Bioresource Technology 110 (2012) 63–70

Contents lists available at SciVerse ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

Anaerobic co-digestion of pig manure and crude glycerol at mesophilic conditions:Biogas and digestate

S. Astals a,⇑, V. Nolla-Ardèvol a,b, J. Mata-Alvarez a

a Department of Chemical Engineering, University of Barcelona, C/ Martí i Franquès, no. 1, 6 th floor, 08028, Barcelona, Spainb Institute for Genome Research and Systems Biology, Center for Biotechnology, Bielefeld University, Universitatsstraße, no. 27, D-33615, Bielefeld, Germany

a r t i c l e i n f o a b s t r a c t

Article history:Received 12 October 2011Received in revised form 10 January 2012Accepted 16 January 2012Available online 26 January 2012

Keywords:Anaerobic digestionCodigestionSwine manureGlycerine

0960-8524/$ - see front matter � 2012 Elsevier Ltd. Adoi:10.1016/j.biortech.2012.01.080

⇑ Corresponding author. Tel.: +34 934039789; fax:E-mail address: sastals@ub.edu (S. Astals).

Crude glycerol derived from biodiesel production is characterized by its high concentration of organiccarbon and its solubility in water; properties that make it a suitable co-substrate to improve the effi-ciency of a manure digester. An increase of about 400% in biogas production was obtained under meso-philic conditions when pig manure was co-digested with 4% of glycerol, on a wet-basis, compared tomono-digestion. The increase in biogas production was mainly a consequence of the increase in organicloading rate. However, the differences could also be related to the synergy between both substrates andthe carbon-to-nitrogen ratio. Moreover, the analysis of the macro-compounds, protein, lipids, carbohy-drates and fibers, showed lower removal efficiencies in the co-digester as the microorganisms obtainednutrients from the soluble carbohydrates provided by the glycerol. The digestate stability, evaluatedthrough a respirometric assay, showed that co-substrate addition does not exert a negative impact ondigestate quality.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Since 1992, biodiesel has been produced at industrial scale inEurope through transesterification of vegetable oil, animal fatand/or used kitchen oil with alcohol (EBB, 2010). At present, morethan 9,000,000 tonnes of biodiesel per year are produced in about120 plants, which have a combined production capacity of over20,000,000 tonnes of biodiesel per year (EBB, 2010). The mainby-product of biodiesel production is crude glycerol, which is about10% of the weight of the initial raw matter. Specifically, crudeglycerol is a mixture of glycerol, alcohol, water, salts, heavy metals,free fatty acids, unreacted mono-, di- and tri-glycerides and methylesters in varying amounts depending on the quality of the rawmatter and the chemical process used to obtain the biodiesel(Pagliaro and Rossi, 2008; Robra et al., 2010). At the present time,in some regions, the glycerol has to be disposed of as waste since(1) the existing glycerol market cannot absorb the large rise inby-product production (Johnson and Taconi, 2007); (2) treatmentand refinement of crude glycerol is too expensive for small andmedium plants (Pachauri and He, 2006); and (3) crude glyceroldoes not have a lot of direct uses due its impurities (Pagliaro andRossi, 2008). In other regions, crude glycerol can be sold for 80 to300 € per tone depending on the regional market availability andthe glycerol purity (Johnson and Taconi, 2007). Within this

ll rights reserved.

+34 934021291.

scenario, many research efforts to develop economical utilizationsof crude glycerol have been made in order to make the cost of thebiodiesel production sustainable in the long term. Among them,the valorisation of this residue as a co-substrate in anaerobic diges-tion (AD) plants is a promising solution, since a renewable source ofenergy is obtained from the treatment. Several successful studies, inbatch and/or continuous experiments, have been published withreference to the benefits of the addition of glycerol to enhancethe AD of agro-wastes (Amon et al., 2006; Anna et al., 2009), cattlemanure (Chen et al., 2008; Mladenovska et al., 2003; Robra et al.,2010), fruit and vegetable wastes (Ma et al., 2008), organic fractionof municipal solid waste -OFMSW- (Fountoulakis and Manios,2009), pig manure (Álvarez et al., 2010; Amon et al., 2006; Astalset al., 2011; Galí et al., 2009), sewage sludge (Fountoulakis et al.,2010), mixture of pig manure and OFMSW (Schievano et al.,2009), mixture of olive mill and slaughterhouse wastewaters(Fountoulakis and Manios, 2009) and mixture of manure andorganic industrial wastes (Holm-Nielsen et al., 2009). Anaerobicco-digestion (AcoD) consists of the anaerobic digestion of a mixtureof two or more substrates with complementary characteristics. As aresult, biogas and organic matter removal yields are enhanced(Mata-Alvarez et al., 2011). Mixing animal manure and glycerol isof interest since (1) the elevated content of water in manure actsas solvent for glycerol; (2) the high alkalinity of manure gives abuffering capacity for the temporary accumulation of volatile fattyacids; (3) the wide range of macro- and micro-nutrients present inthe manure are essential for bacterial growth; and (4) glycerol

64 S. Astals et al. / Bioresource Technology 110 (2012) 63–70

supplies rapidly biodegradable matter (Mata-Alvarez et al., 2000,2011). Even though the AcoD of animal manure has been widelyinvestigated, most of the studies have focused on processperformance and biogas yield whereas little attention has been paidto digestate quality. However, both the biogas and the digestatehave to be managed in an appropriate way in order to make ADplants feasible. Utilization of the digestate as organic fertilizer orsoil conditioner seems to be the best option for its recycling,since it contains considerable amounts of residual organic carbon(Alburquerque et al., 2011; Salminen and Rintala, 2002). However,digestate properties are conditioned by the raw materials used assubstrate and the development of the anaerobic process in thedigester. Furthermore, the introduction of a co-substrate can lead,in some cases, to the production of unstable digestates witch mayexert negative impacts on organic matter mineralization andnutrient turn-over in the plant-soil system (Alburquerque et al.,2011). Aerobic respiration indexes, resulted from short assays,which take less than 24 h, and based on oxygen uptake are reportedto be the most suitable methods for assessing the stability oforganic wastes (Barrena et al., 2006; Ponsá et al., 2008); whereas,long assays, like BOD5d, give more reliable information as digestatesusually maintain a high respirometric activity beyond the first 24 hof testing (Alburquerque et al., 2011).

2. Methods

2.1. Experimental set-up

Two identical 5.5-L semi-continuous stirred tank reactors witha working volume of 4-L were used. Both reactors were equippedwith a pH probe (Metter Toledo HA405) and with an on-line biogasmeasuring device (Ritter MGC-1). The operational temperaturewas ensured by circulating water from a heated water bath (HU-BER 118A-E) through a jacket surrounding the reactor. The digestermedium was continuously stirred at 60 rpm and the hydraulicretention time (HRT) was set at 20 days. The reference digester(D1) was fed only with pig manure while the co-digestion digester(D2) was supplied with several mixtures of pig manure and glyc-erol. The digesters were manually fed and purged once a day,and the co-digestion mixture was prepared daily in order to avoiduncontrolled degradation. Two different batches of pig manurewere used as feed supply: batch 1 was used from day 1 to 99and batch two from day 100 to 196; the glycerol was the samethroughout the experiment.

2.2. Wastes and inoculum origin

Fresh pig manure (PM) and digested pig manure, used as inoc-ulum, were obtained from a centralized plant, which treats themanure anaerobically, located in Lleida (Spain). After collection,pig manure was stored at 4 �C until its utilization. The crude glyc-erol (GLY) was obtained from an industrial plant in Huesca (Spain)which mainly produces biodiesel through the transesterification ofvegetable oils like sunflower, soybean and/or rape. The glycerolwas stored at 4 �C.

2.3. Analytical procedure and methods

The digesters performance was continuously monitored by theon-site pH probe, the biogas measurement device and by periodicalanalysis of total and partial alkalinity. Once constant values ofthese parameters were reached, steady-state conditions wereconsidered to be achieved. At stationary conditions, analyzes tocharacterize influent and effluent were carried out during severalHRTs. The total fraction was determined directly from the raw

samples. For the analysis of the soluble fraction, the sampleswere centrifuged at 4000 rpm for 10 min and the supernatantwas filtered through a 0.45 lm nitrocellulose filter (MilliporeHAWP02500). In order to monitor waste biodegradation duringthe process, total solids (TS), volatile solids (VS), total suspendedsolids (TSS), volatile suspended solids (VSS), total chemical oxygendemand (CODT) and soluble chemical oxygen demand (CODS) weredetermined following the guidelines given by the standardmethods 2540G, 2504D and 5220D, respectively (APHA, 2005).Dissolved organic carbon (DOC) was measured by means of aShimadzu 5055 TOC-VCSN TOC analyzer. The 5-day biochemicaloxygen demand (BOD5d), done following the 5210D standardmethod procedure, was used to determine the stability of bothdigestates (WTW Oxitop� was used as a measuring system). Totalammonia nitrogen (TAN: free ammonia (N-NH3) plus ammonium(N–NH4

+)) and the Total Kjeldahl Nitrogen (TKN) were measuredaccording to standard methods procedures 4500-NH3D and2500-NorgB. The free ammonia concentration was calculated bymeans of Eq. (1) (Anthonisen et al., 1976).

N� NH3 ¼TAN � 10pH

eð6344

273:15þTÞ þ 10pHð1Þ

Protein, lipids, carbohydrates and fibers of the samples wereanalyzed to characterize the organic matter. Protein was estimatedby multiplying the organic nitrogen (TKN minus TAN) by 6.25; lip-ids were analyzed following standard method 5520E procedure;carbohydrates were estimated by subtracting the amount of pro-tein and lipids from volatile solids (Galí et al., 2009), and fibers,which include cellulose, hemicellulose and lignin, were deter-mined according to the Goering and Van Soest (1970) procedure.In addition, alkalinity and volatile fatty acids (VFA) were used asindicators of process stability. Total (TA) and partial (PA) alkalinitywere determined by a titration method at pH 4.3 and at 5.75,respectively and the intermediate alkalinity (IA) by the differencebetween TA and PA (2320B-APHA, 2005). Individual VFAs (acetate,propionate, iso-butyrate, n-butyrate, iso-valerate and n-valerate)were analyzed by a HP 5890-Serie II gas chromatograph equippedwith a capillary column (NukolTM) and a flame ionization detector.Specifically, the chromatograph oven temperature program was asfollows: hold 1.5 min at 85 �C; ramp to 120 �C at 15 �C min�1;ramp to 145 �C at 10 �C min�1; ramp to 175 �C at 20 �C min�1, hold2 min. Injector and detector temperature was set to 280 �C and300 �C respectively, 33 mL min-1 of Helium at 5 psi was used ascarrier gas. The ions (F�, Cl�, PO4

3�, SO42�, Na+, NH4

+, K+, Ca2+

and Mg2+) were determined in an 861 Advanced Compact IC Metr-ohm ionic chromatographer using Metrosep columns. It should benoted that the samples used for VFA and ions analysis were centri-fuged at 4000 rpm for 10 min and filtered through a 0.22 lm nitro-cellulose filter (Millipore GSWP02500).

2.4. Analysis of raw crude glycerol and mixtures with crude glycerol

Since crude glycerol and PM-GLY mixtures have different phys-ical and chemical properties from a water matrix sample (density,viscosity, boiling point, etc.), a series of modifications in the stan-dard methods for the examination of water and wastewater proce-dures (APHA, 2005) were undertaken. Mixtures were establishedon a weight basis since it was difficult to pour the desired mea-sured volume of GLY out of the receptacle due to its viscosity. Totalsolids were measured after an overnight desiccation of the sampleto 105 �C (liquid could still be found in the crucible as GLY’s boilingpoint is about 290 �C). Volatile solids were determined through thefollowing muffle temperature program: room temperature ramp to250 �C at 15 �C min�1, hold 60 min; ramp to 400 �C at 15 �C min�1,hold 30 min; ramp to 550 �C at 15 �C min�1, hold 90 min. Cleaning

2.0

day-1

)

2.0

2.5A I II III IV V

S. Astals et al. / Bioresource Technology 110 (2012) 63–70 65

or replacement of the VFA gas chromatograph liner and silica woolwas done about every 50 injections. Samples containing GLY werenot analyzed for Kjeldahl nitrogen as they burn when they are ex-posed to heat and strong acid conditions.

0.0

0.5

1.0

1.5

0 10 20 30 40 50 60

Time (days)

Biog

as P

rodu

ctio

n (L

L-1

dige

ster

0.0

0.5

1.0

1.5

OLR

(g V

S L-1

day

-1)

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

0 10 20 30 40 50 60

Time (days)

PA a

nd IA

(g C

aCO

3 L

-1)

0.0

0.2

0.4

0.6

0.8

1.0

IA/P

A ra

tio

B

Fig. 1. Start-up of mesophilic anaerobic co-digestion of pig manure and glycerol:(A) Biogas production (�) and organic loading rate (�, secondary axis); (B) partialalkalinity (j), intermediate alkalinity (N) and IA/PA ratio (+, secondary axis).

3. Results and discussion

3.1. Start-up of mesophilic anaerobic co-digestion

As can be seen in Table 1, the acclimatization of the anaerobicmicroorganisms to crude glycerol (GLY), carried out in the co-digestion digester (D2), was divided into four different periods.The percentage of GLY, on a wet-basis (w/w), was increased asthe co-digester showed signs of adaptation to the new influent,i.e. stabilization of the daily biogas production and the intermedi-ate-to-partial alkalinity ratio (IA/PA ratio), since the addition of lowquantities of GLY meant a significant increase in solid and organicmatter content of the feed supply. The stability of the process wasevaluated by the IA/PA ratio instead of the volatile fatty acids-to-total alkalinity ratio (Ferrer et al., 2010). However, both ratios arebased on the same concept: if the acid concentration, estimatedby the IA, exceeds the buffer capacity provided by the HCO3

- spe-cies, determined by the PA, the digester will sour inhibiting themicroorganism’s activity and, specially, affecting methanogens.Therefore, to consider the process stable, the IA/PA ratio has tobe kept below 0.4. Other authors have evaluated the digester sta-bility with the intermediate-to-total alkalinity ratio (IA/TA ratio)(Fernández et al., 2001); however, the IA/TA ratio is less sensitivethan the IA/PA ratio and is not adequate for systems with highalkalinity. At the beginning of the experiment (period I), bothdigesters, D1 and D2, were only fed with pig manure (PM) untilday 11 when both systems showed similar operational parameters(i.e. biogas production, pH and alkalinity). Then in period II, a 1% ofGLY (w/w), was added to the feed supply of D2 while the referencedigester (D1) was kept fed with PM. As expected the addition ofGLY had an important effect in the organic loading rate (OLR)and in the biogas production (Fig. 1a). In contrast, the IA and thePA in D2 showed similar values to the ones obtained in period I(Fig. 1b), whereas only a small reduction of the pH from 8.1 to7.9 could be noticed (data not shown). In period III, the increasein GLY in the influent, from 1% to 3% w/w, had a clear effect onthe IA/PA ratio, which rose from 0.29 to 0.34. This effect was evenclearer in period IV, when the IA/PA ratio achieved values over0.60, which exceeded the critical value (0.4) to assure a stableAD process. At the beginning of each period, the increase in theVFA concentration, and therefore in the IA, due to the increase ofthe GLY content in the feed supply was a result of the VFA turn-over until the anaerobic microorganisms adjust to the new influent(Angelidaki et al., 1997). However, in period IV, where the contentof GLY was augmented to 5% w/w, the system did not show signs ofadaptation. This process instability, which was leading the digesterto failure, was a consequence of several factors: (1) the negligiblealkalinity of the GLY reduced the alkalinity of the mixture and, asa consequence, in the digester; (2) GLY represented a source of rap-idly biodegradable organic matter, which generated large amountsof VFA; and (3) the high OLR as a result of the addition of GLY (as

Table 1Operational mode of the star-up of mesophilic anaerobic co-digestion.

Units I

Proportion PM-GLY (w/w) – 100/0Proportion PM-GLY (TS/TS) – 100/0Proportion PM-GLY (VS/VS) – 100/0Operation time Days 1–11

can be seen in Table 1, in period IV about 80% of the organic matterin the influent was provided by the GLY). After 12 days in period IV(day 49), and to avoid process failure, the percentage of GLY wasreduced from 5 to 4% w/w (period V). The reduction of the GLYcontent had a satisfactory effect on process stability, because aftertwo days (day 51), the IA/PA ratio decreased to values lower than0.4. It should be pointed out that during the whole start-up processthe pH values were stable (between 7.9 and 7.6). However, it isprobable that, if the alkalinity values had been lower, the pH wouldhave dropped more as a result of VFA accumulation. Finally, itshould be noted that biogas production in period IV and in periodV was nearly the same (Fig. 1a), a clear sign of organic overloadingin period IV. Therefore, 4% w/w of GLY in the feed supply was con-sidered to be the limiting concentration to maintain a stable ADprocess. Moreover, this value is similar to the limiting concentra-tions of GLY obtained by other authors that have carried out exper-iments with manure. For instance, Amon et al. (2006) reported a6% w/w of GLY with a mixture of pig manure, maize silage and

II III IV V

99/1 97/3 95/5 96/470/30 43/57 31/69 36/6458/42 31/69 21/79 25/7512–21 22–35 36–48 49–60

2.0

-1)

VI VII

66 S. Astals et al. / Bioresource Technology 110 (2012) 63–70

rapeseed meal; and Robra et al. (2010) proposed a 5% w/w of GLYin a system fed with cattle slurry.

0.0

0.5

1.0

1.5

60 80 100 120 140 160 180 200

Time (days)

Bio

gas

Prod

uctio

n (L

L-1di

gest

er d

ay

Fig. 2. Daily biogas production in the reference (o) and in the co-digestion (�)digester.

3.2. Mesophilic anaerobic co-digestion: first period

The characteristics of the influent of the first stage period (per-iod VI) are reported in Table 2. When the PM (influent of D1) andthe mixture (influent of D2) are compared, the addition of crudeglycerol had an important effect on parameters related to the mat-ter content (TS, VS, COD). The addition of GLY resulted in a 120%increase in TS while the VS and the COD increased by around190%. Moreover, due to its solubility in water, the main impactof GLY was on parameters related to soluble organic matter. Theaddition of GLY led to an increase in the VS/TS ratio from 0.6 to0.8, and the VSS/TSS ratio maintained similar values in both influ-ents (�0.7). In contrast, a decrease in the VSS/VS ratio (from 0.7 to0.3) and an increase in the CODS/CODT ratio (from 0.5 to 0.8) wereobserved. The GLY used in this study was neutral (pH 6.5) and witha negligible concentration of nitrogen compounds and alkalinity. Infact, when both influents were compared, the pH was the samewhile a slight reduction (around 4%) was observed for the nitrogencompounds and alkalinity. The daily biogas production, at standardtemperature and pressure (STP) conditions, of both digesters ispresented in Fig. 2. In period VI, biogas production from D1 wasapproximately 1.2 L day�1 while D2 produced approximately 5.6L day�1, which represents an increase in biogas production of380%. It has to be highlighted that an increase of 380% representsthe highest biogas increase among the studies that have usedGLY as co-substrate, where the average increase vary from 100%to 200% (Amon et al., 2006; Fountoulakis and Manios, 2009;Fountoulakis et al., 2010; Ma et al., 2008). It is clear that the

Table 2Characterisation of influents and effluents of reference and co-digestion digesters in perio

Units Period VI

D1 D2

Influent Effluent Influent

Mixture PM: GLY % (w/w) 100/0 96/4OLR gSV LR

�1 day�1 0.6 ± 0.1 1.9 ± 0.1

Influent and effluent compositionTS g L�1 21.5 ± 1.6 16.9 ± 0.5 47.2 ± 1.6VS g L�1 12.9 ± 1.2 8.3 ± 0.5 37.8 ± 1.6TSS g L�1 12.8 ± 1.2 10.5 ± 0.8 12.6 ± 1.6VSS g L�1 9.5 ± 0.9 6.7 ± 0.4 9.6 ± 1.0CODT g O2 L�1 24.7 ± 2.1 12.6 ± 1.3 71.3 ± 2.2CODS g O2 L�1 12.8 ± 1.5 3.7 ± 0.8 58.3 ± 0.9pH – 7.7 ± 0.1 8.1 ± 0.1 7.8 ± 0.1Partial Alk. g CaCO3g L�1L�1 4.7 ± 0.5 9.1 ± 0.7 4.4 ± 0.6Total Alk. g CaCO3 L�1 9.1 ± 0.7 11.0 ± 0.5 8.9 ± 0.9VFA g L�1 5.1 ± 1.1 0.08 ± 0.06 5.2 ± 1.2- Acetic acid g L�1 4.2 ± 0.9 0.06 ± 0.03 4.1 ± 1.0- Propionic acid g L�1 0.1 ± 0.1 0.04 ± 0.01 0.2 ± 0.1- Butyric acid g L�1 0.5 ± 0.1 0.02 ± 0.01 0.5 ± 0.1- Valeric acid g L�1 0.3 ± 0.1 n.d.* 0.3 ± 0.1N–NH4

+ g L�1 1.2 ± 0.1 1.4 ± 0.1 1.2 ± 0.1N–NH3 g L�1 0.07 ± 0.01 0.18 ± 0.01 0.08 ± 0.0NTK g L�1 1.6 ± 0.2 1.6 ± 0.1 1.5** ± 0.

Removal efficiencyTSremoval % 21.4 ± 4.9 59.1 ± 3.5VSremoval % 35.7 ± 4.7 74.1 ± 3.2CODremoval % 49.0 ± 5.4 78.8 ± 2.5

Biogas characteristicsBiogas production Lbiogas day�1 1.16 ± 0.11 5.58 ± 0.2SBP-VR Lbiogas LR

�1 day�1 0.29 ± 0.03 1.40 ± 0.0SBP-SVadded Lbiogas g VS added

-1 0.45 ± 0.04 0.74 ± 0.0

* n.d. non detected (<0.01 g L�1).** Estimated through mass balance.

difference in biogas yield is mainly a consequence of the increasein the OLR, which increased by 190% (from 0.64 g VS LR

�1 day�1

to 1.88 g VS LR�1 day�1). Nevertheless, the difference observed

between both specific biogas productions, 0.45 Lbiogas g�1 VS inD1 and 0.74 Lbiogas g�1 VS in D2, emphasizes the high biodegrad-ability of glycerol and the synergy between both substrates inthe co-digester medium. The difference can also be related to thecarbon-to-nitrogen (C/N) ratio and the free ammonia nitrogenpresent in the digester.

It is well known that one of the main issues for the co-diges-tion process lies in balancing the C/N ratio. In fact, ideal co-sub-strates for manures, substrates with high nitrogen contents andhigh alkalinity, are wastes which have a high C/N ratio, like

ds VI and VII.

Period VII

D1 D2

effluent influent Effluent Influent Effluent

100/0 96 / 40.5 ± 0.01 1.7 ± 0.1

19.3 ± 0.5 18.8 ± 0.3 14.7 ± 0.1 44.3 ± 1.7 17.2 ± 0.39.8 ± 0.4 10.5 ± 0.2 6.2 ± 0.1 34.9 ± 1.7 7.8 ± 0.211.9 ± 0.7 8.3 ± 0.3 7.8 ± 0.3 8.4 ± 0.3 9.4 ± 0.38.2 ± 0.5 6.0 ± 0.1 4.6 ± 0.2 6.0 ± 0.1 6.3 ± 0.115.1 ± 1.9 21.0 ± 0.7 9.3 ± 0.8 66.9 ± 2.2 11.0 ± 0.94.3 ± 1.2 13.2 ± 0.5 1.9 ± 0.2 56.2 ± 1.1 2.1 ± 0.28.0 ± 0.1 7.7 ± 0.1 7.9 ± 0.1 7.6 ± 0.1 7.8 ± 0.18.7 ± 0.3 4.1 ± 0.2 8.6 ± 0.2 3.9 ± 0.2 8.4 ± 0.310.7 ± 0.3 8.8 ± 0.3 10.4 ± 0.3 8.5 ± 0.2 10.2 ± 0.30.07 ± 0.06 7.4 ± 1.4 0.16 ± 0.09 7.6 ± 1.6 0.17 ± 0.080.05 ± 0.03 5.9 ± 1.3 0.08 ± 0.06 6.0 ± 1.6 0.11 ± 0.040.02 ± 0.01 0.3 ± 0.1 0.05 ± 0.03 0.3 ± 0.1 0.04 ± 0.020.04 ± 0.01 0.7 ± 0.1 0.03 ± 0.02 0.7 ± 0.1 0.02 ± 0.01n.d. 0.3 ± 0.1 n.d. 0.6 ± 0.1 n.d.1.3 ± 0.1 1.0 ± 0.1 1.2 ± 0.1 0.9 ± 0.1 1.0 ± 0.1

1 0.13 ± 0.01 0.05 ± 0.01 0.12 ± 0.01 0.04 ± 0.01 0.08 ± 0.012 1.9 ± 0.2 1.5 ± 0.1 1.5 ± 0.1 1.4** ± 0.1 1.4 ± 0.1

21.8 ± 1.7 61.2 ± 1.841.0 ± 1.7 77.7 ± 1.555.7 ± 3.2 84.9 ± 2.0

0 1.06 ± 0.04 5.44 ± 0.145 0.27 ± 0.01 1.36 ± 0.043 0.50 ± 0.02 0.78 ± 0.02

S. Astals et al. / Bioresource Technology 110 (2012) 63–70 67

crude glycerol (Mata-Alvarez et al., 2011). Moreover, it has beenshown that optimum values for C/N ratio are within the range of20 to 70. The co-digestion digester had a C/N ratio of 48,calculated as COD/TKN (Álvarez et al., 2010), which is withinthe optimum range. In contrast, the reference digester had alow C/N ratio (C/N = 15). Total ammonia nitrogen (TAN) inhibi-tion is especially distinct when digesting manures (Hansenet al., 1998) and a wide range of inhibiting TAN concentrationshave been reported. The differences in TAN can be attributedto the characteristics of the substrates and inoculums, environ-mental conditions (temperature and pH) and adaptation periods(Chen et al., 2008). Since NH3 has been reported to be the maincause of inhibition, especially affecting methanogens, it has tobe pointed out that the NH3 concentration depends basicallyon three parameters: TAN concentration, temperature and pH(Eq. (1)) (Chen et al., 2008; Kayhanian, 1999). In period VI, theTAN concentration in D1 and D2 were similar; however, theNH3 concentration was notably different in both digesters(0.18 g L�1 in D1 and 0.13 g L�1 in D2). The influence of NH3

on the anaerobic process can be described by the inhibitionequation reported in the ADM1. In this model, free ammoniainhibition is a non-competitive inhibition affecting the acetateuptake rate (Eq. (2)), where SNH3 is the free ammonia concentra-tion and KI;NH3 is the inhibition constant (0. 26 g L�1 - Angelidakiet al., 1999). Values from Eq. (2) range from 0 (total inhibition)to 1 (no inhibition).

INH3Xac ¼1

1þ SNH3KI;NH3

ð2Þ

Therefore, D1 ðINH3 ;Xac ¼ 0:59Þ was slightly more inhibited byfree ammonia than D2 ðINH3 ;Xac ¼ 0:66Þ. This fact can be explainedby the dilution effect in the TAN concentration made by the addi-tion of GLY and the slight decrease in pH in D2 (Table 2). However,this inhibition did not lead to an increase in VFA (below 0.2 g L�1 inboth digesters) or process instability, since the interact ion be-tween NH3, VFA and pH led the AD to an ‘‘inhibited steady state’’,which is a condition where the process is running stable but withlower methane yields (Angelidaki and Ahring, 1994). In addition tothe increase in biogas production, a higher VS and CODT removalyield were obtained during co-digestion when compared tomono-digestion. In absolute numbers, D1 degraded 36 and 49%of the available VS and CODT, respectively, while D2 eliminated74% of VS and 79% of the CODT. Although both digestates exhibitedsimilar compositions in terms of solids and COD, the digestate fromD1 exhibited, for all these parameters, lower values than the dige-state from D2. This phenomenon can be explained as a conjunctionof two factors. First, the addition of GLY represented an importantsupply of organic carbon resulting in an increase in biomass (Maet al., 2008), so all the extra organic matter could be degradedand overloading inhibition was avoided. Through a mass balance,where the average anaerobic biomass yield is 0.1 g CODbiomass

g�1 CODeliminated and the biomass growth rate is 0.8 g VSbiomass g�1

CODeliminated, an increase of 0.9 g VSbiomass day�1 (0.2 g VS L�1) wasobtained. However, this biomass growth was not enough to explainthe difference in VS between both digestates, because the differ-ence was 1.5 g VS L�1. Second, the biomass of D2 did not hydrolyzeall the particulate matter supplied by the PM as it used GLY as amajor source of carbon, while the biomass of D1 had to obtainnutrients from the particulate matter as it was the only source ofnutrients. In fact, the biogas potential of the mixture supplied toD2 was 6.1 Lbiogas day�1, where 4.9 Lbiogas day�1 came from theGLY (the theoretical biogas production per gram of glycerol is0.73 Lbiogas gGLY

-1) and 1.2 Lbiogas day�1 came from the PM. How-ever, taking into account that the biogas production of D2 was5.6 Lbiogas day�1 it can be understood that some degradable

compounds remained in the D2 effluent since there is a differenceof about 0.5 Lbiogas day�1 between the potential and the obtainedbiogas production.

3.3. Mesophilic anaerobic co-digestion: second period

Periods VI and VII were run with the same operational condi-tions; consequently similar results were obtained in both periodsin terms of solid removal efficiencies, biogas yield (Table 2) andbiogas production (Fig. 2). The second stage (period VII) wascarried out to study the differences between both processes andtherefore, not only standard parameters were monitored but alsoprotein, lipids, carbohydrates and fibers. It has been reported thatAD of pig manure is limited by its low hydrolysis rate (Bonmatíet al., 2001). This fact is even more significant in very degradedpig manures, like the one used in this study, which was character-ized by an NH4

+/NTK ratio of about 0.7. As can be observed inTable 3, pig manure had large amounts of protein and carbohy-drates, being fibers the main fraction of the carbohydrates, whilethe lipids content was small. In contrast, the 4% w/w mixturewas very rich in carbohydrates, fibers were less than the 10% ofthe fraction, and the amounts of protein and lipids were small.The 4% w/w mixture contained more lipids than PM, probablydue to unreacted glycerides supplied by the GLY. The microorgan-isms from D1 degraded more protein, lipids and fibers than themicroorganisms from D2 (Table 3). These results suggest thatmicroorganisms in D1 had to hydrolyze large quantities of partic-ulate matter to obtain nutrients. In contrast, bacteria in D2 hadplenty of nutrients because of the large amounts of carbohydratesprovided by the GLY and therefore did not need to hydrolyzelarge amounts of particulate matter. Fig. 3, shows that the biomassof D2 produced 25% of the daily biogas production in the first twohours, whereas during the same period, biogas production from D1was very low (10% of the daily biogas production); fact thathighlighted the difference between both feed supplies in terms ofeasy biodegradable organic matter. After this initial period, thedisparity between the degradation rates makes it clear that thePM digester was limited by its hydrolysis rate (0.08 Lbiogas h�1),while the co-digestion digester transformed fluently solublecarbohydrates into biogas (0.2 Lbiogas h�1). Biogas production ofD1 showed a plateau after 18 h (95% of the daily biogas productionwas already produced) as a consequence of organic matter exhaus-tion. In contrast, biogas production of D2 did not show a plateau, asthe biomass of D2 needed more time to consume all the availableorganic matter, which could be obtained with large HRT.

3.4. Digestate stability for agricultural use

The stabilization of organic waste is related to the mineraliza-tion of part of its organic compounds. Many parameters have beenused as indicators of the mineralization of organic streams (AlMomani et al., 2004; Tambone et al., 2009): (1) COD/DOC ratio,where lower ratios imply a higher degree of mineralization; (2)the average oxidation state (AOS – Eq. (3)), which ranges from +4for CO2, the most oxidized state of C, and �4 for CH4, the most re-duced state of C; or (3) the COD/TKN ratio, which decreased due toCOD degradation.

AOS ¼ 4 ðDOC � CODÞCOD

ð3Þ

However, for semi-solid wastes, a respiration index, likeBOD5d, seems more adequate (Alburquerque et al., 2011; Ponsáet al., 2008). As shown in Table 4, the BOD5d of the non-digestedand digested samples highlight the waste stabilization duringthe AD process. Actually, the BOD5d was reduced by about 80%in D1 and more than 90% in D2; values that are similar to the

Table 3Characterisation of protein, lipids, carbohydrates and fibers in influents and effluents of reference and co-digestion digesters in period VII.

Units Period VII

D1 D2

Influent Effluent Influent Effluent

Mixture PM: GLY % (w/w) 100/0 96/4Influent and effluent compositionProtein g L�1 3.5 ± 0.4 1.6 ± 0.2 3.4* ± 0.4 2.5 ± 0.3Lipid g L�1 1.4 ± 0.8 0.4 ± 0.2 2.2 ± 0.2 1.4 ± 0.3Carbohydrates g L�1 5.5 ± 0.8 4.2 ± 0.5 29.3 ± 1.7 3.9 ± 0.3Fiber g L�1 2.3 ± 1.7 1.6 ± 0.9 2.2* ± 1.7 2.0 ± 1.1

Removal efficiencyProteinremoval % 55.5 ± 19.3 25.2 ± 10.8Lipidremoval % 69.9 ± 5.2 34.9 ± 14.1Carbohydratesremoval % 25.4 ± 5.8 86.7 ± 2.4Fiberremoval % 30.3 ± 21.3 11.0 ± 7.8

* Estimated through mass balance.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0 3 6 9 12 15 18 21 24

Time (hours)

Accu

mul

ated

bio

gas

prod

uctio

n (L

L-1di

gest

er)

0.0

0.1

0.1

0.2

0.2

0.3

0.3

0 3 6 9 12 15 18 21 24Time (hours)

Biog

as fl

ow ra

te (L

h-1

)

A

B

Fig. 3. (A) Accumulated biogas production in a day in the reference (o) and in the co-digestion (e) digester. (B) Biogas flow rate in the reference (d) and in the co-digestion(�) digester organic loading rate.

68 S. Astals et al. / Bioresource Technology 110 (2012) 63–70

data reported by other authors, who determined the stabiliza-tion, by aerobic respirometrics techniques, of some organicwaste before and after the AD process (Tambone et al., 2009).Moreover, when the BOD5d of the PM (9.7 g O2 L�1 –7.7 mgO2 VS�1 h�1) was compared with the stability limit value pro-posed by Ponsá et al. (2008) and Alburquerque et al. (2011),(2 mg O2 VS�1 h�1 and 6 g O2 L�1, respectively) for a safety agri-cultural use, the need for a stabilization treatment was accentu-ated. Since the PM exceeded the limit values, it should not bedirectly applied to soil as fertilizer or conditioner. Furthermore,the introduction of GLY as a co-substrate increased the BOD5d

from 9.7 to 32.3 g O2 L�1 as a consequence of the presence ofeasily biodegradable compounds (Barrena et al., 2006). In con-trast, digestate stability of both digesters, in terms of BOD5d,was nearly the same: 1.8 g O2 L�1 (2.4 mg O2 VS�1 h�1) for D1and 2.0 g O2 L�1 (2.1 mg O2 VS�1 h�1) for D2. These values area little higher than those proposed by Ponsá et al. (2008) butlower than the more restrictive limit (<2.5 g O2 L�1) proposedby Alburquerque et al. (2011). Additionally, the latter authorsalso suggested DOC (<1.5 g C L�1) and the DOC/TKN ratio(<1.5 g C g N�1) of digestates as stability indicators for its agri-cultural use. These parameters are of importance since a high

Table 4Digestate quality parameters.

Units Period VII

D1 D2

Influent Effluent Influent Effluent

Mixture PM: GLY % (w/w) 100/0 96/4Influent and effluent characteristicsBOD g O2 L�1 9.7 ± 0.4 1.8 ± 0.3 32.3 ± 0.1 2.0 ± 0.3DOC g C L�1 4.6 ± 0.2 0.8 ± 0.1 17.5 ± 0.3 1.0 ± 0.1Conductivity mS cm�1 17.2 ± 0.8 17.0 ± 0.7 16.6 ± 0.5 17.3 ± 0.4Fluoride g L�1 0.6 ± 0.1 0.6 ± 0.1 n.d. n.d.Chloride g L�1 1.2 ± 0.1 1.2 ± 0.1 1.2 ± 0.1 1.2 ± 0.1Phosphate g L�1 n.d. n.d. n.d. n.d.Sulfate g L�1 0.3 ± 0.1 0.3 ± 0.1 n.d. n.d.Sodium g L�1 0.7 ± 0.1 0.7 ± 0.1 0.7 ± 0.1 0.7 ± 0.1Potassium g L�1 2.2 ± 0.1 2.8 ± 0.1 2.2 ± 0.1 2.8 ± 0.1Calcium g L�1 0.5 ± 0.1 0.5 ± 0.1 0.2 ± 0.1 0.2 ± 0.1Magnesium g L�1 0.2 ± 0.1 0.2 ± 0.1 0.1 ± 0.1 0.1 ± 0.1COD/DOC ratio mol O2 mol C�1 1.08 0.89 1.20 0.75AOS – �0.30 0.44 �0.82 0.85DOC/TN g C g N�1 3.1 0.5 12.5 0.7

⁄ n.d. non detected (<0.1 g L�1)

S. Astals et al. / Bioresource Technology 110 (2012) 63–70 69

percentage of TKN as NH4+ (80% and 70% for D1 and D2 respec-

tively) improve the N-fertiliser potential of the digestate and lowcarbon doses favor carbon mineralization and rapid ammoniumnitrification in the soil–plant system (Riffaldi et al., 1996).

4. Conclusions

Anaerobic co-digestion of pig manure and glycerol is satisfac-tory since the glycerol increased the digester organic loading rate,balanced the carbon-to-nitrogen ratio and decreased the freeammonia concentration in the digester medium. The disparity be-tween organic compounds removal and biogas flow rates madeclear that anaerobic digestion of pig manure is limited by the dis-integration-hydrolysis step while the co-digestion transformed flu-ently soluble carbohydrates into biogas. Finally, the respirationvalues of both digestates were near the more restrictive limit fora safe agricultural use.

Acknowledgements

The research is part of the PROBIOGAS Project (PSE-120000-2007-16), which is supported by the Ministerio de Ciencia e Inno-vación of Spain. The authors are grateful to VAG, S.L. (Valoritza-cions Agroramaderes de les Garrigues) and to Combunet, S.L. forproviding samples and sampling facilities. Sergi Astals Garcia isalso grateful to the Ministerio de Ciencia e Innovación for the doc-toral grant (CTM2008-05986).

References

Alburquerque, J.A., de la Fuente, C., Bernal, M.P., 2011. Chemical properties ofanaerobic digestates affecting C and N dynamics in amended soils. Agric.Ecosyst. Environ. Article in press.

Al Momani, F., Sans, C., Esplugas, S., 2004. A comparative study of the advancedoxidation of 2,4-dichlorophenol. J. Hazard. Mater. 19 (3), 123–129.

Álvarez, J.A., Otero, L., Lema, J.M., 2010. A methodology for optimising feedcomposition for anaeròbic co-digestion of agro-industrial wastes. Bioresour.Technol. 101, 1153–1158.

Amon, Th., Amon, B., Kryvoruchko, V., Bodiroza, V., Pötsch, E., Zollitsch, W., 2006.Optimising methane yield from anaerobic digestion of manure: effects of dairysystems and glycerine supplementation. Int. Congress Ser. 1293, 217–220.

Angelidaki, I., Ahring, B.K., 1994. Anaerobic thermophilic digestion of manure atdifferent ammonia loads: effect of temperature. Water Res. 28 (3), 727–731.

Angelidaki, I., Ellegaard, L., Ahring, B.K., 1997. Modeling anaerobic codigestion ofmanure with olive oil mill effluent. Water Sci. Technol. 36, 263–270.

Angelidaki, I., Ellegaard, L., Ahring, B.K., 1999. A comprehensive model of anaerobicbioconversion of complex substrates to biogas. Biotechnol. Bioeng. 63, 363–372.

Anna, K., Liliana, K., Stanislaw, L., 2009. Anaerobic co-digestion of agriculturalproducts and industrial wastes. Environ. Prot. Eng. 35 (3), 215–224.

Anthonisen, A.C., Loehr, R.C., Prakasam, T.B.S., Srinath, E.G., 1976. Inhibition ofnitrification by ammonia and nitrous acid. J. Water Pollut. Control Fed. 48, 835–849.

Astals, S., Ariso, M., Galí, A., Mata-Alvarez, J., 2011. Co-digestion of pig manure andglycerine: experimental and modelling study. J. Environ. Manage. 92 (4), 1091–1096.

APHA, 2005. Standard Methods for the Examination of Water and Wastewater.American Public Health Association, Washington, ISBN 978-0-87553-047-5.

Barrena, R., Vázquez, F., Sánchez, A., 2006. The use of respiration indices in thecomposting process: a review. Waste Manage. Res. 24, 37–47.

Bonmatí, A., Flotats, X., Mateu, L., Campos, E., 2001. Study of thermal hydrolysis as apre-treatment to mesophilic anaerobic digestion of pig slurry. Water Sci.Technol. 44 (4), 109–116.

Chen, X., Romano, R.T., Zhang, R., Kim, H.S., 2008. Anaerobic co-digestion of dairymanure and glycerine. Am. Soc. Agric. and Biol. Eng. Annu. Int. Meeting 8, 5053–5070.

EBB -European Biodiesel Board., 2010. The EU biodiesel industry. Available from:<http://www.ebb-eu.org/stats.php>.

Fernández, J.M., Omil, F., Méndez, R., Lema, J.M., 2001. Anaerobic treatment offibreboard manufacturing wastewaters in a pilot scale hybrid USBF reactor.Water Res. 35 (17), 4150–4158.

Ferrer, I., Vázquez, F., Font, X., 2010. Long term operation of a thermophilicanaerobic reactor: process stability and efficiency at decreasing sludgeretention time. Bioresour. Technol. 101 (2), 2972–2980.

Fountoulakis, M.S., Manios, T., 2009. Enhanced methane and hydrogen productionfrom municipal solid waste and agro-industrial by-products co-digested withcrude glycerol. Bioresour. Technol. 100, 3043–3047.

Fountoulakis, M.S., Petousi, I., Manios, T., 2010. Co-digestion of sewage sludge withglycerol to boost biogas production. Waste Manage. 30, 1849–1853.

Galí, A., Benabdallah, T., Astals, S., Mata-Alvarez, J., 2009. Modified version of ADM1model for agro-waste application. Bioresour. Technol. 100 (11), 2783–2790.

Goering, H.K., Van Soest, P.J., 1970. Forage fibre analyses (apparatus, reagents,procedures and some applications). United States Agricultural Research ServiceAgricultural Handbook no. 379, Washington, D.C.

Hansen, K.H., Angelidaki, I., Ahring, B.K., 1998. Anaerobic digestion of swinemanure: inhibition by ammonia. Water Res. 32 (1), 5–12.

Holm-Nielsen, J.B., Al Seadi, T., Oleskowicz-Popiel, P., 2009. The future of anaerobicdigestion and biogas utilization. Bioresour. Technol. 100 (22), 5478–5484.

Johnson, D.T., Taconi, K.A., 2007. The glycerin glut: options for the value-addedconversion of crude glycerol resulting from biodiesel production. Environ. Prog.26 (4), 338–348.

Kayhanian, M., 1999. Ammonia inhibition in high-solids biogasification: anoverview and practical solutions. Environ. Technol. 20, 355–365.

Ma, J., Van Wambeke, M., Carballa, M., Verstraete, W., 2008. Improvement of theanaerobic treatment of potato processing wastewater in a UASB reactor by co-digestion with glycerol. Biotechnol. Lett. 30, 861–867.

Mata-Alvarez, J., Macé, S., Llabrés, P., 2000. Anaerobic digestion of organic solidwastes. An overview of research achievements and perspectives. Bioresour.Technol. 74 (1), 3–16.

Mata-Alvarez, J., Dosta, J., Macé, S., Astals, S., 2011. Codigestion of solid wastes: areview of its uses and perspectives including modelling. Crit. Rev. Biotechnol.31 (2), 99–111.

Mladenovska, Z., Dabrowski, S., Ahring, B.K., 2003. Anaerobic digestion of manureand mixture of manure with lipids: biogas reactor performance and microbialcommunity analysis. Water Sci. Technol. 48(6), 271–278.

70 S. Astals et al. / Bioresource Technology 110 (2012) 63–70

Pachauri, N., He, B., 2006. Value-added utilization of crude glycerol from biodieselproduction: a survey of current research activities. American Society of Agriculturaland Biological Engineers Annual International Meeting, Portland (USA).

Pagliaro, M., Rossi, M., 2008. The future of glycerol: new uses of a versatile rawmaterial. The Royal Society of Chemistry, Cambridge, ISBN 978-0-85404-124-4.

Ponsá, S., Gea, T., Alerm, L., Cerezo, J., Sánchez, A., 2008. Comparison of aerobic andanaerobic stability indices through a MSW biological treatment process. WasteManage. 28, 2735–2742.

Riffaldi, R., Saviozzi, A., Levi-Minzi, R., 1996. Carbon mineralization kinetics asinfluenced by soil properties. Biol. Fert. Soils 22, 293–298.

Robra, S., Serpa da Cruz, R., de Olveira, A.M., Almeida-Neto, J.A., Santos, J.V., 2010.Generation of biogas using crude glycerine from biodiesel production as asupplement to cattle slurry. Biomass Bioenergy 34, 1330–1335.

Salminen, E.A., Rintala, J.A., 2002. Semi-continuous anaerobic digestion of solidpoultry slaughterhouse waste: effect of hydraulic retention time and loading.Water Res. 36 (13), 3175–3182.

Schievano, A., D’Imporzano, G., Adani, F., 2009. Substituting energy crops withorganic wastes and agro-industrial residues for biogas production. J. Environ.Manage 90, 2537–2541.

Tambone, F., Genevini, P., D’Imporzano, G., Adani, F., 2009. Assessing amendmentproperties of digestate by studying the organic matter composition and thedegree of biological stability during the anaerobic digestion of the organicfraction of MSW. Bioresour. Technol. 100 (12), 3140–3142.