treatmentChemosphere 50 (2003) 145–153 www.elsevier.com/locate/chemosphere

Biological treatment process of air loaded with an ammonia and hydrogen sulfide mixture
Luc Malhautier a,*, Catherine Gracian a, Jean-Claude Roux a, Jean-Louis Fanlo a, Pierre Le Cloirec b a Ecole des Mines d’Als, Laboratoire du Gnie de l’Environnement Industriel, 6, Avenue de Clavires, 30319 Ales Cedex, France e e e b Ecole des Mines de Nantes, Dpartement Systmes Energtiques et Environnement, La Chantrerie, 4, rue Alfred Kastler, e e e BP 30723, 44307 NANTES Cedex 3, France Received 12 June 2001; received in revised form 10 July 2002; accepted 10 July 2002

Abstract The physico-chemical characteristics of granulated sludge lead us to develop its use as a packing material in air biofiltration. Then, the aim of this study is to investigate the potential of unit systems packed with this support in terms of ammonia and hydrogen sulfide emissions treatment. Two laboratory scale pilot biofilters were used. A volumetric load of 680 g H2 S mÀ3 empty bed dayÀ1 and 85 g NH3 mÀ3 empty bed dayÀ1 was applied for eight weeks to a unit called BGSn (column packed with granulated sludge and mainly supplied with hydrogen sulfide); a volumetric load of 170 g H2 S mÀ3 empty bed dayÀ1 and 340 g NH3 mÀ3 empty bed dayÀ1 was applied for eight weeks to the other called BGNs (column packed with granulated sludge and mainly supplied with ammonia). Ammonia and hydrogen sulfide elimination occur in the biofilters simultaneously. The hydrogen sulphide and ammonia removal efficiencies reached are very high: 100% and 80% for BGSn; 100% and 80% for BGNs respectively. Hydrogen sulfide is oxidized into sulphate and sulfur. The ammonia oxidation products are nitrite and nitrate. The nitrogen error mass balance is high for BGSn (60%) and BGNs (36%). This result could be explained by the denitrification process which would have occurred in anaerobic zones. High percentages of ammonia or hydrogen sulfide are oxidized on the first half of the column. The oxidation of high amounts of hydrogen sulfide would involve some environmental stress on nitrifying bacterial growth and activity. Ó 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Biofiltration; Ammonia; Hydrogen sulfide; Biological oxidation; Elimination efficiency

1. Introduction Ammonia and hydrogen sulfide are emitted into the atmosphere from industries (carcass processing plants, sewage treatment plants, composting works). These emissions, in addition to their own toxicity, constitute a source of olfactory nuisance. More and more, purification processes are based on the ability of some microorganisms to oxidize a variety of inorganic and organic

*

Corresponding author. Fax: +33-466-782701. E-mail address: luc.malhautier@ema.fr (L. Malhautier).

compounds into mineral end-products (Diks and Ottengraf, 1991). Biofiltration as a technique to control air pollution is used mainly for odour removal. In skin manufacture, biofiltration treatment of hydrogen sulfide reached an elimination efficiency of 97%, the contaminated air flow rate was 6000 m3 hÀ1 (Bijl, 1987). Similarly, the treatment of air (20 000 m3 hÀ1 ), coming from oil production manufacture and loaded with hydrogen sulfide, using a biofiltration unit (340 m3 ) gave an elimination efficiency of 89% (Knauf, 1995). Amirhor et al. (1995) studied the treatment, using a biofilter (500 m3 ), of air (40 000 m3 hÀ1 ), coming from a composting work, containing mainly ammonia and reduced sulfur

0045-6535/03/$ - see front matter Ó 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 5 - 6 5 3 5 ( 0 2 ) 0 0 3 9 5 - 8

146

L. Malhautier et al. / Chemosphere 50 (2003) 145–153

compounds. The elimination efficiencies reached were of 98% for ammonia and 80% for reduced sulfur compounds. In biofilters, the packing material commonly used are: peat or different types of peat (Furusawa et al., 1984; Togashi et al., 1986; Martin et al., 1989; Dalouche et al., 1989), compost (Bohn, 1975; Rands et al., 1981; Knauf, 1995), soil, heather branches and bark chips (Carlson and Leiser, 1981; Kurita and Kamata, 1990). Nowadays, peat and compost are usually mixed with bulking agents (Amirhor et al., 1995). Activated carbon is also widely used as microorganism support for removing undesirable molecules from wastewater and gaseous effluents (De Laat et al., 1985; Fanlo et al., 1995; Malhautier et al., 1997, 1998). Sludge used as packing material in biofilters for hydrogen sulfide removal results from several steps applied to digested sludge from sewage treatment plants. The sludge processing consists of dewatering and conditioning as spherical particles which are screened in order to recuperate the median fraction. The packing material has high liquid retention and develops a high mass transfer area without inducing high pressure drops. Sludge is thus favourable to bacterial colonization and nutrient solutions containing sources of carbon, nitrogen and phosphorus do not have to be supplied to maintain the growth and activity of sulfur oxidizing bacteria (Kowal, 1993; Sammani-Vaute et al., 1995). Hence, it is interesting to develop the biofiltration process using sludge and more accurately to study the potential of this support for olfactory nuisance removal. Moreover, in order to better control and optimise biofiltration for a full scale implementation (Bonnin et al., 1990; Amirhor et al., 1995; Michel et al., 1996; Smet et al., 1998), it is important to study the performance of laboratory pilot biofilters whose inlet gaseous effluent is a malodorant pollutants mixture. In this work, the removal of a gaseous effluent constituted of two model compounds (ammonia and hydrogen sulfide) is considered. The aim of this work is on the one hand to determine the performances of the pilot units and on the other to better understand ammonia, and hydrogen sulfide elimination mechanisms.

Convective drying was applied to dehydrated sludge in mixture with dry recycled product using a rotary drum supplied with air–vapour mixture at 450 °C. Then the air–vapour mixture was separated from granulated sludge (93–95% of dryness) using a cyclonic filter and finally the granulated sludge was separated by a vibrating screen. The median fraction is used as the packing material and the others were recycled for the Swiss–Combi process. The physical characteristic studies of the packing are as follows. The medium diameter (dp) varied as a function of the water content (WF) expressed in %, as follows: dpðmmÞ ¼ 1:90 þ 2:35 Â 10À2 WF ð1Þ

This material was dense (qapp ¼ 700 kg mÀ3 ), its void degree (e) was estimated at 0.35. This packing was saturated in humidity for a water content of 40%. From the void degree and the medium diameter, the external specific area (Asp ), which included the mass and energy transfer, could be estimated at 2940 m2 mÀ3 . The chemical characteristics studies have shown that this packing was constituted of organic matter (60–65%), nitrogen (4.3%), and sulfur (0.9%). Ammonia amount in sludge was estimated 4 mg gÀ1 dry packing. 2.2. Pilot unit Two pilot units were run and consisted of glass columns packed with granulated sludge (Fig. 1). The col-

2. Material and methods 2.1. Packing material The packing material came from the sewage treatment plant at Annecy (France). Sludge resulting from the various stages of sewage processing was firstly digested. This microbial process results in stabilization of the organic matter, some destruction of pathogens and in odor control. Swiss–Combi processing was then applied to dehydrated digested sludge (20–25% of dryness):
Fig. 1. Diagram showing pilot unit for ammonia and hydrogen sulfide removal. (1) Nutrient solution, (2) humidification system, (3) percolate waters, (4) air and packing sampling, (5) packed column (sludge), (6) Tedlar bag containing ammonia, (6*) Tedlar bag containing hydrogen sulfide, (7) air, (8) inlet and outlet effluent sampling, (9) pressure drop measurements and (10) pump.

L. Malhautier et al. / Chemosphere 50 (2003) 145–153

147

umn had a diameter of 0.1 m and a height of 1.4 m. Each column was equipped with five outlets, located at intervals of 0.2 m. Ammonia and hydrogen sulfide levels were measured and samples taken at these outlets. The gaseous effluent was obtained by diluting gaseous ammonia or hydrogen sulfide, which were stocked in Tedlar bags of 80 l, in atmospheric air. The packing was kept at constant moisture using a spray nozzle, which regularly sprayed with demineralised water. The percolate waters were collected at the bottom of the column. The pressure drop was followed using a digital differential manometer, 0–25 mbar (Bioblock company, Illkirch, France). 2.3. Sampling 2.3.1. Gas Ammonia and hydrogen sulfide were transferred in aqueous solution by bubbling the gaseous influent in 0.1 l of a solution of hydrogen chloride 0.1 M and 0.1 l of a solution of zinc acetate 1.77 M, respectively. The gas flow rate was adjusted to 1.5 l minÀ1 . The sampling delay was 5 min (Le Cloirec et al., 1988). 2.3.2. Liquid Percolate waters were collected at the bottom of the column three times a week. An aliquot (50 ml) was used to determine nitrogen and sulfur compounds. 2.3.3. Packing The packing material was sampled in order to perform physical, chemical, and biological analysis once a week at the level 3 of the biofilter. At the end of the experiment, the bioreactors were taken down and the packing was collected for every level. Each parcel of sludge was then submitted to a divider in order to constitute some representative material. These samples were then used for the determination of the water content, the nitrogen and sulfur species amounts, and the microorganisms counting. A suspension of sludge packing material was obtained as follows: 2 g of material were mixed with 9 ml of demineralised water and stirring by vortex during 3 min. 2.4. Chemical analysis The ammonia concentration in the hydrogen chloride solution, in the percolate waters, and in the liquid suspension of packing were then evaluated by a spectrophotometric method using Nessler reagent (French Standard Methods, 1975, NF T 90-015). The hydrogen sulfide concentrations in the zinc acetate solution, in the percolate water and in the liquid suspension were then evaluated by the iodometric method (Standard Methods for the Examination, 1995).

Sulfate, thiosulfate, nitrite, and nitrate in the percolate waters and in the liquid suspension of packing were measured by Waters Quanta 4000 capillary electrophoresis (Waters Quanta 4000). Total nitrogen and sulfur amounts analysis were performed by the Service Central dÕAnalyse, CNRS Vernaison, France. Nitrogen of samples, which were previously burning at 1050 °C, was transformed to nitrogen oxides which were reduced to molecular nitrogen before being quantified using a catharometer apparatus (Service Central dÕAnalyse, CNRS Vernaison, France). Sulfur of samples, which were previously burned at 1350 °C, was transformed to sulfur dioxide which was quantified using an infrared detector apparatus (LECO, France). pH (pH-meter Quick, Bioblock company, Illkirch, France) was also measured. 2.5. Biomass counting The biomass measurements were obtained as follows: Most probable number (MPN)-Griess counting for nitrifying bacteria: The protocol used for MPN-Griess counting was based on that of Both et al. (1990). The dispersed packing suspension was subsampled before sedimentation for a series of dilution (1:10) and inoculations of MPN plates with a two-fold Nitrosomonas autotroph medium based on that of Matulewich et al. (1975) with (NH4 )2 SO4 0.50 g lÀ1 or a twofold Nitrobacter autotroph medium (Schmidt et al., 1973) with NaNO2 , 1 g lÀ1 . Samples were incubated for three months at 28 °C in the dark, the Griess reagent was added, and the number of Nitrosomonas or Nitrobacter cells per gram of sludge was determined with CochranÕs tables (Cochran, 1950). MPN counting for thiobacilli: The dispersed packing suspension was subsampled before sedimentation for a series of dilution (1:10) and inoculations of MPN plates with a two-fold Thiobacillus thioparus autotroph medium (Lafleur et al., 1993). Samples were incubated for 15 days at 28 °C in the dark, the reagent bromocresol purple was added, and the number of neutrophil thiobacilli cells per g of sludge was determined with CochranÕs tables. 2.6. Pilot unit procedure The operating conditions of two biofilters were as follows (Table 1). Firstly, they were supplied similarly with a gaseous influent containing an equimolar mixture of ammonia and hydrogen sulfide. Then, one was supplied with a gaseous influent four times more concentrated with ammonia than with hydrogen sulfide (BGNs) during eight weeks, where as the other was supplied with a gaseous influent four times more concentrated with hydrogen sulfide than with ammonia

148

L. Malhautier et al. / Chemosphere 50 (2003) 145–153

Table 1 Operating conditions of two biofilters used for the treatment of an air loaded with a mixture of ammonia and hydrogen sulfide Operating conditions Gas velocity (m hÀ1 ) Humidification (ml dayÀ1 ) Nutrient solution Inoculation Nitrifying bacteria Sulfide oxidizing bacteria Room temperature (°C) BGNs 100 600* Water Yes No 20–30 BGSn 100 600* Water Yes No 20–30

Table 2 Ammonia and hydrogen sulfide applied concentrations and loads for the biofilters Load [H2 S]inlet (mg mÀ3 ) Cv inlet (g H2 S mÀ3 dayÀ1 ) [NH3 ]inlet (mg mÀ3 ) Cv inlet (g NH3 mÀ3 dayÀ1 ) Time (weeks) [H2 S]inlet (mg mÀ3 ) Cv inlet (g H2 S mÀ3 dayÀ1 ) [NH3 ]inlet (mg mÀ3 ) Cv inlet (g NH3 mÀ3 dayÀ1 ) Time (weeks) BGNs 140 340 70 170 0–14 70 170 140 340 15–22 BGSn 140 340 70 170 0–14 280 680 35 85 15–22

BGNs, BGNs: Biofilters packed with sludge; Cv: Volumetric load.

(BGSn) during eight weeks. The applied loads are given in Table 2. The pilot performances are followed according to different factors: lag phase, elimination efficiency and pressure drop. 2.7. Inoculation Equivalent volumes of autotrophy nutrient solution and activated sludge which was sampled from an urban wastewater treatment plant, were mixed and ammonium sulphate was added at a final concentration of 0.5 g lÀ1 . After the incubation period, the sludge in the column was inoculated with 0.5 l of the acclimatized sludge suspension (Malhautier et al., 1997).

3. Results and discussion 3.1. Reactors efficiency The lag phase for nitrite and nitrate detection in percolate waters, according to the detection limits of the

analysis method use, is seven weeks irrespective of the biofilter. The waiting period for sulfate detection in percolate waters, according to the detection limits of the analysis method use, is one week irrespective of the biofilter. The oxidation of hydrogen sulfide is observed at the beginning of the experiment and then it is not necessary to seed the packing material with a suspension enriched with sulfide oxidizing microorganisms. Nitrite and nitrate are detected in percolate waters seven weeks after the beginning of the study whereas the delay is generally lower when ammonia is the only pollutant. Smet and Van Langenhove (2000) observed no microbiological start-up period in the case of the treatment of high concentrated ammonia loaded waste gases in compost biofilters. Hence, the acclimation phase necessary for the ammonia removal of a mixture of ammonia and hydrogen sulfide gas involves a removal efficiency loss. Moreover, in a biofilter, distinct microbial populations frequently interact with each other (Atlas and Bartha, 1997; Devinny et al., 1999), the sulfide oxidizing community could have a negative effect on the nitrifying community growth. The biofilter environmental conditions and a long generation time of 10 h (Bock et al., 1989) could be also unfavourable to the nitrifying community growth. The hydrogen sulfide oxidation is then carried out more easily than the ammonia oxidation. Kowal (1993) showed that hydrogen sulfide oxidation also happens using a chemical pathway which is observed during the adaptation phase of sulfide-oxidizing bacteria. Mac Nevin and Barford (1999) suggested that aerobic sulfide removing biofilters may remove sulfide by a two step process involving initial chemical oxidation to elemental sulfur followed by slower biological oxidation to sulfate. The removal efficiency as a function of time is shown in Fig. 2 for BGNs and BGSn. The H2 S elimination is complete as soon as the experiment begins. As the operating conditions were modified, the BGSn H2 S abatement decreases to 70% (18th week) before reaching the initial value. The BGNs H2 S removal remains the same (100%). NH3 removal reached low values during the first five weeks (6 40%). These results are explained by the adaptation phase of nitrifying bacteria community during this delay. As a steady state is reached, the NH3 oxidation efficiency varied between 60% and 80%. Nevertheless, the operating condition modifications involve an immediate NH3 efficiency decrease corresponding to 30% for BGSn and 10% for BGNs. From week 20, the BGSn efficiency becomes superior to BGNs NH3 performances. These results show on the one hand that the different NH3 proportions have no effect on H2 S efficiency as been shown by Chung and his collaborators (2000). The high concentration of H2 S in the gaseous influent would not decrease the NH3 oxidation efficiency, in return, the NH3 load would be the factor which would limit the elimination capacity.

L. Malhautier et al. / Chemosphere 50 (2003) 145–153

149

670 g H2 S mÀ3 empty bed dayÀ1 and 60–80% for a volumetric load of 335 g NH3 mÀ3 empty bed dayÀ1 . The use of sludge needs no nutritive aqueous solution supply, nor buffer solution as it is the case for activated carbon (Malhautier et al., 1998) or peat (Gracian et al., 1996) use. Subject to appropriate humidification and for hydrogen sulfide concentrations inferior to 140 mg mÀ3 , the pressure drops measured are low (inferior to 600 Pa mÀ1 at a velocity of 100 m hÀ1 ). 3.2. Substrate removal mechanism approaches The study of the mass balance and the degradation product proportions are presented and allow better understanding of the NH3 and H2 S elimination mechanisms by the microbial communities having colonized the sludge. The mass balance for nitrogen and sulfur is realized between the three phases: liquid, gas, and solid: X X ðN À NH3 Þgi þ Nsi ¼ X X X Nsf þ ðN À NO2 Þlf ðN À NH3 Þgf þ X X þ ðN À NO3 Þlf þ ðN À NHþ Þlf ð2aÞ 4 X X ðS À H2 SÞgi þ Ssi ¼ X X X Ssf þ ðS À SO4 Þlf ðS À H2 SÞgf þ X X 0 þ ðS À S Þlf þ ðS À H2 SÞlf ð2bÞ g: gas; s: solid; l: liquid; i: initial; f: final. The error mass balance is defined as follows: X . X  eð%Þ ¼ N inlet À N outlet N inlet à 100%
X X . X

Fig. 2. BGSn and BGNs removal efficiency evolution as a function of time and volumetric load. (j) BGNs; (Ã) BGSn; Load 1: 170 g NH3 mÀ3 dayÀ1 , 340 g H2 S mÀ3 dayÀ1 ; Load 2: BGSn: 85 g NH3 mÀ3 dayÀ1 ; 680 g H2 S mÀ3 dayÀ1 ; Load 2 BGNs: 340 g NH3 mÀ3 dayÀ1 ; 170 g H2 S mÀ3 dayÀ1 .

The pressure drop as a function of time is shown in Fig. 3. It shows that the pressure drop measured for BGSn and BGNs are similar up to the 15th week. The NH3 and H2 S volumetric load modifications involve an increase of the pressure drop up to 2000 Pa mÀ1 for BGSn and the keeping of the pressure drop at 600 Pa for BGNs. The treatment of a large H2 S volumetric load leads to a diminution of the void degree and hence an increase of the pressure drop. Three assumptions could explain this result: an acidification of the reactor which involves a collapse of sludge granules and then a reduction of the bed permeability, a major development of the biomass, due to high substrate quantity, colonizing BGSn and a deposit of sulfur (S0 ), not soluble species intermediate between hydrogen sulfide oxidation and sulfate. The data sets show that simultaneous NH3 and H2 S elimination occurred in biofilters and the efficiencies carried out are very high: 100% for a volumetric load of

eð%Þ ¼

S inlet À

S outlet

ð3aÞ  S inlet à 100% ð3bÞ Fig. 3. Pressure drops evolution in biofilters packed with sludge. (j) BGNs; (Ã) BGSn.

The error mass balance for sulfur is low and corresponds to 5% and 8% for BGNs and BGSn respectively. Nevertheless, an important loss of nitrogen is observed for BGNs (22%) and BGSn (35%). It is probable that ammonia supplying the biofilters is converted into species other than nitrite and nitrate. Two nitrogen species have not been considered: organic nitrogen coming from leached biomass and gaseous nitrogen. The latter is the end product of the denitrification process, occurring in aerophilic or anaerobic zones in biofilters, by which nitrite and nitrate are reduced into gaseous nitrogen. This nitrogen error mass balance is often observed for ammonia removal by biofiltration (Martin et al., 1995; Lau et al., 1996). The outlet nitrogen and sulfur mass balances are calculated for the two biofilters. The results are shown in Figs. 4 and 5. For nitrogen, the results show a different behaviour for the two reactors. Gaseous and organic

150

L. Malhautier et al. / Chemosphere 50 (2003) 145–153

Fig. 4. Ammonia oxidation mass balance for BGNs and BGSn. solved.

––(organic þ gaseous) N; j––Nitrite;

––Nitrate; Ö–NH3 dis-

Fig. 5. Sulfur oxidation mass balance for BGNs and BGSn. ( ) Elemental sulfur; (j) H2 S dissolved; (Ã) sulfate.

nitrogen percentage is higher for BGSn (60%) than for BGNs (36%). According to the mass balance results, it can be assumed that nitrite and nitrate reduction to gaseous nitrogen in BGSn is greater than in BGNs. The pressure drops measured for BGSn are high and could involve serious channel plugging and bypassing as was shown by Sorial et al. (1995) and the creation of anaerobic zones responsible for denitrification. The ammonia dissolved percentage is equivalent for BGNs and BGSn biofilters. Nevertheless, the nitrite and nitrate products percentage is superior for BGNs (28%) than for BGSn (7%). It seems probable that H2 S oxidation involves a decrease of the nitrification activity. This assumption could be supported by the counts realized at the end of the experiment (Fig. 6): the number of nitrifying bacteria having colonized BGNs (103 –104 bacteria gÀ1 packing material) is 10 superior to the BGSn one (102 –103 bacteria gÀ1 ). Moreover, in BGSn, ammonium oxidizing bacteria were not detected at the level 1, and nitrite oxidizing bacteria were not detected, irrespective of the BGSn height. Accumulation of nitrite

is also observed for BGNs (19%) (Fig. 4). Two assumptions could explain this result: • Nitrifiers are sensitive populations (Maier et al., 2000) and physiological conditions are not favourable to the nitrite oxidizing bacteria growth (nitrite and H2 S concentrations, sulfur accumulation); • The kinetic constant of the microbiological oxidation of NH3 to nitrite is higher than the biological oxidation of nitrite to nitrate. It has been shown that the nitrite oxidizing reaction, requiring approximately 100 moles of nitrite to fix 1 mole of carbon dioxide is less efficient than the ammonium oxidizing reaction requiring 34 moles of ammonium to fix 1 mole of carbon dioxide (Maier et al., 2000). H2 S is completely oxidized, the degradation products are mainly sulfate and sulfur. The amount of elemental sulfur that was formed is high (Fig. 5) and the sulfur percentage is greater for BGSn than for BGNs. Usually, sulfide is oxidized all the way to sulfate. Indeed, sulfate

L. Malhautier et al. / Chemosphere 50 (2003) 145–153

151

Fig. 6. Number of nitrifying and sulfide oxidizing bacteria. Z: relative height (H =H0 ); (j) Sulfide oxidizing bacteria; (Ã) Ammonia oxidizing bacteria; ( ) Nitrite oxidizing bacteria.

is the end product of sulfur compound oxidation, but sulfur, sulfite, or polythionates may be accumulated, sometimes transiently, by most species of Thiobacillus colonizing sewage treatment areas and sources of sulfur gases (Holt et al., 1994). In our study, the amount of sulfur did not decrease as a function of time. The mass balance of the sulfur species are comparable to the one established by Fanlo (1994) who studied the treatment of air loaded with hydrogen sulfide using a biofilter packed with sludge: H2 S ¼ ð0:2–0:4Þ À SO2À þ ð0:6–0:8Þ À S0 4 ð4Þ

Another assumption is that the kinetic constant of the microbiological oxidation of sulfide to sulfur is higher than the biological oxidation of sulfur to sulfate. This phenomenon could explain the amount of elemental sulfur that was formed. 3.3. Concentration profiles Fig. 7 represents the experimental concentration profiles obtained for BGSn and BGNs biofilters. For BGNs, the concentration profile shows that H2 S is completely eliminated on the first half of the biofilter, 72% of NH3 inlet load is degraded in the first half of the column and 21% in the second half of the reactor. pH data measured as a function of the height of the column decrease from 8.5 to 7 in the first half of the column, the pH decrease is generated for both nitrification and sulfide oxidizing reactions and ammonia consumption. For BGSn, the concentration profile indicates that 89.5% of H2 S and 71% of NH3 inlet loads are eliminated in the first half of the column, 10.5% and 0% of respectively H2 S and NH3 inlet load are eliminated in the second half of the unit pilot. pH data measured as functions of the

Fig. 7. Concentration profiles in BGNs as a function of Z (17th week): (a) concentration profiles in BGSn as a function of Z (17th week) and (b) BGNs, BGSn: Biofilters (7:85 Â 10À3 m3 ) packed with sludge; *150 ml were sprayed four times a day.

height of the column are of 4.5–5 and corroborate sulfide oxidizing reactions. High percentages of NH3 or H2 S are oxidized on the first half of the column. Nitrification and H2 S oxidation activity are greater in the first half of the column than in the second. These results show that the NH3 oxidation has no negative effect on H2 S removal and the high concentration of H2 S seems to decrease NH3 removal with only 71% of ammonia inlet load (85 g NH3 mÀ3 dayÀ1 ) being treated in the first half of the column BGSn. Chung et al. (2000) studied the biotreatment of H2 S and NH3 containing waste gases by a co-immobilized cells biofilter. They showed

152

L. Malhautier et al. / Chemosphere 50 (2003) 145–153

that the NH3 concentration had little effect on H2 S removal efficiency. Enumerations of nitrifying and sulfide oxidizing bacteria having colonized sludge used as packing material of BGSn supplied with a gas whose the H2 S concentration is 140 mg mÀ3 , indicate that nitrifying bacteria have not been observed at the first level and do not exceed 103 bacteria gÀ1 sludge. Conversely, the counts of sulfide oxidizing bacteria are the greatest at the first level and remain superior to the nitrifying bacteria ones (102 to 105 fold) along the column. It can be suggested that the H2 S oxidation might affect the nitrifying community structure. This has been shown to be the case in the removal of simultaneous H2 S and volatile organic compounds in compost and granular activated carbon biofilters (Webster et al., 1996). The effects of H2 S oxidation involve varying degrees of environmental stress. Moreover, concentrations of contaminant decline as the air passes through the biofilter, so that concentrations are much higher near the influent end than at the effluent end. The characteristics of the biological communities change accordingly. It is probable that, in BGSn and BGNs, the influent end supports a dense biomass and is more prone to acid generation. The microorganisms at the effluent end may be starving and producing little or no biomass (Hugler et al., 1996). High NH3 concentration has no effect on H2 S oxidation efficiency. Even high H2 S concentration does not seem to decrease the NH3 removal efficiency, the concentration profiles study shows that the oxidation of high amounts of H2 S has a negative effect on the nitrifying community growth and activity. Moreover, the denitrification process would be favoured.

concentration profiles as a function of the column height showed that high percentages of NH3 or H2 S are oxidized in the first half of the column. It is probable that, in BGNs and BGSn, the influent end supports a dense biomass and is more prone to acid generation. It is also probable that the oxidation of high amounts of H2 S involves negative effects on nitrifying bacteria community colonization and activity.

References
Amirhor, P., Kuter, G.A., Andrade, M.D., 1995. Biofilters and biosolids. Water Environ. Technol. 3, 44–48. Atlas, R., Bartha, R., 1997. Quantitative ecology: numbers, biomass and activities. In: Microbial ecology: fundamentals and applications. Benjamin/Cummings Science Publishing, CA, USA, pp. 218–280. Bijl, J.J.W., 1987. VT-biofilter, environmental technology applied microbiology biotechnology. Proc. 2nd European Conference on Environmental technology, Amsterdam, The Netherlands, pp. 358–361. Bock, E., Koops, H.P., Harms, H., 1989. Nitrifying bacteria. In: Schlegel, H., Dowien, B. (Eds.), Autotrophic Bacteria. Science Tech. Publishers, Madison, Springer-Verlag, Berlin, pp. 81–96. Bohn, H.L., 1975. Soil and compost filters of malodorant gases. J. Air Pollut. Control Assoc. 25, 953–955. Bonnin, C., Laborie, A., Paillard, H., 1990. Odor nuisances created by sludge treatment: problems and solutions. Water Sci. Technol. 22, 65–74. Both, G.H., Gerards, S., Laanbroek, H.J., 1990. Most probable number of chemolitho-autotrophic nitrite-oxidizing bacteria in well drained grass-land soils: stimulation by high nitrite concentrations. FEMS Microbiol. Ecol. 74, 287–294. Carlson, D.A., Leiser, C.P., 1981. Soil beds for control of sewage odors. J. Water Pollut. Control Feder. 38, 829–840. Chung, Y.C., Huang, C., Tseng, C.P., Pan, J.R., 2000. Biotreatment of H2 S and NH3 -containing waste gases by co-immobilized cells biofilter. Chemosphere 41, 329–336. Cochran, W.G., 1950. Estimation of bacterial densities by means of the most probable number. Biometrics March, 105–106. Dalouche, A., Lemasle, M., Le Cloirec, P., Martin, G., Besson, G., 1989. Utilisation de biofiltres pour lÕpuration de gaz e chargs en composs azots et soufrs. Proc. 8th World Air e e e e Clean Congress, The Hague, The Netherland, September 4, pp. 11–15. De Laat, J., Bouanga, F., Dore, M., Mallevialle, J., 1985. Influence du dveloppement bactrien au sein des filtres e e de charbon actif en grains sur lÕlimination de composs e e organiques biodgradables ou non biodgradables. Water e e Res. 19, 1565–1578. Devinny, J.S., Deshusses, M.A., Webster, T.S., 1999. In: Biofiltration for air pollution control. Lewis Publishers, CRC Press, LLC, FL., USA, pp. 81–109. Diks, R.M., Ottengraf, S.P.P., 1991. Process engineering aspects of biological waste gas purification. Intern. Symposium Environ. Biotechnol., Ostende, Belgique, April, pp. 353–368.

4. Conclusion Simultaneous NH3 and H2 S elimination occurred in biofilters and the efficiencies achieved are very high: 100% for a volumetric load of 680 g H2 S mÀ3 empty bed dayÀ1 and 60–80% for a volumetric load of 340 g NH3 mÀ3 empty bed dayÀ1 . The use of sludge involves no nutritive aqueous solution supply, nor buffer solution. Subject to appropriate humidification and for H2 S concentrations inferior to 140 mg mÀ3 , the pressure drops measured are low (inferior to 600 Pa mÀ1 at a velocity of 100 m hÀ1 ). H2 S is oxidized into sulphate and sulfur and the growth of sulfide-oxidizing bacterial community seems to involve an important diminution of the void degree. The NH3 oxidation products are nitrite and nitrate. A significant nitrogen error mass balance is observed for BGSn (60%) and BGNs (36%). This result could be explained by the denitrification process which would have occurred in aerophilic or anaerobic zones in the reactors and in particular for BGSn. The NH3 and H2 S

L. Malhautier et al. / Chemosphere 50 (2003) 145–153 Fanlo, J.L., 1994. Transfert et transformation dÕhydrogne e sulfur en racteurs biotiques. Application  la dsodorisae e a e tion par biolavage et biofiltration, Thse de Doctorat, e Universit de Montpellier II, Ecole des Mines dÕAls, pp. e e 159–182. Fanlo, J.L., Brandy, J., Leloirec, P., Guey, C., Degorce-Dumas, J.R., 1995. Dsodorisation par biolavage et biofiltration-cas e de lÕhydrogne sulfur. Rcents progrs en gnie des e e e e e Procds 9, 55–60. e e French Standard Methods, 1975. NF T90-015, Recueil de Normes Francaises, part 3, pp. 120–122. ß Furusawa, N., Togashi, I., Hirai, M., Shoda, M., Kubota, H., 1984. Removal of hydrogen sulfide by a biofilter with fibrous peat. Int. J. Ferment. Technol. 62, 589–594. Gracian, C., Fanlo, J.L., Le Cloirec, P., 1996. Ammonia removal using biofiltration and Holland Pale Peat and granulated wastewater sludge as packing material, Eurodeur 96, Paris, June, pp. 95–103. Holt, J.G., Krieg, N.R., Sneath, P.H., Staley, J.T., Williams, S.T., 1994. Aerobic chemolitotrophic bacteria and associated organisms. In: Hensyl, W.R. (Ed.), BergeyÕs Manual of Determinative Bacteriology, 9th edition. Williams & Wilkins, Baltimore, MA, USA, pp. 427–455. Hugler, W.C., Cantu-De la Garza, J.G., Villa-garcia, M., 1996. Biofilm analysis for an odor-removing trickling filter. Proc. 89th Annual Meeting & Exhibition of the Air & Waste Management Association, Air & Waste Management Association, Pittsburgh, PA, 1996. Knauf, S.A., 1995. Biofilter application with high concentrations of hydrogen sulfide in a waste water treatment plant and an oil mill. Proc. 88th Annual Meeting & Exhibition of the Air & Waste Management Association, Pittsburgh, Penna, paper 95-MP9A.03. e a Kowal, S., 1993. Dsodorisation sur biofiltre  support consommable. Application du procd B.S.E pour lÕlimine e e e e e ation de lÕhydrogne sulfur. Thse de doctorat. Universit e e de Provence, Ecole des Mines dÕAls, pp. 109–145. Kurita, M., Kamata, O., 1990. Deodorization systems used at sewage treatment plants in Nagoya city. Sewage Works Japan, 114–119. Lafleur, R., Roy, E.D., Couillard, D., Guay, R., 1993. Determination of iron oxidizing bacteria numbers by a modified MPN procedure. In: Torma, Apel & Brierley (Eds.) Biohydrometallurgical Technologies, vol. II. TMS, pp. 433–441. Lau, A.K., Bruce, M.P., Chase, R.J., 1996. Evaluating the performance of biofilters for composting odor control. J. Environ. Sci. Health A 31 (9), 2247–2273. Le Cloirec, P., Lemasle, M., Martin, G., 1988. Odors: analysis and concentrations in many cases. Pollut. Atm. 2, 107–110. Mac Nevin, D., Barford, J., 1999. Adsorption and biological degradation of ammonium and sulfide on peat. Water Res. 33, 1449–1559. Maier, R., Pepper, I., Gerba, C., 2000. Biogeochemical cycling. In: Environmental Microbiology. Academic Press, San Diego, CA, USA, pp. 332–340.

153

Malhautier, L., Degorce-Dumas, J.R., Degrange, V., Bardin, R., Le Cloirec, P., 1997. Serological determination of Nitrobacter species in a deodorizing granular activated carbon filter. Environ. Technol. 18, 275–283. Malhautier, L., Degrange, V., Guay, R., Degorce-Dumas, J.R., Bardin, R., Le Cloirec, P., 1998. Estimating size and diversity of nitrifying communities in deodorizing filters using PCR and immunofluorescence. J. Appl. Microbiol. 85, 255–262. Martin, G., Le Cloirec, P., Lemasle, M., Cabon, J., 1989. Rtention de produits odorants sur tourbe. Proc. 8th World e Air Clean Congress, The Hague, The Netherland, September 4, pp. 373–378. Martin, G., Lemasle, M., Taha, S., 1995. The control of gaseous nitrogen pollutant removal in a fixed peat bed reactor. J. Biotechnol. 46, 15–21. Matulewich, V.A., Strom, P.F., Finstein, M.S., 1975. Length of incubation for enumerating nitrifying bacteria present in various environments. Appl. Microbiol. 29, 265–268. Michel, M.C., Lessard, P., Fanlo, J.L., Buelna, G., Brandy, J., 1996. Etude dÕun procd de traitement de lÕhydrogne e e e sulfur pour applications en station dÕpuration. Proc. 19th e e International Symposium on Wastewater Treatment, November, pp. 99–105. Rands, M.B., Cooper, D.E., Woo, C., Fletcher, G.C., Rolfe, K.A., 1981. Compost filters for hydrogen sulfide removal from anaerobic digestion and rendering exhausts. J. Water Pollut. Control Feder. 53, 185–189. Sammani-Vaute, L., Fanlo, J.L., Le Cloirec, P., 1995. Elimination de composs odorants sur charbon actif, tourbe et e e boues de station dÕpuration: cas dÕun effluent gazeux dÕquarrissage. Odours & VOCÕs February, pp. 9–13. e Schmidt, E.L., Molina, J.A., Chiang, C., 1973. Isolation of chemo-autotrophic nitrifiers from Moroccansoils. Bull. Ecol. Res. Commun. (Stockholm) 17, 166–167. Smet, E., Lens, P., Langenhove, H., 1998. Treatment of waste gases contaminated with odorous sulfur compounds. Critical Rev. Environ. Sci. Technol. 28, 89–117. Smet, E., Van Langenhove, H., 2000. Abatement of high concentrated ammonia loaded waste gases in compost biofilters. Water Air Soil Pollut. 119, 177–190. Sorial, G.A., Smith, F.L., Suidan, M.T., Biswas, P., Brenner, R.C., 1995. Evaluation of trickle bed biofilter media for toluene removal. J. Air Waste Manage. Assoc. 45, 801– 810. Standard Methods for the Examination of Water and Wastewater, 1995. In: Eaton, A., Clesceri, L., Greenberg, A. (Eds). American Public Health Association, 19th edition, Washington, DC, USA, pp. 4–127. Togashi, I., Suzuki, M., Hirai, M., Skodu, M., Kuboku, H., 1986. Removal of ammonia by a peat biofilter without and with nitrifier. J. Ferment. Technol. 64, 425–432. Webster, T.S., Devinny, J.S., Torres, E.M., Basrai, S.S., 1996. Microbial ecosystems in compost and granular activated carbon biofilters. Biotech. Bioeng. 53, 296–303.