Hydrogen is widely recognized as a clean and efficient resource for the future. However, approximately 95% of commercially produced hydrogen comes from carbon-containing raw materials, primarily fossil in origin. Given these perspectives, biological hydrogen production assumes paramount importance as an alternative energy resource. Despite its relatively lower yields of hydrogen compared with bio-photolysis processes, fermentative hydrogen production, or dark fermentation, is a promising method of bio-hydrogenation due to its higher rate of hydrogen evolution in the absence of any light source as well as the versatility of the substrates used. Fermentative bio-conversion of biomass resources into hydrogen gas is a charming technology in the aspects of production of clean energy as well as the transformation of waste into environmentally sound materials.
Most approaches in environmental engineering have preferred mixed culture as seed material and have been used harsh pretreatment method to harvest hydrogen producing bacteria. However, the suggested methods were expensive and hard to apply in practice so that an economical alternative is necessary. The role and identification of main microorganisms and/or inhibitory microorganisms in a certain system have been paid more attention to by the researchers to understand biological process. In addition, the role and function of fermentative hydrogen producing bacteria and relationship between other microorganisms were not investigated thoroughly until this time. In this study, air purging method was evaluated as an alternative method of well-known pretreatment for fermentative hydrogen production. Moreover, the microbiological observation was kept using molecular microbiological analysis techniques such as terminal restriction fragment lengthy polymorphisms (T-RFLP), denaturing gradient gel electrophoresis (DGGE), and clone library.
In the first approach, the basic factors such as initial pH, VSS concentration and substrate concentration for continuous hydrogen production were investigated using air-purged sludge as seed material via batch tests. When 0.5M phosphate buffer system was applied to keep pH in fixed range, the optimum pH was around 6.0 even hydrogen yields were relatively smaller than those without phosphate buffer system. The optimum pH without phosphate buffer system was around 9.0, which could apply to sequencing batch reactor operation. The optimum initial substrate to biomass concentration was 1.79 at 8gCOD/l and 19.28 at 16gCOD/l at this experiment.
Secondly, start-up of reactors seeded with air-purged sludge (Reactor A) and heat-shocked sludge (Reactor H) was investigated at 12 h HRT with 20 g carbo.-COD/l. T-RFLP analysis was applied as a quantitative and qualitative method to observe the changes of bacterial community in this study. The primer set and restriction enzyme was selected via in-silico simulation of PCR and digestion. As a result, combination of primer set $S-D-Bact-0338-a-A-18^c$ (5’-ACTCCTACGGGAGGCAGC-3’) with S-*-Univ-1392-a-A-15 (5’-ACGGGCGGTGTG TRC-3’) and restriction enzyme (ScrFI,$CC\veeNGG$) was selected as a best option. Reactor H was needed about 2 weeks to obtain stable hydrogen $(0.62 mol H_2/mol hexose_{consumed})$ but Reactor A achieved very fast stabilization within one week and showed good performance $(1.13 mol H_2/mol Hexose_{used})$. Main hydrogen producing bacteria was Clostridium pasteurianum and facultative anaerobic bacteria including spore-forming lactic acid bacteria (SFLAB) performed an oxygen-eating function not only to decrease the hydrogen production yield but also to protect the hydrogen producer from oxygen.
Thirdly, strategy for stable start-up was investigated based on the HRT. Continuous operation at 12 h did not produce hydrogen because HPB were not developed during start-up period even though the seed sludge included the high relative abundance of Clostridium species. Fed-batch operations were conducted to investigate the effects of dilution rate and population dynamics of microbial community were analysed to understand the changes of HPB and non-HPB. Fed-batch operation at $D0.125h^{-1}$ (HRT 8 h) was not enough to make rapid enrichment of HPBs but $D0.5h^{-1}$ (HRT 2 h) made favourable conditions for enrichment of HPBs. Based on the fed-batch experiment, the start-up experiment was carried out with air-purged, heat-shocked and raw anaerobic sludge, respectively. Reactor H was successful for enrichment of HPB but it was very weak to the accidental oxygen exposure. Reactor A was successful for enrichment of HPB and protectable against the oxygen exposure. Continuous operation at short HRT (2 h) made conditions for the enrichment of HPB but other bacterial species were developed together when the seed was not pre-treated at all. DGGE analysis revealed that 10 bands were close to uncultured rumen bacteria and four bands were related with Clostridium species. Also, lactic acid bacteria such as Enterococcus, Streptococcus, Bifidobacterium and Sporolactobacillus were found. Facultative anaerobic bacteria such as Pseudomonas and Alcaligenes were also discovered in the band.
Finally, the operational factors such as HRT and substrate conditions were investigated and guidelines for continuous operation were suggested via interpretation of performance and microbial analysis. Continuous operation for hydrogen production at short HRT (2 and 4 h) achieved more than 1.5mol $H_2/ mol$ Hexose and Clostridium pasteurianum was the main hydrogen producing bacteria. But only small portion of substrate (16% and 22%, respectively) was degraded due to short HRT and SRT. Continuous operation at long HRT (12 h) ceased the production of hydrogen. T-RFLP analysis at 12 h HRT operation revealed that relative abundance of Clostridium pasteurianum decreased and that of Bifidobacterium longum species increased. According to influent substrate concentration, main hydrogen producing bacteria were changed. Dominant HPB in low substrate concentration range (1 to 3g COD/l) was C. butyricum species and dominant HPB in high substrate concentration (20g COD/l) was C. pasteurianum species. Between 5 to 10 g COD/l, both of those species could be a dominant species. The accidental oxygen exposure in substrate concentration experiment resulted in abnormal biomass growth, drop of hydrogen yield, and increase of ethanol, lactate, and propionate production. This abnormal phenomenon disappeared when substrate concentration was below 5g COD/l and appeared again when substrate concentration increased to 10g COD/l. The abnormal phenomenon was supposed to be from the change of metabolic pathway of Clostridium species and from the growth of lactic and propionic acid bacteria. Inhibitory effects of volatile fatty acid were examined. Higher propionate stimulated hydrogen production and hydrogen production rate on the contrary. In cases of other acids, higher concentration decreased the hydrogen production rate but the capacity of hydrogen production was kept. Two concepts were suggested to achieve high hydrogen production rate and high yield. First solution is thorough prevention of the contamination from non-HPB without HRT control. Another solution is co-immobilization of HPB with facultative anaerobic HPB and application of very short HRT.
Moreover, the changes of microbial population were monitored in pilot scale hydrogen fermenter using T-RFLP analysis. Food wastes were alkaline treated at pH 13 for 1 to 2 days after grinding. The pretreated feed was pumped to the feed tank and diluted to 30 g carbo-COD/l with tap water. HRT was changed from 72 to 36 hr after about 1 month later. Start-up took about three weeks. During start-up period, $H_2$ production was 0.87 mol $H_2/mol$ Hexose with high standard deviation (0.369) but it decreased from the $20^{th}$ day of operation. After the change of HRT from 72 to 36 h, yield was decreased for 10 days but it stabilized. The changes of HRT showed little difference indicating that HRT in this region was not critical. Clostridium species were dominant all the time but some Clostridium species might not produce hydrogen rather consume hydrogen. Non hydrogen producing bacteria such as Lactobacilliaceae and Enterobacteriaceae were detected. From the view of start-up, the alkaline treatment of food waste was successful because hydrogen evolution maintained over 1 month. However, from the view of hydrogen yield, hydrogen yield kept low. Exact economic analysis between heat treatment and alkaline treatment was not evaluated on the hydrogen production but both of them need some supplementation. To achieve the economic method, some other method is necessary.
수소는 현재 연구 중인 미래 대체에너지 중 환경친화적이며 효율적인 에너지원으로 각광받고 있다. 그러나 생산되고 있는 수소의 약 95%는 화석연료를 재료로 물리화학적 처리를 거쳐 생산되고 있다. 이러한 생산방법은 환경적인 측면이나 대체 에너지 생산이라는 측면에서 한계점을 지니게 된다. 따라서 생물학적 수소생산은 이러한 측면에서 강점을 지닌다. 발효를 이용한 수소생산방법은 광반응에 비하여 낮은 수율의 단점을 지니고 있지만 빠른 수소생산속도와 다양한 기질로부터 수소를 생산할 수 있다는 장점을 지니고 있다. 또한 폐기물을 처리하는 과정에서 수소를 발생시킨다는 환경 친화적 측면에서 더욱 매력적인 기술이다.
혐기성 수소발효를 하기 위한 방법으로 환경공학에서는 대부분 혐기성소화슬러지와 같은 혐기성 혼합 미생물을 전처리하여 식종균으로 사용하는 방식을 사용한다. 그러나 열이나 화학적 충격을 통하여 수소 생산균을 제외한 대부분의 미생물을 사멸시키는 전처리 방법은 현장에 적용하기에는 경제성과 실용성이 부족하다. 또한 혐기성 수소 생산과정에서 관여하는 수소 생산균의 규명과 그 이외의 다른 미생물과의 상관관계에 대한 연구가 부족하여 아직까지 반응조내의 수소 생산 거동을 충분히 이해하지 못하고 있다. 본 연구에서는 공기처리 방법을 경제적이고 실용적인 전처리 방법의 대안으로 제시하였다. 또한 T-RFLP, DGGE, cloning 등의 분자생물학적 방법을 이용하여 반응조 내 수소생산 미생물과 비수소생산 미생물간의 거동을 분석하였다.
먼저 공기를 이용한 전처리 방법으로서 수소생산 가능성을 평가하기 위하여 회분식 실험을 수행하였다. 또한 연속운전을 위한 기본 운전인자인 초기 pH, VSS 농도, 기질주입농도를 평가하였다. 0.5M의 인산완충용액을 사용하여 수소생산 반응 중 pH의 변화를 억제한 경우 최적의 pH는 6.0이었으며 인산완충 용액을 사용하지 않은 경우 최적의 pH는 9.0이었다. 연속운전을 위한 최적의 초기 기질/ VSS비는 기질 농도 8gCOD/l에서 1.8, 16gCOD/l에서 19.3 이었다.
본 연구에서는 미생물군집변화를 관찰하기 위하여 T-RFLP 기법을 주로 사용하였다. 이때 사용할 PCR primer set과 제한효소를 in silico simulation을 통하여 결정하였다. RDP 데이터베이스를 이용한 시뮬레이션 결과 primer set은 S-D-Bact-0338-a-A-18c (5’-ACTCCTACGGGAGGCAGC-3’) 과 S-*-Univ-1392-a-A-15 (5’-ACGGGC GGTGTGTRC-3’)이 최선의 선택으로 결정되었고 제한효소는 ScrFI (CC∨NGG)으로 결정되었다.
다음으로 공기 전처리한 슬러지를 식종균으로 사용하는 경우에 반응조의 초기 운전 방법에 관하여 연구하였다. HRT 12시간에서는 Clostridium 이외의 다른 미생물군이 성장하여 수소생산에 실패하였으나 HRT 2,4 시간과 같이 짧은 희석률에서는 Clostridium species가 우점되며 수소생산에 성공하였다.
연속운전에서는 HRT와 기질의 농도가 수소 생산에 미치는 영향에 대하여 조사하였다. HRT 2, 4시간과 같이 짧은 HRT에서는 1.5 mol $H_2/mol$ Hexose와 같은 높은 수소생산수율을 얻을 수 있었으나 짧은 체류시간으로 인하여 기질 분해율이 매우 낮았다 (16, 22%). HRT 12시간으로 체류시간을 증가시키자 초기에는 높은 수소 생산수율과 기질 분해율을 얻을 수 있었으나 곧 수소생산이 중단되었다. T-RFLP 분석결과 이 조건에서는 Clostridium pasteurianum의 상대적 비율이 감소하였고 Bifidobacterium longum의 비율이 증가하였다. 유입수 기질의 농도에 따른 영향을 살펴본 실험의 경우 기질의 농도에 따라 수소생산미생물의 변화가 관찰되었다. 기질농도가 1~3gCOD/l에서는 Clostridium butyricum이 우점하였고 20gCOD/l에서는 Clostridium pasteurianum이 우점하였다. 5~10gCOD/l 의 영역에서는 두 미생물이 우점하였다. 이러한 실험결과를 바탕으로 실제 운전에서 수소생산증대를 위한 두 가지 방법을 제시하였다. 첫 번째 방식은 비수소생산균의 오염의 철저하게 차단하는 것이다. 하지만 실폐수나 폐기물에는 다량의 오염원이 존재하여 이를 적용하기 난이하다. 두 번째는 수소 생산균과 통성혐기성 수소 생산균의 고정화 시킨 후 매우 짧은 수리학적 체류시간에서 운전하는 방식이다. 수소 생산균과 통성혐기성 수소 생산균의 고정은 운전실수나 유입수에 포함되어 있는 산소로부터 수소 생산균을 보호하는 역할을 수행할 것이며 짧은 수리학적 체류시간은 기타 비수소생산균의 성장을 억제시켜 높은 수소생산 수율과 분해율을 얻을 수 있을 것으로 판단된다.
마지막으로 음식쓰레기를 처리하여 수소를 생산하는 현장규모의 반응조를 운전하였고 그때의 미생물 군집의 거동을 T-RFLP 방식을 사용하여 관찰하였다. 음식쓰레기는 분쇄기를 이용하여 분쇄한 후 pH 13에서 1~2일간 전처리하였다. 전처리 된 기질은 30g Carbo.-COD/l의 농도로 반응조에 유입되었고 HRT는 72시간으로 운전되었다. 약 한 달 후에 HRT를 36시간으로 감소시켰다. 약 3주간 수소샌은 점차 증가하였으나(0.87 mol $H_2/mol$ Hexose) 20일 이후 급격하게 감소하였다. 운전기간동안 Clostridium species가 우점하였으나 Lactobacilliaceae와 Enterobacteriaceae와 같은 그 이외의 미생물 군집 역시 관찰되었다. 수소생산의 연속성 측면에서 기질의 염기처리방식은 성공적이었으나 수소생산 수율이 매우 낮게 유지된 측면에서 이 방식은 실패하였다. 따라서 높은 수소생산 수율을 얻기 위하여 다른 방식의 접근이 필요하다.
