Laboratory scale examination of the effects of overloading on the anaerobic digestion by glycerol

Download Laboratory scale examination of the effects of overloading on the anaerobic digestion by glycerol

Post on 26-Jun-2016




4 download

Embed Size (px)


<ul><li><p>ts</p><p>en, Ba</p><p>on,</p><p>Anaerobic digestion</p><p>purrydemasu</p><p>erol + 50% acetic acid, and 50% glycerol + 50% thick stillage, (presented in % of 2.60 g COD l1d1 OLR),</p><p>of reneergy iand O</p><p>The inhibitory effect of propionic acid accumulation on metha-nogenesis has been discussed in many publications (Chen et al.,2008; Demirel and Yenigun, 2002; Wang et al., 2009). Tolerablelevels of propionic acid fall in the range of 10006000 mg l1</p><p>(Gallert and Winter, 2008; Ma et al., 2009). The slowly degradable</p><p>tion process in case of drastic OLR raise and co-substrate treatmentin the recovery period. As a co-substrate, feeding acetic acid or ace-tic acid-producing agent (thick stillage) (Kim et al., 2010) couldtherefore have a benecial effect on the conditions of methane pro-duction. It is possible, however, that the positive effect of aceticacid may result not from the improvement of the acetic acid to pro-pionic acid ratio but from another mechanism. Thick stillage is aby-product of the bio-ethanol production and a cost-efcient solu-tion for the industry.</p><p> Corresponding author. Tel.: +36 99 518 176; fax: +36 99 518 249.</p><p>Bioresource Technology 102 (2011) 52705275</p><p>Contents lists availab</p><p>T</p><p>elsE-mail address: (T. Rtfalvi).2007). Biofuels production seems to be a good alternative for coun-tries with high agricultural potential (Gupta et al., 2010; Sordaet al., 2010). The main by-products of biodiesel production are im-pure glycerol and soap water. The rise of the biodiesel industry re-sulted in a surplus on the glycerol market and the price of crudeglycerol decreased (Yazdani and Gonzalez, 2007). Local use of thisby-product for energy production could be a good solution for bio-diesel factories. Glycerol phase (g-phase) is commonly used forbiogas production as a co-substrate but not as a main substrate(Fountoulakis and Manios, 2009; Siles et al., 2010). The anaerobicdigestion of g-phase as a main substrate is problematic becauseof its high potassium content and the low acetic acid to propionicacid ratio of the fermentation sludge (Silez and Martn, 2009).</p><p>increases (Chua et al., 1997). Because acidogenesis is less sensitiveto fermentation conditions, rapid acid accumulation can still ad-versely affect base-inhibited methanogenesis (Rincn et al., 2008;Schoen et al., 2009; Siegert and Banks, 2005). The negative effectis intensied when the ratio of acetic acid to propionic acid dropsbelow the optimal level (Li et al., 2008; Mechichi and Sayadi, 2005).The presence of a high VFA content relating to diminution of thebiogas yield requires an immediately intervention in the operationof the biogas plant. The most common treatment is the drastic de-crease of the OLR (Gallert and Winter, 2008). At the same time,from an economical point of view, the reduction of the recoverystage has a primal importance.</p><p>We present the response of a glycerol fed anaerobic fermenta-GlycerolOverloadVolatile fatty acid</p><p>1. Introduction</p><p>The production and utilisationcould be a solution to increase the enwith few fossil fuel resources (Chum0960-8524/$ - see front matter 2011 Elsevier Ltd. Adoi:10.1016/j.biortech.2011.02.020respectively. The application of co-substrates reduced the recovery period by 5 days compared to feedingwith pure glycerol. When the reactors were loaded with glycerol again (10% OLR raise per day) the pre-viously applied co-substrates had a positive effect on the VFA composition and the biogas yield as well.</p><p> 2011 Elsevier Ltd. All rights reserved.</p><p>wable energy sourcesndependence of regionsverend, 2001; Sovacool,</p><p>iso-fatty acids (isobutyric acid, isovaleric acid), which are producedduring the fermentation process, also have an inhibitory effect onmethane production (Aguilar et al., 1995, Chen et al., 2008).</p><p>A methanogenic process that is inhibited will respond sensi-tively to changes in the organic loading rate (OLR), particularly toKeyword:</p><p>organic loading rate (OLR) of 3.010.5 g COD l d , the concentration of propionic acid increased to62008000 mg l1. Then the inoculum was divided into three parts feeding with 100% glycerol, 50% glyc-Case Study</p><p>Laboratory scale examination of the effecdigestion by glycerol</p><p>Tams Rtfalvi a,, Annamria Tukacs-Hjos b, Leventa Institute of Chemistry, Faculty for Forestry, University of West Hungary, H-9400 SoprobGzInnov Ltd, H-9400 Sopron, Asszonyvsr d}ul}o 31., HungarycCooperation Research Centre Non-prot Ltd., University of West Hungary, H-9400 Sopr</p><p>a r t i c l e i n f o</p><p>Article history:Received 26 November 2010Received in revised form 2 February 2011Accepted 4 February 2011Available online 1 March 2011</p><p>a b s t r a c t</p><p>The anaerobic digestion of0.2, was studied in laborato3000 mg chemical oxygenand daily biogas yield me</p><p>Bioresource</p><p>journal homepage: www.ll rights reserved.of overloading on the anaerobic</p><p>Albert a, Bla Marosvlgyi c</p><p>jcsy-Zs. u.4., Hungary</p><p>Bajcsy-Zs. u. 4., Hungary</p><p>e glycerol, which produces a baseline acetic acid to propionic acid ratio ofscale reactors (3 l working volume) at mesophilic temperature (37 C) withand (COD) l1d1. During the experiment tVFA and C2-C6 VFA analysis</p><p>rement were carried out. Following 10 days of a 15% d1 increase in the1 1</p><p>le at ScienceDirect</p><p>echnology</p><p>evier .com/locate /bior tech</p></li><li><p>Co., USA) (Wagner et al., 2010). The mobile phase was 0.005 M</p><p>(mg mol1) and V is the volume of the sample (ml, at 1 atm,</p><p>echnH2SO4 applied at a 600 ll min1 ow rate and 60 C column tem-perature. The injected analyte volume was 20 ll. Quantitativeanalysis for C2-C6 VFAs (SigmaAldrich Co.) was carried out byWe show that the previously applied co-substrate treatmenthas an inuence on the biogas yield and acid content (tVFA, VFAcomposition) during reload period.</p><p>2. Methods</p><p>2.1. Reactors and anaerobic digestion of glycerol</p><p>During the overloading period, the experimentwas conducted in3 l working volume anaerobic reactor (5000 ml capacity volumethreaded brown bottle, Merck &amp; Co., Germany). Following the max-imum chemical oxygen demand (COD) load, the sludge was dividedinto three parts and drawn through a plastic pipe to three 1 l work-ing volume bottles (2500 ml capacity volume threaded brown bot-tle, Merck &amp; Co., Germany) for the recovery stage. Each of thebottles was equipped with two glass connections, one for loadingof rawmaterials and the other for a gasmeasurement kit. The head-spaces of the digesters were ushed with nitrogen for 4 min afterclosing the screwcaps. The reactorswereoperatedwithoutmechan-ical mixing. The content of the reactors were sporadically (threetimes per day) manually stirred and sedimentation was not ob-served at all. Each digester wasmaintained in awater bath (Memm-ert WNB 14 Basic, Memmert GmbH. &amp; Co.) at constant temperature(37 C). The volume of the biogas produced each day was measuredwith micro gas measurement equipment (Euro Open Ltd., Hungary)connected to each reactor. An oil displacement method was used tomeasure the biogas volume. The biogas yieldwas recalculated to thestandard condition for temperature and pressure.</p><p>2.2. Inoculum and substrates</p><p>Anaerobic digester sludge was obtained from a biogas plant (Su-gar factory, Kaposvr, Hungary). The sludge was adapted to glyc-erol over a six-month period before the beginning of theexperiment. The pure glycerol contained 83% total solids (TS), theCOD value was 1628 g l1, and the pH value was 5.6. The thick stil-lage originated from a bio-ethanol plant (Enviral A.S., Leopoldov,Slovakia) and contained 35.8% TS, 8.1 g l1 N, and 6.9 g l1 P, witha COD value of 604 g l1 and a pH value of 3.6. Acetic acid (usedas a 0.2 M solution, 147 g l1 COD) and urea were obtained fromSigmaAldrich Co. A trace elements supplement solution consist-ing of 1625 mg zinc, 10875 mg manganese, 93 mg boron, 163 mgcopper, 20000 mg cobalt, 138 mg molybdenum and 113 mg sele-nium in organic complex form per litre of solution (42.2% TS),was also used.</p><p>2.3. Analysis methods</p><p>Each day of the experiment, 15 ml samples of inoculum werecollected and centrifuged for 10 min at 3420 RCF (EBA 21, A. Het-tich Co., Germany). The supernatant was divided into two parts.Five (5.0) millilitres were used for determination of the tVFA level,and the rest was centrifuged at 18,111 RCF for 10 min and then l-tered through a 0.2 lm nylon membrane (Pall Co.).</p><p>Volatile fatty acid (VFA) levels were analysed by HPLC. Theinstrument consisted of a Gynkotek M 480 pump, a TOSOH 6040UV detector (210 nm), a Rheodyne 8125 injector with a 20 ll loop,and an Aminex HPX-87H column (300 7.8 mm; 5 lm) (BioRad</p><p>T. Rtfalvi et al. / Bioresource T5-point calibration.The titration method described below is commonly used for</p><p>tVFA determination in Hungarian biogas plants. The sample wassample</p><p>25 C).COD determination was carried out according to the Hungarian</p><p>standard protocol (MSZ ISO 6060). The method based on the oxida-tion of the oxidizable organic matter by an excess amount of potas-sium dichromate solution at the presence of HgSO4 and Ag catalyst.The excess potassium dichromate is titrated with ferrous ammo-nium sulphate. The COD value is calculated by the reduced amountof the Cr3+.</p><p>2.4. Experimental procedure</p><p>The glycerol-adapted sludge was fed with glycerol at a stableOLR of 2.6 g COD l1d1 for 30 days with a 0.2 ratio of acetic acidto propionic acid. Silez and Martn (2009) were able to run stableglycerol fed fermentation with a similar OLR (2.6 g COD l1d1).Beginning with the rst day of the experimental period, the OLRwas increased by 15% d1. When an OLR of 10.5 g COD l1d1 hadbeen reached, the sludge was divided into three equal volumesfor the recovery stage of the experiment. During the recovery stage,each portion of the sludge was treated with a different substratemixture at the original OLR of 2.6 g COD l1d1. The substrate mix-tures (presented in % of the original OLR) were 100% glycerol (A di-gester), 50% glycerol + 50% acetic acid (B digester), and 50% glycerol+ 50% thick stillage (C digester). The recovery stage lasted untilmeasured tVFA values reached 1000 mg l1. In the following stageeach reactor was fed with 100% glycerol at the same OLR as in theprevious stage for 7 days and 2 days in case of B, C and A reactors,respectively. After this short stable phase the OLR was raised againby 10% d1 until the biogas yield exceeded 2.0 l l1 sludge d1. Thisoccurred at an OLR of 4.68 g COD l1d1, after which the OLR wasmaintained at that value for the rest of the experiment (Fig. 1).</p><p>3. Results and discussion</p><p>As an aid to practical interpretation, the results are separated byexperimental period.</p><p>3.1. Experiment 1</p><p>During the 30-day period before the overloading stage, thesludge had a VFA (acetic acid, propionic acid, and isobutyric acid)concentration of 237 mg l1. The average biogas yield during thepre-overload period was 1.2 l l1 sludge d1 (SD = 0.14). Duringthe rst 5 days of experiment 1/a, the biogas yield increased by13% d1, but the VFA concentration was not signicantly altered.From the 7th day onward, the tVFA and VFA concentrations in-prepared by rst adding 45 ml distilled water to the supernatantsample for a total sample volume of 50 ml. The pH value of thissolution was decreased by adding 0.1 M HCl with continuous mix-ing until a pH of 2.2 was reached, followed by 15 min of stirring toeliminate CO2. The pH value was then raised above 5.0 with 0.1 MNaOH. The tVFA was determined with the following equation:</p><p>tVFA mg acetic acid l1 VNaoHpH5:0 VNaoHpH4:0 fNaOH 200Vsample</p><p> 60</p><p>where VNaOHpH5.0 is the volume of the NaOH until pH 5.0 (ml, at1 atm, 25 C); VNaOHpH4.0 is the volume of the NaOH until pH 4.0(ml, at 1 atm, 25 C); fNaOH is the ratio of the actual concentrationand the nominal (0.1 M) concentration of the NaOH solution; 200is an empirical coefcient; 60 is the molar weight of the acetic acid</p><p>ology 102 (2011) 52705275 5271creased rapidly, primarily due to a rise in the concentrations of ace-tic acid and propionic acid, whereas the daily biogas yielddiminished slowly.</p></li><li><p>02</p><p>4</p><p>6</p><p>8</p><p>10</p><p>12</p><p>1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 37 39</p><p>Day</p><p>OLR</p><p> (g</p><p> l-1</p><p> d-</p><p>1 )</p><p>1/b1/a 2/a 2/b</p><p>Fig. 1. Organic loading rate (OLR) periods of the experiment: 1/a, 15% overloading; 1/b, recovery stage; 2/a, 10% loading; 2/b, stable run.</p><p>5272 T. Rtfalvi et al. / Bioresource Technology 102 (2011) 52705275During experiment 1/a, the rst sign of overload was a fast in-crease of the acetic acid concentration, which was followed by anincrease of the propionic acid concentration after a 1 day delay.The concentrations of acetic acid and propionic acid both reachedtheir maxima during experiment 1/b, and the propionic acid con-centration peaked 2 days after the acetic acid concentration. Themaximum amount of propionic acid was 45 times higher thanthe maximum amount of acetic acid (Figs. 2A, 2B and 2C).</p><p>The maxima VFA concentrations on Figs. 2A, 2B, and 2C are sim-ilar to the detected values during the restart of a bio-waste fedindustrial biogas plant (Gallert and Winter, 2008). Although theconcentrations of butyric acid, isobutyric acid, and valeric aciddid rise during experiment 1/a, they did not increase as greatlyor as rapidly as the concentrations of acetic acid and propionic acidin this period (acetic acid: 7644%, propionic acid: 4790%, butyric</p><p>acid: 240%, isobutyric acid: 105%, valeric acid: 435%).</p><p>0</p><p>2000</p><p>4000</p><p>6000</p><p>8000</p><p>10000</p><p>12000</p><p>14000</p><p>1 2 3 4 5 6 7 8 9 10 11 13 15 17 18 19 20 2Da</p><p>Acid</p><p> con</p><p>cent</p><p>ratio</p><p>n (m</p><p>g l-1)</p><p>Acetic acid Propionic acid Isobutyric acid Buty</p><p>Fig. 2A. Gas production yield and organicThe differences between the methods we tested were soonobvious. After a 23-day delay, the effect of the immediate, drasticOLR decrease rst appeared as a reduction in the tVFA and then areduction in the total VFA. At the end of experiment 1/a, the biogasyield from each of the three test digesters immediately dropped;after the acid content values reached their maxima, the biogasyield increased as the acid content decreased.</p><p>In each case in experiment 1/b, the acetic acid concentration de-creased faster than the propionic acid concentration, causing adrop in the acetic acid to propionic acid ratio.</p><p>An analysis of the changes in the acetic acid to propionic acidratio aids understanding of this process (Figs. 3A, 3B and 3C). Weobserved that the acetic acid treatment in the B reactor resultedin a slower decrease of the acetic acid to propionic acid ratio thanwe observed in the other two reactors. In these experiments, the</p><p>higher acetic acid to propionic acid ratio by itself does not indicate</p><p>1 22 23 24 25 26 27 28 29 30 31 32 33 36 37 38 39y</p><p>0</p><p>1</p><p>2</p><p>3</p><p>Biog</p><p>as </p><p>yield</p><p> (l l-1</p><p>)</p><p>ric acid Valeric acid VFA tVFA Biogas yield</p><p>acid concentrations of the A digester.</p></li><li><p>echn8000</p><p>10000</p><p>12000</p><p>14000</p><p>tratin</p><p> (mg l</p><p>-1)</p><p>T. Rtfalvi et al. / Bioresource Ta good condition because accumulation of acetic acid could resultfrom inhibition of the methanogenic bacteria by a high propionicacid concentration. This apparent contradiction can be understoodby studying not only the ratios of the acids but also the exact con-centrations of each acid over the course of the...</p></li></ul>