Anaerobic treatment utilizing the function of activated carbon

e Pergamon PH: 50273-1223(97)00167-4 Wa,. Sci. TlCh. Vol. 35, No.8, pp. 193-201,1997. Ci 1997 IAWQ. PIIblished by Elsevier Science Lid Printed in Oreat Britain 0273-1223197 S17'00 +0-00 ANAEROBIC TREATMENT UTILIZING THE FUNCTION OF ACTIVATED CARBON Keisuke Hanaki*. Toshiaki Saito** and Tomonori Matsuo** • Research Centerfor AdvancedScience and Technology. The University ofTo1cyo. 4-6·1 Komaba, Meguro-ku, To1cyo 153. Japan •• Department ofUrban Engineering. The University ofTo1cyo. 7·]·1 Hongo. Bun1cyo·lcu. To1cyo 11], Japan ABSTRACT Granular activated carbon (GAC) was used as supporting medium for bacteria in anaerobic process. The effectiveness of usina GAC was examined with laboratory·scale experiments. Synthetic wastewater containing phenol was converted to methane more successfully in fluidized bed using GAC. Fluctuation in influent concentration of phenol was intenlionally made to examine the dynamic response of the process. Temporary increase of influent phenol from 400 mg/l to 1000 mg/l. 2000 mg/l. and even to 4000 mg/l for 4 days did not affect the effluent phenol as much as methane production because adso'lltion and desorption capacity by GAC functioned as slllbilizer of the fluctuation. The adsorbed phenol was then gradually degraded into methane. Higher biomass concentration and lower adsorbed phenol were observed in the top part of the fluidized bed than in the bottom part. ~ 1997 IAWQ. Published by Elsevier Science Ltd KEYWORDS Anaerobic process; fluidized bed; activated carbon; granular activated carbon; biological activated carbon; phenol; adsorption; desorption. INTRODUCTION Anaerobic treatment process has advantages in terms of energy saving and recovery, although it also has disadvantages in terms of effluent water quality. treatment rate, and biological inhibition. The disadvantages are mainly caused by slow growth rate of anaerobes. Some of the disadvantages have been solved by using support media, but some are still unsolved. Activated carbon not only functions as media for bacterial attachment. bUI also can also work as adsorbent. Combination of the activated carbon with the biological process provides a possibility of grading up the function of the biological process. Adsorption and desorption are distinctive characteristics of activated carbon. These functions can alter the concentration of substrate in bulk water and at the surface of biofilm. Among various biological processes. the anaerobic process and denitrification process are suitable for this combination. because oxygen supply does not limit the biological reactions. Intensive researches have been achieved by the group of Suidan (Suidan et 01.• 1983; Fox, et aI., 1988; Fox and Suidan, 1993; Nakhala et 193 194 K. HANAKJ ~t al. ai., 1990). These researches have proved \hat the application of activated carbon to anaerobic process enables better degradation of refractory or inhibitory wastewater. However, enhancement of anaerobic degradation through substrate transport between bulk water and activated carbon has not been well examined. Sison et ai., (1995, 1996) found \hat organic carbon storage capacity of the activated carbon makes the operation of denitrification process more flexible. The purposes of the research are to clarify characteristics of biological activated carbon (BAC) process, originating from adsorption/desorption characteristics by giving intentional influent fluctuation, and to examine the possibility of improvement in substrate degradation. Table 1. Components of influent. Phenol lQ0-4000 mgll K2HP04 500 mgll NH4Cl 100 mgll CaCl202H20 20 mgll MgC1206H20 30 mgll FeSQ4.7H20 3.0 mgll CoCl206H20 0.5 mg/l NiCl206H20 0.5 mgll Yeast Extract 10 mgll L-Cystein 10 mgll Na2So9H20 IOmg/1 NaOH for PH adjustment 6SO GAS COtLECt'JON UNrI' MAGNETIC 60 S11RRER Figure I. Anaerobic fluidized bed reactor. EXPERIMENTAL METHODS Fluidized bed and dynamic QperatioQ Three identical anaerobic fluidized bed reactors packed with granular activated carbon (GAC) shown in Fig. 1 were used. Each reactor contained 350 g of GAC (Calgon F-400, mesh 12/20). GAC was fluidized by a Anaerobic treatment utilizing the function of activated carbon 195 -2 recycle pump in an expansion of ratio of 40%. each reactor had been operated for 6 months with feeding of 400 mg/l of phenol as sole carbon source at 1 day of HRT before starting this experiment. Substrate composition is listed in Table 1. Temporary high loading was given to each reactor for 4 days at 1000 (No. 1),2000 (No.2) and 4000 (No.3) mg/l of phenol, respectively, as shown in Fig. 2. Concentration of other components except phenol was maintained constant although concentration of phenol was intentionally fluctuated. ~ .._...•....................._ Nnl 2000 t-'NnL~:2-1 llXXl No.!.. .. 400 ~f~ ~f~ NORMAL ! IDGH : LOADING ; LOADING; PfRlOD PFRIOD 4 days Figure 2. pallern of influent fluctuation. Bjomass measurement Several methods have been proposed to estimate biomass in biological wastewater treatment. However, their applications were limited in this study due to the characteristics of BAC. As activated carbon itself showed ignition loss at 600°C, determination of biomass by volatile solids was not appropriate. As separation of biomass from the activated carbon was difficult, biomass together with activated carbon must be analyzed. Taking account of these limitations, biofilm fonned on the surface of GAC was detennined by measuring organic nitrogen of BAC. Biomass as volatile solids was calculated based on empirical fonnula of protein (C3H70N, nitrogen content: 0.157) and protein content in volatile solids (0.6), namely, Biomass (VS) =organic nitrogen/(O.157xO.6). Estimation of adsorption with Soxhlet extraction Soxhlet extraction of activated carbon was attempted to estimate the substrate adsorbed by the activated carbon. It was found that the adsorbed phenol, which is actually utilized by microorganisms, was not satisfactorily extracted with water. Soxhlet extraction with ethanol with 5 cycles/hour for 10 hours was employed to estimate the adsorbed phenol. RESULTS AND DISCUSSION Response to temporary hi~h loadjn~ Methane Production Rate. Figure 3 shows change of methane production rate. Solid straight line in the figure shows theoretical methane production rate which is calculated with inflow data stoichiometrically. Methane production rate increased gradually during the fluctuation period. Increased methane production rate was maintained even after the fluctuation period, and higher methane production rate than theoretical rate was observed fot 24 (No. 1),29 (No.2), and 47 (No.3) days, respectively. 196 K. HANAKl el al. (m1/day) ~ 1200~ B1(0) 6 ::J 800 § g: Lll Z ~ 400 tii ::E 2 ......... : Theoretical methane production rate*'~ High Loading Period -5 0 5 10 15 20 25 TIME (day) *Theoretical methane production rate corresponding to 400 mg/L of influent concentration. Figure 3. Change of methane production rale. Ejj1uent Phenol Concentration. Eftluent phenol concentration increased gradually during the fluctuation period as shown in Fig. 4. Phenol removal percentage was very high during the fluctuation period, although eftluent phenol concentration increased gradually. Phenol removal was 99.8, 99.6 and 98.4% at reactor No. \. 2 and 3, respectively, during this period. (mglL) -0 0 C lJ D lJo·D: :JjIlA A ... .... ... .. - 0 5 10 15 20 25 TIME (day) ~ Z 0 f= ~ 60 ~ u 40-Z 0 U ....l 0 20-~ 5: v -5 High Loading Period ~....p :::d 0 lJ o No.1 A NO.2 D No.3 Figure 4. Change of efRuenl phenol concentration. Elevated effluent concentration did not decrease promptly after fluctuation period was finished. Even though somehow high concentration of phenol flowed out for long time in the case of No.3, the treatment efficiency was not low. Proportion of phenol in the effluent to total discharged carbon (which is the sum of phenol in the effluent and methane production) was 95%, because the methane production in this period was Anaerobic treatment utilizing the function of activated carbon 197 higher than theoretical production calculated stoichiometrically with inflow data as shown in Fig. 3 and Fig. 7. Effect of GAC and bacterial activity on adsoJ1!tion of excess 10adiOi Fate of influent carbon before and during fluctuation period is shown in Fig. 5. This result shows that the role of GAC adsorption to stabilize the impact of influent fluctuation increases as the loading strength increases. tOO ~ 80 i: Q 2:1I~~I.m~~iJ~~~J~~~ DAY-21-0" DAY~" DAY~•• DAY ~-- No. 1-) No.1 No.2 NO.3 400mR/L lOOOmRiL 2000mRiL 4000mRiL -Avelllge d tine reactors dunng 21 days before influ• ent fluctuation penod. Dunng this period. influent phe• nol conccccentnltion was mainlained at 400mglL. "Avelllgt d 4 daya during influent fluctuation period. During this period. innuent phenol conc«centnltioo were mainlained at 1000. 2000. 4000mRiL. respectively Figure 5. Effect of loading strength on adsorption of influent fluctuation. -I 0 I 2 3 4 TIME day Figure 6. Effect of adsorption on influent fluctuation (No.2). Figure 6 shows relative roles of bacterial activity and GAC adsorption during the 4 days fluctuation period in the case of No.2 as a representative. This figure shows that GAC could stabilize the impact of high loading more quickly than the adaption of bacterial activity, although the bacterial activity also increased with time. As shown in Fig. 3 and Fig. 4, although effluent phenol concentration in the case of No.3 was far higher than others, the methane production activity was not so high and very close to the case of No.2. These results suggest that bacterial activity can not promptly respond to the increased loading and, on the other hand, GAC can stabilize the fluctuation of the loading much more effectively than bacterial activity. 198 K. HANAK! el al. Hi~h Loadin~ Peri (mgC/day) EXCESS LOADING' -1 0 2 3 4 5 6 7 8 TIME (day) 'Solid line displays theoretical loading from influent Excess loading is the loading exceeding theoretical inflow and means the loadin~ from GAC Lhrou~h desorption. '*Theorelicalloading during the fluctuation period (day 0-4) is out of ran~e of this fi~re. Figure 7. Carbon discharging rate exceeding theoretical loading (No.3). Gradual decomposition of adsorbed substrate Figure 7 shows the rate of carbon discharge in the case of No.3 as representative. Solid line in Fig. 7 shows the theoretical rate of loading in this period. Methane production rate increased during the elevated loading period and maintained its high rate for a long period. The sum of methane production and phenol in the effluent exceeded the theoretical loading after the elevated loading returned to the normal loading as shown in this figure. As mentioned before, this phenomena had been observed for a long time. This result suggests that adsorbed substrate was supplied to attached bacteria and decomposed gradually. 10 20 30 40,...--------...,..--------.... o 20 40 ro ~ ~ 0 10 20 30 40 ~ ro Bi()()lMS (mgVS/gGAC) Adsorbedpbenol (mglgGAC) o No.! after infiuent change l:i No.3 after infiuent change D . • No.2 beforeinfiucntcbangeNo.2 after influent change (SlCady stale condition) • Qlher column in similar type (Steady stale coodition) I ~ Port.3 ~ .t: II 8 Port.2 i i Port! ~ "ll .~ .~ l>o Figure 8. Profile of biomass and adsorbed phenol. Anaerobic treatment utilizing the function of activated carbon Profile of biomass and adsorbed substrate 199 Uniform distribution of BAC is expected in a fluidized bed in which constant upflow is provided to maintain each particle of BAC in fluidized condition. However, there actually existed a non-uniform distribution of BAC particles. Amount of biomass and the adsorbed phenol was estimated using the sample of BAC in top and bottom of the fluidized bed after the temporary elevated loading period. Figure 8 shows that biomass was higher in the top than the bottom, whereas adsorbed phenol was lower in the top than the bottom. These tendencies were consistent among the three fluidized beds. Such a gradient in biomass and adsorbed phenol indicates that the adsorption and biodegradation are the major function in the bottom and the top, respectively. Enhancement of microbial activity Theoretical background. It has not been clarified yet whether GAC can supply adsorbed substrate directly to the attached biofilm or the substrate is first released from GAC to the bulk, and then supplied to the biofilm. If a part of GAC surface is covered with bacteria, adsorbed substrate can be utilized not only via bulk liquid but also directly by attached biomass. Substrate gradient among bulk, biofilm, and GAC can be classified into 4 cases as shown in Fig. 9. In the case-(A), adsorption occurs. In the case-(B) and -(C), desorption occurs. In the case-(O), GAC has no available substrate for bacteria and amount of penetrating substrate and amount of substrate decomposed are the same. In this case, there is no difference between GAC and other inactive media. In the case-(B) and -(C), substrate gradient in biofilm increases towards surface of GAC. In contrast, in the case-(A) and -(0), substrate gradient in biofilm decreases monotonously towards surface of GAC. Therefore, specific activity in the case-(B) and -(C) is expected to be apparently higher than in the case-(A) and -(0) as shown in Fig. 10. BllLK B1QElIM Figure 9. Substrate gradient in BAC process. 200 K. HANAKI el a1. ····················::··;,;,·ii··..· --••..•..•..•..•.. CONCENTRATION IN BUlK Figure 10. Relationship between specific activity and bulk concentration Experimental evidence. Figure 11 shows relationship between specific activity and bulk concentration which was derived from data of No.2. This figure shows that reaction rate during the desorption phase was higher than that during the adsorption phase with same substrate concentration in bulk water. In other words, substrate concentration which actually governed the reaction rate was higher than the bulk concentration during the desorption phase. This result suggests that substrate gradient as shown in Case-(C) in Fig. 9 was probably formed, and the substrate from GAC is directly decomposed by the attached bacteria. Similar results were obtained in the case of No. I, No.3 and others, although the phenomenon was not very clear. .,.---.•.. . .......,__ 0 o o doong adsoIpIion phase • dlDing desorption phase 2.0 4.0 PHENOL CONCENTRATION 6.0 mgll Figure II. Specific methane activity at adsorption mode and at desorption mode (No.2). SUMMARY The following results were obtained by evaluating performance of anaerobic GAC fluidized bed reactor against influent fluctuation. I. GAC could stabilize the impact of influent fluctuation much more rapidly and effectively than bacteria. 2. Effluent concentration was maintained at low level mainly by GAC adsorption against elevated influent concentration. 3. Bacterial activity increased gradually with time after receiving the increased loading. 4. The adsorbed substrate was gradually decomposed by attached bacteria through desorption. Anaerobic treatment utilizing the function of activated carbon 201 S. Higher biomass concentration and lower adsorbed phenol were observed in the top part of the fluidized bed than the bottom part. 6. Phenol degradation rate could be apparently raised by supply of substrate from the inside of activated carbon when desorption took place. REFERENCES Fox, P. and Suidan, M. T. (1993). A comparison of expanded-bed GAC reactor designs for the treatment of refractorylinhibitory wastewater. Waler Research, 27, 769-776. Fox, P., Suidan, M. T. and Pfeffer, J. T. (1988). Anaerobic treatment of a biologically inhibitory wastewater. J. Wattr Polliltion Control Federation,lSO, 86-92. Nakhala. G. F., Suidan, M. T. and Pfeffer, 1. T. (1990). Control of anaerobic GAC reactors treating inhibitory wastewater. J. Waler Polliltion Control Federation, 61, 65072. Sison. N. F., Hanald. K. and Matsuo, T. (I99S): "High loading denitrification by biological activated carbon process", Water Research, 29, 2776-2779. Sison, N. F., Hanw. K. and Matsuo, T. (1996). "Denitrification with external carbon source utilizing adsorption and desorption capability of activated carbon", Water Research. 30, 217-227. Suidan, M. T., Strubler, C. E., Kao, S. H. and Pfeffer, J. T. (1983). Treatment of coal gasification wastewater with anaerobic filler technology. J. Water Polliltion Control Federation, 55 (10).