Biotreatment of H2S- and NH3-containing waste gases by co-immobilized cells biofilter

(1)

Biotreatment of H

2

S- and NH

3

-containing waste gases by

co-immobilized cells bio®lter

Ying-Chien Chung

a

, Chihpin Huang

b,*

, Ching-Ping Tseng

c

, Jill Rushing Pan

b

a_{Science and Technology Information Center, National Science Council, Taipei, Taiwan, ROC}

b_{Institute of Environmental Engineering, National Chiao Tung University, 75 Po-ai Street, Hsinchu 3009, Taiwan, ROC} c_{Institute of Biological Science and Technology, National Chiao Tung University, Hsinchu 3009, Taiwan, ROC}

Received 25 June 1999; accepted 7 October 1999

Abstract

Gas mixture of H2S and NH3in this study has been the focus in the research area concerning gases generated from

the animal husbandry and the anaerobic wastewater lagoons used for their treatment. A speci®c micro¯ora (mixture of Thiobacillus thioparus CH11 for H2S and Nitrosomonas europaea for NH3) was immobilized with Ca-alginate and

packed inside a glass column to decompose H2S and NH3. The bio®lter packed with co-immobilized cells was

con-tinuously supplied with H2S and NH3gas mixtures of various ratios, and the removal eciency, removal kinetics, and

pressure drop in the bio®lter was monitored. The results showed that the eciency remained above 95% regardless of the ratios of H2S and NH3used. The NH3concentration has little eect on H2S removal eciency, however, both high

NH3and H2S concentrations signi®cantly suppress the NH3 removal. Through product analysis, we found that

con-trolling the inlet ratio of the H2S/NH3could prevent the bio®lter from acidi®cation, and, therefore, enhance the

op-erational stability. Conclusions from bioaerosol analysis and pressure drop in the bio®lter suggest that the immobilized cell technique creates less environmental impact and improves pure culture operational stability. The criteria for the bio®lter operation to meet the current H2S and NH3 emission standards were also established. To reach Taiwan's

current ambient air standards of H2S and NH3(0.1 and 1 ppm, respectively), the maximum inlet concentrations should

not exceed 58 ppm for H2S and 164 ppm for NH3, and the residence time be kept at 72 s. Ó 2000 Elsevier Science Ltd.

Keywords: Hydrogen sul®de; Ammonia; Bio®lter; Thiobacillus thioparus CH11; Nitrosomonas europaea

1. Introduction

Recently, people in Taiwan have encountered nu-merous air pollution problems, especially those associ-ated with bad odor. It has been reported that more than 300 substances can cause bad odor (Ikeda et al., 1980). Among these substances, NH3 (ammonia) and H2S

(hydrogen sul®de) are the ones that often exist in our

surroundings. Ammonia is colorless, but irritant and smelly, while H2S is corrosive, extremely toxic, and also

smelly. Large amounts of NH3 and H2S are generated

and released from industrial processes, such as petro-chemical re®ning, metallurgy, food preparation, waste-water treatment, and treatment of fuels (Eikum and Storhang, 1986; Ryer-Power, 1991; Yang and Allen, 1994). Excess amounts of NH3 and H2S have to be

re-moved for the sake of safety and health (Buchnan and Gibbons, 1974; Prosser, 1989) and also for the reduction of environmental impacts, such as greenhouse eect, acid rain, and eutrophication. Currently, the Taiwan EPA sets the ambient air standards at 1 and 0.1 ppm for

*_{Corresponding author. Tel.: 35-726-463; fax:}

+886-35-725-958.

E-mail address: cphuang@green.ev.nctu.edu.tw (C. Huang).

(2)

NH3 and H2S, respectively. Based on the cost for the

equipment and operation, biological treatment is believed to be the most economical option for treating NH3 and

H2S. Recently, immobilized-cell technology has been

employed in wastewater treatment. Its advantages include high microorganism content, prevention of the microbial loss, high environmental endurance, and high operational stability. In the study of exhaust gas treatment, Chung et al. (1996a,b) proved the value of the immobilized-cell technology because of its high removal eciency, high removal potential, and high operational stability.

In treating exhaust gas, the selection of the right kind of bacteria is very important. Although activated sludge is often inoculated into the bioreactor, it requires at least three weeks for acclimation (Ottengraf and Van Den Oever, 1983). The use of pure culture has drawn great attention, because it can shorten the start-up time and increase the removal eciencies of the reactor (Sublette and Sylvester, 1987).

Immobilized Thiobacillus thioparus CH11 bio®lter has been successfully applied on removing H2S alone

(Chung et al., 1996a,b). It can treat H2S gas from high

concentration (100 ppm) to low concentration (10 ppm). On the other hand, immobilized Nitrosomonas europaea was discovered capable of treating NH3 gas of various

concentrations (10±100 ppm), and exceeding 97.5% re-moval after a 4-day operation (Chung and Huang, 1998). This high removal eciency can last for three months if the pH value is adequately controlled. As NH3

and H2S often coexist in real situation, high

concentra-tion of these two gases may damage N. europaea and thus decrease the NH3removal eciency (Prakasam and

Loehr, 1972; Hunik et al., 1992; Joye and Holibaugh, 1995). Hence, how to optimize the operation of the immobilized-cell bio®lter to simultaneously remove NH3

and H2S is the key to the success of this technology.

In our previous studies, we have discovered that N. europaea is very eective in removing NH3alone and

T. thioparus CH11 is very good at eliminating H2S alone

(Chung et al., 1996a,b; Chung and Huang, 1998).

There-fore, we adopted N. europaea and T. thioparus CH11 co-immobilized bio®lter in this study. By supplying NH3/H2S

gas mixtures of various ratios, we examined the removal eciency, removal mechanisms, metabolized products, and kinetic parameters of the co-immobilized bio®lter. In addition, the likely encountered operational problems such as pressure drop and contamination by other bac-teria in ®eld applications, as well as the design of optimal operation conditions were also discussed.

2. Materials and methods

2.1. Organism cultivation and medium preparation The original pure-culture strain of autotrophic am-monia oxidizer, N. europaea ATCC 19718 was obtained from the American Type Culture Collection. Their stock cultures were grown in the ammonia medium in the dark at 30°C. Autotrophic sulfur-oxidizing T. thioparus CH11 was isolated from the swine wastewater (Chung et al., 1996b), and the stock cultures were grown in the thio-sulfate medium at 30°C. For all continuous experiments, the in¯ow medium was drawn directly from the nutrient tank. The basic media compositions are listed in Table 1. 2.2. Immobilization procedure

Both N. europaea grown in 100 ml ammonia medium and T. thioparus CH11 grown in 100 ml thiosulfate me-dium were harvested by centrifugation (7500 ´ g for 10 min), and then washed three times with sterile distilled water. These organisms were mixed together with a sterile 4% (w/v) alginate solution. With a syringe, this Na-alginate solution was dropped into a 4% (w/v) CaCl2

so-lution and immediately formed 3 mm-diameter beads. These co-immobilized beads were activated by ¯ushing with sterile buer solution for 5 h. The initial biomass concentrations of each specie in the beads were 105_cells/

g-bead.

Table 1

Composition of basic media for cultivation of microorganisms and continuous experiments

Medium Ammonia medium Thiosulfate medium In¯ow medium Composition g lÿ1 _(NH

4)2SO4 3.3 NH4Cl 0.4 NH4Cl 0.1

MgSO4 0.25 MgCl2 6H2O 0.2 MgCl2 6H2O 0.2

NaH2PO4 0.78 NaH2PO4 1.2 NaH2PO4 0.78

Na2HPO4 0.89 Na2HPO4 1.2 Na2HPO4 0.89

CaCl2a 0.74 Na2S2O3 5H2O 8.0 CaCla2 0.74

FeSO4 7H2Oa 2.5 FeSO4 7H2O 0.01 FeSO4 7H2O 0.01

CuSOa

4 0.08 CuSOa4 0.08

pH 7.5 7.0 7.5

Temp. (°C) 30 30 30

(3)

2.3. Apparatus and H2S/NH3 removal for continuous

operation

A schematic of the experimental set-up of the lab scale bio®lter is shown in Fig. 1. Glass columns (60 mm / ´ 25 cm of working height) were packed with cell-laden Ca-alginate beads on top of a perforated sieve plate ®tted at the bottom of the column to ensure the uniform distribution of the inlet gas. The packed vol-ume, dry weight of beads, and numbers of cells initially packed in each column were 0.7 l, 0.28 kg, and 1010_cells/

g-bead, respectively. Four ports, 12.5 cm in spacing, were drilled along the column for sampling. The ¯ow meter and valve were used for monitoring and control-ling the gas ¯ow through the reactor. The pressure drop across the reactor was measured with a u-tube water manometer. The H2Sgand NH3g, supplied from

sep-arate gas cylinders, were ®rst diluted with compressed air, passed a air ®lter (pore size 0.2 lm, LIDA 3000-06, Made in USA) and ¯owed upwards through the bottom of the bio®lter. An in¯ow medium (composition shown in Table 1) was re-circulated by a peristaltic pump at a ¯ow rate of 25 ml/min to maintain the moisture of the bio®lter and supply nutrient to the co-immobilized cells. The peristaltic pump was connected to a spray nozzle to uniformly spray the medium on the surface of the ®lter bed. During the entire experiments, aluminum foil was wrapped around the column to prevent photo-inhibition.

In the continuous experiment, the simulated H2

S-and NH3-containing waste gas was prepared at 1:1

(60:60), 1:2 (60:120), and 2:1 (120:60) by concentration (ppm/ppm). These mixtures were supplied to the bio®lter at a ¯ow rate of 36 l/h (residence time 72 s) and the operating temperature was controlled at 30°C.

2.4. Bioaerosol analysis

Microorganisms released from the bio®lter were collected by liquid impingement. The air that escaped from the top of the bio®lter was forced through a 250 ml ¯ask containing 80 ml aseptically distilled water at 45 l/min for 10 h. One ml of the collected solution was inoculated to dierent media and the numbers of cells were determined by the serial dilution method. The PDA medium was used for fungi, the nutrient broth medium for heterotrophic bacteria, the thiosulfate medium for non-acidophilic Thiobacilli, and the modi®ed Waksman medium for acidophilic Thiobacilli (Cho et al., 1991). The cell counts of autotrophic ammonia oxidizer were determined by the amount of nitrite produced (Sato et al., 1985). The counts were reported as colony form-ing units in air (CFU/m3_).

2.5. Kinetic analysis

The H2S and NH3removal rate in the

immobilized-cell bio®lter were calculated using the following equa-tion derived from the Michaelis±Menten equaequa-tion (Hirai et al., 1990) 1 R Ks Vm 1 Cln 1 Vm; 1

where R (g-S or g-N/day/kg-bead) ± apparent removal rate; Clnppm C ÿ Ce= lnCo=Ce, logarithmic mean

concentration of H2S or NH3at the inlet and outlet of

the bio®lter; Vm (g-S or g-N/day/kg-bead) ± maximum

apparent removal rate and Ks (ppm) ± apparent

half-saturation constant. From the linear relationship be-tween 1/Clnand 1/R, Vmand Kswere calculated from the

slope and intercept. In this experiment, the ¯ow rates were controlled in the range of 36±72 l/h to minimize the mass-transfer limitation.

When the NH3 oxidation was inhibited due to high

NH3 concentration (>120 ppm), an inhibition constant,

Ki, must be incorporated into Eq. (1) as

1 R Ks Vm 1 Cln 1 Vm Cln Vm Ki: 2

At low inlet concentration (5±65 ppm), the Eq. (2) can be simpli®ed back to Eq. (1). However, at high inlet concentration (120±200 ppm), Eq. (2) becomes Eq. (3): 1 R 1 Vm Cln Vm Ki: 3

2.6. Design criteria of scale-up bio®lte

The concentrations of H2S and NH3 at the bio®lter

outlet were targeted at 0.1 ppm and 1 ppm. The maxi-mum inlet concentrations and critical H2S and NH3

Fig. 1. Schematic of the lab scale bio®lter: (1) air compressor; (2) air ®lter; (3) nutrient tank; (4) ¯ow meter; (5) H2S gas

cyl-inder; (6) NH3 gas cylinder; (7) inlet chamber; (8) sampling

(4)

loads to meet this euent concentration were deter-mined at various space velocity (or residence time) ac-cording to the following equation (Chung et al., 1998). SV _C a oÿ Ce Vm Cln Ks Cln or Coÿ Ce h a Vm Cln Ks Cln; 4 where SV is the dÿ1_{F S}

a Lÿ1; F the gas ¯ow rate

(m3_{/day); S}

a the column cross-section (m2); L the

packing height (m); h the residence time (s); Cothe inlet

concentration (ppm); Cethe outlet concentration (ppm);

and a is the conversion coecient (kg-bead ppm /g-S or g-N). Let Cebe 0.1 or 1 ppm in Eq (4), and the

maxi-mum Co can be estimated at various residence times.

The loading rate (g-S or g-N/m3_{-h) of the bio®lter can be}

obtained from Eq (5).

Load SV C_a o: 5

2.7. Analytical methods

Inlet and outlet H2S gas concentrations of the

bio-®lter were measured either continuously by a single point monitor (MDA Scienti®c) ranging from 50 to 1500 ppb or periodically by gas detector tubes (GASTEC) ranging from 1 to 60 ppm. Inlet and outlet NH3 gas

concentrations were measured either continuously by a single 0 point monitor (MDA Scienti®c) in the range of 0.1±10 ppm, or periodically by gas detector tubes (GASTEC) in the range of 5±100 ppm. In all continuous experiments, H2S/NH3 concentration was recorded as

the variation of H2S/NH3concentration was within 5%

in 2 h. The total 12 data were recorded and then aver-aged to be the H2S or NH3 outlet concentration.

Sam-ples were taken 48 times per day for the periodic measurement with the gas detector tubes. The chemical composition of the cycling solution was also determined. Nitrate, nitrite and sulfate concentrations in the solution were measured by ion chromatography (Dionex 4500i). Ammonium was determined using an ion-speci®c elec-trode. Sul®te was determined by titration using a stan-dard potassium iodide-iodate titrant and a starch indicator (APHA, 1992). Sul®de was determined using an ion-speci®c electrode. Elemental sulfur was deter-mined by reaction with cyanide to produce thiocyanate, which was quantitated as FeSCN3ÿ

6 (Schedel and

Truper, 1980). The pH value in the circulating solution was measured 96 times per day. The data were obtained from two or more duplicate tests.

3. Results and discussion

3.1. H2S/NH3removal eciency in continuous operation

The removal eciencies for dierent ratios (e.g., 1:1, 1:2, and 2:1) of H2S/NH3 gas mixtures at various times

are illustrated in Figs. 2A and B. When H2S and NH3

were mixed in a ratio of 1:1, the optimum set of removal eciency of H2S and NH3 occurred at around the

sev-enth day (e.g., 98% for H2S and 99% for NH3) and then

slightly decreased toward the end of the experiment. The removal eciencies for H2S and NH3 still remained

above 95% at the end of the two-month treatment (data not shown). The reduction in eciency may be caused by the conversion of H2S to SO¸4 by T. thioparus CH11

and the conversion of NH3to NO-2by N. europaea. Both

SO¸

4 and NOÿ2 cause the acidic condition of the system,

resulting in the decrease in the removal eciencies of the bio®lter. During the entire period, the ¯uctuation of pH was between 7.5 and 6.9. When H2S and NH3 were

mixed in a ratio of 1:2, high NH3 concentration (120

ppm), surprisingly, did not inhibit the H2S metabolism

by T. thioparus CH11. On the contrary, it enhanced the

Fig. 2. Relationships between the removal eciencies and op-erating time at dierent ratios (1:1, 1:2, 2:1) of H2S/NH3gas

mixtures. (A) H2S removal eciency (B) NH3 removal

e-ciency. Conditions: 1:1 (60:60), 1:2 (60:120), and 2:1 (120:60) by concentration ppm/ppm; Super®cial gas velocity 36 l/h.

(5)

H2S metabolism when compared to the result of 1:1

ratio, as shown in Fig. 2A. The H2S removal eciency

increased with operating time and reached as high as 99%. The reason for this enhancement may be that the ample supply of nitrogen promotes the metabolic ac-tivity of T. thioparus CH11 (Starkey, 1934). By mea-suring the pH of the bio®lter, we found that the decrease in pH was insigni®cant regardless of the generation of acid products (pH ranging 7.5±7.2). Thus, the high H2S

removal eciency maintained in the bio®lter could be due to the neutralization reaction between NH3aq and

SO¸

4 , which prevents the occurrence of acidi®cation. In

the case of mixing H2S and NH3with a ratio of 1:2, the

removal eciency of NH3is only about 85%. The lower

removal eciency may be caused by the inhibition of the N. europaea at higher NH3concentration (e.g., 120 ppm)

(Hunik et al., 1992). When H2S and NH3were mixed in

a ratio of 2:1, the H2S removal eciency increased

during the operation period and reached a maximum value of 95%. However, at such high H2S concentration

(120 ppm), the NH3 removal was signi®cantly

sup-pressed and decreased down to 72%. This dramatic re-duction in NH3 removal (Fig. 2B) can be explained by

two reasons. The high concentrations of H2S and the

low pH of the bio®lter, which dropped to 6.5 after a 9-day operation, hinder the nitri®cation of N. europaea (Joye and Holibaugh, 1995).

3.2. Product analysis

Mass balances of sulfur and nitrogen in the bio®lter at dierent ratios of H2S/NH3supply are listed in Tables

2 and 3. As indicated in Table 2, when 60 ppm of H2S (in

cases of 1:1 and 1:2) was supplied for 14 days, product varieties and their conversion ratios remained un-changed regardless of the NH3concentration. However,

when the inlet H2S concentration was increased to 120

ppm, the production ratio of S¸_{(8.0%) increased, which}

suggested the incomplete H2S metabolism of T.

thiopa-rus CH11 at high H2S concentration (Table 2). As a

result, the residual S¸_{might suppress the nitri®cation of}

N. europaea (Joye and Holibaugh, 1995). This is con-®rmed by the signi®cantly lower NH3removal eciency

in the case of a H2S/NH3 mixing ratio of 2:1, (Fig. 2B).

Table 3 shows that the main product of NH3oxidation

by the bio®lter was NOÿ

2. Because only NOÿ2 was

pro-duced from the Nitrosomonas, the small amount of NOÿ 2

detected might result from either the contamination of Nitrobacter or the chemical oxidation. Since no Nitrob-acter was detected in the experiment, NOÿ

3 was likely the

oxidation product of NOÿ

2. When the bio®lter was

sup-plied with high NH3 (e.g., 120 ppm), 5% of residual

NH

4/NH3was found in the reactor. This residual NH4/

NH3 may neutralize the acid product of H2S (i.e. SO¸4)

and thus maintain the pH of the operation. This is the reason why high H2S removal eciency was found in the

case of 1:2 mixing ratio (Fig. 2A). In this case, the NOÿ 2

concentration was about 396 mg lÿ1_{. Hence, the low}

NOÿ

2 concentration would not result in toxicity to

Nitrosomonas (Sato et al., 1988). 3.3. Bioaerosol analysis

Applying bio®ltration on deodorized process has been proved to be very promising (Leson and Winer, 1991). However, because bio®lters contain tremendous amounts of microorganisms, it is necessary to assess the environmental risk associated with the bacteria released from the bio®lter when large quantities of waste gases are treated.

Table 4 shows the number of microorganisms in the outlet exhaust when the bio®lter was conducted con-tinuously for two months. Apparently, as microorgan-isms were immobilized in Ca-alginate, the exhaust contained only small amounts of nitrify bacteria (3 CFU/m3_{) and neutrophic Thiobacillus spp. (6±15}

CFU/m3_{) in any cases. Because only Thiobacillus and}

Nitrosomonas were found in the outlet exhaust, it is reasonable to believe that the seeded species remained dominant. Since the immobilized cell bio®lter was free from the contamination by heterotrophic bacteria (Table 4), this system can be considered safe if placed close to populated areas.

3.4. Pressure drop

The relationship between pressure drop and super®-cial gas velocity is an important parameter in deter-mining the operational cost. The in¯uence of super®cial gas velocity on pressure drop is shown in Fig. 3. In this experiment, the super®cial gas velocity was raised gradually from 18 to 180 hÿ1 _{and the temperature was}

Table 2

Sulfur mass balances in the bio®lter at dierent ratios of H2S/NH3supply

Mixture ratio (ppm/ppm) H(g-S/kg-bead)2S Removed SO ¸ 4 Produced (g-S/kg-bead) (%) S 0_Produced (g-S/kg-bead) (%) SO ¸ 3 Produced (g-S/kg-bead) (%) S ¸_Produced (g-S/kg-bead) (%) 1:1a _3.21 _{0.81 (25.2)} _{2.32 (72.3)} _{0.065 (2.0)} _{0.016 (0.5)} 1:2 3.17 0.78 (24.6) 2.28 (72.0) 0.082 (2.6) 0.026 (0.8) 2:1 6.19 1.12 (18.0) 4.45 (72.0) 0.125 (2.0) 0.499 (8.0) a_{1:1 equals 60:60 (ppm/ppm).}

(6)

maintained 30°C. When the operation reached the steady state (about 3 days), a new super®cial gas velocity was selected. Inspection of the ®gure reveals that pres-sure drop of the bio®lter increases with increasing su-per®cial gas velocity and in a good linear behavior. The possible reason is due to no signi®cant biomass accu-mulation from which would slow the growth rate of the autotrophic bacteria. Consequently, the bio®lter exhib-ited excellent dispersion characteristics.

3.5. Kinetic analysis

The apparent kinetic parameters of the maximum removal rate and half-saturation constant to degrade

H2S (5±120 ppm) under the presence of NH3(60 or 120

ppm) are calculated by the Lineweaver-Burk method and the results are shown in Fig. 4. Inspection of the related coecient of the regression equation (i.e., 0.998) shown in Fig. 4 indicates that NH3 does not intervene

the metabolism of H2S by T. thioparus CH11.

Further-more, the apparent maximum removal (1.11 g-S/day/kg-bead) and the apparent half-saturation constant (34.6 ppm) are similar to those reported by Chung et al. (1997) under a NH3-free atmosphere.

Fig. 5A illustrates the kinetic analysis of NH3

re-moval by N. europaea in the range of 5±65 ppm under various H2S concentrations. Compared to the result of

the H2S-free inlet, the in¯uence of 60 ppm H2S in the

inlet on NH3 removal of gas mixture was not obvious.

However, once the concentration of H2S was raised to

Fig. 4. Relationship between 1/R and 1/Cln;H2S of H2S

degra-dation in the bio®lter. Conditions: H2S (5±120 ppm), NH3(60

or 120 ppm), and super®cial gas velocity (36±72 l/h). Fig. 3. Pro®le of pressure drop vs. super®cial gas velocity for

the autotrophic bio®lter. Table 4

Bioaerosol analysis in the outlet exhaust of the autotrophic bio®lter Mixture ratio Medium type

Nutrient PDA Thiosulfate Modi®ed Waksman Nitrifying

1:1 ND ND 9 ND 3

1:2 ND ND 6 ND ND

2:1 ND ND 15 ND NDa

a_{ND < 3 CFU/m}3_{, Unit: CFU/m}3_.

Table 3

Nitrogen mass balances in the bio®lter at dierent ratios of H2S/NH3supply

Mixture ratio (ppm/ppm) NH(g-N/kg-bead)3Removed NH 4/NH3Amount (g-N/kg-bead) (%) NO ÿ 2 Produced (g-N/kg-bead) (%) NO ÿ 3 Produced (g-N/kg-bead) (%) 1:1a _1.67 _{0.017 (1.0)} _{1.62 (97.0)} _{0.03 (2.0)} 1:2 2.98 0.151 (5.0) 2.83 (95.0) ) 2:1 1.49 0.020 (1.4) 1.44 (96.6) 0.03 (2.0) a_{1:1 equals 60:60 (ppm/ppm).}

(7)

120 ppm, a concave-down curve occurred, suggesting that high H2S concentration inhibited the metabolism of

NH3by N. europaea (Julitte et al., 1993). It is dicult to

categorize H2S as a competitive or non-competitive

in-hibitor from the existing experimental evidence. Fig. 5B shows the eect of high NH3 concentration (120±200

ppm) on the NH3removal by N. europaea when the inlet

H2S concentration in the mixture was maintained at 60

ppm. As mentioned in Fig. 2B, high NH3concentration

may inhibit the nitri®cation of N. europaea. Therefore, further research on the inhibition kinetic analysis at high NH3concentration is essential. The inhibition coecient

(Ki) was calculated as 80.5 ppm from Eq. (3) using linear

regression.

3.6. Criteria for design of scale-up bio®lters

A complete removal for H2S and NH3 can be

achieved only at less than critical loading rate. The loading rate of the system is de®ned as the amount of inlet gas per unit of time, per volume of packing material (g-S or-N/m3_{-h). Thus, inlet gas concentrations play an}

important role in the design of a scale-up bio®lter when constant packing material volume and space velocity are used. Finding the maximum inlet concentration and the optimal loading rate, therefore, is crucial for the bio®lter operation. The relationship between the maximum inlet concentration and residence time (space velocity) for H2S and NH3 removal is shown in Fig. 6. From the

operational point of view, to reach the ambient air standards of H2S and NH3(0.1 and 1 ppm, respectively),

the maximum inlet concentrations were 58 ppm for H2S

and 164 ppm for NH3, as speci®ed in the ®gure, at the

residence time of 72 s. Under such condition, the max-imum loading rate was determined as 3.8 g-S/m3_{-h for}

H2S and 5.6 g-N/m3-h for NH3, respectively (data not

shown).

4. Conclusions

It is demonstrated that the removal eciency remains above 95% when treating H2S and NH3 gas mixtures of

various ratios with a co-immobilized cell bio®lter. Hy-drogen sul®de appears to be an inhibitory substrate for NH3 removal, while NH3concentration has little eect

on H2S removal. From our studies, appropriate inlet

ratios of H2S/NH3 (e.g., <2) and excellent dispersion

characteristics can maintain the operational stability of the bio®lter. Maximum inlet concentrations to reach the ambient air standards are 58 ppm for H2S and 164 ppm

for NH3 at residence time of 72 s. These inlet

concen-trations are in the range of gas emission from the live-stock farming and the accompanying wastewater treatment.

Fig. 6. Relationship between maximum inlet concentration and residence time for H2S and NH3removal by the autotrophic

bio®lters.

Fig. 5. Relationship between 1/R and 1/Cln;NH3 of NH3

degra-dation in the bio®lters. (A) Eect of H2S concentrations on

NH3removal by N. europaea at NH3(5±65 ppm) and H2S (0,

60 or 120 ppm). (B) Eect of NH3concentrations on NH3

re-moval by N. europaea at NH3(120±200 ppm) and H2S 60 ppm.

(8)

Acknowledgements

Funding for this work was provided partially by the National Science Council, ROC.

References

APHA, 1992. American Public Health Association, Standard Method: Examination of Water and Wastewater, 18th ed. APHA, New York.

Buchnan, R.E., Gibbons, N.E., 1974. Bergey's Manual of Determinative Bacteriology, 7th ed. Williams and Wilkins Co, Baltimore, MD, pp. 52±54.

Cho, K.S., Zhang, L., Hirai, M., Shoda, M., 1991. Removal characteristics of hydrogen sul®de and methanethiol by Thiobacillus spp. isolated from peat in biological deodor-ization. Journal of Fermentation and Bioengineering 71, 44±49.

Chung, Y.C., Huang, C., 1998. Biotreatment of ammonia in air by an immobilized Nitrosomonas europaea bio®lter. Envi-ronmental Progress 17, 70±76.

Chung, Y.C., Huang, C., Tseng, C.P., 1996a. Biodegradation of hydrogen sul®de by a laboratory-scale immobilized Pseudo-monas putida CH11 bio®lter. Biotechnology Progress 12, 773±778.

Chung, Y.C., Huang, C., Tseng, C.P., 1996b. Operation optimization of Thiobacillus thioparus CH11 bio®lter for hydrogen sul®de removal. Journal of Biotechnology 52, 31±38.

Chung, Y.C., Huang, C., Tseng, C.P., 1997. Removal of hydrogen sulphide by immobilized Thiobacillus sp. strain CH11 in a bio®lter. Journal of Chemical Technology and Biotechnology 68, 58±62.

Chung, Y.C., Huang, C., Tseng, C.P., 1998. Advanced study of H2S removal by Thiobacillus novellus CH3 bio®lter in

autotrophic and mixotrophic environments. Journal of Environmental Engineering-ASCE 124, 362±367.

Eikum, A.S., Storhang, R., 1986. Odour Prevention and Control of Organic Sludge and Livestock Farming. Elsevier Applied Science Publishers, London, pp. 12±18.

Hirai, M., Ohtake, M., Shoda, M., 1990. Removal kinetic of hydrogen sulphide, methanethiol and dimethyl sulphide by peat bio®lters. Journal of Fermentation and Bioengineering 70, 334±339.

Hunik, J.H., Meijer, J.G., Tramper, J., 1992. Kinetic of Nitrosomonas europaea at extreme substrate, product and salt concentrations. Applied Microbiology and Biotechnol-ogy 37, 802±807.

Ikeda, H., Asaba, H., Takeuchi, Y., 1980. Removal of H2S,

CH3SH and (CH3)3N from air by use of chemically treated

activated carbon. Japan Journal of Chemical and Engineer-ing 21, 91±97.

Joye, S.B., Holibaugh, J.T., 1995. In¯uence of sul®de inhibition of nitri®cation on nitrogen regeneration in sediments. Science 270, 623±625.

Julitte, L.Y., Michael, R., Daniel, J.A., 1993. Inhibition of ammonia oxidation in Nitrosomonas europaea by sulfur compounds. Applied and Environmental Microbiology 59, 3718±3727.

Leson, G., Winer, A.M., 1991. Bio®ltration: an innovative air pollution control technology for VOC emission. Journal of the Air & Waste Management Association 41, 1045±1054. Ottengraf, S.P.P., Oever, A.H.C., 1983. Kinetics of organic

compound removal from waste gases with a biological ®lter. Biotechnology and Bioengineering 25, 3089±3102. Prakasam, T.B.S., Loehr, R.C., 1972. Microbial nitri®cation

and denitri®cation in concentrated wastes. Water Research 6, 859±867.

Prosser, J.I., 1989. Autotrophic nitri®cation in bacteria. Ad-vances in Microbial Physiology 30, 125±181.

Ryer-Power, J.E., 1991. Health eects of ammonia. Plant/ Operations Progress 10, 228±232.

Sato, C., Leung, S.W., Schnoor, J.L., 1988. Toxic response of Nitrosomonas europaea to copper in inorganic medium and wastewater. Water Research 22, 1117±1127.

Sato, C., Schnoor, J.L., McDonald, D.B., Huey, J., 1985. Test medium for the growth of N. europaea. Applied and Environmental Microbiology 32, 1101±1107.

Schedel, M., Truper, H.G., 1980. Anaerobic oxidation of thiosulfate and elemental sulfur in Thiobacillus denitrifans. Archives of Microbiology 2-3, 205±210.

Starkey, R.L., 1934. Cultivation of organisms concerned in the oxidation of thiosulfate. Journal of Bacteriology 28, 365±386.

Sublette, K.L., Sylvester, N.D., 1987. Oxidation of hydrogen sul®de by continuous cultures of Thiobacillus denitri®cans. Biotechnology and Bioengineering 29, 753±758.

Yang, Y., Allen, E.R., 1994. Bio®ltration control of hydrogen sul®de. 1 design and operational parameters. Journal of the Air & Waste Management Association 44, 863±868.