Applied Energy 116 (2014) 409–415

Contents lists available at ScienceDirect

Applied Energy

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

The pressure effects on two-phase anaerobic digestion

0306-2619/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.apenergy.2013.11.012

Abbreviations: COD, chemical oxygen demand; ODM, organic dry matter; SBP,specific biogas production; SMY, specific methane yield; TAN, total ammonianitrogen; TIC, total inorganic carbon; VFA, volatile fatty acid.⇑ Corresponding author. Tel.: +49 711 459 23348; fax: +49 711 22111.

E-mail addresses: Yuling.Chen@uni-hohenheim.de (Y. Chen),Benjamin.Roessler@uni-hohenheim.de (B. Rößler), Simon.Zielonka@uni-hohen-heim.de (S. Zielonka), Andreas.Lemmer@uni-hohenheim.de (A. Lemmer), Anna-Maria.Wonneberger@kit.edu (A.-M. Wonneberger), Thomas.Jungbluth@uni-hohen-heim.de (T. Jungbluth).

Yuling Chen a,⇑, Benjamin Rößler a, Simon Zielonka a, Andreas Lemmer a, Anna-Maria Wonneberger b,Thomas Jungbluth a

a State Institute of Agricultural Engineering and Bioenergy, University Hohenheim, Garbenstraße 9, D-70599 Stuttgart, Germanyb DVGW – Research Center at the Engler-Bunte-Institute, Karlsruhe Institute of Technology (KIT), Engler-Bunte-Ring 1, D-76131 Karlsruhe, Germany

h i g h l i g h t s

� The pressure effect on anaerobic digestion up to 9 bar was examined.� Increasing pressure decreased pH value in the anaerobic filter.� Increasing pressure increased methane content.� Increasing pressure decreased specific methane yield slightly.� The pressurized methane reactor was very stable and performed well.

a r t i c l e i n f o

Article history:Received 26 February 2013Received in revised form 2 September 2013Accepted 3 November 2013Available online 27 November 2013

Keywords:Anaerobic digestionPressureTwo-phaseSubstitute natural gasBiogasBiomethane

a b s t r a c t

Two-phase pressurized anaerobic digestion is a novel process aimed at facilitating injection of the producedbiogas into the natural gas grid by integrating the fermentative biogas production and upgrading it to substi-tute natural gas. In order to understand the mechanisms, knowledge of pressure effects on anaerobic digestionis required. To examine the effects of pressure on the anaerobic digestion process, a two-phase anaerobicdigestion system was built up in laboratory scale, including three acidogenesis-leach-bed-reactors and onepressure-resistant anaerobic filter. Four different pressure levels (the absolute pressure of 1 bar, 3 bar, 6 barand 9 bar) were applied to the methane reactor in sequence, with the organic loading rate maintained atapproximately 5.1 kgCOD m�3 d�1. Gas production, gas quality, pH value, volatile fatty acids, alcohol, ammo-nium-nitrogen, chemical oxygen demand (COD) and alkaline buffer capacity were analyzed. No additionalcaustic chemicals were added for pH adjustment throughout the experiment. With the pressure increasingfrom 1.07 bar to 8.91 bar, the pH value decreased from 7.2 to 6.5, the methane content increased from 66%to 75%, and the specific methane yield was slightly reduced from 0.33 lN g�1COD to 0.31 lN g�1COD. Therewas almost no acid-accumulation during the entire experiment. The average COD-degradation grade wasalways more than 93%, and the average alkaline buffering capacity (VFA/TIC ratio) did not exceed 0.2 atany pressure level. The anaerobic filter showed a very stable performance, regardless of the pressure variation.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction at the biogas plant, a great amount of cogenerated heat is dissi-

Anaerobic digestion is a biological process, in which organicwastes and energy crops can be degraded and converted into bio-gas, mainly containing CH4 and CO2 [1]. The most common practicein Germany is to transfer biogas into a local combined heat andpower plant for electricity and heat production. However, withthe exception of digester thermal control and heating purposes

pated into the air and wasted. Alternatively, biogas can be injectedinto the gas grid as substitute natural gas after desulfurization,CO2-removal, drying and pressurization [2]. Since it temporallyand spatially separates biogas production from utilization, theoverall energy utilization efficiency is largely increased. In addi-tion, the existing gas distribution and storage infrastructure canbe used without modification. Therefore, this application has re-ceived increased attention in recent years [3]. Nevertheless, biogaspurification and upgrading is usually energy demanding [4]. Two-phase pressurized anaerobic digestion is a possible solution. It isintended to directly remove CO2 and H2S in the digester, makinguse of their high gas solubility under pressure. Thus, biogas pro-duction, purification and pressurization are integrated in one sys-tem, and the expenses involved in the subsequent treatment canbe considerably reduced.

410 Y. Chen et al. / Applied Energy 116 (2014) 409–415

To better understand pressurized anaerobic digestion, it isessential to first gain knowledge of how pressure affects the pro-cess. Although various operational parameters (e.g. temperature,pH, hydraulic retention time, organic loading rate and mixingmode) have been thoroughly studied and reviewed [5–8], therehave been few discussions on the effect of pressure. The pressurevariable is not given enough recognition in anaerobic digestion,mainly due to the limitations in the available techniques and facil-ities suitable for experimental investigation on pressure effects [9].In addition, pressure change induces complicated interactionsamong operational conditions and microorganism activity in areactor. For the sake of easy management, the total gas pressurein a common anaerobic digester is maintained slightly above atmo-spheric pressure (up to 0.02 bar overpressure) [10]. Compared toother operational parameters, pressure is a constant, rather thana variable, in actual application. In order to improve anaerobicdigestion performance, efforts are primarily focused on the optimi-zation of more adjustable and controllable parameters, and as a re-sult, the investigation of pressure effects on anaerobic digestionhas been overlooked.

As a matter of fact, anaerobic digestion for methane productionunder pressure is not rare in natural ecosystems or in the wastewa-ter treatment industry. It is common in marine sediments, hun-dreds of meters deep [11], in landfills [12] and at the lower partof anaerobic digestion towers or biogas tower reactors [13] thatare used in wastewater treatment to save ground space. Based onthe pressure adaptability, microorganisms can be divided intothree categories: piezosensitive, piezotolerant and piezophilic mi-crobes. Piezosensitive microbes have optimal growth at atmo-spheric pressure and stop reproduction around 500 bar [14]. Bothpiezotolerant and piezophilic microbes are bacteria that are ableto grow and proliferate up to a pressure of 1000 bar [15]. The opti-mal growth rate of piezotolerant microorganisms, however, occursat atmospheric pressure [16]. Since most microbes in anaerobicdigesters are inoculated from sewage slurry, excrement or waste-water treatment sludge under atmospheric pressure, they are nor-mally not piezophilic. That means, their growth rates are hardlyinhibited by pressure up to 10 bar, according to the research sum-mary of Abe and Horikoshi and Abe [16]. This offered a theoreticalbasis for the experiment of anaerobic digestion under pressure.

This study examined the effects of pressure on anaerobic diges-tion by testing four different pressures (the absolute pressure of1 bar, 3 bar, 6 bar and 9 bar) in a two-phase anaerobic digestionsystem. Gas production, gas quality, pH value, volatile fatty acids(VFAs), chemical oxygen demand (COD) degradation grade, buffercapacity and ammonium were analyzed and compared.

2. Materials and methods

2.1. Reactors

The flow diagram of the two-phase anaerobic digestion systemis shown in Fig. 1. The hydrolysis-acidification was performed inthree parallel-operated acidogenesis-leach-bed-reactors at 55 �C.Each reactor had 50-liter volume, and was alternately fed with10 kg (fresh mass) maize silage from the Field-test station of theUniversity Hohenheim (Unterer Lindenhof, Eningen, Germany) ata time interval of seven days. In order to avoid the deficiency ofthe nutrients necessary for microbial growth and biological processdisturbances, micronutrients were also added once a week. Thedosage and the composition of the micronutrients was based onthe recommendation of Vintiloiu et al. [17]. The maize silage wasloosely stacked on a perforated grate. In the acidogenesis-leach-bed-reactors, the substrate was gradually decomposed and a leach-ate rich in organic acids, as well as alcohols was produced. Every

week, approximately 20 l of leachate from each acidogenesis-leach-bed-reactor was introduced into a tank (Tank 1 in Fig. 1)for storage and homogenization. Every six hours, a certain amountof the leachate was pumped from the tank into an anaerobic filter,which was operated under pressure for further degradation. Thefeeding amount was only dependent on the influent COD concen-tration, since the organic loading rate of the methane reactor re-mained unchanged. For the stable working volume, the sameamount of liquid was eluted from the methane reactor to the othertank (Tank 2 in Fig. 1) for storage under atmospheric pressure. Dueto the pressure difference, the dissolved CO2 could be released, andthe liquid in Tank 2 was distributed evenly to the three acidogen-esis-leach-bed-reactors once a week. A liquid level sensor(Endress + Hauser, Liquicap T FMI21) constantly controlled theworking volume of the anaerobic filter.

The upflow-operated anaerobic filter was running at 37 �C. Thereactor was composed of a fixed bed, a three-phase separator and agas chamber at the top. The 20 l fixed bed was packed with sin-tered glass (Sera Siporax) as a carrier material. With 270 m2 l�1

biologically effective surfaces, the sintered glass helped the micro-organisms’ immobilization and biofilm development. Despite thefixed bed, there was still a certain amount of biomass suspendedin the fluid. The three-phase separator prevented the suspendedbiomass from leaving the anaerobic filter with the effluent. In addi-tion, the gas bubbles formed in the process could also be collectedwithout significant foaming by the separator.

The produced biogas did not immediately leave the anaerobicfilter, but accumulated therein, till the desired pressure in theanaerobic filter was reached. At that point, the valve of the gas out-let automatically opened, and the produced biogas was injected toa gasbag. As soon as the gas was released, the pressure of theanaerobic filter started to drop, and the gas outlet was closed again,allowing the auto-generated biogas pressure to increase to the de-sired value. By this means, the anaerobic filter could be set under acertain operating pressure. The entire process was controlled by apressure sensor (Endress + Hauser Ceraphant T PTC31) and a con-trol valve (Bürkert 2712).

In addition to the anaerobic filter, the acidogenesis-leach-bed-reactors also had gas outlets. The gas outlet of each reactor wasconnected to a gasbag for gas-quality and -quantity measurement.Furthermore, both the anaerobic filter and the acidogensis-leach-bed-reactors were equipped with pumps for internal circulation(about 1.5 l min�1), mixing for five minutes every ten minutes.

2.2. Experiment procedure

The anaerobic filter was seeded with the effluent from anotherfixed-bed anaerobic reactor, which had been fed with leachate ofgrass silage [18]. The reactor start-up period and preliminary testslasted approximately four months. After that, the reactors reacheda steady state, and the experiment on the pressure effects on two-phase anaerobic digestion began.

The experiment was divided into four runs. With the exceptionof the working pressure of the anaerobic filter, all the operatingparameters were maintained at a constant. The influent COD con-centration stayed at 23 ± 0.9 g l�1. The organic loading rate of5.1 ± 0.1 kgCOD m�3 d�1 was applied to the anaerobic filter. Fourdifferent working pressures on the anaerobic filter were tested (Ta-ble 1). No additional caustic chemicals were added for pH adjust-ment throughout the experiment, so that the pressure effects onanaerobic digestion could be clearly examined.

2.3. Analytical methods and data acquisition

In this study, pH, pressure and temperature of the anaerobic fil-ter were monitored in real-time (pH-sensor: Endress + Hauser

Fig. 1. Schematic diagram of the two-phase pressurized anaerobic digestion system.

Table 1Operating parameters of each run of the experiment.

Run Duration (days) Absolute pressure of anaerobic filter (bar)

Target value Daily mean Min. Max.

1 21 1 1.07 ± 0.01 1.06 1.102 27 3 2.97 ± 0.003 2.93 3.003 18 6 5.95 ± 0.004 5.86 6.004 55 9 8.91 ± 0.04 8.79 9.00

Y. Chen et al. / Applied Energy 116 (2014) 409–415 411

CPS11D; pressure sensor: Endress + Hauser Ceraphant T PTC31;temperature sensor: Greisinger GTF 103 Pt100), and the data werelogged in Labview 11.0.1 (National Instruments). The produced gascomposition and volume were measured under atmospheric pres-sure every day (Gas meter: Ritter TG20/5; Gas analyzer: Sick Mai-hak S710). The gas volume was corrected to dry gas at a standardtemperature and pressure (0 �C and 1 atm).

The process liquid in the anaerobic filter was sampled onceevery other day shortly before feeding. It was analyzed for COD,VFAs and the content of sugar, alcohol, total inorganic carbon(alkaline buffer capacity) and ammonium nitrogen (NHþ4 –N). Thecollected leachate from the acidognesis-leach-bed-reactors re-ceived the same analyses once a week. The content of organicdry matter (ODM), VFAs, sugar and alcohol in the maize silagewas evaluated on a weekly basis. ODM was assessed by specifyingthe ash content of dry samples in a muffle furnace at 550 �C. CODwas determined with a specific COD analysis system (Hach LangeCompany), including pre-dosed reagents (LCK014, 1000–10,000 mg l�1), a high temperature thermostat (HT 200 S) and asensor array photometer (LASA 20). The concentration of acetate,propionate, n- and iso-butyrate, n- and iso-valerate as well as capr-onate was measured with capillary column gas chromatography(Varian CP-3800). DL-lactic acid, formic acid, sucrose, glucose, fruc-

tose, ethanol and propylene glycol content were analyzed withhigh-performance liquid chromatography (Bischoff Company).Before gas chromatography and high-performance liquid chroma-tography analyses, maize silage was treated by solid–liquidshake-flask extraction. Total inorganic carbon was measured witha titrator (785 DMP Titrino – Metrohm Filderstadt). Ammoniumnitrogen was determined by distillation (Vapodest 50, GerhardtCompany) and back titration. All the acquired data were statisti-cally analyzed with ANOVA in SAS (v8.0).

3. Results

3.1. Substrate characteristics

During the experiments, the maize silage was fed into the aci-dogenesis-leach-bed-reactors for leachate production. The col-lected leachate flowed through a 100 lm filter and was used asan intermediate substrate for further acetogenesis and methano-genesis. Therefore, the leachate containing little solid was analyzedwith respect to COD concentration, while the fresh maize silagewas determined with its ODM content. The chemical characteris-tics of the maize silage and the leachate from the acidogenesis-leach-bed-reactors are shown in Table 2.

In spite of different analytical parameters (ODM and COD), itwas clearly observed that the leachate had much less organic frac-tion than the fresh maize silage. This means that a large part of or-ganic substance was still hidden in a solid form. Compared to theadded maize silage, the leachate contained a lower concentrationof acetic acid, DL-lactic acid, sugar and alcohol, but a higher con-tent of other VFAs (C3–C6). Nevertheless, acetic acid, n-Butyric acid,and DL-lactic acid were dominant in the organic fraction of theleachate, which was ready for further degradation in the pressur-ized anaerobic filter.

Table 2Chemical characteristic of the substrates used for the pressurized two-phaseanaerobic digestion experiments.

Maize silage Leachate

ODM (g kg�1) 312.6 ± 12.7 /COD (g l�1) / 23.2 ± 0.9Acetic acid (g kg�1) 9.1 ± 0.2 2.7 ± 0.4Propionic acid (g kg�1) 0.03 ± 0.01 0.4 ± 0.04iso-Butyric acid (g kg�1) 0 0.03 ± 0.01n-Butyric acid (g kg�1) 0.05 ± 0.01 3.3 ± 0.3iso-Valeric acid (g kg�1) 0 0.05 ± 0.01n-Valeric acid (g kg�1) 0 0.1 ± 0.02Capronic acid (g kg�1) 0 0.4 ± 0.1DL-lactic acid (g kg�1) 16.1 ± 0.9 2.7 ± 0.4Fructose (g kg�1) 1.3 ± 0.2 0.1 ± 0.01Ethanol (g kg�1) 3.5 ± 0.02 0.9 ± 0.07Propylene glycol (g kg�1) 4.5 ± 0.2 0.1 ± 0.04

Table 3Acid and ammonium concentration in the effluent of the anaerobic filter operated atdifferent pressure levels.

Pressure (bar) 1.07 ± 0.01 2.97 ± 0.003 5.95 ± 0.004 8.91 ± 0.04Acetic acid (g kg�1) 0 0 0.02 ± 0.008 0.06 ± 0.03Propionic acid (g kg�1) 0 0.003 ± 0.003 0.01 ± 0.007 0.02 ± 0.01n-Butyric acid (g kg�1) 0 0 0 0n-Valeric acid (g kg�1) 0 0 0 0Capronic acid (g kg�1) 0 0 0 0NHþ4 –N (g kg�1) 0.66 ± 0.01 0.71 ± 0.01 0.73 ± 0.01 0.74 ± 0.02

412 Y. Chen et al. / Applied Energy 116 (2014) 409–415

3.2. Control parameters

Fig. 2 shows that the pH value of the anaerobic filter wasstrongly affected by the pressure. The daily average pH value fellfrom approximately 7.2 ± 0.04 at 1.07 ± 0.01 bar to 6.5 ± 0.05 at8.91 ± 0.04 bar. However, the pH reduction rate decreased as pres-sure increased.

The VFA concentrations in the anaerobic filter directly beforethe next feeding are summarized in Table 3. Almost no VFAsaccumulated in the anaerobic filter throughout the experiment.The acetic acid and propionic acid concentration were a littlehigher in the case of higher working pressure. A slight increasewas also observed in ammonium nitrogen concentration, from0.66 ± 0.01 g kg-1 to 0.74 ± 0.02 g kg�1 when the pressure in-creased from 1.07 ± 0.01 bar to 8.91 ± 0.04 bar. Furthermore,neither ethanol nor propylene glycol was detected in the effluentfrom the anaerobic filter. The COD concentration was alwaysbelow 1.5 g l�1. The COD degradation grade of the anaerobic fil-ter remained more than 94% until reaching the operational pres-sure of 5.95 ± 0.004 bar (Fig. 3). Although the difference in CODdegradation grade between 8.91 ± 0.04 bar and the previouslytested pressures reached the statistical significance, the averageCOD degradation grade at 8.91 ± 0.04 bar was still above 93%.As a guide value for assessing the anaerobic digestion stability,the VFA/TIC ratio stayed between 0.1 and 0.2 at all pressurelevels.

Fig. 2. pH Value measured in the anaerobic filter at different pressure levels. The boxmaximum value. The significant differences among the pressure levels are marked with

3.3. Biogas production of the anaerobic filter

Fig. 4 illustrates gas quantity under different working pressurelevels of the anaerobic filter. In this study, the specific methaneyield (SMY) and specific biogas production (SBP) represent thestandardized methane and biogas produced from one gram fedCOD, respectively. In contrast to the SMY, the SBP was more sensi-tive to the variations of pressure. The rising pressure resulted in aremarkable decrease of the specific biogas production. The specificbiogas production at 2.97 ± 0.003 bar, 5.95 ± 0.004 bar and8.91 ± 0.04 bar was reduced by 6%, 12% and 24%, respectively, incomparison with that at the atmospheric pressure. By contrast,the specific methane yield stayed in the same range (the averageSMY was 0.33 ± 0.01 lN g-1CODadded) at a pressure of 1.07 ± 0.01 to5.95 ± 0.004 bar. The increase of the working pressure from5.95 ± 0.004 bar to 8.91 ± 0.04 bar led to a marginal drop of SMYto 0.31 ± 0.02 lN g�1CODadded.

The biogas composition was also influenced by pressure varia-tion (Fig. 5). The methane content increased from 65% to 75%,while the carbon dioxide content decreased from 35% to 25%, aspressure rose from 1.07 ± 0.01 bar to 8.91 ± 0.04 bar. The CH4:CO2

ratio in the collected biogas changed from 1.9 to 2.9. The methanewas enriched under higher pressure.

4. Discussion

4.1. Effect of pressure on pH value

The pH value decreased considerably when pressure increased.Since Henry’s Law can be used for modeling the solubility of CO2 inwater for pressure up to 100 bar [19], it can be explained in thisway: with the working pressure of the anaerobic filter rising, the

plot is characterized with the minimum, low quartile, median, upper quartile anddifferent letters (p < 0.05, LSD test).

Fig. 3. COD degradation grade at different pressure levels. The box plot is characterized with the minimum, low quartile, median, upper quartile and maximum value. Thesignificant differences among the pressure levels are marked with different letters (p < 0.05, LSD test).

Fig. 4. Specific biogas production and specific methane yield of the anaerobic filter at different pressure levels. The box plot is characterized with the minimum, low quartile,median, upper quartile and maximum value. The significant differences among the pressure levels are marked with different letters (p < 0.05, LSD test).

Y. Chen et al. / Applied Energy 116 (2014) 409–415 413

partial pressure of CO2 also increases. Based on Henry’s Law, thesolubility of CO2 is directly proportional to the partial pressure,and thus it is also increased. For example, it was reported thatthe solubility of CO2 in water at 30 �C was 13.5 g CO2 kg�1 H2O at10 bar partial pressure of CO2, ten times as that at 1 bar at the sametemperature [20]. The more CO2 that is dissolved, the more car-bonic acid (H2CO3) will be formed. Carbonic acid might dissociateto bicarbonate (HCO�3 ) and carbonate (CO2�

3 ). When those protonsare liberated in the water, the pH value will be decreased. Since thepH value is defined as the decimal logarithm of the proton concen-tration, it does not vary linearly with the pressure. As shown in theresults, the pH value had a slower downward trend as the pressureincreased.

4.2. Anaerobic filter process stability under pressure

Throughout the experiment period, there was very little acidaccumulation detected in the reactor. At 8.91 bar, prior to the nextfeeding, there was only 0.06 g kg�1 acetic acid and 0.02 g kg�1 pro-pionic acid in the anaerobic filter, much lower than the critical

threshold level (acetic acid: 1 g kg�1 and propionic acid:0.25 g kg�1) [2]. The VFA/TIC ratio was also well below the processinstability threshold value which is reported to be 0.5 [21]. Thisindicated that the pressurized anaerobic filter was constantly sta-ble. Moreover, it can be speculated that the pressurized anaerobicfilter has potential to deal with an even higher organic loading rate(more than 5 kgCOD m�3 d�1) under pressure of 8.91 bar. The COD-degradation grade of higher than 93% further verified that theanaerobic digestion in the pressurized anaerobic filter accom-plished in an efficient way.

It is widely accepted that the optimal pH value for methanogen-esis should be between 6.8 and 7.4 [22]. At a lower pH value, or-ganic acids are prone to be in an undissociated state. Since acidsin uncharged form more easily penetrate the microbial cell wall,too much undissociated acids accumulated in the digester will ex-ert a strong toxic effect [2,23]. Therefore, pH value is directly re-lated to the process stability of the anaerobic digestion underatmospheric pressure. High acid concentration and even fermentersouring generally accompany an anaerobic digester with low pHvalue. However, the results in this study seem contradictory to

Fig. 5. Methane content and carbon dioxide content measured at different pressure levels. The box plot is characterized with the minimum, low quartile, median, upperquartile and maximum value. The significant differences among the pressure levels are marked with different letters (p < 0.05, LSD test).

414 Y. Chen et al. / Applied Energy 116 (2014) 409–415

the common phenomena. Almost all the fatty acids could be de-graded in time and few acids accumulated in the reactor. Thelow pH value of the pressurized anaerobic filter was exclusivelycaused by the dilution of CO2, instead of organic acid accumulation.Even at a pH value of 6.5, the COD-degradation grade remained at ahigh level of 93%. This might imply that the pH value itself seemsto have less effect on the methanogens than high organic acidconcentration.

Moreover, the reduction of pH value induced by pressure in-crease favors the recombination of protons and anions in the solu-tion. At a low-pH value, the ammonia nitrogen appears more indissociated form in the digester. Vavilin, et al. [24] have reportedthat ammonium nitrogen tended to increase with pressure rising.In this study, the upward tendency of ammonium nitrogen wasnot significant, especially at a pressure of more than 2.97 bar. Be-cause the pH value of the reactor fell to 6.5–6.8, and almost allammonia nitrogen turns into ammonium nitrogen when pH valueis below 7.0 [2]. Therefore, the ammonium nitrogen concentrationin the anaerobic filter at a pressure of 2.97–8.91 bar was at the samerange (0.71–0.74 g kg�1), and represents the total ammonia nitrogen(TAN) concentration. Compared with the reported inhibitory TANconcentration (1.7–14 g l�1) [25], the values in this study were quitelow. Since the inhibition concentration of ammonia to methanogensis higher in undissociated form [26], the pressurized anaerobic filtercan reduce the risk of ammonia inhibition to some extent.

4.3. Effect of pressure on biogas production

Throughout the experiment, the average SMY of the pressurizedanaerobic filter ranged from 0.31 to 0.33 lN g�1CODadded, whichcorresponds with other studies [27,28]. Although the SMY at apressure of 8.91 bar was only a little lower than at other pressures,statistically it was a significant difference. There are two possiblereasons for that. The first possibility is that there was a smallamount of acids accumulated in the anaerobic filter at the pressureof 8.91 bar, which indicated that the acids were not completelyconverted into methane at this pressure level. The second possibil-ity is that the methane solubility changed with pressure. Since thepressure dependency of Henry coefficient of methane in the lowpressure range (1–10 bar) is negligible [29], according to Henry’s

Law, with pressure rising from 1.07 bar to 8.91 bar, the concentra-tion of methane increases correspondingly about nine fold, from0.016 g CH4 kg�1 H2O to 0.15 g CH4 kg�1 H2O. The calculation isbased on the Henry coefficient given by Wilhelm et al. [29].Although the CH4 concentration is fairly low in terms of the abso-lute value, it can still be inferred that more produced CH4 tends toremain in the solution at higher pressure. As a result, the SMY wasreduced a little at 8.91 bar.

As previously mentioned, compared with CH4, CO2 dissolvesmuch more readily in water. Due to this significant difference in sol-ubility between the two gases under pressure, the biogas quality isalso affected. During anaerobic digestion, the gases are producedin the liquid. Once the saturation limit is reached, gases escape fromthe liquid in the form of gas bubbles and rise into the gas phase.With more CO2 remaining in the solution at higher pressure, lessCO2 enters the gas phase. Thus, the specific biogas production is re-duced with increased pressure and CH4 tends to be dominant in thebiogas. As shown in this study, the CH4 content was significantly in-creased. A similar observation was also documented in other studies[9,30]. Nevertheless, the CH4 content in this study was lower thanthat in Hayes et al.’s research, in which a 93% methane gas was pro-duced [30]. The primary cause of the difference lies in the varied pHvalue. In Hayes et al.’s experiment, the pH value of the anaerobic fil-ter was maintained at 7.5, while the anaerobic filter in this studywas operated with the pH value down to 6.5. Like pressure, thepH value also influences the amount of dissolved CO2. More CO2

tends to be dissolved in water as pH rises. Wonneberger et al. ex-pressed the relation between the Henry coefficient of CO2 and pHvalue in a formula [31]. With the proper dissociation constant ofCO2 [32] and pH value substituted, the Henry coefficient, as wellas the solubility of CO2 can be calculated. It is found that at the samepressure, the solubility of CO2 at a pH value of 7.5 is about seventimes as much as that at a pH value of 6.5. Therefore, the methaneenrichment in that study was more apparent.

5. Conclusion

The study has examined the pressure effects on the anaerobicfilter in a two-phase anaerobic digestion, in terms of pH value, bio-gas production and process stability. The results presented in this

Y. Chen et al. / Applied Energy 116 (2014) 409–415 415

study demonstrate that an increase in pressure leads to a decreasein pH value in the reactor and methane enrichment in the biogas.Based on the COD-degradation grade, the concentration of VFAsand the specific methane yield, the pressurized anaerobic filter inthis study was characterized by a stable performance with workingpressure up to 8.91 bar, despite the low pH value. The increasingmethane content in the raw biogas can largely reduce the costfor biogas purification and upgrading. In addition, the compressedbiogas can be injected into the public gas grid without additionalpressurization. In light of the good process stability, the two-phasepressurized anaerobic digestion might allow operation under evenhigher pressure (more than 9 bar), as well as higher organic load-ing rate (more than 5 kgCOD m�3 d�1). Although low pH value in-duced by rising pressure might not cease the anaerobic digestion, itcounteracts the pressure effect on CO2-solubility, and thus partlyweakens the methane enrichment. Therefore, an effective and eco-nomical method for maintaining pH value in the pressurized reac-tor should be developed and further verified.

Policy and ethics

We declare that the work described in the article has been car-ried out in accordance with the Code of Ethics of the World MedicalAssociation (Declaration of Helsinki) for experiments involving hu-mans; EU Directive 2010/63/EU for animal experiments; UniformRequirements for manuscripts submitted to Biomedical journals.

Conflict of interest

We declare that we have no financial and personal relationshipswith other people or organizations that can inappropriately influ-ence our work; there is no professional or other personal interestof any nature or kind in any product, service and/or company thatcould be construed as influencing the position presented in, or thereview of, the manuscript entitled.

Submission declaration and verification

We declare that the work described in the article has not beenpreviously published, and it is not under consideration for publica-tion elsewhere. Its publication is approved by all authors and tac-itly or explicitly by the responsible authorities where the work wascompleted, and if accepted, it will not be published elsewhere inthe same form, in English or in any other language, including elec-tronically without the written consent of the copyright holder.

Role of the funding source

This study was financed by the German Ministry of Educationand Research (BMBF). The funding source was not involved instudy design, the collection, analysis and interpretation of data,the writing of the report or the decision to submit the article forpublication.

Acknowledgement

The authors gratefully acknowledge the financial support of thework by the German Ministry of Education and Research withinthe joint research project B2G.

References

[1] Madsen M, Holm-Nielsen JB, Esbensen KH. Monitoring of anaerobic digestionprocesses: a review perspective. Renew Sustain Energy Rev2011;15(6):3141–55.

[2] Frank Graf, Bajohr S. Biogas – Erzeugung, Aufbereitung, Einspeisung.Oldenbourg Industrieverlag: München; 2011.

[3] Jury C, Benetto E, Koster D, Schmitt B, Welfring J. Life cycle assessment ofbiogas production by monofermentation of energy crops and injection into thenatural gas grid. Biomass Bioenergy 2010;34(1):54–66.

[4] Köppel W, Götz M, Graf F. Biogas upgrading for injection into the gas grid. Gwf-Gas/Erdgas 2009;150:26–35.

[5] Weiland P. Biogas production: current state and perspectives. Appl MicrobiolBiotechnol 2010;85(4):849–60.

[6] Amani T, Nosrati M, Sreekrishnan TR. Anaerobic digestion from the viewpointof microbiological, chemical, and operational aspects – a review. Environ Rev2010;18 (NA):255–78.

[7] Sakar S, Yetilmezsoy K, Kocak E. Anaerobic digestion technology in poultry andlivestock waste treatment – a literature review. Waste Manage Res2009;27(1):3–18.

[8] Kowalczyk A, Harnisch E, Schwede S, Gerber M, Span R. Different mixingmodes for biogas plants using energy crops. Applied Energy 2013;112:465–72.

[9] Friedmann H, Märkl H. Der Einfluß von erhöhtem hydrostatischen Druck aufdie Biogasproduktion. Wasser Abwasser 1993;134(12):689–98.

[10] Chynoweth DP, Ron I. Anaerobic digestion of biomass. Illustrated ed. Springer;1987. p. 296.

[11] Sivan O, Schrag DP, Murray RW. Rates of methanogenesis and methanotrophyin deep-sea sediments. Geobiology 2007;5(2):141–51.

[12] Cherubini F, Bargigli S, Ulgiati S. Life cycle assessment (LCA) of wastemanagement strategies: landfilling, sorting plant and incineration. Energy2009;34(12):2116–23.

[13] Märkl H, Reinhold G. Der biogas-turmreaktor, ein neues reaktorkonzept für dieanaerobe abwasserreinigung. Chem Ing Tech 1994;66:534–6.

[14] Fang J, Zhang L, Bazylinski DA. Deep-sea piezosphere and piezophiles:geomicrobiology and biogeochemistry. Trends Microbiol 2010;18(9):413–22.

[15] Aertsen A, Meersman F, Hendrickx MEG, Vogel RF, Michiels CW. Biotechnologyunder high pressure: applications and implications. Trends Biotechnol2009;27(7):434–41.

[16] Abe F, Horikoshi K. The biotechnological potential of piezophiles. TrendsBiotechnol 2001;19(3):102–8.

[17] Vintiloiu A, Lemmer A, Oechsner H, Jungbluth T. Mineral substances andmacronutrients in the anaerobic conversion of biomass: an impact evaluation.Eng Life Sci 2012;12(3):287–94.

[18] Zielonka S, Lemmer A, Oechsner H, Jungbluth T. Energy balance of a two-phaseanaerobic digestion process for energy crops. Eng Life Sci 2010;10(6):515–9.

[19] Carroll JJ, Mather AE. The system carbon dioxide-water and the Krichevsky–Kasarnovsky equation. J Solution Chem 1992;21(7):607–21.

[20] Carroll JJ, Slupsky JD, Mather AE. The solubility of carbon dioxide in water atlow pressure. J Phys Chem Ref Data 1991;20(6):1201–9.

[21] Lebuhn M, Liu F, Heuwinkel H, Gronauer A. Biogas production from mono-digestion of maize silage-long-term process stability and requirements. WaterSci Technol 2008;58(5):1645–51.

[22] Leslie Grady CP, Daigger GT, Lim HC. Biological wastewater treatment:principles and practice. 2nd ed. Marcel Dekker New York; 1999. p. 1076.

[23] Wang Y, Zhang Y, Wang J, Meng L. Effects of volatile fatty acid concentrationson methane yield and methanogenic bacteria. Biomass Bioenergy2009;33(5):848–53.

[24] Vavilin VA, Vasiliev VB, Rytov SV. Modelling of gas pressure effects onanaerobic digestion. Bioresour Technol 1995;52(1):25–32.

[25] Chen Y, Cheng JJ, Creamer KS. Inhibition of anaerobic digestion process: areview. Bioresour Technol 2008;99(10):4044–64.

[26] Rittmann BE, McCarty PL. Environmental biotechnology: principles andapplications. McGraw-Hill; 2001.

[27] Gunaseelan VN. Anaerobic digestion of biomass for methane production: areview. Biomass Bioenergy 1997;13(1–2):83–114.

[28] Amon T, Amon B, Kryvoruchko V, Zollitsch W, Mayer K, Gruber L. Biogasproduction from maize and dairy cattle manure—influence of biomasscomposition on the methane yield. Agric, Ecosystems amp Environ2007;118(1–4):173–82.

[29] Wilhelm E, Battino R, Wilcock RJ. Low-pressure solubility of gases in liquidwater. Chem Rev 1977;77(2):219–62.

[30] Hayes TD, Isaacson HR, Pfeffer JT, Liu YM. In situ methane enrichment inanaerobic digestion. Biotechnol Bioeng 1990;35(1):73–86.

[31] Anna-Maria Wonneberger, Frank Graf, Siegfried Bajohr, Andreas Lemmer,Reimert R. Process development of two-phase pressure fermentation –influence of gas solubilities. International Congress on Progress in Biogas.March 30, 2011-April 01. Stuttgart, Germany; Part 1 2011. p. 176–182.

[32] Edwards TJ, Maurer G, Newman J, Prausnitz JM. Vapor-liquid equilibria inmulticomponent aqueous solutions of volatile weak electrolytes. AIChE J1978;24(6):966–76.