Removal of toluene in waste gases using a biological trickling filter

The removal of toluene from waste gas was studied in a trickling biofilter. A high level of water recirculation (4.7 m h^–1) was maintained in order to keep the liquid phase concentration constant and to achieve a high degree of wetting. For loads in the range from 6 to 150 g m^–3 h^–1 the maximum volumetric removal rate (elimination capacity) was 35±10 g m^–3 h^–1, corresponding to a zero order removal rate of 0.11±0.03 g m^–2 h^–1 per unit of nominal surface area. The surface removal was zero order above the liquid phase concentrations of approximately 1.0 g m^–3, corresponding to inlet gas concentrations above 0.7–0.8 g m^–3. Below this concentration the surface removal was roughly of first order. The magnitude of the first order surface removal rate constant, k_1A , was estimated to be 0.08–0.27 m h^–1 (k_1A a=24–86 h^–1). Near-equilibrium conditions existed in the gas effluent, so mass transfer from gas to liquid was obviously relatively fast compared to the biological degradation. An analytical model based on a constant liquid phase concentration through the trickling filter column predicts the effluent gas concentration and the liquid phase concentration for a first and a zero order surface removal. The experimental results were in reasonable agreement with a very simple model valid for conditions with an overall removal governed by the biological degradation and independent of the gas/liquid mass transfer. The overall liquid mass transfer coefficient, K_La, was found to be a factor 6 higher in the system with biofilm compared to the system without. The difference may be explained by: 1. Difference in the wetting of the packing material, 2. Mass transfer occurring directly from the gas phase to the biofilm, and 3. Enlarged contact area between the gas phase and the biofilm due to a rough biofilm surface.     

The Rotary Trickle‐Bed Reactor – A New Reactor Concept for Biological Gas Purification

A new type of reactor employed to the biological gas purification is presented. The avoidance of clogging in the carrier packing is achieved by i) the use of a structured, rotating carrier packing, ii) a definite liquid irrigation regime during start‐up, operation and clean‐up time phases, iii) an on‐line determination and control of the fixed biofilm mass. A uniform biofilm thickness is generated by an optimized liquid irrigation of the carrier packing with spray nozzles. The detachment of the fixed biomass is accomplished by liquid shear forces generated with jet nozzles. The time‐scheduled operation regime of the reactor is founded on the on‐line quantification of the immobilized biomass, which results in a new quality of process governing of biotrickling reactors applied to gas purification. This is proved by the experimental results of pressure drop, dynamic liquid holdup as well as the volumetric degradation rates. The degradation of styrene was investigated in laboratory and field experiments showing a maximal volumetric degradation rate of 150 g m^–3 h^–1 at a pollutant load of 200 g m^–3 h^–1. The feasibility of this reactor prototype is demonstrated by employing it to the elimination of industrial waste gas.     

Exhaust gas purification using biocatalysts (fixed bacteria monocultures) — the influence of biofilm diffusion rate (O2) on the overall reaction rate

Summary This paper presents results of experiments on the influence of O₂ and substrate (pollutant) concentration on the overall reaction rate of a trickle-bed reactor used for biological waste gas purification. The biocatalyst was a pollutant-specific bacterial monoculture fixed on porous glass carriers. The conversion of acetone and propionaldehyde, as model pollutants that are easily soluble in water, was measured. Under constant hydrodynamic conditions (gas and liquid flow rates) the inlet pollutant concentration was varied. The O₂ partial pressure in the model gas was increased to investigate the influence of O₂ supply on pollutant conversion. At higher pollutant concentrations (>117 mg acetone.m^-3 gas and > 150 mg propionaldehyde.m^-3 gas) higher concentrations of dissolved O₂ led to a significant rise in the maximum degradation capacity of the reactor. This maximum reaction rate was independent of the pollutant mass flow. It seems that the diffusion of O₂ in the biofilm is rate-determining. The reaction rate at lower inlet concentrations was not affected by the improved O₂ supply. Here the external mass transfer through the liquid film limits the reaction rate and the maximum separation efficiency of about 80% at a residence time of 1.2s (space velocity 3000h^-1) is achieved.     

Microbial cleaning of waste gas containing volatile organic compounds in a bioreactor system with a closed gas circuit

The cleaning of the exhaust gases of a bioreactor containing volatile hydrocarbons in a bioreactor system with a closed gas circuit is described. The bioreactor system consisted of three different reactor types: a stirred tank which was filled with hydrocarbon-containing waste water to simulate the exhaust gases of a remediation process; a trickle-bed reactor for aerobic treatment of the exhaust gas from the stirred tank; and a photoreactor containing an algae culture which assimilated CO₂ from the trickle-bed reactor and also produced O₂. With this bioreactor system, it was possible to efficiently remove volatile organic compounds (VOC) from the waste gases. Depending on the type of waste water investigated, elimination rates of 41% to 93% of BTEX (benzene, ethylbenzene, toluene, xylene) and 29% to 53% of VCH (volatile chlorinated hydrocarbons) were obtained. Due to the photosynthesis of the algae in the system's photoreactor, oxygen concentrations between 12% and 18% [v/v], equivalent to about 57% to 83% DOT, were obtained. This concentration permitted the aerobic degradation to be carried out without having to add fresh air. The trickle-bed reactor and the photoreactor worked continuously, whereas the waste water in the stirred bioreactor was replaced in different batches. The accumulation of toxic compounds in the nutrient solutions of the trickle-bed (EC-50 > 30 g/l) and of the photoreactor (EC-50 > 35 g/l) was low. Carbon dioxide concentrations in the gas flow were higher than in fresh air (1% to 3% [vol/vol]), but no long-term accumulation of CO₂ occurred. This means that the algae in the photoreactor were active enough to assimilate the CO₂ which had been produced. They were also able to produce sufficient oxygen for aerobic hydrocarbon degradation. The system described is a first step towards treating waste gases which results from the bioremediation of hydrocarbon-contaminated media in a closed gas circuit without any emission (e.g. VOC, CO₂, germs).     

Propene removal from synthetic waste gas using a hollow-fibre membrane bioreactor

 Hollow-fibre modules containing microporous membrane material were evaluated as bioreactors for waste gas treatment. The reactors were inoculated with the propene-utilizing strain Xanthobacter Py2, which formed a biofilm on the inner side of the fibres. The removal of the poorly soluble volatile propene from synthetic waste gas was monitored for up to 170 days. The maximum removal rates were 70–110 g propene per m³ reactor per hour. A gas residence time of 80 s was required to remove 95% of an initial propene concentration of 0.84 g/m³. The presence of ammonium in the liquid medium resulted in the development of an additional population of nitrifying organisms. Therefore, nitrate was used as the source of nitrogen in later experiments. During long-term operation, the propene removal rates gradually decreased. At low liquid velocities (1–5 cm/s) clogging of individual fibres with excess biomass was observed. Elevation of the liquid velocity in the fibres to 90 cm/s resulted in the formation of a dense biofilm and prevented clogging of the fibres. However, also at this high liquid velocity a gradual decrease in propene removal rate was observed. These results suggest that aging of biofilms is a very important factor in long-term operation of hollow-fibre bioreactors. Received: 24 November 1995 / Received revision: 14 February 1996 / Accepted: 20 February 1996     

A bioactive foamed emulsion reactor for the treatment of benzene-contaminated air stream

An adapted bioactive foamed emulsion bioreactor for the treatment of benzene vapor has been developed. In this reactor, bed clogging was resolved by bioactive foam as a substitute of packing bed for interfacial contact of liquid to gaseous phase. The pollutant solubility has been increased using biocompatible organic phase in liquid phase and this reactor can be applied for higher inlet benzene concentration. Experimental results showed a benzene elimination capacity (EC) of 220 g m⁻³ h⁻¹ with removal efficiency (RE) of 85% for benzene inlet concentration of 1–1.2 g m⁻³ at 15 s gas residence time in bioreactor. Assessment of benzene concentration in liquid phase showed that a significant amount of transferred benzene mass has been biodegraded. By optimizing the operational parameters of bioreactor, continuous operation of bioreactor with high EC and RE was demonstrated. With respect to the results, this reactor has the potential to be applied instead of biofilter and biotrickling filters.     

Gas-phase methyl ethyl ketone biodegradation in a tubular biofilm reactor: microbiological and bioprocess aspects

A novel type of bioreactor was designed to clean VOCs-containing air.The operation of this reactor consists in mixing the polluted gas and a mistof nutrient solution in the presence of microorganisms in order to maximizecontact and transfer between gas, liquid and microorganisms and to promotethe degradation kinetics and the relative removal efficiency of thepollutant. A bacterial consortium acclimatized to MEK and containing apreponderance of Alcaligenes denitrificans was established under non-axenicconditions. On the tubular reactor's glass walls, a continuous biofilm wasdeveloped. This biofilm was rapidly contaminated by two fungi able todegrade MEK: Geotrichum candidum and Fusarium oxysporum. Their abundance inthe reactor is probably linked to the acidic conditions inside the biofilmand to their broader tolerance for low pH values concomitant with MEKdegradation. In the reactor, a maximum volumetric degradation rate of 3.5 kgMEK/m³ _reactor·d was obtained for arelative removal efficiency of 35%, whereas the latter was maintainedat 70% for more modest applied loadings of 1.5 kgMEK/m³ _reactor ·d. In liquid batchcultures, a biomass originating from the biofilm was able to degrade 0.40g_MEK/g_DCW·h at the optimal pH of 7. Aregular cycle of detachment-recolonization was observed during the operationof the bioreactor. The maximal degradation activity was obtained with a thinbiofilm and was not increased as the biofilm grew in thickness. The overalldegradation rate of the process did not appear to be limited by thediffusion of oxygen inside the biofilm. Over short periods of time, the MEKtransfer from the gaseous phase to the biofilm was neither affected by thepresence of the mist nor by the wetting of the biofilm. A better control ofthe biofilm pH led to improved performance in terms of removal rate but notin terms of relative elimination efficiency.     

Steady-state and transient-state performance of a biotrickling filter treating chlorobenzene-containing waste gas

Biotrickling filter (BTF) technology was applied for the treatment of waste gas containing a mixture of chlorobenzene and 1,2-dichlorobenzene. An adapted microbial community was immobilised on a structured packing material. The strategy followed was to reach high removal efficiencies at initially low mass loading rates followed by an increase of the latter. This procedure was successful and resulted in a short start-up period of only 2 weeks. A 3-month operation under steady-state conditions showed good performance, with >95% removal efficiency at a mass loading rate of 1,800 g m^–3 day^–1. Dimensionless concentration profiles showed that the chlorobenzenes were simultaneously degraded. Low dissolved organic carbon of 15 mg l^–1 and stoichiometric chloride concentrations in the trickling liquid indicated complete mineralisation of the pollutant. Transient-state experiments with five times higher mass loading rates caused a decrease in the removal efficiency that recovered rapidly once the mass loading rate returned to its original steady-state level. A progressive increase of the mass loading rate in a long-term performance experiment showed that the removal efficiency could be kept stable between 95 and 99% at loads of up to 5,200 g m^–3 day^–1 over several days. Above this mass loading rate, the elimination capacity did not increase any further. These results demonstrated that with a well-adapted inoculum and optimal operation parameters, a BTF system with excellent performance and stability that efficiently removes a mixture of cholorobenzene vapours from air can be obtained.     

Use of activated carbon as a buffer in biofiltration of waste gases with fluctuating concentrations of toluene

Fluctuations in contaminant concentrations often adversely influence the effectiveness of bioreactors for waste gas treatment. Application of an adsorbent to minimize such fluctuations could improve the overall process. Therefore the buffer capacity of a number of activated carbons and other adsorbents was tested. The buffer capacity of the adsorbents depends on the desired concentration range of the contaminants entering the bioreactor and on the time available for desorption. When fluctuations between 0 and 1000 mg toluene m^–3 were applied to a biofilter this resulted in significant concentrations of toluene leaving the biofilter. Using one selected type of activated carbon it was demonstrated that these fluctuations could be decreased to a value of about 300 mg m^–3, which subsequently completely degraded in the biofilter.     

Styrene degradation by <Emphasis Type="Italic">Pseudomonas</Emphasis> sp. SR-5 in biofilters with organic and inorganic packing materials

Pseudomonas sp. SR-5 was isolated as a styrene-degrading bacterium. In liquid culture containing 1% (v/v) styrene, more than 90% styrene was degraded in 53 h and the doubling time of SR-5 was 2 h. The removal of styrene gas was investigated in biofilters for 31 days using an organic packing material of peat and an inorganic packing material of ceramic inoculated with SR-5. The maximum-styrene-elimination capacities for peat and ceramic packing materials were 236 and 81 g m^–3 h^–1, respectively. The percentage of styrene converted to low molecular weight compounds including CO₂ in the peat and ceramic biofilters during a 10-day operation were estimated to be 90.4 and 36.7%, respectively. As the pressure drop in the peat bioflter at the end of experiment was significantly higher than that in ceramic biofilter, a biofilter using a mixture of peat and ceramic was tested. We determined that the maximum elimination capacity was 170 g m^–3 h^–1 and the production of low molecular weight compounds was 95% at a low pressure drop for this mixed packing material filter.     

Sustained degradation of n-pentane and isobutane in a gas-phase bioreactor

Summary Microorganisms were able to remove hydrocarbons (pentane and isobutane) from air by biological action in a columnar bioreactor with ceramic packing. The reactor was operated in a liquid continuous mode with gas recirculation and a slow addition of the organic-containing air. After a period of acclimation, the reactor has operated for 12 months with only pentane and isobutane as carbon sources. The gaseous hydrocarbons have been degraded throughout this period. The hydrocarbon removal rates measured between 1 and 2 g h^–1 m^–3. The microbes were shown to be able to degrade these gaseous hydrocarbons completely in a closed bioreactor without any additional nutrients.Research supported by the Advanced Industrial Concepts Division-Biological and Chemical Technologies Research. U.S. Department of Energy, under contract DE-AC05-84OR21400 with Martin Marietta Energy Systems. Inc.     

Mathematical modeling of wastewater decolorization in a trickle-bed bioreactor

This work focuses on mathematical modeling of removal of organic dyes from textile industry waste waters by a white-rot fungus Irpex lacteus in a trickle-bed bioreactor. We developed a mathematical model of biomass and decolorization process dynamics. The model comprises mass balances of glucose and the dye in a fungal biofilm and a liquid film. The biofilm is modeled using a spatially two-dimensional domain. The liquid film is considered as homogeneous in the direction normal to the biofilm surface. The biomass growth, decay and the erosion of the biofilm are taken into account. Using experimental data, we identified values of key model parameters: the dye degradation rate constant, biofilm corrugation factor and liquid velocity. Considering the dye degradation rate constant 1×10?? kg m?3 s?1, we found optimal values of the corrugation factor 0.853 and 0.59 and values of the liquid velocity 5.23×10?3?m?s?1 and 6.2×10?3?m?s?1 at initial dye concentrations 0.09433 kg m?3 and 0.05284 kg m?3, respectively. A good agreement between the simulated and experimental data using estimated values of the model parameters was achieved. The model can be used to simulate the performance of laboratory scale trickle-bed bioreactor operated in a batch regime or to estimate values of principal parameters of the bioreactor system.     

Evaluation of the performance of structured mixed packing and inert packing materials in toluene biotrickle-filtration

Packing materials play a key role in waste gas treatment. Organic and inert packing materials have their disadvantages, which may be minimized by mixed packing. In this study, various operating conditions were applied to evaluate the performance of structured mixed packing and inert packing materials in toluene biotricklefiltration. Four biotrickle filters were packed with structured mixed packing materials, namely, ceramic pall rings, ceramic rashig rings, and lava rock. Their toluene removal capacity was studied for 217 day using a laboratory-scale reaction under various operating conditions. The key elimination capacity (removal efficiency > 95%) ranking of the biotrickle filters was as follows: Structured mixed packing (306.20 ± 7.90 g/m³/h) > pall ring (156.71 ± 7.84 g/m³/h) > rashig ring (153.31 ± 6.14 g/m³/h) > lava rock (150.32 ± 9.19 g/m³/h). The structured mixed packing and inert packing resulted in excellent toluene-degrading biofilter performance under long-term operation. The structured mixed packing provided a more rapid startup rate and better process robustness than the inert packing did. The biotrickle filter with mixed packing materials had a high elimination capacity which makes it suitable for various real-life applications, whereas the capability of the inert packing material was more suitable for treating a steady low toluene load.     

A new bioreactor with a semi-fixed packing: investigation of degradation of phenol

A new bioreactor using a semi-fixed packing of frames with “sacks” made of a fabric of “Raschell” type stretched on them is proposed. The construction provides not only a large surface area of the biofilm carrier per unit volume of the apparatus, but also the possibility for an easy removal of the biomass after reaching a certain thickness of the biofilm increasing the gas velocity. Aerobic degradation of phenol in the new bioreactor, using microorganisms of the strain Pseudomonas putida, was studied. The experiments are carried out using water containing 0.7?g/l of phenol at a temperature of 27–30?°C. Different specific surface area of the packing (within 176 and 387?m²/m³) are studied. Degradation rates from 60 to 140?mg/(1?h) are attained. A retardation of the process at the end is observed, probably due to inhibition effect. This rate is 5–6 times higher than the rate observed when using free cells. At air velocity of 0.03–0.035?m/s (related to the total cross section of the bioreactor) the vibrations of the packing material lead to destruction and removal of the old biomass.     

Inhibition of the metabolism of nitrifying bacteria by direct electric current

The nitrifying bacteria in activated sludge and biofilms consisting of the bacteria immobilized on polypropylene packing were subjected to an electric current via two electrodes. In activated sludge, the metabolism of nitrifying bacteria was inhibited when the applied current was over 2.5 A m^–2, whilst in biofilms, inhibition began when the current reached 5 A m^–2. At 15 A m^–2, the nitrification rate of NH₄ ⁺-N in a biofilm with a bacterial density of 1.62 g total solids, dry wt m^–2 decreased to about 80% of its initial value. Ninety-two % of the initial biomass on the packing was retained after 36 h.     

Biotrickling filtration of isopropanol under intermittent loading conditions

This paper investigates the removal of isopropanol by gas-phase biotrickling filtration. Two plastic packing materials, one structured and one random, have been evaluated in terms of oxygen mass transfer and isopropanol removal efficiency. Oxygen mass transfer experiments were performed at gas velocities of 104 and 312 m h^?1 and liquid velocities between 3 and 33 m h^?1. Both materials showed similar mass transfer coefficients up to liquid velocities of 15 m h^?1. At greater liquid velocities, the structured packing exhibited greater oxygen mass transfer coefficients. Biotrickling filtration experiments were carried out at inlet loads (IL) from 20 to 65 g C m^?3 h^?1 and empty bed residence times (EBRT) from 14 to 160 s. To simulate typical industrial emissions, intermittent isopropanol loading (16 h/day, 5 day/week) and intermittent spraying frequency (15 min/1.5 h) were applied. Maximum elimination capacity of 51 g C m^?3 h^?1 has been obtained for the random packing (IL of 65 g C m^?3 h^?1, EBRT of 50 s). The decrease in irrigation frequency to 15 min every 3 h caused a decrease in the outlet emissions from 86 to 59 mg C Nm^?3 (inlet of 500 mg C Nm^?3). The expansion of spraying to night and weekend periods promoted the degradation of the isopropanol accumulated in the water tank during the day, reaching effluent concentrations as low as 44 mg C Nm^?3. After a 7-week starvation period, the performance was recovered in less than 10 days, proving the robustness of the process.     

Biotechniques for air pollution control

This paper gives an overview of present biological techniques for the treatment of off-gases and the techniques that are being developed at the moment. The characteristics, advantages, disadvantages, costs and application area are discussed and compared. Biological off-gas treatment is based on the absorption of volatile contaminants in an aqueous phase or biofilm followed by oxidation by the action of microorganisms. Biofilters, bioscrubbers and biotrickling filters are used for elimination of odour and bioconvertable volatile organic and inorganic compounds and are enjoying increasing popularity. This popularity is a result of the low investment and operational costs involved compared to physico-chemical techniques and the elimination efficiencies that can be obtained. The operational envelop is still extending to higher concentrations and gas flow rates (exceeding 200,000 m³ h^–1) and a broader spectrum of degradable compounds. Research and development on the use of membranes and the addition of activated carbon or a second liquid phase to the biological systems may lead to a more efficient elimination of hydrophobic compounds and buffering of fluctuating loads. Shorter adaptation periods can be obtained by inoculation with specialized microorganisms. Improved design and operation are made possible by the growing insights in the kinetics and microbiology and supported by the development of models describing biological off-gas treatment. In conclusion, biotechniques are efficient and cost effective in treating off-gases with concentrations of biodegradable contaminants up to 1–5 g/m³. They could play a justified and important role in air pollution control in the coming years.Abbreviations VOC volatile organic compound - NO _x gaseous oxides of nitrogen     

Long term performance of biotrickling filters removing a mixture of volatile organic compounds from an artificial waste gas: dichloromethane and methylmethacrylate

Two problems still hamper the widespread industrial application of biotrickling filters (BTFs) for waste gas treatment in practice: clogging of the filters at higher carbon loads and a decrease in the elimination of a target compound when more than one compound is present in the waste gas. To investigate these phenomena three identical BTFs removing dichloromethane (DCM) from an artificial waste gas were operated counter-current wise for 12 months at a DCM load of 0.94 Cmole-DCM/(m_r³ · h). After five months of operation methylmethacrylate (MMA) was added to the waste gas. Three different MMA loads were applied: 0.5, 1.0 and 1.5 Cmole-MMA/(m_r³ · h). Although the elimination of DCM in all three BTFs decreased after the introduction of MMA to the air stream, it stabilised at a lower steady-state value than before the MMA addition. MMA was completely degraded during the applied standard conditions. In all three filters biomass accumulation eventually caused clogging of the packing. In the filter with the lowest MMA load the first signs of clogging were observed only after 7 months of stable operation, illustrating the need for long term studies to evaluate process stability. Short term experiments have provided information about the system's dynamics and showed that an accumulation of intermediates and a subsequent adaptation of the biomass in the BTF will occur upon a step increase in MMA load. To evaluate whether a stable BTF operation without clogging is possible, a novel process parameter (the rate of Carbon Conversion per unit void packing Volume) is introduced which possibilities and limitations are discussed.     

Biofiltration of waste gases with the fungi Exophiala oligosperma and Paecilomyces variotii

Two biofilters fed toluene-polluted air were inoculated with new fungal isolates of either Exophiala oligosperma or Paecilomyces variotii, while a third bioreactor was inoculated with a defined consortium composed of both fungi and a co-culture of a Pseudomonas strain and a Bacillus strain. Elimination capacities of 77 g m^–3 h^–1 and 55 g m^–3 h^–1 were reached in the fungal biofilters (with removal efficiencies exceeding 99%) in the case of, respectively, E. oligosperma and Paecilomyces variotii when feeding air with a relative humidity (RH) of 85%. The inoculated fungal strains remained the single dominant populations throughout the experiment. Conversely, in the biofilter inoculated with the bacterial–fungal consortium, the bacteria were gradually overgrown by the fungi, reaching a maximum elimination capacity around 77 g m^–3 h^–1. Determination of carbon dioxide concentrations both in batch assays and in biofiltration studies suggested the near complete mineralization of toluene. The non-linear toluene removal along the height of the biofilters resulted in local elimination capacities of up to 170 g m^–3 h^–1 and 94 g m^–3 h^–1 in the reactors inoculated, respectively, with E. oligosperma and P. variotii. Further studies with the most efficient strain, E. oligosperma, showed that the performance was highly dependent on the RH of the air and the pH of the nutrient solution. At a constant 85% RH, the maximum elimination capacity either dropped to 48.7 g m^–3 h^–1 or increased to 95.6 g m^–3 h^–1, respectively, when modifying the pH of the nutrient solution from 5.9 to either 4.5 or 7.5. The optimal conditions were 100% RH and pH 7.5, which allowed a maximum elimination capacity of 164.4 g m^–3 h^–1 under steady-state conditions, with near-complete toluene degradation.     

Toluene degradation in the recycle liquid of biotrickling filters for air pollution control

Pollutant degradation in biotrickling filters for waste air treatment is generally thought to occur only in the biofilm. In two experiments with toluene degrading biotrickling filters, we show that suspended microorganisms in the recycle liquid may substantially contribute to the overall pollutant removal. Two days after reactor start up, the overall toluene elimination capacity reached a maximum of 125 g m⁻³ h⁻¹, which was twice that found during prolonged operation. High biodegradation activity in the recycle liquid fully accounted for this short-term peak of pollutant elimination. During steady-state operation, the toluene degradation in the recycle liquid was 21% of the overall elimination capacity, although the amount of suspended biomass was only 1% of the amount of immobilized biomass. The results suggest that biotrickling filter performance may be improved by selecting operating conditions allowing for the development of an actively growing suspended culture. Received: 16 June 1999 / Received revision: 17 November 1999 / Accepted: 15 December 1999