1.2 Operation of the Reactors

Small — and large-scale, complete-mixing, composting reactors were used in this study. Three composting reactors for household use (“Namagomi-eater” TK400-H, Matsushita Electric Works, Japan) were used as the small-scale, FBC reactors. The working volume was 15 L. The biomass carrier (or bulking agent) comprised about 5 L of wood chips with a size range of 0.5-2.0 mm, or plastic bottle flakes with a size range of 2.0-10.0 mm. An artificial organic waste sample, made up of 500 g wet wt dog food (VITA-ONE, Nihon Pet Food, Japan) containing about 90% water, was loaded daily into each reactor. The contents in the reactors were gently mixed by automated paddles for 1 min each hour. Mechanical heating was used to maintain the temperature in all the reactors above 35°C to accelerate biodegradation. In the small-scale reactor, three experimental conditions (reactor A, B, and C) were used. In reactor A and C, a high decomposition rate of organic materials was maintained by “partial washing” [14, 16] as follows. Approximately 10% of the contents (0.75 L) were taken out every three days, mixed with 10 L of water, and then filtered on a 35 ^m mesh filter. Upon filtration, the solid part retained on the mesh filter was dried in an oven at 60°C for 48 h, and then re-loaded into the reactor. This process prevents not only a decrease in decomposition rate but also aggregation of the contents in the FBC reactor [14]. In reactor B, there was no maintenance, except for moisture content, where spontaneous aggregation was allowed in the decomposition process. The moisture content in each reactor was kept at 40-50% by the addition of distilled water. Samples were obtained from each small-scale reactor after 60 days of operation.

To compare the difference of the scale of the reactor, large-scale composting reactor was also used (O-1, which we constructed). The working volume was 4 m3. The bulking agent comprised about 2 m3 of plastic bottle flakes with a size range of 2-10 mm. 600 kg wet weight food waste derived from a school cafeteria with 73% moisture content, was loaded into the reactor at the start of the experiment. This is termed batch operating. The contents of the reactor were gently mixed by automated paddles at 1.5 r. p.m. for 30 min once a day. Temperature was not regulated. Samples were obtained from the large-scale reactor, once a day, for 25 days. In the small-scale reactor (reactors A, B, and C), 16S rDNA-clone analysis was performed to compare difference in the bacterial communities under each set of conditions, and in the large-scale reactor, both 16S rDNA-DGGE and clone analysis were performed to analyze how bacterial community succession changes day by day.