Appl Microbiol Biotechnol DOI 10.1007/s00253-015-6486-4

BIOENERGY AND BIOFUELS

Effects of feedstock carbon to nitrogen ratio and organic loading on foaming potential in mesophilic food waste anaerobic digestion Musa Idris Tanimu & Tinia Idaty Mohd Ghazi & Mohd Razif Harun & Azni Idris

Received: 23 October 2014 / Revised: 12 February 2015 / Accepted: 15 February 2015 # Springer-Verlag Berlin Heidelberg 2015

Abstract Foaming problem which occurred occasionally during food waste (FW) anaerobic digestion (AD) was investigated with the Malaysian FW by stepwise increase in organic loading (OL) from 0.5 to 7.5 g VS/L. The FW feedstock with carbon to nitrogen (C/N) ratio of 17 was upgraded to C/N ratio of 26 and 30 by mixing with other wastes. The digestion which was carried out at 37 °C in 1-L batch reactors showed that foam formation initiated at OL of 1.5 g VS/L and was further enhanced as OL of feedstock was increased. The digestion foaming reached its maximum at OL of 5.5 g VS/L and did not increase further even when OL was increased to 7.5 g VS/Ld. Increase in the C/N ratio of feedstock significantly enhanced the microbial degradation activity, leading to better removal of foam causing intermediates and reduced foaming in the reactor by up to 60 %. Keywords Organic loading . Anaerobic digestion . Foaming . Carbon to nitrogen ratio . Biogas methane

Introduction Anaerobic digestion (AD) has been widely accepted as an economic treatment option for food waste (FW) and other high moisture content (MC) organic wastes. This is because of its energy recovery benefits in the form of biogas methane yield. However, biologically mediated foaming during AD could result to environmental pollution, poor digestion performance, and colossal damage to the reactor, pipeline, pump, and other plant equipment. The exact mechanism of foam initiation and its stability is still not well known (Frayer et al. 2011; Heard et al. M. I. Tanimu : T. I. Mohd Ghazi (*) : M. R. Harun : A. Idris Department of Chemical and Environmental Engineering, Faculty of Engineering, Universiti Putra Malaysia, 43400 Serdang, Selangor, Malaysia e-mail: [email protected]

2008). In the past, almost every foaming event in the industry has been linked to the presence of foam causing microbial cells. However, fundamental research carried out on the basic mechanisms of stable foam formation in AD revealed that the foaming results from the close interactions of three phenomena (Griffiths and Stratton 2010; Pagilla et al. 1997; Ganidi et al. 2011, 2009; Kougias et al. 2013). These include, first, the presence of gas bubbles. This could be the biogas produced during AD or air entrained in influents to the anaerobic digesters especially if it is emanating from the secondary treatment unit in an activated sludge process in wastewater treatment plants. Second, the presence of surface active substances (surfactants or biosurfactants), e.g., oil, grease, lipid volatile fatty acids, detergents, proteins, glycolipids, lipoproteins, phospholipids, polysaccharide-lipid complexes (Kougias et al. 2013; Ganidi et al. 2009; Nitschke and Pastore 2006; Ron and Rosenberg 2002). These surface active agents which are both hydrophobic and hydrophilic could be present both in the feedstock or produced as intermediates during AD. The surfactants largely serve as food source to bacteria cells in digester media (Ganidi et al. 2009). The third requirement for foaming is the presence of foaming bacteria cells in the digester medium such as Microthrix parvicella or Nicordia (Marneri et al. 2009). Filamentous bacteria has been identified by researchers as a foaming cause due to large amount of activated sludge present in the influent to digesters especially in sewage treatment plants (Ganidi et al. 2011, 2009; Pagilla et al. 1997). In conventional digesters such as those treating FW, foaming has been attributed to overfeeding or inconsistent feeding (Pagilla et al. 1997). This factor border on the concentration of the organic content or OL of feedstock digested. Ganidi et al. (2011) investigated the foaming initiation during the AD of municipal wastewater sludge and found OL of 2.5 g VS /L as the foam initiation threshold and OL of 5 g VS /L resulting in persistent foaming. Similarly, Kougias et al. (2013) investigated the effect of feedstock composition and organic loading rate (OLR) on foaming of cattle manure-based feedstock. The researchers observed the

Appl Microbiol Biotechnol

formation of foam at an OLR of 4.2 g VS/Ld. They concluded that the OLR and organic composition of the cow manure (such as protein and lipid content) and the biogas produced during AD were the main contributors to the foaming observed. Experimental evidence on the identification of OL for foam initiation in a conventional FW AD is unclear. FW AD process is aimed at biogas recovery through the use of numerous classes of bacteria for the solubilization and degradation of the different FW components such as carbohydrates or sugars, protein, and lipids (Zhang et al. 2012a, b; Izumi et al. 2010). The presence of all these components in the feedstock and the synthesis of surfactants and biosurfactants as intermediate products during AD make FW AD more prone to foaming (Kougias et al. 2014). Researchers have identified that accumulation of volatile fatty acid (VFA), ammonia, and the presence of lipids and protein during AD may contribute more or less to foaming in digesters (Kougias et al. 2013; Boe et al. 2012). Although the presence of foam causing cells have been identified in virtually all cases of foaming investigated (Kougias et al. 2014; Dalmau et al. 2010; Marneri et al. 2009; Griffin et al. 1998), however, the mere presence of foaming bacteria without the presence of the intermediate food source to the cells during AD is unlikely to cause significant foaming event (Ganidi et al. 2011; Pagilla et al. 1997). Therefore, foaming in conventional AD could be stemmed by rapid removal or degradation of the intermediate biosurfactants, which serve as food source to bacteria cells (Nitschke and Pastore 2006). In our earlier work (Tanimu et al. 2014a) on batch AD of the FW, it was established that balanced microbial activity during AD could enhance the delicate balance between the activities of acidogens and methanogens resulting in rapid utilization of foam enhancing intermediates. Therefore, adequate nutrition to the bacterial cells involved in AD through codigestion of the feed with complementary wastes could improve the C/N ratio and enhance the activities of methanogens (Tanimu et al. 2014b; Zeshan and Karthikeyan 2012; Jiang et al. 2012; Zhang et al. 2012a, b). In this study, the influence of increasing the FW C/N ratio on its foaming potential during AD was studied by stepwise increase in OL in the batch anaerobic digesters under mesophilic condition.

Materials and method Source and nature of the digested food waste FWs were collected from a FW AD treatment plant in Taman Sri Serdang, Selangor, Malaysia. The AD plant which treats a mixture of FW obtained from the commercial restaurants and markets within Serdang area (population of over 200,000 people) is part of an integrated waste management system for the city of Serdang. Sample of the FW was obtained for foaming analysis after a persistent foaming was observed in the plant. The main components of the mixed FW feedstock (F1)

include chicken meat/beef (5 %); rice, bread, and noodles (77 %); leafy vegetables/salad (6 %); soup (5 %); seafood, fish, egg, and egg shell (7 %). Two other feedstock mixtures (F2 and F3) were formulated using F1 as the base feedstock by adding fruits, vegetables, and chicken meat wastes as cosubstrates to adjust their C/N ratio from 17 to 26 and 31, respectively (Table 1). The average compositions of the vegetable wastes include baby corn (5.0 %), lettuce (23.6 %), carrot (4.5 %), broccoli (18.2 %), and green leafy vegetables (48.7 %) all on wet weight (WW) basis. The average compositions of the fruit wastes include papaya (26.8 %), orange (18.6 %), pineapple (38.8 %), water melon (10.6 %), and berries (5.2 %) all on WW basis. Table 1 presents the characteristics and composition of the feedstock.

Feedstock and inoculum preparation The FW (F1), chicken meat, fruits, and vegetable wastes were ground separately using a heavy duty blender (model 39BL11 Table 1

Feedstock characteristics

Detail

F1 (C/N=17) F2 (C/N=26) F3 (C/N=30)

Physical characteristics Moisture content (%)a TS (%)a TS (g/L) VS (g/L)a VS/TS (%) Chemical characteristics C (% TS) K (% TS) N (%TS) Na (%TS) P (%TS) C/N COD (g/L) pH Lipids (%TS) Lignin (%TS) ADF (%TS) NDF (%TS) GE (cal/g) CF (%TS) CP (%TS) Feed composition FW (%)a

92.43 7.55 15.20 14.64 96.30

95.15 4.82 12.80 12.19 95.20

90.63 9.35 19.64 18.58 94.60

33.53 0.48 2.02

44.55 3.33 1.73

43.50 0.44 1.45

1.06 0.66 16.60 273 4.40 28.18 26.81 39.19 42.77 5452 15.65 19.31

0.05 0.39 25.75 282 4.83 13.13 19.31 49.04 68.77 3816 28.78 10.55

0.17 1.57 30.00 294 5.14 34.49 12.30 8.33 29.30 6918 0.29 41.70

100.00

50.00 40.00 10.00

77.30

a

Fruit waste (% ) Vegetable waste (%)a Chicken meat waste (%)a a

Based on percentage of wet weight

22.70

Appl Microbiol Biotechnol

Waring commercial, USA) and sieved to remove large particles using a 1-mm sieve size. They were prepared in bulk and stored at −20 °C for later use. F2 and F3 were made by mixing part of F1 with portions of fruits, vegetables, and chicken meat wastes. The composition of FW, chicken meat, fruits, and leafy vegetables added to achieve C/N ratio of 26 and 30 in F2 and F3, respectively, is presented in Table 1. Inoculum was sourced from the FW treatment plant digester. The proportion of feed mixture to inoculum was 7:3 by volume (Liu et al. 2009). The data in Table 1 are the mean values of eight different samplings within a period of 30 days. The characteristics of the inoculum are presented in Table 2. The percentage of TS and VS in Table 2 represents the percentage of total solids (TS) and volatile solids (VS), respectively, contained in the inoculum used. The percentage of TS in Table 2 gives the fractional weight of the entire amount of solids contained in a known volume of the inoculum after drying at 105 °C in an oven. The VS is the fraction of the TS that will disappear when a known weight of the TS is ignited in a furnace at 550 °C for 2 h. Both TS and VS values can be calculated on dry weight basis but mostly calculated on wet weight basis by researchers. Experimental procedure Thirty glass bottles, each of 1-L capacity and a working volume of 0.8 L, were used as the batch digesters. Each OL investigated in the batch AD experiments was performed in duplicate for each feedstock. The digestion of F1 was carried out in reactor R1 to R10. Similarly, F2 was digested in reactor R11 to R20, and F3 was digested in reactors R21 to R30. Each bottle was filled with the feedstock formulated and diluted with tap water to attain the OL of 0.5-, 1.5-, 3.5-, 5.5-, and 7.5-g volatile solids (VS)/L, respectively. These OLs were selected based on the performance data reported for FW in literatures (Jiang et al. 2012; Zhang et al. 2012a, b; Forster-Carneiro et al. 2008; Ganidi et al. 2011; Liu et al. 2009; Zupancic et al. 2008; Bolzonella et al. 2005). These loadings were achieved based on each feed’s volatile solid (VS) content with different water dilution amount. The ratios of feedstock to water added to attain each of the OL above were approximately 1:30, 1:9, 1:3, 1:1.5, and 1:0.8, respectively. The bottles were tightly covered with rubber bungs and sealed with adhesive. Nitrogen gas at low pressure was bubbled through the content of Table 2 Inoculum characteristics

Parameter

Unit

Value

Moisture content TS TS VS VS/TS NH3-N pH

% mg/L % mg/L % mg/L

82.57 19.27 17.43 17.80 92.37 625 7.3

each of the bottles for 5 min before it was connected with silicone tubing to a Tedlar gas bag (SKC Inc., USA) for biogas collection. The bottles were placed in an oven at 37 °C, and its content was shaken daily after observing the foam formation. Pressure in the bottle and the gas bag was not measured during the batch digestion and gas production; however, the gas volumes reported were expressed and corrected to standard temperature and pressure (STP) of 0 °C and 101.325 kPa. Two openings or holes were bored in the rubber bung cover through which 30-mm length each of the silicone tubing was inserted. The end of the tubing inserted through the first opening in the digester cover was positioned to collect the biogas produced from the bottle digester headspace only. The outside end of the biogas collection tubing was fitted to the gas entry point connector of the Tedlar bag. The tube inserted into the second hole was for sampling purpose. The end of the sampling tube inserted into the digester was dipped into the digestion medium in the bottle digester while the outside end was properly clipped to prevent air entry. Each time, sample was extracted from inside the bottles by fitting a syringe at the clipped end of the sampling tube. The sample required was drawn into the syringe by temporarily removing the clip from the sampling tube. When the syringe is filled to capacity, the sampling tube was clipped, and the sample drawn into the syringe was emptied into a measuring cylinder. This sample extraction method was repeated until enough samples for analysis were obtained. In this way, the digestion bottle was kept under anaerobic condition throughout the period of the batch digestion. Biogas production and other parameters such as TS, VS, COD removal, VS destruction, VFA, ammonia-nitrogen (NH3-N), and alkalinity were monitored every other day. Potential foaming tests were carried out twice weekly because of limited sample. F1, F2, and F3 blanks and inoculum were each digested separately to observe any biogas production, but no further tests were carried out on them during the digestion. Digestion was carried out for 30 days. Foaming tests Foaming tendency of the digestion medium was determined by the method of aeration (air bubbling or circulation into the sampled digestion sludge obtained from the batch digester) using a bubble column apparatus. The apparatus, similar to that used by Frayer et al. (2011), consists of a 60-mm diameter glass cylinder of 1-L capacity with a galvanized wire placed at the bottom of the cylinder as air diffuser. Fifty milliliters of sample was obtained from the reactor medium into the apparatus. The maximum foaming height and volume obtained after aeration of the sample sludge for 10 min at 60 mL/min were recorded (Kougias et al. 2013; Ganidi et al. 2011; Deshpande and Barigou 2000; Marneri et al. 2009). Foaming potential of the feedstock was evaluated by observing the foam formation in the reactor, the foaming tendency, and foam stability. The foaming tendency was obtained by dividing the foam volume (mL) immediately after aeration

Appl Microbiol Biotechnol

by the flow rate of air (mL/min). Foam stability was obtained as the percentage of foam persisting after 1 h of aeration. This was obtained by comparing the remaining volume (mL) of foam after 1 h of aeration with the volume of foam obtained immediately after aeration. Test for potential foaming and foam stability was carried out in duplicate. The content of the digesters was also monitored daily for any foam formation. Foaming volume in the digester was determined by multiplying the digester surface area by the height of foam observed. Each time, after the foaming in the reactor has been recorded, the digesters were shaken until the observed foam disappear.

dimension 30.0 m×250 μm×0.25 μm. Helium gas was used as carrier at a flow rate of 154 mL/min, with a split ratio of 100 to give a column flow rate of 1.5 mL/min and a purge rate of 150 mL/min. The oven temperature was set to increase from 40 to 230 °C in 17.5 min. Temperature of the injector and detector was 200 and 320 °C, respectively. Samples were acidified to pH 2.0 using 3 M phosphoric acid after filtration with a syringe type filter of pore size 0.45 μm. Standard VFA solutions containing acetic, propionic, iso-butyric, n-butyric, and iso-valeric (Sigma-Aldrich, UK) were used as standards for VFA.

Analytical methods

Results

Physiochemical characteristics tests such as MC, total solids (TS), and volatile solids (VS) were carried out according to standard methods (Eaton et al. 2005). Ultimate analysis for carbon and nitrogen was carried out using the CHNS machine (model CHNS 932, Leco Corporation, USA). Portions of the digestion medium were collected and centrifuged with a centrifuging machine (Model 2420, Kubota Corporation, Japan) at 5000 rpm for 15 min prior to VFA and NH3-N tests. Analysis on alkalinity was performed by titration of 10 mL of the sample against 1 N H2SO4. Ammonia-nitrogen (NH3-N) was performed by Nessler’s method. Volume of biogas collected in the gas bags were measured using the water displacement method. The biogas composition was obtained using a gas chromatograph machine (Agilent 6890 N network GC system) equipped with the carboxen 1010 plot column with dimension 30.0 m×530.0 μm×0.0 μm nominal (Supelco 25467, USA). The injector and detector temperature were 200 and 230 °C, respectively, with the oven temperature initially set at 40 °C. Argon gas was used as the carrier gas at a flow rate of 3.0 mL/ min. pH was determined using a combined pH meter and thermometer instrument (Trans Instruments, Singapore). Gross energy (GE) was determined using model C2000 oxygen bomb calorimeter (ThermoFisher Scientific Australia). The crude fiber (CF) in the FW was extracted in a heated digestion apparatus using 0.13 mol/L of boiling sulfuric acid and 0.23 mol/L boiling sodium hydroxide in a successive extraction and washing process. The residue was vacuum-filtered and dried in the oven at 130 °C to a constant weight and finally ashed at 475 °C in a muffle furnace for about 30 min. Crude protein (CP) was determined by Kjeldahl’s digestion method with sulfuric acid (98 %) and selenium catalyst. Lignin, acid detergent fiber (ADF), and neutral detergent fiber (NDF) were analyzed using Fibertec 2010 auto fiber analysis system (Foss Analytical, Denmark). Calorific value (CV) was obtained using a bomb calorimeter (Parr 6300 calorimeter, USA). Trace elements were measured using the ICP-OES model DV5300 (Perkin-Elmer, USA). VFA test was carried out in an Agilent 6890 series gas chromatography using a flame ionization detector and a capillary column type HP-INNOWax polyethylene glycol capillary with

Effect of OL variation on feedstock foaming in the reactor Figure 1 presents the plots of foaming and biogas yield at five different OL tested (0.5, 1.5, 3.5, 5.5, and 7.5 g VS/L) during the digestion of feedstock F1, F2, and F3. Figure 1a shows the foaming and biogas yield obtained during the digestion of F1, F2, and F3 at OL of 0.5 g VS/L. At this OL, there was no visible foam produced during the digestion of the three feedstocks. However, visible foaming initiated at OL of 1.5 g VS/ L (Fig. 1b) feedstock F1, with C/N=17 was most prone to foaming with maximum foaming of 0.0, 0.14, 0.22, 0.26, and 0.26 L/Lreactor recorded at OL of 0.5, 1.5, 3.5, 5.5, and 7.5 g VS/L, respectively. The corresponding foaming recorded in F2 (C/N=26) was 0.0, 0.10, 0.8, 0.16, and 0.16 L/Lreactor. The least foaming of 0.0, 0.6, 0.8, 0.1, and 0.1 L/Lreactor was obtained during the digestion of F3 (C/N=30) at the same OL. These results show a significant reduction in foaming as feedstock C/N ratio increased from 17 to 26 and then finally to 30. Although the foaming generally increased with digestion time, the foaming rate was highest in F1 followed by F2 and then the least was at F3. For example, while foaming volume of F1 increased sharply from 0.15 to 0.26 L/Lreactor within 4 days of digestion at OL of 5.5 g VS/L (Fig. 1d), that of F2 increased more slowly from 0.09 to 0.16 at a loading of 7.5 g VS/L within the same number of days (Fig. 1e). F3 had the least foaming of 0.1 L/Lreactor at OL of 7.5 g VS/L (Fig. 1e). Therefore, the foaming observed during the digestion of F2 in this case was less than F1 by 38 % in spite of the higher OL of 7.5 g VS/L. Similarly, the foaming in F3 reduced significantly by up to 60 % when compared with F1. Effect of biogas production on foaming Figure 1 indicates that there was an increase of biogas yield as the OL of all the feedstock digested increased (except during the digestion of F1 at OL of 7.5 g VS/L which recorded the highest foam formation of 0.26 L/Lreactor and no biogas was collected in the gas bag). Foaming in the reactor also increased with increase in the biogas yield. This seems similar to the

Appl Microbiol Biotechnol

Foaming volume (L/Lreactor)

1

0.35 0.3

0.8

0.25 0.6

0.2

0.4

0.15 0.1

0.2

0.05

0

0 0

5

10

15

20

25

Cumulative biogas (L/gVS)

(a) OL = 0.5g VS/L

Fig. 1 Effects of C/N ratio and biogas production on foam formation at OL of a 0.5 g VS/L, b 1.5 g VS/L, c 3.5 g VS/L, d 5.5 g VS/L, and e 7.5 g VS/L

30

(b) OL = 1.5g VS/L Foaming volume (L/Lreactor)

0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0

0.8 0.6 0.4 0.2 0 0

5

10

15

20

25

Cummulative biogas (L/gVS)

Time (days)

30

(c) OL = 3.5g VS/L 0.8

Foaming volume (L/Lreactor)

0.25 0.2

0.6

0.15 0.4 0.1 0.2

0.05 0

0 0

5

10

15

20

25

Cummulative biogas (L/gVS)

Time (days)

30

Time (days)

1 0.8

0.2

0.6 0.4

0.1

0.2 0

0 0

5

10

15

20

25

Cumulative biogas (L/gVS)

Foaming volume (L/Lreactor)

(d) OL= 5.5 gVS/L 0.3

30

Foaming volume (L/Lreactor)

(e) OL = 7.5g VS/L 0.3

1 0.8

0.2

0.6 0.4

0.1

0.2 0

0 0

5 FV(C/N=17) Biogas yield (C/N=17)

findings of some researchers that increased biogas production may also increase foaming tendencies (Varley et al. 2004;

10

15 Time (days) FV (C/N=26) Biogas yield (C/N=26)

20

25

30

Cummulative biogas (L/gVS)

Tme (days)

FV (C/N=30) Biogas yield (C/N=30)

Griffiths and Stratton 2010; Kougias et al. 2013). It was observed that the case of digestion of F1 at OL of 7.5 g VS/L

Appl Microbiol Biotechnol Fig. 2 VFA and NH3-N concentration during the batch digestion at OL of a 0.5 g VS/L, b 1.5 g VS/L, c 3.5 g VS/L, d 5.5 g VS/L, e 7.5 g VS/L

(a) OL= 0.5 gVS/L

VFA (mg/L)

4000

3000

3000 2000 2000 1000

NH3-N (mg/L)

5000

4000

1000 0

0 0

5

10

15 Time (day)

20

25

30

5000

5000

4000

4000

3000

3000 2000

2000

1000

1000

NH3-N (mg/L)

VFA (mg/L)

(b) OL= 1.5 gVS/L 6000

0

0 0

5

10

15 Time (days)

20

25

30

5000 4000 3000 2000 1000

NH3-N (mg/L)

VFA (mg/L)

(c) OL= 3.5 gVS/L 8000 7000 6000 5000 4000 3000 2000 1000 0

0 0

5

10

15 Time (days)

20

25

30

5000

10000

4000

8000

3000

6000 2000

4000

1000

2000

NH3-N (mg/L)

VFA (mg/L)

(d) OL= 5.5g VS/L 12000

0

0 0

5

10

15 Time (days)

20

25

30

(e) OLR= 7.5 gVS/L VFA (mg/L)

4000 10000

3000 2000

5000

1000

NH3-N (mg/L)

5000

15000

0

0 0

5 VFA (C/N=17) Ammonia-N (C/N=17)

10

15 20 Time (days) VFA (C/N=26) Ammonia-N (C/N=26)

25

30 VFA (C/N=30) Ammonia-N (C/N=30)

Appl Microbiol Biotechnol

(a)

Fig. 3 a pH and b cumulative methane yield during the batch AD at OL of 0.5–7.5 g VS/L

7.5 7 6.5

pH

6 5.5 5 4.5 4 3.5 3 0

5

10

15

20

25

20

25

30

Time (days)

(b) 0.35

Cummulative CH4 (L/g VS)

0.3 0.25 0.2 0.15 0.1 0.05 0 0

5

10

15

30

Time (days)

OL= 0.5g VS/L (C/N=17) OL= 5.5g VS/L (C/N=17) OL= 1.5g VS/L (C/N=26) OL= 7.5g VS/L (C/N=26) OL= 3.5g VS/L (C/N=30)

where no biogas was collected in the gas bag during digestion was due to oversaturation of the liquid phase with the biogas produced. Some of the biogas locked within the digestion medium was observed to move into the gas phase as the foam formation disappeared when the reactor was shaken. The foam covering was observed to become even thicker with increase in OL thereby affecting the equilibrium between gas and liquid phase and requiring more vigorous shaking for the foams to disappear. The same observation was made during the digestion of F2 at OL of 7.5 g VS/L which resulted to relatively lower biogas production and its eventual seizure on day 10 (Fig. 1e). This has partly contributed to the low biogas yield during the digestion. With the above observations, it may appear that increased biogas production is linked to the foaming. However, during the digestion of F2 and F3 at OL of 0.5 g VS /L (Fig. 1a), there was no visible foam formation at a cumulative biogas production of 0.262 and 0.305 L/g VS, respectively. These values compare with cumulative gas production of 0.300 and 0.399 L/g VS in F1 and F2 at OL of 1.5 g VS/L (Fig. 1b) at which point foaming was first initiated.

OL= 1.5g VS/L (C/N=17) OL= 7.5g VS/L (C/N=17) OL= 3.5g VS/L (C/N=26) OL= 0.5g VS/L (C/N=30) OL= 5.5g VS/L (C/N=30)

OL= 3.5g VS/L (C/N=17 OL= 0.5g VS/L (C/N=26) OL= 5.5g VS/L (C/N=26) OL= 1.5g VS/L (C/N=30) OL= 7.5g VS/L (C/N=30)

Higher and lower biogas production volumes than stated above were also recorded during foam formation as the OL of the feedstock digested were increased (Fig. 1c to e). This therefore suggests that increase or decrease in the biogas production alone may not trigger foaming in this case. Effects of intermediate VFA during digestion The VFA values in Fig. 2 show that the range of concentration of VFA produced in the medium during the digestion was lowest at OL of 0.5 g VS/L (500–3800 mg/L). Although the sudden increase in VFA from 2.7 g/L on day 28 to 3.8 g/L on day 30 (Fig. 2a) is rather high especially at OL of 0.5 g VS/L; however, this is likely an indication of acetic acid accumulation especially when the acetoclastic methanogens are not active. This observation is supported by the reduction in pH from neutral to 6 (Fig. 3a) during the same period. During this period, NH3-N, which is also known to act as a buffer against the acidic VFA effect (Lahav and Morgan 2004), also decreased slightly from 3 to 2.8 g/L (Fig. 2a). In a similar AD

Appl Microbiol Biotechnol Average range of parameters obtained during the 30-day digestion of F1, F2, and F3 at different organic loadings and 35 °C

OL (g VS/L) F1 (C/N=17) 0.5 1.5 3.5 5.5 7.5 F2 (C/N=26) 0.5 1.5 3.5 5.5 7.5 F3 (C/N=30) 0.5 1.5 3.5 5.5 7.5

Alkalinity range (mg/L)

COD removal (%)

VS destruction (%)

CH4 in biogas (%)

Max. Cum. CH4 yield (L/g VS)

8000–10,000 8200–12,000 7500–12,800 5430–13,600 3160–10,810

50.1 43.9 33.3 21.9 14.9

49 41 43.2 43 17

35–40 36–43 38–50 35–41 -

0.090 0.146 0.165 0.1055 -

9850–14,000 9280–15,000 8540–15,420 7800–15,000 6000–11,300

52.1 45.8 40.5 36 19.6

58.5 44.3 56.7 36.5 35.2

38–42 38–50 40–58 39–55 35–49

0.133 0.185 0.238 0.200 0.150

10,566–15,800 11,000–16,800 10,000–16,000 9500–16,000

64.8 58.6 48.4 42.4

68.9 70.2 59.5 45.2

40–49 45–58 43–70 40–70

0.150 0.211 0.268 0.334

7820–10,000

37.2

42.5

42–58

0.250

the methanogens to biogas. It has been found that when hydrophobic biosurfactants accumulate in gas interfaces, breaking of gas bubble is difficult (Ganidi et al. 2011). Therefore, the presence of these biosurfactants during the digestion together with the biogas produced is likely to have contributed to foam formation. Under this condition, stable foam formation could result especially at high loading (5.5–7.5 g VS/L) as there are now sufficient accumulation of the hydrophobic biosurfactants in the medium which also serve as source of

Fig. 4 Maximum foaming (FV) and foam stability (FS) attained during the AD of F1, F2, and F3

Maximum foaming (L/Lreactor)

of FW mixtures, Zeshan and Karthikeyan (2012) obtained a VFA of up to 3.6 g/L at OLR of 0.65 g VS/L. This is less than the VFA of 6–8 g/L reported by Propasert (2007) as inhibitory. At higher OL range of 1.5–7.5 g VS/L, increasingly higher VFA concentration (1200–10,700 mg/L) was produced. The VFA results showed that the presence of higher VFAs (propionic, butyric, and valeric acids only) than acetic acid during the foaming occurrence. Higher VFAs need to be further broken down to acetic acid before it can be converted by

0.3

60

0.25

50

0.2

40

0.15

30

0.1

20

0.05

10

0

0 0

1

2

3

4

5

6

7

OL ( g VS/L)

FV (C/N=17)

FV (C/N=26)

FV (C/N=30)

FS (C/N=17)

FS (C/N=26)

FS (C/N=30)

8

Foam stability (%)

Table 3

Appl Microbiol Biotechnol

food for the foam producing organisms (the presence of foaming organisms was not verified during this digestion). No microbial analysis or identification of foam causing cells was done in this study. Since the presence of foam causing cells cannot be separated from most foaming events (Ganidi et al. 2011; Dalmau et al. 2010; Heard et al. 2008), the main focus of this study was to find out how feed composition, OL, and C/N ratio of the FW could influence foaming events of which foaming cells are already an inherent part of.

two parameters during the digestion of the feedstock at C/N ratio 17 (1), 26 (2), and 30 (3), respectively:  F r ¼ 9:914lnðLÞ þ 8:311 R2 ¼ 0:98 ð1Þ

 F r ¼ 5:582lnðLÞ þ 4:769 R2 ¼ 0:85

ð2Þ

Alkalinity effect

 F r ¼ 3:714lnðLÞ þ 3:319 R2 ¼ 0:96

ð3Þ

Table 3 presents the increasing range of alkalinity concentrations in the digestion medium with increase in OL and foaming. Alkalinity during AD helps to provide buffering effects in the digestion medium for optimal pH and biological activity. A minimum alkalinity concentration of 5000 mg/L at pH of 7.4 was found to produce adequate buffering effect during the digestion of a Korean FW (Lee et al. 2009). The high MC and very low pH characteristics of FW referred to by Seo et al. (2004) as a typical Korean FW is similar to that of the FW digested in this study. Most of the alkalinity values recorded during this study (Table 3) were not less than this, except at OL of 7.5 g VS/L during the digestion of F1 where lower alkalinity of 3000 mg/L was recorded. This lack of enough buffering in F1 (as indicated by a sharp reduction in pH from 7.2 to 3.5 during the digestion, Fig. 3a) in the presence of increasing VFA range of 1600–10,730 mg/L is likely to have resulted to microbial imbalance and shock (Hansen et al. 1998). This shock loading is likely responsible for the lack of biogas production during the digestion of F1 at OL of 7.5 g VS/L.

Discussion Effect of OL variation on feedstock foaming in the reactor The maximum foaming recorded during the digestion of F1 was at OL of 5.5 and 7.5 g VS/L (Fig. 4). Comparison of the foaming of feedstock F1 with F2 at the OLs tested shows a decrease in foam formation of 28.57, 63.63, and 38.46 % at OL of 1.5, 3.5, and 5.5 g VS/L, respectively. A further reduction in foaming of 57.14, 63.63, and 61.54 % was obtained at the same OL when the foaming of F1 is compared with F3. When foaming of F2 is compared with F3, a decrease in foaming of 30 % was obtained at OL of 1.5 g VS/L and 37.5 % at OL of 5.5 and 7.5 g VS/L. This means that for the same OL, the tendency to foaming decreases with increase in feedstock C/N ratio. However, within the same feedstock C/N ratio, foaming increased with increase in OL. Correlation between OL and reactor foaming (Fig. 4) showed a logarithmic trend. The equations below show the relationship between the

where Fr =foam formation in the reactor (L/Lreactor); L= organic loading of feedstock at 37 °C (g VS/L). Foam stability during the digestion of the 3 feedstock showed a decreasing trend with increase in C/N ratio (Fig. 4). A reduction in foam stability of 50, 64, 60, and 62 % was obtained during the digestion of F3 at OL of 1.5, 3.5, 5.5, and 7.5 g VS/L, respectively, when compared with the corresponding values obtained during the digestion of F1. Effect of OL and C/N ratio on methane yield The maximum yield of CH4 obtained after 30 days of feedstock digestion were 0.165 L/gVS at OL of 3.5 g VS/L for F1; 0.238 L/g VS at OL of 3.5 g VS/L for F2, and 0.334 L/g VS at OL of 5.5 g VS/L for F3. The methane composition in Table 3 indicates better activity of the methanogens as C/N ratio of feedstock increased. F3 produced higher methane than F1 and F2 at the same OL. The average range of methane composition obtained during the digestion of F1 was 35–50 %. Range of methane composition during F2 and F3 digestion was 38– 58 and 40–70 %, respectively. The range of methane composition demonstrates increasingly better methanogenic activities in F3 than F1 and F2. This could be responsible for obtaining the highest yield of methane as well as the reduced foaming effect observed during the digestion of F3. Better methanogenic activity in this case has contributed to increased methane yield and removal of the foam causing digestion intermediates such as the VFAs. The highest methane yield obtained in this study is about 60 % of the theoretically possible (0.572 L/g VS). The lower yield in this study is likely due to the foam formation and the difference in composition of feedstock. Assessment of other parameters Increase in VFA within the digestion medium as OL increased was observed to result in the lowering of pH below the neutral bound (Fig. 3). Reduction in pH during digestion may result to poor methanogenic activities, which may lead gradually to

Appl Microbiol Biotechnol 0.3

(a)

0.25

50

0.2

40

0.15

30

0.1

20

0.05

10

0

0 2

4 6 Organic Loading (g VS/L)

8 25

0.18

(b)

0.16

20

0.14 0.12

15

0.1 0.08

10

0.06 0.04

5

Foam stability (%)

Reactor foam (L/Lreactor) Foam tendency (Lfoam/Lair/min)

0

0.02 0

0 0

1

2

3 4 5 6 Organic Loading (g VS/L)

7

8

0.12 Reactor foam (L/Lreactor) Foam tendency (Lfoam/Lair/min)

Foam stability (%)

60

40

(c)

0.1

35 30

0.08

25 20

0.06

15

0.04

10 0.02

Foam stability (%)

Reactor foam (L/Lreactor) Foam tendency (Lfoam/Lair/min)

Fig. 5 Correlation between foaming tendency, foam stability, and the actual foaming in the reactor during the batch mesophilic digestion of feedstock a F1, b F2, and c F3

5 0

0 0

2

4

6

8

Organic loading (gVS/L) Reactor foam Poly. (Reactor foam)

inhibition (Izumi et al. 2010; Griffin et al. 1998). For proper performance of methanogenic bacteria, it has been reported that the pH of the digestion medium should be within the range of 6.8 to 7.8 (Zeshan and Karthikeyan 2012; Lahav and Morgan 2004). Although there was more rapid pH drop during the digestion of F1, however, the pH was still within neutral bounds of 6.5 to 7.2 during the first 10 days of digestion before decreasing below pH 6. F2 and F3 provided better buffering capacity (Fig. 3a) than F1. In spite of the increased loading from 0.5 to 7.5 g VS/L, the pH of F2 and F3 mostly remained within the range of 7.5–6.5 in the first 16 and 22 days, respectively, of the 30-day digestion period. This showed better pH stability in the digestion of F2 and F3 over

Foam tendency Poly. (Foam tendency)

Foam stability Poly. (Foam stability)

F1. Therefore, the increase in the feedstock C/N ratio provided adequate buffering for the digestion medium against the accumulation effect of VFA, which is the main source of pH decrease. During this digestion, NH3-N production ranged from 984 to 3871 mg/L during the digestion of F1, 1120 to 3144 mg/L during the AD of F2, and 1200 to 2761 mg/L in the case of F3 (Fig. 2). These values show a decrease of about 30 % in the concentration of NH3-N released during the digestion as the C/N ratio of the feedstock increased from 17 to 30. Also, the range of NH3-N concentration released during the digestion of F1, F2, and F3 was 65, 53, and 46 %, respectively, of the inhibiting amount (6000 mg/L) reported by Hansen et al.

Appl Microbiol Biotechnol

(1998) during the digestion of swine manure. Therefore, F1 was potentially more inhibited than F2 and F3 in that order. Zeshan and Karthikeyan (2012) attributed the reduction in NH3-N produced during a similar FW AD to the reduced protein solubilization rate as the C/N ratio of the feed was increased. Zhang et al. (2012a, b) also found that the dilution of feed or biomass waste mixture, as done in this case, could help upgrade its C/N ratio and reduce NH3-N inhibition. It was observed that the methane yield was higher at low NH3-N values. For example, F3 at OL of 5.5 g VS/L yielded the highest methane of 0.334 L/g VS (Fig. 3d) and at the same time produced the lowest range of NH3-N concentration of 1200–2115 mg/L. Although F3 contained the highest protein content (Table 1), however, it did not lead to the production of high concentration of NH3-N to inhibiting level. The characteristics of the feed (Table 1) have shown that the protein and lipid content of F1 was higher than F2. It is therefore expected that F1 should contain more of the foam causing surface active biosurfactants than F2. This may have been responsible for the higher foaming volume recorded during the digestion of F1. The higher NH3-N content of F1 than F2 is likely due to the higher protein solubilization in F1. In addition, the higher range of VFA released in F1 indicates that the condition during the digestion of F1 was more stressed than F2. This is further confirmed by the lower methane yields recorded in F1 compared with F2 as stated earlier. High concentration of NH3-N and VFA during AD as identified during the digestion process of F1 indicates imbalanced or inhibited microbial activity, which may have enhanced foam formation (Resch et al. 2011). The lower microbial activity observed in F1 is likely responsible for its lower methane yield than F2 at the same OL (Fig. 3b). It is expected that the digestion of F1 should yield higher theoretical methane than F2 as it contains more lipid content. However, this was not so as a result of the higher stress in F1. Despite having the highest protein content (Table 1), F3 yielded the highest methane (0.440 L/g VS) and the least foam formation range (0.06–0.10 L/Lreactor). This can be attributed to a more balanced and stable microbial activity as evidenced by the improved methanogenic activity (49– 70 % methane content) during its digestion. Volatile solid (VS) destruction in all feedstock decreases as OL was increased (Table 3). The ranges of VS destruction during the digestion of F1, F2, and F3 were 17–49, 35–58, and 42–70 %, respectively. This has shown an improved microbial degradation with the increase in C/N ratio of feedstock from 17 to 26 and finally to 30. Similarly, removal of chemical oxygen demand (COD) during digestion (Table 3) showed a progressively better removal of organic content with the increase in C/N ratio from 17 to 30. This, together with the improved degradation, indicates a progressively better acidogenic and methanogenic balance during AD as the feedstock C/N ratio increased. Therefore, the less foam formation observed with increase in C/N ratio during the digestion was

likely due to the increasingly better removal or degradation of the foam causing intermediates. Figure 5 shows the correlation between foaming tendency, foam stability, and the actual foaming in the reactor. The foaming values considered in each case were the respective maximum values of foam recorded at each of the OLs tested. The foaming tendency showed relatively similar pattern or correlation with the actual foaming produced in the reactors at the OLs tested and in all the feedstock digested. However, there was no clear correlation between the foam stability test and the other foaming parameters in all the feedstock digested. This showed that the foaming tendency test was more reliable in predicting the actual foaming in the reactor than the foam stability test. The foaming parameters show that feedstock F1 was most prone to potential stable foaming incidence (Fig. 5a). Feedstock F3 indicates the highest potential resistance to stable foaming incidence that may occur during the AD (Fig. 5c). Foaming in each feedstock increased with increase in OL from 1.5 to 5.5 g VS/L. At low OL of 0.5 g VS/L, there was no foam formation in spite of comparable biogas production to those of higher OL which resulted to foaming. Therefore, increase in biogas production alone may not cause foaming. Feedstock F1 (C/N=17) was most prone to potential foaming incidence due to the presence of sufficient foam causing intermediate biosurfactants (e.g., VFAs) during the digestion. However, improved microbial activity due to the increase in feedstock C/N ratio to 30 significantly reduced the foaming by 60 %. Acknowledgments The authors would like to thank Universiti Putra Malaysia for the financial assistance and facilities through the Research University Grant Scheme (RUGS). We would also like to thank the management of FW treatment plant in Sri Serdang, Selangor, Malaysia, for supplying the sample.

References Eaton AD, Clesceri LS, Greenberg AE, Franson MAH, American Public Health Association, American Water Works Association, Water Environment Federation (2005) Standard Methods for the Examination of Water and Waste Water, 21st edn. APHA, Washington Boe K, Kougias PG, Pacheco F, O-Thong S, Angelidaki I (2012) Effects of substrates and intermediate compounds on foaming in manure digestion systems. Water Sci Technol 66:2146–2154 Bolzonella D, Pavan P, Battistoni P, Cecchi F (2005) Mesophilic anaerobic digestion of waste activated sludge: Influence of the solid retention time in the waste water treatment process. Process Biochem 40: 1453–1460 Dalmau J, Comas J, Rodriguez-Roda I, Pagilla K, Steyer JP (2010) Model development and simulation for predicting risk of foaming in anaerobic digestion systems. Bioresour Technol 101:4306–4314 Deshpande NS, Barigou M (2000) Mechanical suppression of the dynamic foam head in bubble column reactors. Chem Eng Process 39:207– 217

Appl Microbiol Biotechnol Forster-Carneiro T, Perez M, Romero L (2008) Influence of total solid and Inoculum contents on performance of anaerobic reactors treating food waste. Bioresour Technol 99:6994–7002 Frayer M, O’Flaherty E, Gray NF (2011) Evaluating the measurement of activated sludge foam potential. Water 3:424–444 Ganidi N, Sean T, Elise C (2009) Anaerobic digestion foaming causes-a review. Bioresour Technol 100:5546–5554 Ganidi N, Sean T, Elise C (2011) The effect of organic loading rate on foam initiation during mesophilic anaerobic digestion of municipal wastewater sludge. Bioresour Technol 102:6637–6643 Griffin ME, Mcmahon KD, Mackie RI, Raskin L (1998) Methanogenic population dynamics during start-up of anaerobic digesters treating municipal solid waste and biosolids. Biotechnol Bioeng 57:342–355 Griffiths P, Stratton H (2010) Foaming organisms in sewage treatment friend or foe: Victim of bad publicity. 35th Annual Qld Water Industry Operators Workshop, Rockampton, UK Hansen KH, Angelidaki I, Ahring BK (1998) Anaerobic digestion of swine manure inhibition by ammonia. Water Res 32:5–12 Heard J, Harvey E, Johnson B, Wells J, Angove M (2008) The effect of filamentous bacteria on foam production and stability. Colloids Surf Biointerf 63:21–26 Izumi K, Yu-ki O, Norio N, Chiaki N, Shuichi Y, Tatsuki T (2010) Effects of particle size on anaerobic digestion of food waste. Int Biodegrad 64:601–608 Jiang Y, Heaven S, Banks C (2012) Strategies for stable anaerobic digestion of vegetable waste. Renew Energy 44:206–214 Kougias PG, Boe K, Angelidaki I (2013) Effect of organic loading rate and feedstock composition on foaming in manure-based biogas reactors. Bioresour Technol 144:1–7 Kougias PG, Boe K, Tsapekos P, Angelidaki I (2014) Foam suppression in overloaded manure-based biogas reactors using antifoaming agents. Bioresour Technol 153:198–205 Lahav O, Morgan B (2004) Titration methodologies for monitoring of anaerobic digestion in developing countries-a review. J Chem Technol Biotechnol 79:1331–1341 Lee DH, Behera SK, Kim JW, Park HS (2009) Methane production potential of leachate generated from Korean food waste recycling facilities: A lab scale study. Wasman 29:876–882 Liu G, Zhang R, El-Mashad HM, Dong R (2009) Effect of feed to inoculum ratio on biogas yields of food and green wastes. Bioresour Technol 100:5103–5108

Marneri M, Mamais D, Koutsiouki E (2009) Microthrix parvicella and Gordona amarae in mesophilic and thermophilic anaerobic digestion systems. Environ Technol 30:5–15 Nitschke M, Pastore GM (2006) Production and properties of a surfactant obtained from Bacillus subtilis grown on cassava wastewater. Bioresour Technol 97:336–341 Pagilla KR, Craney KC, Kido WH (1997) Causes and effects of foaming in anaerobic sludge digesters. Water Sci Technol 36:463–470 Resch C, Worl A, Waltenberger R, Braun R, Kirchmayr R (2011) Enhancement options for the utilization of nitrogen rich animal by-products in anaerobic digestion. Bioresour Technol 102:2503– 2510 Ron EZ, Rosenberg E (2002) Bio-surfactants and oil bioremediation, Curr Opin Biotechnol 13:249–252 Seo JY, Heo JS, Kim TH, Joo WH, Crohn DM (2004) Effect of vermiculite addition on compost produced from Korean food waste. Waste Manag 24:891–987 Tanimu MI, Mohd Ghazi TI, Harun MR, Idris A (2014a) Effect of feed loading on biogas methane production in batch mesophilic anaerobic digesters treating food waste. Int J Chem Environ Eng 5:39– 41 Tanimu MI, Mohd Ghazi TI, Harun MR, Idris A (2014b) Effect of carbon to nitrogen ratio of food waste on biogas methane production in a batch mesophilic anaerobic digester. Int J Innov Manag Technol 5(2):116-119 Varley J, Brown AK, Boyd JWR, Dodd PW, Gallagher S (2004) Dynamic multipoint measurement of foam behaviour for a continuous fermentation over a range of key process variables. J Biochem Eng 20:61–72 Zeshan O, Karthikeyan CV (2012) Effect of C/N ratio and ammonia-N accumulation in a pilot-scale thermophilic dry anaerobic digester. Bioresour Technol 113:294–302 Zhang Y, Heaven S, Banks CJ (2012a) Strategies for stable anaerobic digestion of Vegetable waste. Renew Energy 44:206–214 Zhang Y, Banks CJ, Heaven S (2012b) Co-digestion of source segregated domestic food waste to improve process stability. Bioresour Technol 114:168–178 Zupancic GD, Uranjek-Zevart N, Ros M (2008) Full-scale anaerobic codigestion of organic waste and municipal sludge. Biomol Bioeng 32: 162–167

Effects of feedstock carbon to nitrogen ratio and organic loading on foaming potential in mesophilic food waste anaerobic digestion.

Foaming problem which occurred occasionally during food waste (FW) anaerobic digestion (AD) was investigated with the Malaysian FW by stepwise increas...
506KB Sizes 1 Downloads 8 Views