In a co-digestion system running with rapeseed oil and sewage sludge, an extremely fast increase in the organic loading rate was studied to develop a procedure to allow for flexible and demand-driven energy production. The over-acidification of the digestate was successfully prevented by calcium oxide dosage, which resulted in granule formation. Mineralogical analyses revealed that the granules were composed of insoluble salts of long chain fatty acids and calcium and had a porous structure. Long chain fatty acids and calcium formed the outer cover of granules and offered interfaces on the inside thereby enhancing the growth of biofilms. With granule size and age, the pore size increased and indicated degradation of granular interfaces. A stable biogas production up to the organic loading rate of 10.4 kg volatile solids m−3 d−1 was achieved although the hydrogen concentration was not favorable for propionic acid degradation. However, at higher organic loading rates, unbalanced granule formation and degradation were observed. Obviously, the adaption time for biofilm growth was too short to maintain the balance, thereby resulting in a low methane yield.

INTRODUCTION

The digestion process relies on the degradation of organic materials under anaerobic conditions by acidogenic, acetogenic and methanogenic microorganisms and leads to the formation of biogas. The process requires low energy input for operation (Steyer et al. 2002). The different groups of microorganisms differ widely in nutritional needs, growth kinetics and sensitivity to environmental conditions (Pohland & Ghosh 1971). The lack of balance between those groups of microorganisms caused by organic overloads is the primary reason for process failure, i.e., so-called over-acidification (Demirel & Yenigün 2002). Due to the risk of process destabilization and failure, organic loading rates (OLRs) for continuously stirred tank reactors (CSTRs) used in practice are still quite low and do not exceed 4 kg of volatile solids (VS) m−3 d−1 (Dupla et al. 2004; Deublein & Steinhauser 2008).

The increase in efficiency of wastewater treatment through the formation of granules is one of the most important phenomena discovered in upflow anaerobic sludge blanket (UASB) reactors. It has been noted that biological sludge granulation has a positive impact on the process stability in wastewater treatment (Hulshoff Pol et al. 2004) or on process recovery in the case of failure during biogas production in a CSTR after the addition of calcium oxide (CaO) (Kleyböcker et al. 2012a). It was discovered that conditions inside a UASB reactor favor microbial biomass accumulation. The adhering microorganisms are active in the formation of granules, which are typically a few millimetres in size. The microorganisms that are attached to the surface are exposed to high concentrations of nutrients within the reactor (MacLeod et al. 1990). The close association of different groups of microorganisms living in syntrophic pathways ensures a high level of metabolic activity, which is essential for process stability (Demirel & Yenigün 2002) and is required for syntrophic degradation (Boone & Bryant 1980).

Granulation processes described for UASB reactors are dependent on many other factors, such as wastewater composition, availability of nutrients, pH, alkalinity, temperature, stirring intensity, and thus shear force (e.g., Tay & Yan 1996). In a previous study, we observed sustainable biogas process recovery by CaO addition after the excess accumulation of fatty acids during the co-digestion of sewage sludge and rapeseed oil in a CSTR (Kleyböcker et al. 2012a). CaO was successfully used to stabilize the biogas process while maximizing the space-time yield by moderately increasing the OLR to 9.5 kg VS m−3 d−1 (Kleyböcker et al. 2014). CaO dosage was regulated using an early warning indicator (EWI), which is defined by the ratio of the concentrations of volatile fatty acids (VFAs) and calcium (Kleyböcker et al. 2012b). We assumed that calcium formed insoluble salts with long chain fatty acids (LCFAs) and phosphate. Hereby, the phosphate was very likely released from phosphorus-accumulating organisms (PAOs) during their uptake of VFAs (Kleyböcker et al. 2012a). We speculated that the precipitation of LCFA-Ca favored the accretion of granules and formation of microhabitats. It may be an interesting option to prevent the disturbance of the biogas formation as suggested by Koster (1987).

The aim of this study was to increase the OLR every 2 to 3 days, which is three times as fast as in Kleyböcker et al. (2014), to develop a procedure for a demand-driven biogas production and to study its unfavorable side effects. Thus, a four-fold increase in the OLR with regard to typical OLRs used at full-scale CSTR plants within three hydraulic retention times was applied, and large amounts of CaO were added to investigate the role of granules on biofilm formation and on process stability and efficiency.

Granules were extracted from the digestate and subjected to detailed characterization to study their structure and mineral composition using scanning electron microscopy (SEM), with energy dispersive spectrometry (EDS), X-ray powder diffraction (XRD), and Fourier transform infra red (FTIR) spectroscopy.

MATERIALS AND METHODS

Laboratory-scale reactor

The reactors contained 23 L of sludge. The operational temperature was 50 °C, and the sludge was mixed pneumatically. The substrates were excess and primary sludge with 28 g VS L−1 from a wastewater management plant applying enhanced biological phosphorus removal and rapeseed oil. The OLR was increased from 4.9 to 14.1 kg VS m−3 d−1, as described by Kleyböcker et al. (2014). The hydraulic retention time ranged between 17 and 20 days. As a response to the warning of the EWI, CaO was dosed seven times. In addition, CaO was dosed 34 times as a preventive measure to stabilize the process of biogas formation when the OLR was increased or the methane yield decreased. The pH was kept constant at a neutral level. Prior to analysis, granules were separated manually from the digested sludge and dried under the infrared radiator for 12 hours.

A more detailed description of the laboratory reactor and the sewage sludge is given by Kleyböcker et al. (2012a).

Wet chemical and gas analyses

The temperature and pH of the digested sludge were measured each day. Total solids and VS were analyzed according to the German guideline DIN 38409-1. Total VFAs (LCK 365), phosphate (LCK 350), and Ca (LCK 327) were analyzed photometrically (Hach-Lange DR2800), while the short chain fatty acids were determined by ion chromatography. The proportion of LCFA was estimated by subtracting the sum of short chain fatty acids in acetate equivalents from the total VFAs. The gas composition was analyzed by gas chromatography. The measured gas components were hydrogen, oxygen, nitrogen, methane, and carbon dioxide.

To determine the expected methane yield, the methane yield of each substrate was multiplied with its OLR and the sum of these products was divided by the total OLR (Equation (1)). 
formula
1

For the upper and lower limits, the highest values from the literature and the values from our own measurements were taken, respectively (Table 1).

Table 1

Values used to calculate the range of the expected methane yield

Substrate Biogas yield [m³ (kg VS)−1CH4 content [%] Literature/Source 
Sewage sludge 0.2–0.75 65–75 Deublein & Steinhauser (2008)  
 0.53 68 Own measurement 
Rapeseed oil 1.435 70 Buswell & Müller (1952)  
Substrate Biogas yield [m³ (kg VS)−1CH4 content [%] Literature/Source 
Sewage sludge 0.2–0.75 65–75 Deublein & Steinhauser (2008)  
 0.53 68 Own measurement 
Rapeseed oil 1.435 70 Buswell & Müller (1952)  

Because the expected methane yield for rapeseed oil is higher than for sewage sludge, the total expected methane yield of both substrates increased according to the increase in the rapeseed oil fraction. More detailed descriptions of the wet chemical and gas analyses are given by Kleyböcker et al. (2012a).

Methods used for granule characterization

To identify the main mineral components a Philips X'Pert diffractometer (APD type) with a PW 3020 vertical goniometer was used. The analytical range for XRD analyses was 2–64°θ, Cu Kα radiation, a step size of 0.02° and a time of 1 s/step. For phase identification, the Philips X'Pert Graphics and Identify software with the PDF2 database was used.

The identification of mineral components with a low degree of crystallinity and the general identification of organic compounds by determining their functional groups were performed using FTIR spectroscopy with a BioRad FTS 135 spectrometer. FTIR spectra were recorded in the range of 4,000–400 cm−1 with a 2 cm−1 resolution. The interpretation of the spectra was based on the literature focusing on sludge characterization by FTIR, notably the works of Ellerbrock & Kaiser (2005), Christy & Egeberg (2006), Huang et al. (2006), and Pokorna et al. (2009).

To characterize the inner structure and the spatial relationship between the components of granules and biofilms, both Hitachi S-4700 and Ultra 55 Plus (Carl Zeiss SMT) SEM microscopes were used. The quantitative analyses of elements were performed with EDS and identified using the analytical software Thermo Noran NSS. The samples were coated with carbon or gold and examined with an SEM operating at accelerating voltages of 15 and 20 kV, using secondary electrons and backscattered electrons signals.

RESULTS AND DISCUSSION

The EWI gave a warning five times during the first three weeks (Figure 1(a)). Four times, the CaO was added as a countermeasure, and, one time, the OLR was not further increased to prevent process failure. When the OLR was increased every 2 to 3 days, CaO was added daily, and the dosage was increased to prevent over-acidification. Until the OLR was below 6.0 kg VS m−3 d−1, the methane yield was in the expected range (Figure 1(b)), and it then decreased to 43% of its expected value. After the CaO addition was tripled, the methane yield increased to 70% of its expected value for the OLR of 10.4 kg VS m−3 d−1. The further increase in the OLR resulted in a methane yield decrease to 21%. A temporary interruption of CaO dosage on day 54 led to a seven-fold increase in the EWI, thus indicating the necessity of adding CaO.

Figure 1

(a) OLR increase and CaO dosages, depending on the trend of the EWI; (b) comparison of the measured methane yield with its expected range at an increasing OLR; (c) fatty acid concentration below critical values for over-acidification; (d) phosphate and calcium concentration.

Figure 1

(a) OLR increase and CaO dosages, depending on the trend of the EWI; (b) comparison of the measured methane yield with its expected range at an increasing OLR; (c) fatty acid concentration below critical values for over-acidification; (d) phosphate and calcium concentration.

In particular, after multiple dosing of CaO, LCFAs decreased in their concentration. The methane yield was very low at OLRs above 12 kg VS m−3 d−1, whereas the VFA concentration remained below 1,300 mg L−1 (Figure 1(c)) and no propionic acid accumulated. However, the hydrogen partial pressure was two to 13 times too high to allow for the degradation of propionic acid under standard conditions.

Until the CaO dosage was tripled, the phosphate concentration ranged between 200 and 280 mg L−1 and the calcium concentration remained at a low level of 14 mg L−1, although CaO was charged. Then, the phosphate concentration decreased to 60 mg L−1 and the calcium concentration in the liquid phase increased to almost 200 mg L−1. During the further increase in the OLR, the calcium concentration decreased again, first to a range between 100 and 150 mg L−1 and later on to almost 50 mg L−1 (Figure 1(d)). In the digestate, particularly after the CaO dosage was tripled, small granules were found.

After 64 days of operation, the reactor was opened. The liquid phase had been separated from the solid phase. Several large granules, ca. 10 cm in diameter, remained in the reactor because they were too large to leave the reactor through its outlet during operation.

The time intervals of 2 to 3 days between the OLR increases were very short compared to 8 days in a prior experiment (Kleyböcker et al. 2014). Although the adaption phase to higher OLRs was drastically shortened and the system was overloaded with rapeseed oil, the process of biogas formation was stable because of the high CaO dosages. Due to an imbalance in the production of LCFAs and their conversion to biogas, the excess LCFAs were not degraded, instead they precipitated with calcium and enriched in the solid phase when CaO was dosed. Thus, the inhibitory effect of LCFAs (Angelidaki & Ahring 1992) was significantly reduced, and the VFAs thus did not increase, despite the exceedingly high OLRs. The pH remained neutral due to the CaO additions and LCFA precipitation and thus the milieu was favorable for acid degradation and LCFA conversion to biogas. In a previous study, we showed that the precipitated LCFAs were degraded to biogas during process stabilization (Kleyböcker et al. 2012a). The reduced methane yield in this experiment revealed that the adaption phase of 2 to 3 days between two OLR increases was too short to establish the required amount of biofilms on the granule surfaces to degrade the majority of precipitated LCFAs. A suboptimal relation between the formation and degradation of granules led to the reduced methane yield and, after day 45, to a separation of the solid and liquid phases in the reactor. Additionally, the very high content of granules disturbed the mixing of the reactor contents. Thus, the substrate was not distributed adequately in the reactor, and the methane yield subsequently decreased drastically to 21% at an OLR of 14.1 kg VS m−3 d−1. Thus, at that time, 79% of the organic fraction was retained in the solid phase corresponding to the observed precipitation. The low methane yield, despite favorable VFA and LCFA concentrations in the liquid phase, can be considered further proof of the need to limit the calcium dosage. Consequently, time intervals between OLR increases should be longer than 3 days to achieve a lower demand of CaO dosage. Kleyböcker et al. (2014) showed that 8 days was sufficient to stabilize the process with moderate CaO additions only. Thus, sufficient time will be offered for biofilm growth to allow for LCFA degradation and a high methane yield. However, the space-time yield was twice as high as expected for typical OLRs in practice, although the methane yield was significantly reduced. The fast increase to 10.4 kg VS m−3 d−1 and the relatively high methane yield of 73% showed that a demand-driven energy production can be run more flexibly if the process is regulated by the EWI and CaO dosage.

Macroscopically, granules were white to yellowish in color. The diameter of granules increased with the increase in the OLR and CaO dosage and varied from 1 mm to 1.5 cm (Table 2). Granules were delicate and soft with low resistance to crushing, which corresponded to low density and high porosity. They were usually rounded in shape but were often broken, layered, and composed of tiny needle-like material, which surrounded dense fragments of organic material.

Table 2

Granule and pore size during the experimental run

Time (d) Range of size (mm) Average granule size (mm) Range of size (μm) Average pore size (μm) 
10 1.0–1.1 1.0 0.2–2 1.0 
20 2.3 2.3 0.2–17.3 4.1 
32 1.7–2.7 2.2 0.25–4.7 2.3 
40 4.0 4.0 1.4–13.2 5.1 
44 10–15 12.5 33–35 32.0 
45 1.3–2.1 1.7 0.4–6.2 2.0 
Time (d) Range of size (mm) Average granule size (mm) Range of size (μm) Average pore size (μm) 
10 1.0–1.1 1.0 0.2–2 1.0 
20 2.3 2.3 0.2–17.3 4.1 
32 1.7–2.7 2.2 0.25–4.7 2.3 
40 4.0 4.0 1.4–13.2 5.1 
44 10–15 12.5 33–35 32.0 
45 1.3–2.1 1.7 0.4–6.2 2.0 

SEM imaging of the surface of granules showed their porous structure, with a pore size differing over a wide range throughout the time of the experiment. The average size of granules and pores is listed in Table 2. Figure 2 shows an example of the porous surface of the granule after 40 days of experiments. On the surface of granules, also cracks and fissures were present (Figure 2(b)).

Figure 2

Granule surface. (a) Example of the granule surface with numerous pores; (b) surface of a granule, with pores and cracks, composed of organic substance which covers microorganisms.

Figure 2

Granule surface. (a) Example of the granule surface with numerous pores; (b) surface of a granule, with pores and cracks, composed of organic substance which covers microorganisms.

The relationship between the size of granules and the size of pores was noted. The larger the granules were, the larger the pore size was (Figure 3). On day 45, when the calcium loading rate reached its highest value, a very large granule appeared with a size of ca. 12.5 mm and an average pore size of 32 μm. It should be noted that after day 45, the VS in the digestate were approximately 80% lower than expected. However, the methane yield showed that the organic fraction was only degraded to 21%; the rest was precipitated and retained in the reactor. Consequently, the average granule size after day 45 was bigger than before. The increase in the pore diameter with granule size and subsequently also the granule age are considered indicators for the advanced degradation of the internal granule components, as suggested by Ahmad et al. (2011), occurring as flakes. Within granules, the gas was released through pores. Thus, the pores themselves favored mass transfer by causing flow processes that supported the supply of the biofilms with substrate and nutrients.

Figure 3

Ratio of the average granule size and the pore size.

Figure 3

Ratio of the average granule size and the pore size.

The main minerals determined in the granules using XRD were sal ammoniac (NH4Cl), halite (NaCl), calcite (CaCO3), mica (X2Y4–6Z8O20(OH, F)4) (X: K, Na, or Ca; Y: Al, Mg, Fe, Mn, Cr, Ti, or Li, etc.; Z: Si or Al), quartz (SiO2) and natrolite (Na2Al2Si3O10·2H2O). Peaks in the low-angle range up to 10°2θ were related to fatty acids or other organic compounds, though precise identification was not possible (Figure 4). The functional groups of organic compounds and inorganic components were determined by FTIR (Figure 5). The location of indicator bands and their assignment to functional groups are listed in Table 3.

Table 3

Location of indicator bands in granules and their assignment to functional groups based on the literature

Location wavenumber (cm−1Vibration and functional group or component 
3,400 stretching vibrations of O—H (OH groups or water molecules) (Pokorna et al. 2009)  
3,000–2,800 C—H stretching vibrations of aliphatic groups of long carbon chains (Pokorna et al. 2009)  
1,740 C=O typical for carboxylic acids (Ellerbrock & Kaiser 2005)  
1,640 C=O stretching vibration of COO, or it can be related to adsorbed water (Ellerbrock & Kaiser 2005)  
1,540–1,570 N—H vibrations (Pokorna et al. 2009)  
1,470–1,430 stretching and as well as deformation bands of C—H (Pokorna et al. 2009)  
1,380 CH (CH3) symmetric bending (Christy & Egeberg 2006)  
1,120–1,050 C—O stretching vibrations of C—O—C groups or Si—O stretching and Si—O—Si vibrations related to the presence of alumina-silicates as quartz, clay minerals, and silica (Ellerbrock & Kaiser 2005)  
720 CH2 stretching vibrations (Pokorna et al. 2009)  
471 Si–O–Si bending vibration (Huang et al. 2006)  
Location wavenumber (cm−1Vibration and functional group or component 
3,400 stretching vibrations of O—H (OH groups or water molecules) (Pokorna et al. 2009)  
3,000–2,800 C—H stretching vibrations of aliphatic groups of long carbon chains (Pokorna et al. 2009)  
1,740 C=O typical for carboxylic acids (Ellerbrock & Kaiser 2005)  
1,640 C=O stretching vibration of COO, or it can be related to adsorbed water (Ellerbrock & Kaiser 2005)  
1,540–1,570 N—H vibrations (Pokorna et al. 2009)  
1,470–1,430 stretching and as well as deformation bands of C—H (Pokorna et al. 2009)  
1,380 CH (CH3) symmetric bending (Christy & Egeberg 2006)  
1,120–1,050 C—O stretching vibrations of C—O—C groups or Si—O stretching and Si—O—Si vibrations related to the presence of alumina-silicates as quartz, clay minerals, and silica (Ellerbrock & Kaiser 2005)  
720 CH2 stretching vibrations (Pokorna et al. 2009)  
471 Si–O–Si bending vibration (Huang et al. 2006)  
Figure 4

Example of X-ray pattern of a granule (CuKα). Sal – sal ammoniac; Qz – quartz; Mi – mica; Cal – calcite; Hl – halite; Ntr – natrolite; FA – organic components.

Figure 4

Example of X-ray pattern of a granule (CuKα). Sal – sal ammoniac; Qz – quartz; Mi – mica; Cal – calcite; Hl – halite; Ntr – natrolite; FA – organic components.

Figure 5

Example of FTIR spectra of a granule.

Figure 5

Example of FTIR spectra of a granule.

Granules were primarily composed of organic matter. SEM-EDS analysis revealed high C content with up to 30 wt% which was also proved by the presence of XRD peaks in the low-angle position, indicating a high content of organic carbon. The bands obtained in FTIR analysis within the range of 3,000–2,800 cm−1, 1,470–1,430 cm−1 and 722 cm−1 are typical for hydrocarbon chains (C–H stretching vibrations, deformation and CH2 vibrations stretching bands, respectively) that are characteristic of structural moieties of fatty acids (Pokorna et al. 2009). EDS spot analyses also indicated the presence of calcium on the surface of granules and in the inner part. SEM-EDS analyses showed a spatial relationship between calcium and organic carbon. The organic compounds are interpreted as LCFA and/or EPS on the surface and in the inner part of granules. These results are consistent with those obtained by Liebrich et al. (2012), who considered the fluorescence microscopy of granules formed in a similar co-digestion experiment with the same sludge and co-substrates. Granules were stained for calcium, oil and/or LCFA, and proteins and indicated a strong spatial relationship between calcium and LCFAs. This finding supports the assumption that LCFAs precipitated with calcium or were adsorbed on the surface of granules. Calcium was bonded or incorporated in the inner structure of organic material present either in the form of fatty acids or EPS that determined the strengthening of the surface and granule walls as well as internal structures through the crystallization of flakes (Figure 6). The precipitated LCFA and calcium is commonly interpreted as insoluble salt of LCFA and calcium.

Figure 6

Close-up of the flakes which built the inner part of granules.

Figure 6

Close-up of the flakes which built the inner part of granules.

Flakes were randomly oriented and compacted, thereby forming a three-dimensional structure composed of carbon (organic compound) and calcium. Several authors investigated the granules that formed in UASB reactors. They noticed that calcium has a positive impact on granule strength and resistance (e.g., Ren et al. 2008). The granulation process was favored by the high availability of LCFA and calcium and the low mineral fraction content.

Different phosphates were formed during the biogas production process. The most common were iron phosphate (Figure 7(a)) and aluminum phosphate (Figure 7(b)) but also scarce calcium phosphate. Shapeless phosphate accumulations were also present in samples, primarily in the inner parts of granules. These are also one of the most common phases in the sludge and granules described in literature (e.g., Mañas et al. 2011). Some phosphates exhibited partly euhedral shapes, which were most likely originally present in the sludge. Other phosphates were rounded, which is considered a hint to their microbiological origin.

Figure 7

Phosphates found in granules. (a) Example of iron phosphate present on the granule surface; (b) example of rounded crystals of Al-phosphate (in the circle).

Figure 7

Phosphates found in granules. (a) Example of iron phosphate present on the granule surface; (b) example of rounded crystals of Al-phosphate (in the circle).

The cations required for phosphate crystallization were captured from the sludge during the granule formation. This process can also be strongly related to the metabolic activity of PAOs described by Liebrich et al. (2012) In that study, the increase in phosphate concentration was attributed to PAO activity in the beginning of OLR increases. Through VFA uptake, the PAOs reduced the concentration of the acids inside the granules and in the sludge, as noted by Barat et al. (2008). This led to a release of phosphate and its precipitation with aluminum, iron, and calcium present in the sludge. Interestingly, expected calcium-rich minerals did not precipitate in a higher amount, though CaO was added.

The precipitation of calcium carbonate was barely observed and only a few calcium phosphate crystals were found in granules, though a decrease of phosphate in the sludge was noted after CaO dosage. It should be noted that the substrate load containing LCFAs was a factor of 10 higher than that of sewage sludge with PAOs containing stored polyphosphate. Therefore, we observed a very low precipitation of phosphate, whereas more than 90% of precipitates consisted of LCFAs and calcium. Binding of calcium and fatty acids was a favored process and thus prevented the precipitation of calcium carbonates. The addition of CaO did not affect the mineral composition but did influence the size and structure of the granules. Granule size increased with CaO addition, and their pore size became larger with increasing age. Due to adsorption and precipitation of LCFA both in and on granules, over-acidification was avoided; as a result, the methane-producing organisms were not inhibited.

Microorganisms formed colonies or biofilms that were occasionally covered by a layer of amorphous, thin material rich in carbon and were detected on the granule surface (Figure 2(b)). However, higher accumulations of microorganisms were detected in the inner part of the granules. SEM observations indicated the presence of various rods, cocci, chain-forming cocci and filamentous microorganisms (Figure 8(a) and 8(b)) that formed colonies or attached to the flakes. Microorganisms usually participated in the construction of the dimensional network of flakes.

Figure 8

Microbial diversity. (a) Variation of microorganism cells present in granules: agglomeration of cocci and rods microorganisms; (b) dense, porous wall-like structure built up by filamentous microorganisms.

Figure 8

Microbial diversity. (a) Variation of microorganism cells present in granules: agglomeration of cocci and rods microorganisms; (b) dense, porous wall-like structure built up by filamentous microorganisms.

The biofilm enabled further degradation, though the hydrogen partial pressure was 13 times too high to allow for the degradation of propionic acid under standard conditions. A similar study was performed by Ahmad et al. (2011), who showed that the layered structure in a granule protected the methanogens in the core zone from unfavorable conditions, such as a low pH. Layered structures in granules containing different zones/milieus have also been described by other authors, such as MacLeod et al. (1990), Hulshoff Pol et al. (2004) and Del Nery et al. (2008).

Thus, the experiment showed that even at very high loading rates and subsequently high hydrogen partial pressure the biogas formation process stayed stable due to the addition of CaO and resulted in granule formation, which provided microhabitats.

CONCLUSION

Precipitation of insoluble salts of LCFA and calcium served as building material for granules and offered interfaces inside the granules, thereby enhancing the growth of biofilms and LCFA degradation. The precipitation and adsorption of LCFAs reduced their toxicity, and the VFA concentration remained low, despite OLRs above the degradation capacity of the acidogenic, acetogenic and methanogenic microorganisms.

The porous structure of granules promoted mass transfer. The exchange of nutrients between granules and the sludge due to flow processes that were promoted by gas release supported the effective metabolic activity of microorganisms. As a consequence, CaO addition, provided that the dosage is limited, is an excellent additive for preventing over-acidification and sustaining the methane yield. It is an appropriate tool for stabilizing the biogas production process and will help produce energy from organic residues and waste upon market demand. Therefore, the time interval for increasing the OLR should be longer than three days to offer sufficient time for biofilm growth on granules and allow for moderate CaO additions to improve process efficiency and methane yields.

ACKNOWLEDGEMENTS

The results were gained during the ‘Cofermentation’ project financed by the Volkswagen Foundation (II/80 703) and the ‘Optgas’ project (03KB018A) funded by the Federal Ministry for the Environment, Nature Conservation and Nuclear Safety. Thanks to Tobias Lienen, Marietta Liebrich, Sebastian Teitz and Rona Miethling-Graff for helpful suggestions.

REFERENCES

REFERENCES
Angelidaki
I.
Ahring
B. K.
1992
Effects of free long-chain fatty acids on thermophilic anaerobic digestion
.
Applied Microbiology and Biotechnology
37
,
808
812
.
Barat
R.
Montoya
T.
Borras
L.
Ferrer
J.
Seco
A.
2008
Interactions between calcium precipitation and the polyphosphate-accumulating bacteria metabolism
.
Water Research
42
(
13
),
3415
3424
.
Boone
D. R.
Bryant
M. P.
1980
Propionate-degrading bacterium, Syntrophobacter wolinii sp. nov. gen. nov., from methanogenic ecosystems
.
Applied and Environmental Microbiology
40
,
626
632
.
Buswell
A.
Müller
H.
1952
Mechanism of methane fermentation
.
Industrial and Engineering Chemistry
44
(
3
),
550
552
.
Demirel
B.
Yenigün
O.
2002
Two-phase anaerobic digestion processes: a review
.
Journal of Chemical Technology and Biotechnology
77
,
743
755
.
Deublein
D.
Steinhauser
A.
2008
Biogas from Waste and Renewable Resources
.
Wiley-VCH Verlag GmbH & Co. Kg
,
Weinheim
.
Dupla
M.
Conte
T.
Bouview
J. C.
Bernet
N.
Steyer
J. P.
2004
Dynamic evaluation of a fixed bed anaerobic-digestion process in response to organic overloads and toxicant shock loads
.
Water Science and Technology
49
(
1
),
61
68
.
Huang
G. F.
Wu
Q. T.
Wong
J. W. C.
Nagar
B. B.
2006
Transformation of organic matter during co-composting of pig manure with sawdust
.
Bioresource Technology
97
,
1834
1842
.
Hulshoff Pol
L. W.
de Castro Lopes
S. I.
Lettinga
G.
Lens
P. N. L.
2004
Anaerobic sludge granulation
.
Water Research
38
,
1376
1389
.
Liebrich
M.
Kleyböcker
A.
Kasina
M.
Seyfarth
D.
Würdemann
H.
2012
Influence of the microbial community and mineral formation on the process performance of biogas production
.
Fourth International Symposium on Energy from Biomass and Waste
,
Venice
2012
.
MacLeod
F. A.
Guiot
S. R.
Costerton
J. W.
1990
Layered structure of bacterial aggregates produced in an upflow anaerobic sludge bed and filter reactor
.
Applied and Environmental Microbiology
56
,
1598
1607
.
Pohland
F. G.
Ghosh
S.
1971
Developments in anaerobic stabilization of organic wastes – the two-phase concept
.
Environmental Research Letters
1
,
255
266
.
Pokorna
E.
Postelmans
N.
Jenicek
P.
Schreurs
S.
Carleer
R.
Yperman
J.
2009
Study of bio-oils and solids from flash pyrolysis of sewage sludges
.
Fuel
88
,
1344
1350
.
Ren
T. T.
Liu
L.
Sheng
G. P.
Liu
X. W.
Yu
H. Q.
Zhang
M. C.
Zhu
J. R.
2008
Calcium spatial distribution in anaerobic granules and its effects on granule structure, strength and bioactivity
.
Water Research
42
(
13
),
3343
3352
.
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
,
495
502
.