anaerobic digestion of glycerol derived from biodiesel manufacturing

Bioresource Technology 100 (2009) 5609–5615

Contents lists available at ScienceDirect

Bioresource Technology

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

Anaerobic digestion of glycerol derived from biodiesel manufacturing

José Ángel Siles López *, María de los Ángeles Martín Santos, Arturo Francisco Chica Pérez,Antonio Martín MartínDepartamento de Química Inorgánica e Ingeniería Química, Facultad de Ciencias, Universidad de Córdoba, Campus Universitario de Rabanales, Edificio C-3, Ctra. Madrid-Cádiz,km 396, 14071 Córdoba, Spain

a r t i c l e i n f o

Article history:Received 6 March 2009Received in revised form 1 June 2009Accepted 6 June 2009Available online 9 July 2009

Keywords:Biodiesel manufacturingGlycerol-containing wasteAnaerobic digestionKinetic constantsBiodegradability

0960-8524/$ - see front matter � 2009 Elsevier Ltd. Adoi:10.1016/j.biortech.2009.06.017

* Corresponding author. Tel.: +34 957 218586; fax:E-mail address: [email protected] (J.Á. Siles López).

a b s t r a c t

The anaerobic digestion of glycerol derived from biodiesel manufacturing, in which COD was found to be1010 g/kg, was studied in batch laboratory-scale reactors at mesophilic temperature using granular andnon-granular sludge. Due to the high KOH concentration of this by-product, H3PO4 was added to recoverthis alkaline catalyst as agricultural fertilizer (potassium phosphates). Although it would not be econom-ically viable, a volume of glycerol was distilled and utilised as reference substrate. The anaerobicrevalorisation of glycerol using granular sludge achieved a biodegradability of around 100%, while themethane yield coefficient was 0.306 m3 CH4/kg acidified glycerol. Anaerobic digestion could be a goodoption for revalorising this available, impure and low priced by-product derived from the surplus of bio-diesel companies. The organic loading rate studied was 0.21–0.38 g COD/g VSS d, although an inhibitionphenomenon was observed at the highest load.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Concern has recently risen over the use of fossil resources, theircost, their sustained availability and their impact on globalwarming and pollution (Hansen et al., 2005). Since biodiesel fuelshave various advantages such as a complement to petroleum-based fuel, renewable fuel, a favourable energy balance, less harm-ful emissions and non-toxic fuel, these have drawn much attentionrecently (Ito et al., 2005).

During the biodiesel production process, oils/fats (triglycerides)are mixed with methyl alcohol and alkaline catalysts to produce freefatty-acid esters, with glycerol as a primary by-product (Chi et al.,2007). The production of 100 kg of biodiesel yields approximately10 kg of impure glycerol, with 55–90% glycerol (Hazimah et al.,2003). Glycerol is generated as a by-product not only when biodie-sel fuels are produced chemically, but also when they are manufac-tured enzymatically (Du et al., 2003; Vicente et al., 2004) and duringthe production of bioethanol (Yazdani and Gonzalez, 2007). The dra-matic growth of the biodiesel industry has created a surplus of glyc-erol that has resulted in a dramatic 10-fold decrease in crudeglycerol prices over the last few years and generated environmentalconcerns associated with contaminated glycerol disposal (Yazdaniand Gonzalez, 2007; Sandun et al., 2007).

At present, due to its properties, pure glycerol has more than2000 different applications (Elvers et al., 1990). However, the pro-

ll rights reserved.

+34 957 218625.

duction of crude glycerol exceeds the present commercial demandfor purified glycerol. Furthermore, the purification of glycerol (witha view to being sold) that is generated during biodiesel manufactur-ing is not a viable option for the biodiesel industry (Chi et al., 2007).Although crude glycerol can be burnt, with the consequent ener-getic advantages, the setting up of biorefineries that co-produceproducts of higher economic value along with biofuels has been pro-posed as a solution for the economic viability of this product. Severalstrategies based on chemical and biological transformations arebeing pursued to convert glycerol into more valuable products (Yaz-dani and Gonzalez, 2007). An example of some of these includes:

– The conversion of glycerol into propylene glycol and acetone,through thermo-chemical processes (Chiu et al., 2006; Dasariet al., 2005).

– The etherification of glycerol with either alcohols (e.g. methanolor ethanol) or alkenes (e.g. isobutene) and the production ofoxygen-containing components, which could have suitableproperties for use in fuel or solvents (Karinen and Krause, 2006).

– The microbial conversion (fermentation) of glycerol to 1,3-pro-panediol, which can be used as a basic ingredient of polyesters(Barbirato et al., 1998; Ito et al., 2005).

– Other products such as butanol (Biebl, 2001), propionic acid(Bories et al., 2004), ethanol and formate (Jarvis et al., 1997),succinic acid (Lee et al., 2001), dihydroxyacetone (Bories et al.,1991; Claret et al., 1994), polyhydroxyalkanoates (Koller et al.,2005), or hydrogen and ethanol (Ito et al., 2005) were alsoobtained using glycerol as a carbon source.

http://dx.doi.org/10.1016/j.biortech.2009.06.017

mailto:[email protected]

http://www.sciencedirect.com/science/journal/09608524

http://www.elsevier.com/locate/biortech

Nomenclature

Alk alkalinity (mg CaCO3/L)CODremoved removed chemical oxygen demand (mg/L)CODsoluble soluble chemical oxygen demand (mg/L)COD STO added chemical oxygen demand (mg/L)CODtotal total chemical oxygen demand (mg/L)G cumulative methane volume (mL)Gm cumulative methane volume at infinite time (mL)GT experimental maximum methane volume (mL)KG specific methane production kinetic constant (L/g VSS h)K 0G apparent kinetic constant (1/h)MS total mineral solids (mg/L)MSS mineral suspended solids (mg/L)OLR organic loading rate (g COD/g VSS d; kg COD/m3 d)

r0 specific rate of methane production (mL CH4/g VSS h)Sremoved removed substrate (g COD)STO added substrate (g COD)t time (h)TS total solids (mg/L)TSS total suspended solids (mg/L)VA volatile acidity (mg acetic/L)VS total volatile solids (mg/L)VSS volatile suspended solids (mg/L)X biomass concentration (mg VSS/L)YCH4=Sremoved

methane yield coefficient (mL CH4/g COD removed)

5610 J.Á. Siles López et al. / Bioresource Technology 100 (2009) 5609–5615

Anaerobic digestion is another possible way to revalorise abun-dant and low-priced glycerol streams. This process may be definedas the biological conversion of organic material to a variety of endproducts including ‘biogas’ whose main constituents are methaneand carbon dioxide (Gujer and Zehnder, 1983; Olthof andOleszkiewick, 1982; Speece, 1983; Wheatley, 1990). The advanta-ges of anaerobic digestion include low levels of biological sludge,low nutrient requirements, high efficiency and the production ofmethane, which can be used as an energy source. The stoichiome-try of the anaerobic digestion of glycerol can be summarised as fol-lows (Christensen and McCarty, 1975; McCarty, 1975):

C3H8O3 þ aNH3 ! bCH4 þ cCO2 þ dC5H7NO2 þ eNH4HCO3

The products of the reaction are methane, carbon dioxide, bio-mass and ammonic bicarbonate, where a, b, c, d and e are stoichi-ometric coefficients. In order to obtain the value of thesecoefficients and determine the methane yield of the anaerobicdigestion of glycerol, a mass balance was made taking into consid-eration a biomass yield of anaerobic bacteria of 0.05 (w/w). Thevalues of a, b, c, d and e were found to be 0.663, 1.648, 0.526,0.041 and 0.622 mol, respectively. In contrast, when an electronbalance was carried out assuming that every electron was usedin methane generation, b reached a value of 1.750. A simple calcu-lation showed that the theoretical methane yield is 94.2%, which isa useful compound due to its caloric power (Lower Caloric Power):35,793 kJ/m3, equivalent to 9.96 kW h/m3. Revalorising glycerol isof special interest as it gives the highest reduced carbon with thecost advantage of anaerobic processes. The aim of this work wasto evaluate the performance and the stability of anaerobic diges-tion process of glycerol-containing waste derived from biodieselmanufacturing. The study was carried out in six batch labora-tory-scale reactors at mesophilic temperature (35 �C).

2. Methods

2.1. Experimental set-up

The experimental set-up used for the anaerobic digestion of thebiodiesel-derived glycerol consisted of six 1-L Pyrex reactors withfour connections to allow for the loading of feedstock, the ventila-tion of the biogas, the injection of inert gas (nitrogen) to maintainthe anaerobic conditions and the removal of effluent. The contentof the reactors was magnetically stirred and temperature wasmaintained by means of a thermostatic jacket containing waterat 37 �C. The volume of methane produced during the processwas measured by using 1-L Boyle–Mariotte reservoirs connectedto each reactor. To remove the CO2 produced during the process,tightly closed bubblers containing a NaOH solution (6 N) were con-

nected between the two elements. The methane volume displacedan equal measurable volume of water from the reservoir.

The reactors were inoculated with methanogenically-activegranular biomass obtained from a full-scale anaerobic reactor usedto treat brewery wastewater from the Heineken S.A. Factory (Jaen,Spain) and non-granular sludge from a full-scale anaerobic reactorused to treat urban wastewater in Jerez de la Frontera (Cadiz,Spain). The granular sludge contained 37,500 mg VSS/L and31,875 mg MSS/L, while the non-granular sludge contained28,400 mg VSS/L and 20,330 MSS/L. The inocula were selected onthe basis of their high methanogenic activity (Field et al., 1988)with values ranging from 0.87 to 0.99 g COD/g VSS d.

2.2. Glycerol

The raw material used as substrate was the glycerol-containingwaste discharged after the biodiesel manufacturing process at theBIDA S.A. Factory in Fuentes de Andalucia (Seville, Spain). In general,this waste contained glycerol, water, methanol, salts and fatty acids.

2.3. Raw material pre-treatment

The substrate was previously treated in two different ways: (a)acidification with phosphoric acid and centrifugation in order to re-cover the catalyst used in the transesterification reaction (KOH) asagricultural fertilizer (potassium phosphates). Additionally, metha-nol and water were removed by vacuum distillation. We call thissubstrate ‘‘acidified glycerol”. (b) Acidification followed by distilla-tion (rectification). After the same acidification process, a rectifica-tion at laboratory-scale (135–140 �C; 1.6–2.0 � 10�3 atm) wascarried out. Subsequently, the organic impurities in the distillatewere removed by liquid/liquid extraction with hexane, which waseliminated by vacuum distillation. We call this substrate ‘‘distilledglycerol”. Table 1 shows the characteristics and features of acidified,distilled and commercially available pure glycerol (Elvers et al.,1990). Due to its high COD, the acidified and distilled glycerol werediluted using distilled water until reaching 81.6 and 85.7 g COD/L,respectively, and neutralized by adding sodium hydroxide. Finally,several nutrients and alkalinity (NaHCO3) were added to the dis-tilled glycerol to provide the necessary nutrients for the appropriatemetabolism of the anaerobic microorganisms (DiStefano andAmbulkar, 2006). Table 2 shows the composition of the nutrientand trace element solutions added to the distilled glycerol.

2.4. Anaerobic digesters. Experimental procedure

The anaerobic reactors were initially loaded with 12 g VSS ofgranular sludge as inoculum, and the anaerobic digestion of

Table 1Composition and features of the acidified, distilled and pure glycerol.

Parameter Acidifiedglycerol

Distilledglycerol

Pureglycerol

Density at 20 �C (g/mL) 1.044 1.260 1.261Refraction index at 20 �C 1.4440 1.4728 1.4746COD (g/kg) 1010 1155 1217Dynamic viscosity at 50 �C (mPa s) 57 150 152Colour Brown Colourless Colourless

Table 2Composition of the nutrient and trace element solutions added to distilled glycerol.

Nutrient solution Trace element solution

Compound g/La Compound g/L

NH4CI 0.200 MnCl2�4H2O 0.100K2HPO4�3H2O 0.100 CoCl2�6H2O 0.170KH2PO4 0.055 ZnCl2 0.100MgCl2�6H2O 0.200 CaCl2 0.200Resazurine 0.001 H3BO4 0.019FeCl2�4H2O 0.100 NiCl2�6H2O 0.050NaS�9H2O 0.500 Na2MoO4�H2O 0.100NaHCO3 5.000 Ad 10 mL of trace element solution per litre of diluted

glycerol

a Concentration of each compound per litre of diluted glycerol.

Table 3pH and volatile acidity/alkalinity ratio values in the effluents of the reactors obtainedfor the different substrates and sludge studied.

Substrate and sludge type pH VA (eq acetic)/Alk(eq CaCO3)

Acidified glycerol and granular sludge 7.72 ± 0.24 0.05 ± 0.02Acidified glycerol and non-granular sludge 7.74 ± 0.30 0.03 ± 0.01Distilled glycerol and granular sludge 7.81 ± 0.17 0.05 ± 0.01

J.Á. Siles López et al. / Bioresource Technology 100 (2009) 5609–5615 5611

acidified and distilled glycerol was studied. Another set of reactorswere then loaded with non-granular sludge and the anaerobicdigestion of acidified glycerol was studied. In all cases, the nutrientand trace element solutions described by Fannin (1987) and Fieldet al. (1988) were added when the sludge was loaded. Bothsolutions are very important for activating bacterial growth andmetabolism at the beginning of the process.

In order to activate the biomass prior to the experiments, thereactors were first fed with a synthetic solution composed of glu-cose, sodium acetate and lactic acid (GAL solution) at concentra-tions of 50 g/L, 25 g/L and 20.8 mL/L, respectively. During thisinitial period, the organic load added to the reactors was graduallyincreased from 0.25 to 1.00 g COD over a 16-day period. After thisprevious stage, biomass acclimatization was carried out. The reac-tors were fed with 1 g COD, in which the percentage of glycerolused in the COD was increased from 25% to 100% after four loads.During this acclimatization period, the volume of methane wasmeasured as a function of time. The maximum duration of each as-say was 48 h; the time interval required for the complete biometh-anization of each load. Once this preliminary acclimatization stepwas finished, a series of batch experiments were carried out usingboth sludge types in addition to the acidified and distilled glycerolas substrates. During each set of experiments, the organic loadadded to the reactors was gradually increased from 1.0 to 1.5and 2.0 g COD with distilled glycerol and from 1.0 to 1.5, 2.0 and3.0 g COD with acidified glycerol. In all cases, the volume of meth-ane was measured as a function of time and samples were takenand analysed before and after feeding. The duration of each exper-iment was equal to the time interval required for maximum gasproduction and COD removal. Each glycerol solution load was car-ried out at least in duplicate and the results expressed as means.

2.5. Chemical analyses

The following parameters were determined in the effluents ofeach load: pH, COD total, COD soluble, TS, MS, VS, TSS, MSS, VSS,volatile acidity (VA) and alkalinity (Alk). All analyses were carriedout in accordance with the Standard Methods of the APHA (APHA,1989).

2.6. Software

Sigma-Plot software (version 9.0) was used to create graphs,perform the statistical analysis and fit the experimental data pre-sented in this work.

3. Results and discussion

3.1. Stability

The stability of the process is evaluated based on the evolutionof the pH, alkalinity, volatile acidity and volatile acidity/alkalinityratio (VA/Alk) during the anaerobic digestion process of the differ-ent substrates. Table 3 shows the mean value and standard devia-tion of the pH and volatile acidity/alkalinity ratio in the effluents ofthe reactors with the different substrates and sludge types. The pHwas approximately constant across experiments, with a meanvalue of 7.72 ± 0.24 using acidified glycerol-granular sludge,7.74 ± 0.30 using acidified glycerol-non-granular sludge and7.81 ± 0.17 with distilled glycerol-granular sludge. These values re-mained within the optimal range for methanogenic bacteria(Fannin, 1987; Wheatley, 1990). On the other hand, the volatileacidity/alkalinity ratio was always found to be lower than 0.30–0.40, thus indicating that the process operated favourably withoutthe risk of acidification (Balaguer et al., 1992).

3.2. Methane yield coefficient and biodegradability of glycerol

Methane yield coefficient was determined from the experimen-tal maximum methane volume produced (GT) and the final and ini-tial COD, which were known in all loads. By fitting (GT, CODremoved) value pairs to a straight line (Fig. 1), the methane yieldcoefficient coincides with the slope of the regression line andwas found to be 292 mL CH4/g COD removed (at 1 atm, 25 �C) usinggranular sludge-acidified glycerol, 288 mL CH4/g COD removedusing non-granular sludge-acidified glycerol and 356 mL CH4/gCOD removed using granular sludge-distilled glycerol. Similar val-ues have been described in the literature. Gross and Lanting (1988),for example, studied the anaerobic digestion of wastewater gener-ated during the manufacturing of fuel ethanol from corn, whichcontained a highly soluble mixture of weak organic acids such asacetic, propionic, lactic and butyric acid. The organic strength ofthe wastewater typically ranged from 3000 to 5000 mg/L COD sol-uble. The methane yield coefficient was found to be 330 mL CH4/gCOD removed.

According to Wheatley (1990), and considering the biomassgrowth and cell maintenance null, 382 mL of methane are theoret-ically produced (at 1 atm, 25 �C) per gram of COD removed. Exper-imentally, the effectiveness of the process in each case was: 76%using granular sludge-acidified glycerol, 75% using non-granularsludge-acidified glycerol and 93% with granular sludge-distilledglycerol. When only taking into account these results, granularsludge-distilled glycerol would be the best option. Studying the re-moved COD percentage is as important as evaluating the methaneproduction coefficient in order to determine waste biodegradabil-ity. The high biodegradability of glycerol can be demonstrated by


plotting the amount of substrate removed against the substrateadded for each experiment. Fig. 2 shows these data in such a man-ner that the slope of the straight line denotes the percentage bio-degradability of the waste, which was found to be around 100%using granular sludge-acidified glycerol, 75% using non-granularsludge-acidified glycerol and 85% using granular sludge-distilledglycerol. These values remained constant throughout the loads.The biodegradability obtained with glycerol was high and similarto bioethanol wastewater, which achieved 86% biodegradability(Gross and Lanting, 1988).

In all the cases, the reactors contained soluble and non-biode-gradable COD before adding the substrate. In the same figures,the initial non-biodegradable COD value was found to be 0.502 gCOD when granular sludge-acidified glycerol was used, 0.093 g

Granular sludge-Acidified glycerol

GT (

mL

CH

4 )

200

400

600

800

1000

YCH4/S removed = 292 mL CH4 / g COD removed

Non-granular sludge-Acidified glycerol

GT (

mL

CH

4 )

0

200

400

600

800

1000


Granular sludge-Distilled glycerol

S removed (g COD)

0 1 2 3 4

GT (

mL

CH

4 )

0

200

400

600

800

1000


Fig. 1. Variation of the experimental maximum methane volume produced (GT) (at1 atm, 25 �C) with the COD consumed to obtain the methane yield coefficient of theprocess using granular sludge-acidified glycerol, non-granular sludge-acidifiedglycerol and granular sludge-distilled glycerol.

COD when non-granular sludge-acidified glycerol was added and0.665 g COD for granular sludge-distilled glycerol.

Taking into account both the methane production percentageand the waste biodegradability, the granular sludge-acidified glyc-erol would be the best choice. All the substrate concentrationwould be removed, thus producing methane and biomass growthand obtaining 100% biodegradability and 76% effectiveness in gen-erating methane. This fact would simultaneously ensure cell main-tenance and non-waste accumulation in the reactors.

3.3. Kinetics of methane production

Fig. 3 shows the variation in the volume of methane accumu-lated (G) as a function of time for the loads with the different sub-strates and sludge types. The variable concerned is the amount of

S re

mov

ed (

g C

OD

)

m = 1.01 g COD removed / g COD STO

r2 = 0.9946

S re

mov

ed (

g C

OD

)


r2 = 0.9542

STO (g COD)

0 1 2 3 4 5

S re

mov

ed (

g C

OD

)


r2 = 0.8925

x = 0.665 g COD

0.0

0.5

1.0

1.5

2.0

2.5


3.0

3.5

x = 0.093 g COD

0.5

1.0

1.5

2.0

2.5

3.0

3.5


3.5

2.0

1.5

1.0

2.5

3.0

0.5

x = 0.502 g COD


Fig. 2. Plot of the amount of substrate removed against the substrate added for allthe experiments to obtain the biodegradability percentage of the by-product usinggranular sludge-acidified glycerol, non-granular sludge-acidified glycerol andgranular sludge-distilled glycerol.

G (

mL

CH

4 )

0

200

400

600

800

1000

1200

1400

1.0 g COD 2.0 g COD 3.0 g COD

G (

mL

CH

4 )

0

200

400

600

800

1000

1200

1400

1.0 g COD1.5 g COD 2.0 g COD3.0 g COD



Time (h)0 10 20 30 40 50 60

G (

mL

CH

4 )

0

200

400

600

800

1000

1200

1400


1.0 g COD1.5 g COD 2.0 g COD

Fig. 3. Variation of the methane volume accumulated (G) as a function of time forthe loads of 1.0, 1.5, 2.0 and 3.0 g COD using granular sludge-acidified glycerol, non-granular sludge-acidified glycerol and granular sludge-distilled glycerol.


substrate (g COD) added to the reactors. The results show that alarger volume of methane was produced when the substrate loadincreased and the time to complete the biodegradable fraction re-moval was 40 h for both sludge types.

In order to characterize each set of experiments kinetically, andthus facilitate comparisons, the methane production first order ki-netic model described by Borja et al. (1995) was used to fit theexperimental data. According to this model, the volume of meth-ane accumulated (G) (mL, at 1 atm, 25 �C) at a given time t (h) fitsthe following equation:

G ¼ Gm½1� expð�K 0G � tÞ� ð1Þ

where Gm is the maximum methane volume accumulated at an infi-nite digestion time; G is zero at t = 0 and the rate of gas productionbecomes zero at t equal to infinite.

K 0G is an apparent kinetic constant for methane production (h�1)which included the biomass concentration:

K 0G ¼ KG � X ð2Þ

where KG is the specific methane production kinetic constant (L/gVSS h) and X is the biomass concentration (g VSS/L).

As it was previously mentioned, a larger amount of gas was pro-duced as the load increased. Moreover, the slopes of the curves alsodecreased with increasing time. This decrease in the slopes withtime can be explained by the gradual decrease in the concentrationof biodegradable substrate. Eq. (1) shows a good fit with the exper-imental data. Thus, it is correct to apply the proposed kinetic modelto all the loads studied. The values of K 0G and Gm for each load werecalculated numerically from the experimental data obtained bynon-linear regression using Sigma–Plot (version 9.0) (Borja et al.,1995). Table 4 shows the Gm and K 0G values and their standard devi-ations obtained in the experiments. The standard deviations of Gm

and K 0G values were lower than 5% and 15%, respectively, for all theloads studied. This suggests that the proposed model adequatelyfits the experimental data.

Once the biomass concentration values were determined, thevalues of the specific methane production kinetic constant, KG,were calculated using Eq. (2). Table 4 also shows these KG values.The specific constant, KG, decreased with substrate concentrationfor all the experiments, showing an inhibition process. Specifically,in the case of acidified glycerol, the KG value decreased by 42%when loading 1.0–3.0 g COD using granular sludge and by 58%using non-granular sludge. Surprisingly, a 76% decrease was ob-tained when loading between 1.0 and 2.0 g COD of distilled glyc-erol. This marked decrease in the KG value when using distilledglycerol could be explained by the accumulation in the reactorsof the alkalinity that was added as a nutrient. In contrast, the inhi-bition observed when using acidified glycerol could be the result oftwo causes: the increasing load or the accumulation of phospho-rous due to pre-treatment or even both effects simultaneously. Inorder to avoid the phosphorous concentration effect, CaO wasadded to the reactors during the experiments until a final concen-

Table 4Gm, K 0G and KG values obtained for all the loads studied using acidified glycerol-granular s

Substrate and sludge type Load (g COD) Gm (mL CH4)

Acidified glycerol and granular sludge 1.0 323 ± 52.0 592 ± 53.0 857 ± 7

Acidified glycerol and non-granular sludge 1.0 318 ± 111.5 451 ± 82.0 590 ± 153.0 789 ± 20

Distilled glycerol and granular sludge 1.0 375 ± 61.5 578 ± 52.0 766 ± 15

tration of 25 ± 4 mg PO3�4 =L was obtained. The occurrence of a sim-

ilar inhibitory phenomenon has been described in the literature(Ito et al., 2005; Barbirato et al., 1998).

ludge, acidified glycerol-non-granular sludge and distilled glycerol-granular sludge.

K 0G (h�1) R2 KG (Lg�1 h�1)

0.2075 ± 0.0117 0.9676 0.03840.1858 ± 0.0053 0.9909 0.03380.1316 ± 0.0036 0.9798 0.0246

0.1655 ± 0.0177 0.9527 0.02030.1175 ± 0.0151 0.9960 0.01780.1216 ± 0.0091 0.9698 0.01900.0749 ± 0.0059 0.9716 0.0128

0.2006 ± 0.0105 0.9548 0.01310.1372 ± 0.0048 0.9784 0.01010.0829 ± 0.0046 0.9697 0.0068

Gm(mL CH4 )

0 200 400 600 800 1000

GT e

xper

imen

tal (

mL

CH

4)

0

200

400

600

800

1000

m = 0.9942Confidence intervals of 95%

Fig. 4. Comparison between the experimental maximum methane production (GT)values and the theoretical values (Gm) predicted by Eq. (1).


Moreover, the theoretical values of maximum methane produc-tion (Gm) were calculated using Eq. (1) and plotted against theircorresponding experimental values (Fig. 4). The deviations ob-tained were lower than 5% in practically all cases, again suggestingthat the proposed model can be used to accurately predict thebehaviour of the process and that the kinetic parameters obtainedrepresented the microorganism behaviour affecting the anaerobicdigestion of this by-product.

In order to compare the different substrates and sludge types,the specific methane production rate, r0 (mL CH4/g VSS h), was de-

1.0 g COD

r' (

mL

CH

4 /g

VSS

·h)

0

2

4

6

8

10

12

14

Granular sludge-Acidified glycerolNon-granular sludge-Acidified glycerolGranular sludge-Distilled glycerol

2.0 g COD

Time (h)

r' (

mL

CH

4 /g

VSS

·h)

0

5

10

15

20

25

0 10 20 30 40 50 60

Fig. 5. Variation of the specific methane production rate, r0 , as a function of time for all t

fined as the volume of methane generated per gram of volatile sus-pended solid and per hour for each set of experiments. As can beobserved in Fig. 5, the r0 values were always higher when usinggranular sludge and acidified glycerol at the same time. The lowestvalues were reached when using distilled glycerol with a high alka-linity concentration.

3.4. Organic loading rate (OLR)

The organic loading rate (OLR) for each substrate and sludgetype was determined using the experimental data. The OLR (kgCOD/m3 d) values were not excessively high for any set of experi-ments (0.92–2.00 kg COD/m3 d) when compared to other biode-gradable substrates. For food, organic chemical, soft drink andbakery wastes, which range from 5000 to 11,000 mg/L in rawCOD, removals of 68–93% were obtained at organic loadings of3–27 kg COD/m3 d (Jeris, 1983).

The OLR values, expressed as g COD/g VSS d, take into accountboth the biomass concentration and the organic load, thus makingthis an interesting variable for comparing the results. The highestvalues were obtained using granular sludge-acidified glycerol:0.21–0.38 g COD/g VSS d. On the contrary, when non-granularsludge-acidified glycerol was used, the ORL reached a value rang-ing from 0.12 to 0.26 g COD/g VSS d and 0.075 to 0.080 g COD/gVSS�d when granular sludge-distilled glycerol was used. Fig. 6shows the evolution of the specific methane production kineticconstant, KG (L/g VSS�h), with the organic loading rate (g COD/gVSS�d) for each substrate and sludge type studied. A strong de-crease in the KG value was observed when the distilled glycerolOLR was increased. Although the constant also decreased whenusing acidified glycerol, the kinetic constant slowdown was much

1.5 g COD

3.0 g COD

Time (h)

70 0 10 20 30 40 50 60 70

he loads using granular and non-granular sludge and acidified and distilled glycerol.

OLR (g COD/g VSS·d)

KG (

L/g

VSS

·h)

Granular sludge-Acidified glycerolNon-granular sludge-Acidified glycerolGranular sludge-Distilled glycerol

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.0 0.1 0.2 0.3 0.4 0.5

Fig. 6. Variation of the specific kinetic constant KG with the Organic Loading Rate(OLR), using granular and non-granular sludge and acidified and distilled glycerol.


lower. Additionally, this substrate has allowed the reactors to oper-ate in a stable way with higher OLR. Finally, according to economiccriteria, acidified glycerol is the target substrate. Given that thegranular sludge KG shows values that are twice as high as thenon-granular sludge, the glycerol anaerobic treatment could becarried out using granular sludge in addition to acidified glycerol.

However, one option to improve the OLR could be to mix glyc-erol with another organic waste such as wastewater generatedduring the biodiesel manufacturing process and study its co-diges-tion. This research would form part of what is known as the Bior-efinery concept; a relatively new term referring to the conversionof feedstock into a host of valuable chemicals and energy. Indeed,this is considered an important strategy for achieving sustainabledevelopment.

4. Conclusions

The results obtained through this research study reveal thatglycerol-containing waste derived from biodiesel production afterpre-treatment has a high level of anaerobic biodegradability andthat a substantial quantity of methane can be obtained this way.The highest methane yield coefficient was obtained using granularsludge-distilled glycerol, but a strong inhibition was observedwhen the load was increased. The use of granular sludge and acid-ified glycerol was found to be the best option for revalorising glyc-erol anaerobically given that biodegradability was found to bearound 100% and the methane yield coefficient was 0.306 m3

CH4/kg acidified glycerol (at 1 atm, 25 �C).

Acknowledgements

The authors are very grateful to the BIDA S.A. Factory (Fuentesde Andalucía, Seville, Spain) and to the Spanish Ministry of Scienceand Innovation for funding José Ángel Siles López through ProjectCTM2005-01293 and Grant BES2006-14074. This study was co-fi-nanced by the European Social Fund. We also wish to expressour gratitude to the laboratory technician Inmaculada BellidoPadillo for her help.

References

APHA (American Public Health Association), 1989. Standard Methods for theExamination of Water and Wastewater, 17th ed. Washington, DC, USA.

Balaguer, M.D., Vicent, M.T., Paris, J.M., 1992. Anaerobic fluidized bed reactor withsepiolite as support for anaerobic treatment of vinasses. Biotechnol. Lett. 14,433–438.

Barbirato, F., Himmi, E.H., Conte, T., Bories, A., 1998. 1,3-Propanediol production byfermentation: an interesting way to valorise glycerine from the ester andethanol industries. Ind. Crops Prod. 7, 281–289.

Biebl, H., 2001. Fermentation of glycerol by Clostridium pasteurianum – batch andcontinuous culture studies. J. Ind. Microbiol. Biotechnol. 27, 18–26.

Bories, A., Claret, C., Soucaille, P., 1991. Kinetic study and optimisation of theproduction of dihydroxyacetone from glycerol using Gluconobacter oxydans.Process Biochem. 26, 243–248.

Bories, A., Himmi, E., Jauregui, J.J.A., Pelayo-Ortiz, C., Gonzales, V.A., 2004. Glycerolfermentation with Propionibacteria and optimisation of the production ofpropionic acid. Sci. Aliments. 24, 121–135.

Borja, R., Martín, A., Alonso, V., García, C.J., Banks, C.J., 1995. Influence of differentpretreatments on the kinetics of anaerobic digestion of olive mill wastewater.Water Res. 29, 489–495.

Chi, Z., Pyle, D., Wen, Z., Frear, C., Chen, S., 2007. A laboratory study of producingdocosahexaenoic acid from biodiesel-waste glycerol by microalgalfermentation. Process Biochem. 42, 1537–1545.

Chiu, C.W., Dasari, M.A., Sutterlin, W.R., Suppes, G.J., 2006. Removal of residualcatalyst from simulated biodiesel’s crude glycerol for glycerol hydrogenolysis topropylene glycol. Ind. Eng. Chem. Res. 45, 791–795.

Claret, C., Salmon, J.M., Romieu, C., Bories, A., 1994. Physicology of Gluconobacteroxydans during dihydroxyacetone production from glycerol. Appl. Microbiol.Biotechnol. 41, 359–365.

Christensen, D.R., McCarty, P.L., 1975. Multi-process biological treatment model. J.Water Pollut. Contr. Fedn. 47, 2652–2664.

Dasari, M.A., Kiatsimkul, P.P., Sutterlin, W.R., Suppes, G.J., 2005. Low-pressurehydrogenolysis of glycerol to propylene glycol. Appl. Catal. A-Gen. 281, 225–231.

DiStefano, T.D., Ambulkar, A., 2006. Methane production and solids destruction inan anaerobic solid waste reactor due to post-reactor caustic and heat treatment.Water Sci. Technol. 53 (8), 33–41.

Du, W., Xu, Y., Liu, D., 2003. Lipase-catalysed transesterification of soya bean oil forbiodiesel production during continuous batch operation. Biotecnol. Appl.Biochem. 38, 103–106.

Elvers, B., Hawkins, S., Weinheim, G.S., 1990. Ullmann’s Encyclopaedia of IndustrialChemistry, fifth ed. VCH, NY, Basel (Switzerland), Cambridge, New York.

Fannin, K.F., 1987. Start-up, operation, stability and control. In: Chynoweth, D.P.,Isaacson, R. (Eds.), Anaerobic Digestion of Biomass. Elsevier, London, pp. 171–196.

Field, J., Sierra, R., Lettinga, G., 1988. Ensayos anaerobios. In: Fdz-Polanco, F., García,P.A., Hernando, S. (Eds.), 4� Seminario de Depuración Anaerobia de AguasResiduales. Valladolid, Spain, Secretariado de Publicaciones, Universidad deValladolid, pp. 52–82.

Gross, L.R., Lanting, J., 1988. Anaerobic wastewater treatment of a fuel ethanolfacility. In: Torpy, M.F. (Ed.), Anaerobic Treatment of Industrial Wastewaters.Pollution Technology Review, vol. 152 . Noyes Data Corporation, Park Ridge,New Jersey, USA, pp. 23–34.

Gujer, W., Zehnder, A.J., 1983. Conversion process in anaerobic digestion. Water Sci.Technol. 15, 123–167.

Hansen, A.C., Zhang, Q., Lyne, P.W.L., 2005. Ethanol-diesel fuel blends – a review.Bioresour. Technol. 96, 277–285.

Hazimah, A.H., Ooi, T.L., Salmiah, A., 2003. Recovery of glycerol and diglycerol fromglycerol pitch. J. Oil Palm. Res. 15, 1–5.

Ito, T., Nakashimada, Y., Senba, K., Matsui, T., Nishio, N., 2005. Hydrogen and Ethanolproduction from glycerol-containing wastes discharged after biodieselmanufacturing process. J. Biosci. Bioeng. 100 (3), 260–265.

Jarvis, G.N., Moore, E.R.B., Thiele, J.H., 1997. Formate and ethanol are the majorproducts of glycerol fermentation produced by a Klebsiella planticola strainisolated from red deer. J. Appl. Microbiol. 83, 166–174.

Jeris, J.S., 1983. Industrial wastewater treatment using anaerobic fluidized bedreactors. Water Sci. Technol. 15, 169–176.

Karinen, R.S., Krause, A.O.I., 2006. New biocomponents from glycerol. Appl. Catal. A-Gen. 306, 128–133.

Koller, M., Bona, R., Braunegg, G., Hermann, C., Horvat, P., Kroutil, M., Martinez, J.,Neto, J., Pereira, L., Varila, P., 2005. Production of polyhydroxyalkanoates fromagricultural waste and surplus materials. Biomacromol. 6 (2), 561–565.

Lee, P.C., Lee, W.G., Lee, S.Y., Chang, H.N., 2001. Succinic acid production withreduced by-product formation in the fermentation of Anaerobiospirillumsucciniciproducens using glycerol as a carbon source. Biotechnol. Bioeng. 72,41–48.

McCarty, P.L., 1975. Stoichiometry of biological reactions. Prog. Water Technol. 7,157–172.

Olthof, M., Oleszkiewick, J., 1982. Anaerobic treatment of industrial wastewater.Chem. Eng. 15, 1321–1326.

Sandun, F., Sushil, A., Kiran, K., Ranjitha, B., 2007. Glycerol based automotive fuelsfrom future biorefineries. Fuel., doi: 10.1016/j.fuel.2007.03.030.

Speece, R.E., 1983. Anaerobic biotechnology for industrial wastewater treatment.Environ. Sci. Technol. 17, 416–427.

Vicente, G., Martinez, M., Aracil, J., 2004. Integrated biodiesel production: acomparison of different homogeneous catalyst systems. Bioresour. Technol.92, 297–305.

Wheatley, A., 1990. Anaerobic Digestion: A Waste Treatment Technology. Elsevier,London.

Yazdani, S.S., Gonzalez, R., 2007. Anaerobic fermentation of glycerol: a path toeconomic viability for the biofuels industry. Curr. Opin. Biotech. 18, 213–219.

http://dx.doi.org/10.1016/j.fuel.2007.03.030

anaerobic digestion of glycerol derived from biodiesel manufacturing

Documents