practically liquid state in the thermophilic

digester, whereas closer to a solid state in the mesophilic

digester (See Supporting Information).

3.2.5. Fiber-to-lipid ratio, rather than lipid loading rate e themain control of stability in the thermophilic CSADThe second upset of the thermophilic digester occurred after

the overall OLR was increased from 1.5 to 2 g VS/L-

d (0.36e0.49 g lipid/L-d) (Table 5). LCFAs accumulated to high

levels within one HRT cycle in the thermophilic digester

(Fig. 2), and acetate and propionate increased as the pH

concurrently dropped, bringing biogas production to a halt

(Fig. 1). To recover the digester, only cow manure was fed on

day 352 and for the following three feeding cycles; dog food

was re-introduced on day 360 at a decreased portion of 25%,

which forced the premature start of Period III for the ther-

mophilic CSAD (see Fig. 1). Within 24 h of manure-only

feeding (day 354), the concentration of LCFAs decreased by

50% (Fig. 2). A similar behavior was observed by Hanaki et al.

(1981), who reported accumulation of LCFAs onto the solid

matrix (and decrease from the aqueous phase) within a day of

exposure to neutral fats. Indeed, the fact that acetate or biogas

production did not increase during the 24-h period, also sug-

gests that LCFAs were adsorbed on the manure matrix.

Particularly, the non-polar, hydrophobic surfaces of the

manure fibers have been reported to effectively reduce LCFAs

from solution, which in turn reduce their bioavailability and

therefore inhibitory effects (Nielsen and Ahring, 2006). It was

apparent that for Periods I and II of this study, the ratio of

manure-to-lipid-containing substrate (i.e., dog food) was

insufficient for the amount of hydrolyzed lipids in the ther-

mophilic CSAD (Table 5). This was evidenced by the accu-

mulation of LCFAs in the liquid phase, which produced

inhibition, instability, and ultimately digester failure. An

excess of adsorbent material (e.g., manure fibers) should

decrease the bioavailability of free (non-adsorbed) LCFAs,

which in turn would reduce possible inhibition and increase

the stability of the process. If not enough adsorbentmaterial is

available for the amount of adsorbate (i.e., LCFAs), the adsor-

bent will saturate and LCFAs will accumulate in the liquid

phase. Indeed, even though the lipid loading rates of Periods I

(unstable) and IV (stable) were comparable, no accumulation

of LCFAs was observed in Period IV e i.e., the higher fiber-to-

lipid ratio provided enough adsorbent surface area to main-

tain the thermophilic reactor exceptionally stable compared

to Periods I and II (Table 5). Therefore, it is evident that the

amount of LCFAs that can be absorbed onto the manure fibers

will primarily depend on the proportion betweenmanure and

the lipid-rich material, rather than the lipids loading rate.

Furthermore, it is apparent that relatively higher manure-to-

lipid ratios should be required at thermophilic temperatures

as compared to mesophilic temperatures due to the increased

hydrolysis and b-oxidation rates at thermophilic tempera-

tures. The increase of the manure-to-lipid ratio is not the only

factor improving the resilience of the thermophilic digester in

periods III and IV; the continuous exposure of the digester to

fats from the beginning of the study could also have modified

the microbial community structure to be more efficient at

degrading fats, particularly LCFAs.

4. Conclusions

Results of this study confirm that digester temperature has an

important role on determining process performance and sta-

bility by directly influencing key physical properties of the

system and the substrate. Specifically, increased hydrolysis

rates of thermophilic digesters can produce significant accu-

mulation of LCFAs over time when lipid-rich wastes are co-

digested with inadequate amounts of manure. Rather than

the lipid loading rate, it is the fiber-to-lipid ratio that appears to

control the stability of the process. Manure fibers provide the

surface area for LCFAs to adsorb and decrease their concen-

tration in solution, thereby providing stability to the thermo-

philic process. In addition, mixing intensities of CSADs are

significantly higher at thermophilic temperatures; hence,

excessive mixing should be avoided not only to minimize

operational costs, but also because it can increase the extent of

homogenization of lipid emulsions, and thus the rate of hy-

drolysis. The process can be further exacerbated by the fact

that LCFAs are in a more liquid state at thermophilic temper-

atures, and the bioenergetics for b-oxidation are more favor-

able, making thermophilic digesters more prone to inhibition

and instability due to molecular hydrogen accumulation and

consequential shock loads of acetate and propionate.

Overall, provided that lipid-containing wastes are co-

digested with adequate amounts of manure, thermophilic

co-digestion presents marginal, yet statistically-significant,

advantages over mesophilic co-digestion at high OLRs (i.e.,

15- and 10-d HRTs), by increasing biomethane production

rates, substrate utilization efficiency, and organic matter

stabilization.Mesophilic digestion appears to be amore robust

and stable process, regardless of the OLR and substrate

chemical composition. Given the typical HRTs of commercial

manure-based co-digestion CSADs (i.e., 20e25 d), conven-

tional mesophilic digestion appears to be a more reliable and

cost-effective alternative considering the additional energy

required for thermophilic systems.

Acknowledgments

We would like to thank the New York State Energy Research

and Development Authority (NYSERDA) and CONICYT, Chile,

wat e r r e s e a r c h 5 3 ( 2 0 1 4 ) 2 4 9e2 5 8 257

for partial financial support in completing this study. Kristen

Vitro, an undergraduate student at Cornell University, is

recognized for her continued assistance and thorough labo-

ratory analyses in conjunction with this study.

Appendix A. Supplementary data

Supplementary data related to this article can be found at

http://dx.doi.org/10.1016/j.watres.2014.01.035.

r e f e r e n c e s

Angelidaki, I., Ahring, B.K., 1992. Effects of free long-chain fatty-acids on thermophilic anaerobic-digestion. Appl. Microbiol.Biotechnol. 37 (6), 808e812.

Angelidaki, I., Ahring, B.K., 1993. Thermophilic anaerobic-digestion of livestock waste e the effect of ammonia. Appl.Microbiol. Biotechnol. 38 (4), 560e564.

Angelidaki, I., Ahring, B.K., 1994. Anaerobic thermophilicdigestion of manure at different ammonia loads e effect oftemperature. Water Res. 28 (3), 727e731.

APHA, 1995. In: Eaton, A.D., Clesceri, L.S., Greenberg, A.E. (Eds.),Standard Methods for the Examination of Water andWastewater. American Public Health Association, AmericanWater Works Association, Water Environment Federation,Washington, D.C.

Biey, E.M., Musibono, E.D., Verstraete, W., 2003. Start-up of amulti-stage system for biogas production and solid wastetreatment in low-tech countries. Water Sci. Technol. 48 (4),239e243.

Chen, Y., Cheng, J.J., Creamer, K.S., 2008. Inhibition of anaerobicdigestion process: a review. Bioresour. Technol. 99 (10),4044e4064.

Gaudy, A.F., 1962. Colorimetric determination of protein andcarbohydrate. Indust. Wat. Wastes 7, 17e22.

Gavala, H.N., Yenal, U., Skiadas, I.V., Westermann, P., Ahring, B.K.,2003. Mesophilic and thermophilic anaerobic digestion ofprimary and secondary sludge. Effect of pre-treatment atelevated temperature. Water Res. 37 (19), 4561e4572.

Gottschalk, G., 1986. Bacterial Metabolism. Springer-Verlag, NewYork, NY.

Hanaki, K., Nagase, M., Matsuo, T., 1981. Mechanism of inhibitioncaused by long-chain fatty-acids in anaerobic-digestionprocess. Biotechnol. Bioeng. 23 (7), 1591e1610.

Hansen, K.H., Angelidaki, I., Ahring, B.K., 1998. Anaerobicdigestion of swine manure: inhibition by ammonia. Water Res.32 (1), 5e12.

Hansen, K.H., Angelidaki, I., Ahring, B.K., 1999. Improvingthermophilic anaerobic digestion of swine manure. Water Res.33 (8), 1805e1810.

Hoffmann, R.A., Garcia, M.L., Veskivar, M., Karim, K., Al-Dahhan, M.H., Angenent, L.T., 2008. Effect of shear onperformance and microbial ecology of continuously stirredanaerobic digesters treating animal manure. Biotechnol.Bioeng. 100 (1), 38e48.

Hwu, C.S., Lettinga, G., 1997. Acute toxicity of oleate to acetate-utilizing methanogens in mesophilic and thermophilicanaerobic sludges. Enzyme Microb. Technol. 21 (4), 297e301.

Jeris, J.S., McCarty, P.L., 1965. The biochemistry of methanefermentation using C14 tracers. J. Water Pollut. Control Fed. 37(2), 178e192.

Kayhanian, M., 1994. Performance of a high-solids anaerobicdigestion process under various ammonia concentrations.J. Chem. Technol. Biotechnol. 59 (4), 349e352.

Khanal, S.K., Environmental, Water Resources Institute, BiofuelTask, 2010. Bioenergy and Biofuel from Biowastes andBiomass. American Society of Civil Engineers, Reston, Va.

Kim, M., Ahn, Y.H., Speece, R.E., 2002. Comparative processstability and efficiency of anaerobic digestion; mesophilic vs.thermophilic. Water Res. 36 (17), 4369e4385.

Koster, I.W., Cramer, A., 1987. Inhibition of methanogenesis fromacetate in antigranulocytes sludge by long-chain fatty-acids.Appl. Environ. Microbiol. 53 (2), 403e409.

Koster, I.W., Lettinga, G., 1988. Anaerobic-digestion at extremeammonia concentrations. Biol. Wastes 25 (1), 51e59.

Labatut, R.A., 2012. Anaerobic Biodegradability of ComplexSubstrates: Performance and Stability at Mesophilic andThermophilic Conditions. Ph.D. Dissertation. CornellUniversity, Ithaca.

Leven, L., Eriksson, A.R.B., Schnurer, A., 2007. Effect of processtemperature on bacterial and archaeal communities in twomethanogenic bioreactors treating organic household waste.FEMS Microbiol. Ecol. 59 (3), 683e693.

Loehr, R.C., Rohlich, G.A., 1962. A wet method for grease analysis.In: Proc. 17th Ind. Waste Conf. Purdue University, pp. 215e232.

McCarty, P.L., Mosey, F.E., 1991. Modeling of anaerobic-digestionprocesses (a discussion of concepts). Water Sci. Technol. 24(8), 17e33.

McCarty, P.L., Smith, D.P., 1986. Anaerobic waste-water treatment4. Environ. Sci. Technol. 20 (12), 1200e1206.

McInerney, M.J. (Ed.), 1988. Anaerobic Hydrolysis andFermentation of Fats and Proteins. John Wiley and Sons, NewYork.

McMahon, K.D., Stroot, P.G., Mackie, R.I., Raskin, L., 2001.Anaerobic codigestion of municipal solid waste and biosolidsunder various mixing conditions e II: microbial populationdynamics. Water Res. 35 (7), 1817e1827.

Mertens, D.R., Allen, M., Carmany, J., Clegg, J., Davidowicz, A.,Drouches, M., Frank, K., Gambin, D., Garkie, M.,Gildemeister, B., Jeffress, D., Jeon, C.S., Jones, D., Kaplan, D.,Kim, G.N., Kobata, S., Main, D., Moua, X., Paul, B., Robertson, J.,Taysom, D., Thiex, N., Williams, J., Wolf, M., 2002. Gravimetricdetermination of amylase-treated neutral detergent fiber infeeds with refluxing in beakers or crucibles: collaborativestudy. J. AOAC Int. 85 (6), 1217e1240.

Moller, J., 2009. Gravimetric determination of acid detergent fiberand lignin in feed: interlaboratory study. J. AOAC Int. 92 (1),74e90.

Nelson, D.L., Lehninger, A.L., Cox, M.M., 2008. LehningerPrinciples of Biochemistry. W.H. Freeman, New York.

Neves, L., Oliveira, R., Alves, M.M., 2009a. Fate of LCFA in theco-digestion of cow manure, food waste and discontinuousaddition of oil. Water Res. 43 (20), 5142e5150.

Neves, L., Pereira, M.A., Mota, M., Alves, M.M., 2009b. Detectionand quantification of long chain fatty acids in liquid and solidsamples and its relevance to understand anaerobic digestionof lipids. Bioresour. Technol. 100 (1), 91e96.

Nguyen, P.H.L., Kuruparan, P., Visvanathan, C., 2007. Anaerobicdigestion of municipal solid waste as a treatment prior tolandfill. Bioresour. Technol. 98 (2), 380e387.

Nielsen, H.B., Ahring, B.K., 2006. Responses of the biogas processto pulses of oleate in reactors treating mixtures of cattle andpig manure. Biotechnol. Bioeng. 95 (1), 96e105.

Novak, J.T., Carlson, D.A., 1970. Kinetics of anaerobic long chainfatty acid degradation. J. Water Pollut. Control Fed. 42 (11),1932.

O’Rourke, J.T., 1968. Kinetics of Anaerobic Waste Treatment atReduced Temperatures. Stanford University.

Palatsi, J., Illa, J., Prenafeta-Boldu, F.X., Laureni, M., Fernandez, B.,Angelidaki, I., Flotats, X., 2010. Long-chain fatty acidsinhibition and adaptation process in anaerobic thermophilicdigestion: batch tests, microbial community structure and

wat e r r e s e a r c h 5 3 ( 2 0 1 4 ) 2 4 9e2 5 8258

mathematical modelling. Bioresour. Technol. 101 (7),2243e2251.

Pavlostathis, S.G., Giraldo-Gomez, E., 1991. Kinetics of anaerobictreatment e a critical-review. Crit. Rev. Environ. Control 21(5e6), 411e490.

Raskin, L., Poulsen, L.K., Noguera, D.R., Rittmann, B.E., Stahl, D.A.,1994. Quantification of methanogenic groups in anaerobicbiological reactors by oligonucleotide probe hybridization.Appl. Environ. Microbiol. 60 (4), 1241e1248.

Rinzema, A., Boone, M., Vanknippenberg, K., Lettinga, G., 1994.Bactericidal effect of long-chain fatty-acids in anaerobic-digestion. Water Environ. Res. 66 (1), 40e49.

Schmidt, J.E., Ahring, B.K., 1993. Effects of hydrogen and formateon the degradation of propionate and butyrate in thermophilicgranules from an upflow anaerobic sludge blanket reactor.Appl. Environ. Microbiol. 59 (8), 2546e2551.

Song, Y.C., Kwon, S.J., Woo, J.H., 2004. Mesophilic andthermophilic temperature co-phase anaerobic digestioncompared with single-stage mesophilic- and thermophilicdigestion of sewage sludge. Water Res. 38 (7), 1653e1662.

Speece, R.E., Boonyakitsombut, S., Kim, M., Azbar, N., Ursillo, P.,2006. Overview of anaerobic treatment: thermophilic andpropionate implications. Water Environ. Res. 78 (5), 460e473.

Stroot, P.G., McMahon, K.D., Mackie, R.I., Raskin, L., 2001.Anaerobic codigestion of municipal solid waste and biosolids

under various mixing conditions e I. Digester performance.Water Res. 35 (7), 1804e1816.

Usack, J.G., Spirito, C.M., Angenent, L.T., 2012. Continuously-stirred anaerobic digester to convert organic wastes intobiogas: system setup and basic operation. J. Vis. Exp. 65, e3978.

USEPA, 2000. Guide to Field Storage of Biosolids and OtherOrganic By-products Used in Agriculture and for Soil ResourceManagement. U.S. Environmental Protection Agency, p. 134.

VanLier, J.B., Martin, J.L.S., Lettinga, G., 1996. Effect oftemperature on the anaerobic thermophilic conversion ofvolatile fatty acids by dispersed and granular sludge. WaterRes. 30 (1), 199e207.

Vavilin, V.A., Angelidaki, I., 2005. Anaerobic degradation of solidmaterial: importance of initiation centers for methanogenesis,mixing intensity, and 2D distributed model. Biotechnol.Bioeng. 89 (1), 113e122.

Watanabe, H., Kitamura, T., Ochi, S., Ozaki, M., 1997. Inactivationof pathogenic bacteria under mesophilic and thermophilicconditions. Water Sci. Technol. 36 (6e7), 25e32.

Wilson, C.A., Murthy, S.M., Fang, Y., Novak, J.T., 2008. The effectof temperature on the performance and stability ofthermophilic anaerobic digestion. Water Sci. Technol. 57 (2),297e304.

Zinder, S.H. (Ed.), 1986. Thermophilic Waste Treatment Systems.Wiley-Interscience, New York.