Abstract
This report describes the results from anaerobic batch acidification of chicken manure as a mono-substrate studied under mesophilic conditions. The manure was diluted with tap water to prevent methane formation during acidification and to improve mixing conditions by reducing fluid viscosity; no anaerobic digester sludge has been added as an inoculum. Highest acidification rates were measured at concentrations of 10 gVS L−1 and 20 gVS L−1; the pH value remained high (pH 6.9–7.9) throughout the test duration and unexpected fast methane formation was observed in every single batch. At substrate concentrations of 10 gVS L−1 there was a remarkable methane formation representing a value of 82% of the respective biochemical methane potential of chicken manure. Increasing substrate concentrations did not supress methane formation but impaired acid production. Consequently, the liquor cannot be stored over longer periods but should immediately be used in a digestion process.
Introduction
During our research we investigated the batch acidification of chicken manure (CM) with the aim of producing a strong liquor containing high concentrations of volatile fatty acids (VFA) as a substrate for anaerobic digesters. CM is an interesting substrate for anaerobic digestion. With chicken farming being one of the most intensive operations in agriculture (Abouelenien et al., 2009), large amounts of CM with a considerable biogas potential are locally available. The manure is known for its high nitrogen content and the fermentation of CM has been described as difficult (Abouelenien et al., 2009). Therefore, Abouelenien et al. (2010) suggested ammonia removal to improve digestion conditions. Nonetheless, the toxicity of ammonia derived from CM, in reality, prevents the use of CM for digestion. Several recent publications, including Niu et al. (2013) and Fotidis et al. (2014), are dealing with this problem. Other authors, Safley et al. (1985), Webb and Hawkes (1985) and Bujoczek et al. (2000), have described in detail the digestion process of CM. They reported that the anaerobic digestion of CM is difficult at higher loadings of total solids (TS) (higher than 10% TS) and that an optimal concentration range is between 4% and 10% influent TS feed concentration. Webb and Hawkes (1985) suggested the optimisation based on a two-stage process. Current works from Fu and Holtzapple (2011), Liu et al. (2012), Yan et al. (2014) or Jie et al. (2014) give insight into the optimisation of acidification conditions in anaerobic digestion (AD) and the advantage of separately controlled acidification processes.
Although our work includes research into the two-stage process, the main focus is on the acidification step with CM as a mono-substrate. The key objective of this research was to produce high-strength liquor, rich in VFA by maximising the solubilisation of nutrients and production of VFA in a separately operated batch acidification step.
Material and methods
Substrate
CM was collected from a local poultry farm near Jena/Thuringia (Germany). The fresh CM, as described in Table 1, was dried at 40 °C in an oven with forced ventilation, ground in a ball mill, sieved through a 1 mm screen and thoroughly mixed to provide a homogeneous substrate. The drying procedure led to a 6.8% reduction of total nitrogen as compared with the fresh manure owing to ammonia losses. Table 1 shows substrate characteristics of the CM.
Chemical characteristics of dried CM (mean ± standard deviation).
The biochemical methane potential (BMP) of CM was measured in batch experiments at 37 °C and over 21 days in accordance with the German standard method of VDI 4630 (2006). The final methane yield was 200 L CH4/kg VS (corresponding to 56% methanisation of initially added chemical oxygen demand (COD)).
Batch experiments
Initial tests included short-term acidification experiments, which have been performed in 0.5 L SIMAX-bottles (duration between 4 to 5 days). Substrate concentrations were set up at 10, 20, 30, 40, 50, 60 and 100 gVS L−1. Corresponding amounts of CM were suspended in 0.5 L of tap water stirred by a magnetic stirrer and heated to 37 °C. The bottles were then flushed with nitrogen to ensure anaerobic conditions. Each bottle had been connected to an eudiometer (liquid displacement system) for measurement of biogas formation.
Investigations into the long-term conditions for acidification formed the second part of our research. The experimental procedure was similar to the short-term batch acidification, but with an extended incubation time of 41 days. The CM was again suspended in tap water with initial substrate concentrations of 10, 20 and 40 gVS L−1. The first acidification batches were set up without the addition of inoculation sludge, but this was changed for later batches, where we added 100 ml of the suspension from the previous batch (total volume 500 ml). Sampling occurred every 3 or 4 days through the bottle sampling port.
To ensure constant and accurate sampling conditions, gas and liquid samples were taken from two identical sets of bottles, one set for gas sampling and the other set for liquid sampling. This approach of sampling was chosen to avoid withdrawal of nutritions that could affect gas production rates. Biogas samples were taken from the eudiometers with a gas tight syringe and transferred into headspace vials displacing the barrier solution (saturated saline solution; pH 2).
Analytical methods
The wet samples were dried overnight at 105 °C (TS). Volatile solids (VS) content was estimated as the loss of ignition by dry matter combustion at 525 °C. For all pH measurements, a pH-meter (WTW, Germany) with a glass electrode (Schott/Germany) was used.
The concentration of individual volatile fatty acids (acetate, propionate, butyrate, isobutyrate, valeriate, isovaleriate and caproate) was determined by gas chromatography with a Shimadzu gas chromatograph/flame ionisation detector and equipped with a DB-1701 column (Macherey-Nagel, Germany). For determination of soluble COD, the samples were passed through a 0.45-µm-pore-size membrane filter. The COD of the filtrate was determined using a COD-Spectroquant test kit (Merck, Germany) and a digital photometer SQ 118 (Merck, Germany). Gas composition (CH4, CO2) was analysed with a Combimass GA-m (Bender, Germany) multi-gas monitor.
Results and discussion
Optimal CM concentration during acidification
The amount of total volatile fatty acids (TVFA) produced from dry CM was optimised in a first acidification step. Several initial substrate concentrations were investigated (Figure 1). Optimal acid production occurred at concentrations of between 10 gVS L−1 and 20 gVS L−1. The rate of acid formation decreases substantially at low concentrations. CM-concentrations around 10 gVS L−1 show a four times higher acidification rate than the initial CM-concentration of 1 gVS L−1 (a decrease from 2.8 mgTVFA gVS−1*h for the 10 gVS L−1 to 0.75 mgTVFA gVS−1*h for the 1 gVS L−1). At higher concentrations than 20 gVS L−1, the acidification rate decreases successively and drops to 1.28 mgTVFA gVS−1*h at a concentration of 100 gVS L−1.
Characterisation of the acidogenesis: Acid-production (TVFA) per hour under conditions of different concentrations of VS. TVFA are given as acetic acid equivalent (AC).
Although unexpected, we observed already during the first days methane formation. Therefore the methane formation during batch acidification was investigated in more detail in subsequent experiments.
Characterisation of the produced VFA-liquor
During long-term batch experiments, the hydrolysis and acidification of CM were investigated at a concentration of 10 gVS L−1 (equivalent to 12870 mg L−1 of COD). CM was rapidly degraded to VFA under anaerobic conditions. Maximum TVFA concentration (4663 mg L−1) was achieved at Day 4–5, indicating an acidification efficiency of 46% of COD (Figure 2(a)). Acetic acid was the predominant VFA species (3400 mg L−1) followed by comparably low concentrations of propionate and butyrate (<500 mg L−1). The maximum TVFA concentration coincided with the peak of soluble COD (4680 mg L−1) (Figure 2(c)). In total, 36.4% of the initially added COD was solubilised. Another parameter monitored was the reduction of organic nitrogen to ammonium (NH4-N) (Figure 2(b)). The conversion took place within the first 3–5 days, with the final concentration close to total-N concentration in the feed. These data indicate a rapid and nearly complete degradation of organic nitrogen into soluble NH4-N. This opens the possibility for a subsequent ammonia removal from the digestate prior to full methanisation.
Long-term batch digestion of CM at an initial concentration of 10 g L−1 VS. (a) Acid production, (b) total nitrogen and NH4-N, (c) solubilised COD and (d) pH value.
In spite of intense VFA formation from CM, the pH of the batches did not drop below pH 6.9 (Figure 2(d)). The stabilisation of pH in a neutral range can give rise to methanogenic activity counteracting the accumulation of VFA. Hence, after 5 days TVFA and soluble COD decreased slowly, reaching values as low as a third of the respective maximum value (Figure 2(a) and (c)) at Day 26. It has been concluded that methane formation was the cause for VFA reduction.
Methane formation during acidification
Gas production during batch CM acidification started at Day 2, followed by a short lag-phase until Day 5, indicating a diauxic growth curve (Figure 3). At a concentration of 10 gVS L−1 the continued anaerobic incubation of batches led to continuous production of methane starting on Day 2 and accelerating on Day 6. On Day 41, a cumulative methane production of 189 L CH4/kg VS was measured. These values approximate the BMP of 200 L CH4/kg VS. The results indicate that under the chosen conditions of CM batch acidification, a significant methane production will occur. The early appearance of methane during the acidification process interferes with VFA accumulation and is expected to prevent a temporary storage of the liquor rich in VFA at ambient temperatures. Additional attempts to stabilise TVFA-accumulation and to supress methane formation failed. Using tap water instead of anaerobic digester sludge (ADS) for inoculation of batches was not sufficient to inhibit methane formation. Higher CM-concentrations led to impaired acidification and a delayed gas production, but did neither prevent methane formation. This is consistent with the observation of Bujoczek et al. (2000), who described 40 days of lag phase for digestion of undiluted CM. We concluded that elevated buffer capacity (ammonium and carbonate buffer systems) led to pH stabilisation of the liquid phase with values above pH 7, and these conditions were sufficient to maintain methanogenic activity in the VFA-liquor.
Analysis of the gas-production during batch-acidogenesis: (a) biogas formed per kgVS; (b) methane formed per kgVS.
Conclusions
The results from our trials show optimal substrate concentration for CM batch acidification between 10 and 20 gVS L−1 with an maximum acidification rate of 2.8 mgTVFA gVS−1*h (acidification yields of 47%). The highest TVFA accumulation in the liquid phase corresponds to 450 mgVFA gVS−1. A rapid hydrolysis of CM into soluble products is indicated by peaks in soluble COD, TVFA and the complete conversion of organic nitrogen into ammonia-N at Day 5. However, after Day 5, soluble COD and TVFA decrease significantly, accompanied by increasing methane formation, leading to consumption of VFA by methanogens. After 41 days, most of the CM is consumed for biogas formation. The approach to produce VFA-liquor from CM in an anaerobic batch process was successful with respect to rapid hydrolysis and acidification, and the use of an ADS-free process. However, the system failed in stabilising the accumulated VFA in the liquid phase and preventing methanogenic conditions. Therefore, the liquor should immediately be used in a digestion process.
Footnotes
Acknowledgements
We thank Manfred Schmid and Ursula Kepp for proofreading of the article and William Than SP (SP Multitech Malaysia) for the collaboration and regular access to his CM fed two-stage plant.
Declaration of conflicting interests
The authors declare that there is no conflict of interest.
Funding
This work was supported by a grant of the German Federal Ministry of Economics and Technology [ProInno II, grant VP2531901ST9 / ZIM, grant 16KN017626].