The slaughterhouse wastewater used in this experiment was obtained from the slaughterhouse in Braunschweig, Germany. This wastewater consisted of two different flows: one from the slaughtering process and the other from the tripery. The wastewater from the tripery passed through a grease trap. This wastewater was brought together in a sump before discharge to the municipal sewage system. Then the slaughterhouse wastewater was taken from the sump during cattle slaughtering, when all processes were running simultaneously. The COD wastewater was diluted with tap water and ended up to an average COD concentration of 7,500 mg/L which used for the anaerobic fixed bed reactor. This average concentration was the result of the continuous monitoring for 72 h. The size distribution of the solid particles in the slaughterhouse wastewater is shown in Fig. 1. 20–50 percent of the solid particles were retained on a screening whose opening was less than 0.1 mm in diameter.

The inoculum used in this pilot experiment was the effluent from a cattle manure continuously stirred tank reactor (CSTR) (100 m³ volume & HRT = 12.5 d). Before filling the fixed-bed reactor, the course solid particles were removed. Screen analyses showed that only 6% of the solid particles were retained on a screening whose diameter was greater than 1 mm and 66% of the solid particles were retained on a screening whose diameter was greater than 0.06 mm. The screened inoculum contained TCOD of 50.3 g/L (SCOD = 19.5 g/L) with 2.8% total solids (its TVS = 73% of TS). The pH was 7.8.

Bamboo rings were used as support media. Bamboo was obtained from Taiwan, and air dried (its TS = 89.5%). The bamboo sticks were cut into rings whose size was 2.5 cm in length, 0.4–0.7 cm in width and 6–10 cm in inside diameter. Bamboo rings were randomly packed in the reactor. Characteristics of bamboo rings and commercially available support media are compared in Table 1. The specific surface and porosity of bamboo rings are relatively much lower than those of commercial packing media. Bamboo rings as bio-carriers are very affordable and almost free compared to the cost of other packing materials (about 100–1,000 $/m³) [11] in underdeveloped countries of Southeast Asia.

Two batch reactors were operated to examine changes in physical characteristics of slaughterhouse wastewater during the intermediate storage. Each reactor consisted of 20 liter volume and was open at constant temperatures of 10°C and 20°C. Samples were taken at 2–3 d intervals for the parameter analyses.

The schematic flow diagram of a 2.8 m³ pilot scale fixed-bed reactor filled with bamboo rings is illustrated in Fig. 2. Process relevant data on the fixed-bed reactor are given in Table 2. The content of the storage tank were mixed before and during feeding. The HRT was approximately 2–8 d, depending on the experimental schedule. The reactor was fed quasi-continuously every hour on the hour. The recirculation pump for effluent recycle was regulated so that the rising velocity of the reactor content was kept constant 1.0 meter per hour. Following completion of the start-up phase, the reactor was run in upflow and downflow operation for approximately one year each, the loading rate being stepwise increased.

All parameters were analyzed according to Standard Methods [4] except volatile organic acids and alkalinity. For the alkalinity measurement, sample were centrifuged to remove solids, and then titrated with 0.1 N H₂SO₄ to pH 5.0 and further titrated down to pH 4.4, respectively. Each concentration of volatile fatty acids (VFA) was also measured by a gas chromatograph (Mega 5340, Carlo Erba Instruments, France). The gas composition was determined with an infra-red gas analyzer (Maihak, SICK MAIHAK, Inc., Germany). All gas measurements were expressed in terms of standard temperature and pressure (0°C, 760 m Hg).

The batch tests on the intermediate storage of slaughterhouse wastewater at the simulated temperature of 10°C and 20°C showed a low degree of acidification. As shown in Fig. 3, total volatile fatty acids (TVFA) rose from 247/mg/L as acetic acid at the beginning of the experiment to a maximal value of 1,019 mg/L as acetic acid concentration at 20°C on day 3. The production of VFA exceeded the extent which buffering capacity was formed during the first week of operation. Then the buffering capacity began to overtake VFA production and increased continuously to 1,000 mg/L as CaCO₃ at the end of the experiment. The production rate of the buffering capacity at 20°C was approximately two times higher than that at 10°C during the entire storage period. The continual increase in buffering capacity was mainly due to the methanation and mineralization of protein, which eventually led to the dynamics of carbonate and ammonia buffer systems in all storage reactors. These phenomena were also reflected by the ratio of TVFA and alkalinity (TVFA/ALK). The initial ratio of 1.41 sharply increased to a ratio of 2.91 within 3 d when the reactor was operated at 20°C and then dropped continuously until the end of storage period. The storage reactor took 9 d to recover the initial TVFA/ALK ratio. While the storage reactors when operated at 10°C took 13 d.

Whatever anaerobic treatment processes are to be selected, the intermediate storage of waste and wastewater is required prior to feed to the main treatment process. It is therefore important to know about the changes in characteristics that occur in the substrate during the storage period. In the batch tests only a low degree of acidification was found during the first week of storage that is in a normal operational range of the hydraulic retention time for the anaerobic fixed-bed reactor with slaughterhouse wastewater. An obvious question still remains – the question whether such a low degree of acidification is beneficial to the methanation of slaughterhouse wastewater. However as the first stage of anaerobic fermentation, the controlled storage of slaughterhouse wastewater does not seem to be necessary for a highly efficient bamboo ring mesophilic anaerobic fixed-bed reactor.

Initially the reactor was randomly packed with bamboo rings and then filled with 2.1 m³ of tap water up to the overflow outlet which was 2.8 m³. The bamboo rings had a porosity of 89% as measured by liquid volume in the media section of the reactor and a surface to volume ratio of 88%. After 1 d soaking, 1 m³ of leaching water was drained, and 1 m³ of screened liquid manure was replaced as an inoculum. The gas production from inoculum itself was gradually increased and then reached stationary phase on day 11 which produced 1.37 m³ of biogas per day with a CH₄ content of 68%. The TVFA/ALK ratio at this point was 0.23 and thus indicated a good buffering capacity. Soluble organics leached from the soaking process might contribute to the reactor start-up period. The reactor loading started with 1 kg COD/m³-d in an upflow mode. Reactors with upflow mode were documented to be faster in biofilm development than downflow reactors due to the lower washout effect of suspended biomass. The successful start-up operation was achieved within 2–3 weeks mainly because of larger size of inoculum utilized. When the reactor started with the OLR of 1 kg COD/m³-d, the substrate to inoculum ratio was approximately 0.06 as a total COD mass basis.

The COD removal efficiencies as a function of OLR are compared between upflow and downflow fixed bed reactors as illustrated in Fig. 4. It became clear that the COD removal efficiency of the upflow mode showed slightly higher than that of the downflow mode. However, the results were not comprehensive enough to draw any conclusion which operational mode is more suitable for the full scale plant operation of slaughterhouse wastewater. In general, the COD removal efficiencies decreased gradually as the organic loading rates increased. Conversely the COD removal efficiencies increased steadily at longer HRT as shown in Fig. 5. The maximum COD removal of 95% was achieved when the upflow fixed-bed reactor was operated at an OLR of 1 kg COD/m³-d with its corresponding HRT of 7.5 d. Contrarily the lowest COD removal efficiencies of 74% for the upflow mode and 72% for the downflow mode were obtained, respectively at an OLR of 4.0 kg COD/m³-d with its corresponding HRT of 1.9 d.

The average methane gas yield (methane volume produced per reactor volume per day, v/v d) in both upflow and downflow reactors increased from 0.35 m³/m³-d at an OLR of 1 kg COD/m³-d to 1.5 m³/m³-d when the OLR was raised to 4 kg COD/m³-d as illustrated in Fig. 6. The methane content of the biogas ranged from 72–75% during the entire operation. Independently of the OLRs (OLR = 1–4 kg COD/m³-d) or their corresponding HRTs (HRT = 2–8 d), the reactor performance and its relevant operational behavior during the respective upflow and downflow operations are quite similar, as illustrated in Table 3. Reactor pH varied from 7.3 to 7.9 initially, but stayed above 7.7 for most of the operational period in both modes. The TVFA/ALK ratio of slaughterhouse wastewater was 1.45 with its alkalinity concentration of 590 mg/L as CaCO₃. This ratio was reduced down to 0.15 with its corresponding alkalinity of 1,830 mg/L as CaCO₃ at the end of operation. This indicates that over 90% of the reactor organic acids were completely buffered if the relative magnitude of organic acids and buffering capacity produced during the operational period were considered. The TVFA/ALK ratio depends on substrate characteristics, OLRs, solids retention times and overall reactor capability to convert organic acids to CO₂ and CH₄. However, this result shows that the ratio less than 0.3 has been a good indicator for the stable reactor operation with slaughterhouse wastewater [8]. The concentration of total volatile acids in the reactor influent and the acid spectrum were similar to those of batch test results. As shown in Table 4, the average total volatile fatty acid concentration of the reactor effluent was 156 mg/L as acetic acids concentration which remained much lower concentration. This could illustrate that hydrolysis reactions were rate limiting in the complex conversion process. Lower ratios of TVFA/ALK (0.04–0.33) observed during the entire operation also showed the hydrolysis limiting. The total nitrogen in the slaughterhouse wastewater consisted of approximately 71% of organic nitrogen and 29% of inorganic forms. During the anaerobic degradation process, about 85% of organic nitrogen was mineralized. This explained that why the ammonia nitrogen concentration increased from 95 mg/L in the influent to 289 and 285 mg/L in the effluents of downflow and upflow reactors, respectively. The ratio of COD: total nitrogen: total phosphorus in the treatment effluent was much narrower than that of influent mainly because of COD degradation in the reactor. However, there was no reduction in the total nitrogen during the operational period. Generally the magnitude of the ratio became narrower as the HRTs increased or the OLRs decreased. The BOD/TKN ratio is regarded as one of the most important parameters in designing a subsequent biological denitrification process. A minimum ratio 3.0 for the denitrification of domestic sewage was documented in ATV report [6]. The COD/TKN ratio of 6.0 in the effluent when the reactor OLR increased over 3.0 kg COD/m³-d was only marginal acceptable range. Otherwise carbon supplementation is required at a lower OLR.