Instability mechanisms and early warning indicators for mesophilic anaerobic digestion of vegetable waste

Accepted Manuscript Instability mechanisms and early warning indicators for mesophilic anaerobic digestion of vegetable waste Dong Li, Lin Chen, Xiaofeng Liu, Zili Mei, Haiwei Ren, Qin Cao, Zhiying Yan PII: DOI: Reference:

S0960-8524(17)31214-2 http://dx.doi.org/10.1016/j.biortech.2017.07.098 BITE 18518

To appear in:

Bioresource Technology

Received Date: Revised Date: Accepted Date:

15 June 2017 18 July 2017 20 July 2017

Please cite this article as: Li, D., Chen, L., Liu, X., Mei, Z., Ren, H., Cao, Q., Yan, Z., Instability mechanisms and early warning indicators for mesophilic anaerobic digestion of vegetable waste, Bioresource Technology (2017), doi: http://dx.doi.org/10.1016/j.biortech.2017.07.098

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Instability mechanisms and early warning indicators for mesophilic anaerobic digestion of vegetable waste

Dong Lia, Lin Chena, Xiaofeng Liua, Zili Mei b*, Haiwei Renc, Qin Caoa, Zhiying Yana

a

Key Laboratory of Environmental and Applied Microbiology, Environmental

Microbiology Key Laboratory of Sichuan Province, Chengdu Institute of Biology, Chinese Academy of Sciences, Chengdu, 610041, China b

Key Laboratory of Development and Application of Rural Renewable Energy,

Ministry of Agriculture, Biogas Institute of Ministry of Agriculture, Chengdu, 610041, China c

School of Life Science and Engineering, Lanzhou University of Technology,

Lanzhou, 730050, China

* Corresponding author. Address: No. 13, Section 4, Renmin Nan Road, Chengdu, Sichuan, P. R. China Tel.:+86 28 85230701 Fax:+86 28 85230701 E-mail address: [email protected]

1

Abstract In order to elucidate the instability mechanism, screen early warning indicators, and propose control measures, the mesophilic digestion of vegetable waste (VW) was carried out at organic loading rates (OLR) of 0.5, 1.0, and 1.5 g volatile solid (VS)/(L·d). The process parameters, including biogas components, volatile fatty acids (VFA), ammonia, pH, total alkalinity (TA), bicarbonate alkalinity (BA), and intermediate alkalinity (IA), were monitored every day. Digestion was inhibited at OLR of 1.5 g VS/(L·d). The primary causes of instability are a high sugar and negligible ammonia content, in addition to the feed without effluent recirculation, which led to BA loss. The ratios of CH4/CO2, VFA/BA, propionate, n-butyrate and iso-valerate were selected as early warning indicators. In order to maintain the digestion of VW at a high OLR, control measures including effluent recirculation and trace element addition are recommended. Keywords: vegetable waste; anaerobic digestion; early warning; instability mechanisms; indicators

2

1. Introduction The China Statistical Abstract 2015 reported that 760 million tons of vegetables were produced in 2014 in China. More than 30% of vegetables were lost as vegetable waste (VW) during harvest, transportation, storage, marketing, and processing (Liu, 2000). It was estimated that more than 80% of VW were either dumped or transported to landfill sites and incineration plants with other municipal solid wastes (MSW) (Liu et al., 2012). Consequently, the high water content (>80%), high organic content (>95% dry basis), and easy biodegradability of VW have led to serious negative effects on the landfill and incineration systems, such as an abundant production of leachate from landfill and unstable combustion during incineration (Cheng & Hu, 2010). However, VW with these characteristics are suitable for anaerobic digestion (AD) for recovering energy. Many studies have confirmed that the specific methane yield of VW is higher than any other MSW (Gunaseelan, 1997). However, carbohydrate-rich substrates such as VW are easily perishable. They have a tendency to accumulate volatile fatty acids (VFA), which can lead to acidification, low pH, and process inhibition. Several studies indicate that the anaerobic digestion of VW can be stably operated only at low organic loading rates (OLR). A preliminary study by Knol et al. (1978) showed that the OLR was < 1.6 g volatile solid (VS)/(L·day) for the stable digestion of VW, producing average biogas yields of 0.30–0.58 L/ g VS. Mata-Alvarez et al. (1992) found that the maximum OLR that could be achieved was < 3 g VS/(L·d) for mesophilic single-stage digestion of fruit and vegetable wastes (FVW). These studies suggested that there might be

3

some instability in the anaerobic digestion of VW or FVW at a high OLR. In order to operate at high OLR, two stage digestion and co-digestion were used to treat VW(Shen et al., 2013; Zuo et al., 2014). Effective monitoring and diagnosis of processes is a great challenge for anaerobic digestion reactors, which limits their stable operation. The effective process parameters are the basis for ensuring the process monitoring and control. A number of parameters have been frequently adopted in the process monitoring, including pH, redox potential, ammonia, alkalinity, VFAs, biogas production rate, biogas composition, microbial community and activity (Boe, 2006). Howeover, most of these parameters can only reveal the current reactor status, but it is often too late for an effective process control once the threshold values are reached. In order to devolop an effective method for AD reactor diagnosis and risk prediction, intensive studies have focused on screening early warning indicators. Such indicators should be ideally accurate and sensitive to environmental fluctuations, reveal the change dynamics of reactor status, and be adaptive to online monitoring, autoalert, and control systems. Li et al. (2014) carried out AD of food waste using a continuous stirred tank reactor (CSTR), and concluded that a combination of total VFA, the ratio of VFA to total alkalinity (VFA/TA) and the ratio of bicarbonate alkalinity to total alkalinity (BA/TA) can reflect the metabolism of the digesting system and realize rapid and effective early warning. Dong et al. (2011) proposed two indexes, i.e., stability index S (Eq.1) and auxiliary index a (Eq.2), which incorporate both gas- and liquid-phase parameters for upflow anaerobic sludge blanket (UASB).

4

S =

a =

1 dVFA × × 100 QCH4 dt dQCH4

dt

(1)

( 2)

where (dVFA)/(dt) (mmol/L/h) is the change rate of total VFA concentration at time t and QCH (mmolCH4/L/h) is methane production rate. 4 Boe et al. (2010) considered that a combination of acetate, propionate and biogas production is an effective early warning indicator for CSTR using cattle manure as substrate. According to the running of a hybrid USBF (UASB +anaerobic filter) pilot plant (Molina et al., 2009a), the best indicators both for the carbohydrate-based and protein-based wastewaters, considering both process steady states and organic load perturbations are: methane flow rate and hydrogen concentration in the gas phase, volatile fatty acids and partial alkalinity in the liquid phase. Even though lots of researches on early warning indicators have been conducted, the proposed early warning indicators are only effective when applied to these specific substrates and operating conditions, while less effective in systems with different substrates and operation conditions. Unfortunately, there are no reports regarding the screening of early warning indcators for anaerobic digestion of VW using CSTR. In this study, early warning indicators were screened, instability mechanisms were analyzed, and control measures were recommended for stable mesophilic single-stage digestion of VW.

2. Material and methods 2.1 Substrates and inoculum 5

VW was obtained from the Chengdu HIGREEN wholesale vegetable market. Impurities such as plastic bags were manually removed. The main constituents of the VW were cabbage leaves, potatoes, lettuce leaves, Benincasa hispida, and pumpkins. The collected and sorted VW was chopped into particles 4–5 mm in size, uniformly mixed, and stored at 4°C. The characteristics of VW are listed in Table 1. The inoculum was the digested residue, filtered by a 1-mm sieve. The digested residue was taken from an anaerobic digester fed with pig manure. The inoculum was acclimated for 20 days using VW as a substrate, until the methane content was above 60%. The pH of the acclimated inoculum was 7.6.

6

Table 1 Composition of vegetable wastes and other materials used for co-digestion

Characteristics

VW

VW

VW

VW

FVW

CS

CM

PM

CS

CM

TS (g/kg)

105.7

54.0

91.2

77.4

167.0

118.5

375.0

334.3

125.4

249.5

VS (% of TS)

92.1

90.7

86.7

93.5

93.4

88.5

185.0

70.4

73.8

73.8

Carbon (% of TS) Nitrogen (% of TS) C/N Ammonia nitrogen (mg/kg) pH Carbohydrate (% of TS) Soluble sugar (% of TS) Crude fiber (% of TS) Crude fat (% of TS) Crude protein (kJ/kg TS)

44.5

41.5

35.5

47.8

-

-

-

55.8

49.9

47.6

2.64

3.8

2.9

4.3

-

-

-

4.8

3.7

8.7

17.1

10.9

12.2

11.0

-

-

-

11.7

13.6

5.5

0.2

-

-

-

＜10

1483

9900

1844

747

4891

5.60

-

-

5.8

4.2

7.8

7.3

-

-

-

73.6

-

67.5

-

-

-

-

-

-

-

63.0

-

60.0

-

-

-

-

-

-

-

11.9

30.2

7.5

23.88

-

-

-

-

-

-

3.4

-

-

-

-

-

-

-

-

-

13.2

-

-

-

-

-

-

-

-

-

(Verrier et

(Jiang et

(Callaghan

(Callaghan

(Callaghan

et al.,

et al.,

(Li et al.,

(Li et al.,

al., 2012)

et al.,

(Li et al.,

al., 1987)

2002)

2002)

2002)

2016)

2016)

2016)

(Zuo et

Ref.

This study

al., 2014)

Note：VW: vegetable waste; FVW: fruit and vegetable waste; KW: kitchen waste; CS: cattle slurry; CM: chicken manure; PM: pig manure.

7

2.2 Experimental setup and design The continuous experimental system consisted of three parts, including the digester, online gas monitoring, and online liquid monitoring (Fig.1). The total volume of the reactor was 70 L. To avoid pipe blockage due to the expansion of the digested material, the loading volume was restricted to 55 L. AD was carried out at 35 ± 2°C. The contents of the reactor were mixed 8 times per day at 40 rpm for 30 min. The mesophilic AD of VW was carried out at OLRs of 0.5, 1.0 and 1.5 g VS/(L·d). The hydraulic retention time (HRT) was fixed at 20 d by restricting the total feed to 2750 g, with different substrate concentrations. The daily feed and discharge for different OLRs are listed in Table 2. Table 2 Operating condition for the mesophilic digestion of vegetable waste Daily feed (g)

Running time

OLR

HRT

(d)

(g VS/(L·d))

(d)

VW

Water

(g)

Initial period

1–30

0.5

20

282

2468

2750

Stable period

31–60

1.0

20

564

2186

2750

Overloading period

61–90

1.5

20

846

1904

2750

Period

Daily discharge

2.3 Analytical methods The TS and VS were determined using standard techniques (APHA, 1998). Analyses for C and N were conducted using a Vario EL element analyzer (Elementar Analysensysteme GmbH, Germany). The compositions of carbohydrate, soluble sugar, crude fiber, crude lipid, and crude protein were determined based on the Chinese Standard (GB/T 5009-2003). Biogas production was monitored online by a gas flowmeter (Beijing Sevenstar Electronics Co., Ltd, China). Biogas components (0–100%

8

CH4, 0–100% CO2, 0–5% H2, 0–3% CO) were detected online by an automatic biogas analyzer (Wuhan Cubic Optoelectronics Co., Ltd, China). Liquid samples were centrifuged at 12000 rpm for 10 min and subsequently filtered with a 0.45 µm membrane filter to analyze the ammonia nitrogen level, alkalinity, and VFA. Ammonia nitrogen was analyzed using a DR-1900 spectrophotometer (HACH, USA). Total alkalinity (TA), partial alkalinity (PA), bicarbonate alkalinity (BA), and intermediate alkalinity (IA), were analyzed using a ZDJ-4B Automatic Potentiometric Titrator (Shanghai Precision & Scientific Instrument Co., Ltd, China), in accordance with Anderson and Yang (1992). The pH end-points for PA, IA, and TA titration were 5.75, 4.3, and 3.8. BA is PA multiplied by 1.2 (Anderson & Yang, 1992). Each VFA (including acetate, propionate, n-butyrate, iso-butyrate, n-valerate and iso-valerate) was analyzed on a gas chromatograph equipped with a DB-FFAP capillary column and a flame ionization detector (Dong et al., 2010).

3. Results and discussion 3.1 Feedstock characteristics Table 1 lists the characteristics of VW or FVW and other materials used for co-digestion. The C/N ratio of 17.1 is similar to that from other studies. Maintaining a suitable C/N ratio is essential for sustainable digestion, with the optimum being in the range of 25–30 (Hartmann & Ahring, 2006). The C/N ratio of both, vegetable waste and animal manure used as co-substrate, were less than 20. Therefore, maintaining the C/N ratio at 25–30 could not have been the promotion mechanism.

9

In general, hydrolysis is the rate-limiting step if the substrate is rich in cellulose (Veeken & Hamelers, 1999). However, the anaerobic degradation of cellulose-poor and sugar-rich wastes, like VW, is limited by methanogenesis rather than hydrolysis. Sugars can be easily degraded to VFA by fermentative bacteria, which would lead to a decrease in pH. Compared to fermentative bacteria, methanogens grow slowly and cannot tolerate low pH conditions. The rapid acidification of VW will stress and inhibit the activity of methanogenic bacteria. Sugar accounted for more than 60% of the total dry matter of the VW. Ammonia is the main source of alkalinity, reflecting the buffer capacity of the digestion system. An extremely low ammonia nitrogen concentration was found in VW. However, the animal manure usually used for co-digestion with VW has a higher ammonia nitrogen concentration. A preliminarily conclusion is that the instability during the mono-digestion of VW was the result of a high sugar and low ammonia content. Moreover, co-digestion with animal manure supplied the necessary ammonia nitrogen, which increased the buffer capacity.

3.2 Monitoring of process parameters 3.2.1 Biogas production Biogas production is the most common parameter measured. It can be expressed in terms of volumetric biogas production rate (VBPR) (L/(L·d)) and biogas yield (L/kg VS). The VBPR is shown in Fig. 2a. Since the HRT was 20 days, each OLR maintained steady-state operation during the latter 10 days. The average steady-state VBPRs for OLRs of 0.5 and 1.0 g VS/(L·d) were 0.15 and 0.28 L/(L·d), respectively. When the OLR increased to 1.5 g VS/(L·d), the VBPR sharply declined, with the biogas 10

production ceasing on day 89. The biogas yield is shown in Fig. 2b. The biogas yield was 240-310 L/kg VS during the steady-state period. 3.2.2 Biogas components During the starting and stable periods, the CH4 content was much higher than the CO2 content (Fig. 2c). The CH4 content stabilized at about 55%. During the overloading period, a sudden decrease in the CH4 content was observed on day 72. Subsequently, the CH4 content decreased, while the CO2 content increased, until the CH4 content was lower than the CO2 content. Hydrogen is important, both as an intermediate and an electron carrier, in the digestion process. Hydrogen metabolism involves fermentation, syntrophic acetogenesis, homoacetogenesis, syntrophic acetate-oxidizing, and hydrogenotrophic methanogenesis (Processes, 2002). The hydrogen concentration affects the thermodynamics and degradation pathway of the anaerobic process. High hydrogen concentration can inhibit VFA degradation, resulting in VFA accumulation. Thus, hydrogen accumulation has been suggested as an early warning indicator for process imbalance (Steyer et al., 2002). Carbon monoxide is a possible intermediate in the metabolic pathway of syntrophic acetogens, homoacetogens, syntrophic acetate-oxidizing bacterium and methanogens. It showed a good potential for indicating organic overloads in sewage sludge digester. The level of gaseous carbon monoxide has been reported to be directly related to the acetate concentration and inversely related to the methane concentration (Switzenbaum et al., 1990). In this study, no hydrogen or carbon monoxide was detected in the gas phase during the entire digestion process. This could be because hydrogen is formed in the

11

liquid phase. The sensitivity of hydrogen in biogas is also limited by the liquid-to-gas mass transfer rate (Bjornsson et al., 2001a; CordRuwisch et al., 1997; Kramer & Conrad, 1993; Strong & Cord-Ruwisch, 1995). Dissolved hydrogen should be monitored in future studies. 3.2.3 Volatile fatty acids (VFAs) VFAs, including acetate, propionate, n-butyrate, iso-butyrate, n-valerate, and iso-valerate, are products of hydrolysis and acidogenesis and substrates for methanogenesis. Stable VFA concentrations indicate a balance between hydrolysis/acidogenesis and methanogenesis. For stable AD, the VFA concentration stabilizes at a lower value, since VFAs produced from hydrolysis and acidogenesis can be consumed by methanogens over time. VFA accumulation during process imbalance directly reflects a kinetic uncoupling between the acid producers and consumers. The total VFAs concentration has been highly recommended for monitoring anaerobic digesters (Feitkenhauer et al., 2002; Molina et al., 2009b). However, several studies have pointed out that individual VFA concentrations can provide significant information as early warnings before process failure (Li et al., 2016; Li et al., 2015). The levels of iso-butyric and iso-valeric acids have been suggested as indicators of the stress level in advance of process failure (Cobb & Hill, 1991), while Ahring et al. (Ahring et al., 1995) concluded that the concentrations of n-butyric and iso-butyric acids would be better indicators. Propionic acid is known to be thermodynamically the most unfavorable to degrade; therefore, some authors used propionic acid as the sole process indicator (Hansson et al., 2002).

12

In this study, individual VFA`s were monitored everyday by gas chromatography, which is shown in Fig. 3. During the stable period, the acetate and propionate concentrations were maintained below 200 mg/L and 50 mg/L, respectively. The concentrations of n-butyrate, iso-butyrate, n-valerate, and iso-valerate were less than the detection limit. During the overloading period, the concentrations of acetate, propionate, n-butyrate, iso-butyrate, n-valerate, and iso-valerate suddenly increased on days 69, 70, 75, 75, 75, and 72, respectively. Detailed changes are listed in Table 3. During the overloading period, n-butyrate and iso-valerate were the main isomers of butyrate and valerate, respectively. Table 3 Significant changes in parameter values during the overloading period. Parameters

Day of sudden change (d)

Sudden change

Warning time* (d)

CH4 (%)

72

52→48

17

CO2 (%)

72

43→46

17

Acetate (mg/L)

69

99→312

20

Propionate (mg/L)

70

18→80

19

n-Butyrate (mg/L)

75

0→4

14

iso-Butyrate (mg/L)

75

0→10

14

n-Valerate (mg/L)

75

0→3

14

iso-Valerate (mg/L)

72

0→6

17

pH

85

6.27→6.18

4

ORP (mV)

85

-456→-450

4

TA

/

No

0

BA

/

No

0

IA

/

No

0

CH4/CO2

72

1.22→1.06

17

VFA/BA

69

0.07→0.19

20

VFA/TA

69

0.06→0.18

20

IA/BA

/

No

0

BA/TA

/

No

0

* Warning time equals the date on which the biogas production ceased (89 d) minus the date of sudden change.

3.2.4 Ammonia 13

Ammonia, including free ammonia (NH3) and its ionized form ammonium (NH4+), originate from the degradation of proteins, peptides, and amino acids. It is an important source of nitrogen for the growth of biogas-producing microorganisms, in addition to being a key pH-stabilizing molecule for the neutralization of VFAs. The concentration of ammonia nitrogen in VW was only 0.2mg/kg. In this study, the ammonia nitrogen concentration declined from day 11 (Fig. 4a) since there was no effluent recirculation, which resulted in the net loss of ammonia. 3.2.5 pH Each group of microorganisms has different optimal pH ranges. Methanogenic archaea can function within a very narrow pH range, with the optimal level being 6.5–8.0. Fermentative bacteria can function within a wider pH range of 4–8.5 (Hwang et al., 2004). The pH level also affects the acid-base equilibrium in the digester. In a mixed-culture anaerobic digester, the optimal pH range is 6.8–8.0 (Poschl et al., 2010). The use of pH as a process indicator is normally based on the fact that a pH drop corresponds to VFA accumulation. In this study, the pH value declined during the entire digestion process, slowly during initial stages but sharply from day 85 (Fig. 4b). Based on the evolution of VFAs, ammonia, and pH, it can be concluded that the decrease in pH during the starting and stable periods resulted from the loss of ammonia, rather than the accumulation of VFAs. 3.2.6 Oxidation-reduction potential (ORP) Oxidation-reduction potential reflects changes in oxidizing or reducing agents. In theory, the measurement of ORP would provide a good way of accurately describing

14

variations in the intermediate product composition in an anaerobic digester. Switzenbaum et al. (1990) found a good correlation between the ORP and VFA accumulation. In this study, ORP and pH were negatively correlated (Fig. 4b). The increase in ORP, as pH decreased, may be due to the reaction NADH+H + ↔ NAD+ +H 2 . However, the exact reason needs to be determined in the future, with studies on quantitatively determining the NAD+ concentration. 3.2.7 Alkalinity In an AD system, a change in pH does not necessarily reflect metabolic activity, due to the presence of strong buffers such as ammonia (NH3), bicarbonate (HCO3-), and ionized volatile fatty acid (VFA-). The buffering capacity reflects the capacity of the liquid to withstand acidification without lowering the pH. Alkalinity, which is an index of the buffering capacity, is a better alternative than pH for indicating VFA accumulation, since increased VFA will directly reduce the alkalinity before causing any major changes to the pH. TA normally includes both HCO3- and VFA-, since the sample is titrated to pH 3.8 (Anderson & Yang, 1992). In this study, TA proved to be insensitive, since increasing VFA concentrations can also lead to an increase in TA (Bjornsson et al., 2001b). By changing the end-point of the alkalinity titration to 5.75, it is possible to exclude IA from the alkalinity measurement, thereby truly characterizing the presence of bicarbonate. BA is empirically correlated to VFA accumulation (Hawkes et al., 1994). However, this relationship is not observed during VFA accumulation in response to an ammonia overload, as ammonia adds alkalinity to the system (Bjornsson et al., 2001b). In this study, the ammonia nitrogen in VW was negligible, while the ammonia

15

nitrogen in the digester continuously declined. The main buffer substances were HCO3and VFA- in the digestion system. As shown in Fig. 4c, the curves of BA and TA nearly overlapped before day 72, since the IA level was very low. After day 72, the increase in IA and the decrease in BA resulted in a stable TA level. During days 69–84, VFA accumulation did not result in a sharp decline in pH, since there was sufficient BA. However, VFA accumulation led to a sharp decline in pH when the level of BA was low (＜800 mg/L) during days 85–89. This result is in agreement with Murto’s viewpoint. In a low-buffer system, pH, BA and VFA measurements are useful for process monitoring, whereas in a highly buffered system only VFA is reliable for indicating process imbalances (Murto et al., 2004). The result proved once again that only BA was an effective indicator of the buffering capacity of an anaerobic digestion system.

3.3 Screening for early warning indicators Biogas production rate is an important parameter as it indicates the overall performance of the process. However, it cannot be used to indicate process imbalance, since changes in the biogas production rate depends on HRT, OLR, and feed composition. Moreover, it has low sensitivity to overloading compared to other process indicators, with a decrease in biogas production often occurring after the process is severely inhibited or already broken down. Therefore it is not an effective early warning indicator (Boe, 2006). In this study, biogas production ceased on day 89. The warning time was defined as the day on which the biogas production ceased minus the day of sudden change. The days of the sudden change and corresponding changes are listed in

16

Table 3 for all process parameters. Switzenbaum suggested the ratio of VFA/TA as a good indicator (Switzenbaum et al., 1990). Instead of single parameters, coupling parameters such as the ratios of CH4/CO2, VFA/BA, VFA/TA, IA/BA, and BA/TA were used as potential indicators for process imbalance in this study. Fig. 5 illustrated the evolution of coupling indicators. During the overloading period, it was difficult to observe a sudden change in TA, BA, IA, IA/BA, BA/TA. The early warning ability of pH and ORP was very poor. The warning times of acetate, propionate, VFA/BA, and VFA/TA were longer than those of CH4, CO2, CH4/CO2, and iso-valerate, as well as much longer than those of n-butyrate, iso-butyrate, and n-valerate. Although there was a sudden increase in acetate on day 69, it decreased from day 72. The unstable trend of acetate is the comprehensive result of acidogenesis, syntrophic acetogenesis, homoacetogenesis, syntrophic acetate-oxidizing, and hydrogenotrophic methanogenesis. Therefore, acetate is not suitable as an early warning indicator. Compared to iso-butyrate and n-valerate, n-butyrate and iso-valerate showed a greater response to inhibition. Since the slope of VFA/BA is steeper than that of VFA/TA, the ratio of VFA/BA was selected as an early warning indicator. The values of total VFA and BA could be determined in field laboratory using a rapid, simple, and accurate method proposed by Lahav et al. (2002) or by Wu et al (2015). In general, the CH4 and CO2 concentrations were monitored online by an automatic biogas analyzer in the biogas plant. When the AD system was acidified, the CH4 content would decrease while the CO2 content would increase. The ratio of CH4/CO2 had a greater response value due

17

to the superposition effect. For the engineering application of anaerobic digestion, the ratios of CH4/CO2 and VFA/BA were recommended as early warning indicators. With technological progress and cost reduction in the online determination of individual VFA by gas chromatography (Boe et al., 2007) or near-infrared spectroscopy (Jacobi et al., 2009), propionate, n-butyrate, and iso-valerate concentrations can be used as early warning indicators. The thresholds of instability and severe instability for the mesophilic digestion of VW, based on the early warning indicators, are listed in Table 4. Table 4 Proposed early warning indicators for the anaerobic digestion of vegetable waste Indicators

Threshold of instability

Threshold of severe instability

Propionate (mg/L)

> 80

> 200

n-Butyrate (mg/L)

> 10

> 150

iso-Valerate (mg/L)

> 10

> 50

CH 4/CO2

< 1.20

< 0.90

VFA/BA

> 0.15

> 0.40

3.4 Analysis of instability mechanism Fig. 6 illustrates the instability mechanism of the mesophilic digestion of VW. In this study, the sugar content of VW was very high, while its ammonia content was negligible. Moreover, the feed was diluted by water without effluent recirculation, leading to total ammonia loss. This resulted in BA reduction, since bicarbonate is the main source of bicarbonate alkalinity and ammonia contributes to bicarbonate. The consumption of bicarbonate increased the H+ concentration, based on the reaction equilibrium, and the decline in pH promoted a shift from ionized acetic acid (HAc-) to free acetic acid (HAc). Compared with ionized VFA, free VFA has been suggested as the main cause of inhibition, since it is freely membrane-permeable. Free VFA primarily 18

inhibits acetoclastic methanogens, although it has a lesser effect on hydrogenotrophic methanogens or acidogens (Kroeker et al., 1979). Before day 68, the increased free acetate gradually repressed the acetotrophic methanogens, even though the total acetate level remained stable. On day 69, the total acetate concentration increased since the acetotrophic methanogens were inhibited. Subsequently, propionate, butyrate and valerate started to accumulate due to the feedback inhibition on syntrophic VFA-oxidizing bacteria. The increase in free VFA would consume BA, producing equivalent IA. Bicarbonate thus consumed would be released to the gas phase as CO2; therefore, an increase in the CO2 content in biogas was observed after day 70. Moreover, the continuous decline in pH and increase in total VFA caused the total free VFA concentration to increase continuously. Eventually, acetotrophic methanogens, as well as hydrotrophic methanogens, were inhibited. In addition, the dilution by water also led to the loss of trace elements (TE), which were required by methanogens. For stable mesophilic single-stage digestion of VW, control measures, including effluent recirculation and TE addition, are recommended. The addition of TE, including Ni, Co, Fe, W, Cu, Mn, Zn, V, Mo, Se, could improve the activity of enzymes involved in anaerobic reactions and transformation, such as hydrogenase, CO-dehydrogenase, methyltransferase, and formate dehydrogenase (Fermoso et al., 2009). Recirculation could recover alkalinity and methanogens (Cavinato et al., 2011; Kobayashi et al., 2012).

4. Conclusions

19

VW rich in sugar content, with a negligible ammonia content, has a tendency to accumulate VFA, which can lead to acidification and methanogenesis inhibition during mesophilic single-stage mono-digestion. Dilution by water, without effluent recirculation, led to total ammonia loss; this resulted in BA reduction, eventually inhibiting both acetotrophic and hydrotrophic methanogens. For the engineering application of AD of VW, the ratios of CH4/CO2, VFA/BA, propionate, n-butyrate and iso-valerate are recommended as early warning indicators. In order to maintain the digestion of VW at a high OLR, control measures including effluent recirculation and trace element addition are also recommended.

Acknowledgements This research was supported by the National Natural Science Foundation of China (21476222), Youth Innovation Promotion Association CAS (2017423), Program of Strategic Resource Service Network CAS (ZSYS-009), the Agricultural Science and Technology Innovation Program (CAAS-ASTIP-2015- BIOMA), and the Key Laboratory of Environmental and Applied Microbiology, Chengdu Institute of Biology CAS (KLCAS-2016-10).

References 1.

Ahring, B., Sandberg, M., Angelidaki, I. 1995. Volatile fatty acids as indicators of process imbalance in anaerobic digestors. Applied Microbiology and Biotechnology, 43(3), 559-565.

2.

Anderson, G.K., Yang, G. 1992. Determination of bicarbonate and total volatile acid concentration in anaerobic digesters using a simple titration. Water Environment Research, 64(1), 53-59.

3.

APHA. 1998. Standard Methods for the Examination of Water and Wastewater. 21st Edition[R], American Public Health Association.

20

4.

Bjornsson, L., Hornsten, E.G., Mattiasson, B. 2001a. Utilization of a palladium-metal oxide semiconductor (Pd-MOS) sensor for on-line monitoring of dissolved hydrogen in anaerobic digestion. Biotechnology and Bioengineering, 73(1), 35-43.

5.

Bjornsson, L., Murto, M., Jantsch, T.G., Mattiasson, B. 2001b. Evaluation of new methods for the monitoring of alkalinity dissolved hydrogen and the microbial community in anaerobic digestion. Water Research, 35(12), 2833-2840.

6.

Boe, K. 2006. Online monitoring and control of the biogas process Technical University of Denmark.

7.

Boe, K., Batstone, D.J., Angelidaki, I. 2007. An innovative online VFA monitoring system for the anerobic process, based on headspace gas chromatography. Biotechnology and Bioengineering, 96(4), 712-721.

8.

Boe, K., Batstone, D.J., Steyer, J.-P., Angelidaki, I. 2010. State indicators for monitoring the anaerobic digestion process. Water Research, 44(20), 5973-5980.

9.

Callaghan, F.J., Wase, D.A.J., Thayanithy, K., Forster, C.F. 2002. Continuous co-digestion of cattle slurry with fruit and vegetable wastes and chicken manure. Biomass & Bioenergy, 22(1), 71-77.

10.

Cavinato, C., Bolzonella, D., Fatone, F., Cecchi, F., Pavan, P. 2011. Optimization of two-phase thermophilic anaerobic digestion of biowaste for hydrogen and methane production through reject water recirculation. Bioresource Technology, 102(18), 8605-8611.

11.

Cheng, H., Hu, Y. 2010. Municipal solid waste (MSW) as a renewable source of energy: Current and future practices in China. Bioresource Technology, 101(11), 3816-3824.

12.

Cobb, S.A., Hill, D.T. 1991. Volatile Fatty-Acid Relationships in Attached Growth Anaerobic Fermenters. Transactions of the Asae, 34(6), 2564-2572.

13.

CordRuwisch, R., Mercz, T.I., Hoh, C.Y., Strong, G.E. 1997. Dissolved hydrogen concentration as an on-line control parameter for the automated operation and optimization of anaerobic digesters. Biotechnology and Bioengineering, 56(6), 626-634.

14.

Dong, F., Zhao, Q.B., Li, W.W., Sheng, G.P., Zhao, J.B., Tang, Y., Yu, H.Q., Kubota, K., Li, Y.Y., Harada, H. 2011. Novel Online Monitoring and Alert System for Anaerobic Digestion Reactors. Environmental Science & Technology, 45(20), 9093-9100.

15.

Dong, L., Zhenhong, Y., Yongming, S., Longlong, M. 2010. Evaluation of pretreatment methods on harvesting hydrogen producing seeds from anaerobic digested organic fraction of municipal solid waste (OFMSW). International Journal of Hydrogen Energy, 35(15), 8234-8240.

16.

Feitkenhauer, H., von Sachs, E., Meyer, U. 2002. On-line titration of volatile fatty acids for the process control of anaerobic digestion plants. Water Research, 36(1), 212-218.

17.

Fermoso, F.G., Bartacek, J., Jansen, S., Lens, P.N.L. 2009. Metal supplementation to UASB bioreactors: from cell-metal interactions to full-scale application. Science of the Total Environment, 407(12), 3652-3667.

18.

Gunaseelan, V.N. 1997. Anaerobic digestion of biomass for methane production: A review. Biomass & Bioenergy, 13(1-2), 83-114.

19.

Hansson, M., Nordberg, A., Sundh, I., Mathisen, B. 2002. Early warning of disturbances in a laboratory-scale MSW biogas process. Water Science and Technology, 45(10), 255-260.

20.

Hartmann, H., Ahring, B.K. 2006. Strategies for the anaerobic digestion of the organic fraction of municipal solid waste: an overview. Water Science and Technology, 53(8), 7-22.

21.

Hawkes, F.R., Guwy, A.J., Hawks, D.L., Rozzi, A.G. 1994. Online Monitoring of Anaerobic-Digestion Application of a Device for Continuous Measurement of Bicarbonate Alkalinity. Water Science and Technology, 30(12), 1-10.

22.

Hwang, M.H., Jang, N.J., Hyun, S.H., Kim, I.S. 2004. Anaerobic bio-hydrogen production from ethanol fermentation: the role of pH. Journal of Biotechnology, 111(3), 297-309.

21

23.

Jacobi, H.F., Moschner, C.R., Hartung, E. 2009. Use of near infrared spectroscopy in monitoring of volatile fatty acids in anaerobic digestion. Water Science and Technology, 60(2), 339-346.

24.

Jiang, Y., Heaven, S., Banks, C.J. 2012. Strategies for stable anaerobic digestion of vegetable waste. Renewable Energy, 44(0), 206-214.

25.

Knol, W., Van Der Most, M.M., De Waart, J. 1978. Biogas production by anaerobic digestion of fruit and vegetable waste. A preliminary study. Journal of the Science of Food and Agriculture, 29(9), 822-830.

26.

Kobayashi, T., Xu, K.Q., Li, Y.Y., Inamori, Y. 2012. Effect of sludge recirculation on characteristics of hydrogen production in a two-stage hydrogen-methane fermentation process treating food wastes. International Journal of Hydrogen Energy, 37(7), 5602-5611.

27.

Kramer, H., Conrad, R. 1993. Measurement of Dissolved H-2 Concentrations in Methanogenic Environments with a Gas-Diffusion Probe. Fems Microbiology Ecology, 12(3), 149-158.

28.

Kroeker, E.J., Schulte, D.D., Sparling, A.B., Lapp, H.M. 1979. Anaerobic Treatment Process Stability. Journal Water Pollution Control Federation, 51(4), 718-727.

29.

Lahav, O., Morgan, B.E., Loewenthal, R.E. 2002. Rapid, Simple, and Accurate Method for Measurement of VFA and Carbonate Alkalinity in Anaerobic Reactors. Environmental Science & Technology, 36(12), 2736-2741.

30.

Li, D., Mei, Z.L., He, W., Yuan, Y.X., Yan, Z.Y., Li, J., Liu, X.F. 2016. Biogas production from thermophilic codigestion of air-dried rice straw and animal manure. International Journal of Energy Research, 40(9), 1245-1254.

31.

Li, L., He, Q., Ma, Y., Wang, X., Peng, X. 2015. Dynamics of microbial community in a mesophilic anaerobic digester treating food waste: Relationship between community structure and process stability. Bioresource Technology, 189, 113-120.

32.

Li, L., He, Q.M., Wei, Y.M., He, Q., Peng, X.Y. 2014. Early warning indicators for monitoring the process failure of anaerobic digestion system of food waste. Bioresource Technology, 171, 491-494.

33.

Liu, Q.S. 2000. Current situation and development trend of vegetable processing and circulation in China. Grain oil processing and food machinery (03), 1-4.

34.

Liu, X., Gao, X., Wang, W., Zheng, L., Zhou, Y., Sun, Y. 2012. Pilot-scale anaerobic co-digestion of municipal biomass waste: Focusing on biogas production and GHG reduction. Renewable Energy, 44, 463-468.

35.

Mataalvarez, J., Llabres, P., Cecchi, F., Pavan, P. 1992. Anaerobic-Digestion of the Barcelona Central Food Market Organic Wastes - Experimental-Study. Bioresource Technology, 39(1), 39-48.

36.

Molina, F., Castellano, M., Garcia, C., Roca, E., Lema, J.M. 2009a. Selection of variables for on-line monitoring, diagnosis, and control of anaerobic digestion processes. Water Science and Technology, 60(3), 615-622.

37.

Molina, F., Ruiz-Filippi, G., Garcia, C., Lema, J.M., Roca, E. 2009b. Pilot-Scale Validation of a New Sensor for On-Line Analysis of Volatile Fatty Acids and Alkalinity in Anaerobic Wastewater Treatment Plants. Environmental Engineering Science, 26(3), 641-649.

38.

Murto, M., Bjornsson, L., Mattiasson, B. 2004. Impact of food industrial waste on anaerobic co-digestion of sewage sludge and pig manure. Journal of Environmental Management, 70(2), 101-107.

39.

Poschl, M., Ward, S., Owende, P. 2010. Evaluation of energy efficiency of various biogas production and utilization pathways. Applied Energy, 87(11), 3305-3321.

40.

Processes, I.T.G.f.M.M.o.A.D. 2002. Anaerobic Digestion Model No.1 (ADM1). The International Water Association Publishing, London SW1H 0QS, UK.

41.

Shen, F., Yuan, H., Pang, Y., Chen, S., Zhu, B., Zou, D., Liu, Y., Ma, J., Yu, L., Li, X. 2013. Performances of anaerobic co-digestion of fruit & vegetable waste (FVW) and food waste (FW): Single-phase vs.

22

two-phase. Bioresource Technology, 144(0), 80-85. 42.

Steyer, J.P., Bouvier, J.C., Conte, T., Gras, P., Sousbie, P. 2002. Evaluation of a four year experience with a fully instrumented anaerobic digestion process. Water Science and Technology, 45(4-5), 495-502.

43.

Strong, G.E., Cord-Ruwisch, R. 1995. An in situ dissolved-hydrogen probe for monitoring anaerobic digesters under overload conditions. Biotechnology and Bioengineering, 45(1), 63-68.

44.

Switzenbaum, M.S., Giraldo-Gomez, E., Hickey, R.F. 1990. Monitoring of the anaerobic methane fermentation process. Enzyme and Microbial Technology, 12(10), 722-730.

45.

Veeken, A., Hamelers, B. 1999. Effect of temperature on hydrolysis rates of selected biowaste components. Bioresource Technology, 69(3), 249-254.

46.

Verrier, D., Roy, F., Albagnac, G. 1987. Two-phase methanization of solid vegetable wastes. Biological Wastes, 22(3), 163-177.

47.

Wu, S.B., Chen, X.Y., Liu, L., Guo, J.B., Dong, R.J. 2015. Automatic Titrator for Measurement of Total Volatile Fatty Acids and Total Inorganic Carbon in Scaled Biogas. Transactions of the Chinese Society for Agricultural Machinery, 46(9), 217-222.

48.

Zuo, Z., Wu, S., Zhang, W., Dong, R. 2014. Performance of two-stage vegetable waste anaerobic digestion depending on varying recirculation rates. Bioresource Technology, 162(0), 266-272.

23

Figure 1. 70 L Experimental set-up Figure 2. (a) Volumetric biogas production rate, (b) biogas yield and (c) biogas content Figure 3. Evolution of volatile fatty acids Figure 4. Evolution of (a) total ammonia nitrogen (b) pH, ORP and (c) alkalinity Figure 5. Evolution of coupling indicators Figure 6. Instability mechanism for the anaerobic digestion of vegetable waste

24

1. Feed inlet; 2. Heating layer; 3. Mixer; 4. Online pH meter; 5. Online ORP meter; 6. Online electrical conductivity meter; 7. Thermocouple probe; 8. Biogas outlet; 9. Biogas flowmeter; 10. Discharge outlet; 11. Instrument cabinet; 12. Automatic gas analyzer Figure 1

25

Volumetric biogas production rate (L/(L·day))

0.5

OLR=0.5 Starting period

OLR=1.0 Stable period

OLR=1.5 Overloading period

0.4

0.3

0.2

0.1

0.0

0

15

30

45

60

75

90

Time (day) (a) 500

OLR=0.5 Starting period

OLR=1.0 Stable period

OLR=1.5 Overloading period

Biogas yield (L/kg VS)

400

300

200

100

0 0

15

30

45

Time (day)

(b)

26

60

75

90

70

OLR=0.5

OLR=1.0

Starting period

OLR=1.5

Stable period

Overloading period

Biogas content (%)

60

CH4 CO2

50

40

30

20

0

15

30

45

Time (day) (c) Figure 2

27

60

75

90

1200

OLR=0.5

OLR=1.0

Starting period

OLR=1.5

Stable period

1000

1200

Overloading period

OLR=0.5

OLR=1.0

Starting period

Stable period

OLR=1.5 Overloading period

Propionate (mg/L)

Acetate (mg/L)

1000 800

600

400

200

0

350

30

Time (day)

60

400

Stable period

250

0

90

0

30

Time (day)

60

90

120

OLR=1.5

OLR=1.0

OLR=0.5 Starting period

300

Butyrate (mg/L)

600

200

OLR=0.5

OLR=1.0

Starting period

Overloading period

n-butyrate iso-butyrate

200

150

OLR=1.5

Stable period

100

Valerate (mg/L)

0

800

Overloading period

n-valerate iso-valerate

80

60

40

100 20

50 0

0

0

30

Time (day)

60

90

Figure 3

28

0

30

Time (day)

60

90

Total ammonia nitrogen (mg/L)

2000

OLR=0.5

OLR=1.0

Starting period

OLR=1.5

Stable period

Overloading period

1600

1200

800

400

0

0

15

30

45

60

75

90

Time (day) (a)

-375

OLR=0.5

OLR=1.0

Starting period

7.5

OLR=1.5

Stable period

Overloading period

pH ORP

pH

7.0

-400 -425

6.5

-450

6.0

-475

5.5

-500

5.0

0

15

30

45

Time (day) (b)

29

60

75

-525

90

ORP (mV)

8.0

OLR=0.5 Starting period

12000

OLR=1.0 Stable period

OLR=1.5 Overloading period

Total alkalinity Bicarbonate alkalinity Intermediate alkalinity

10000 8000 6000

2000

1600

1200

800

4000 400 2000 0

0

15

30

45

Time (day)

(c) Figure 4

30

60

75

0 90

Intermediate alkalinity (mg/L)

Total and bicarbonate alkalinity (mg/L)

14000

Ratio

4.0

OLR=0.5

3.6 1.8

OLR=1.0

Starting period

OLR=1.5

3.6 Overloading period 1.8

Stable period

1.6

1.6

1.4

1.4

1.2

1.2

1.0

1.0 CH4/CO2 BA/TA IA/BA VFA/BA VFA/TA

0.8 0.6 0.4

0.8 0.6 0.4

0.2 0.0

0.2 0

15

30

45

60

Time (day) Figure 5

31

75

0.0 90

Ratio

4.0

Figure 6

32

Highlights •

Vegetable waste was digested at different organic loading rates (OLR).

•

The digestion was inhibited and failed when the OLR increased to 1.5 g VS/(L·d).

•

The primary cause of instability was the loss of ammonia and bicarbonate alkalinity.

•

The ratios of CH4/CO2 and VFA/BA were selected as early warning indicators.

•

Effluent recirculation and TE addition are recommended for stable digestion.

33