Experimental studies on a kinetic model for design and operation of activated sludge processes

Previous experimentation in our laboratory has shown that the classical theory developed for continuous growth of pure cultures in completely mixed aerobic systems in which the recycle cell concentration factor, c (where c = X_R/X), is a selectable system constant, did not provide a suitable model for the heterogeneous (natural) populations of the activated sludge process. Another model was derived in which the recycle cell concentration, X_R was employed as a system constant instead of c, and computational analysis was performed. Laboratory pilot plant experimentation was undertaken in order to determine whether a “steady state” in aerator biological solids concentration, X?, and substrate concentration, S?, could be approached under this mode of operation. Studies were performed at various organic feed concentrations holding dilution rate, D, at 0.125 hr^?1, hydraulic recycle ratio, α, at 0.25, and X^R at 10,000 mg/liter. Also, values of maximum specific growth rate, μ_max, and saturation constant, K_s were determined. It was found that the model approached the steady state condition with heterogeneous populations more closely than did the classical model, and the high degree of treatment efficiency predicted by the model was demonstrated experimentally.     

The application of constant recycle solids concentration in completely mixed oxygen activated sludge process

For better operational control of the completely mixed oxygen activated sludge process (CMOAS), a study concerning the kinetics, performance, and operational stability of the Ramanathan-Gaudy model was conducted. Short-term experiments were conducted at various dilution rates (1/9, 1/6, 1/3, 1/1.5, and 1/1.0 hr^?1) by using two recycle solids concentration values (5000 and 10,000 mg/liter). The influent substrate was an actual industrial organic wastewater (soft drink waste) and its concentration was maintained at 1000 mg/liter COD. The hydraulic recycle ratio, α, was maintained at 0.30. It was found that for CMOAS system with constant recycle cell concentration, a “steady state” with respect to reactor biological solids and effluent COD at different dilution rates could be attained. No appreciable dilute-out of reactor biological solids and substrate was observed up to the dilution rate of 1 hr^?1 for both systems of different X_R (5000 and 10,000 mg/liter). For the system of X_R = 5000 mg/liter, except the dilution rate of hr^?1, the effluent filtrate COD was lower than 100 mg/liter, the aerator biological solids concentration was about 1550 mg/liter, and the COD removal efficiency was higher than 90% for all dilution rates. For the system of X_R = 10,000 mg/liter, the effluent filtrate COD was lower than 71 mg/liter, the aerator biological solids concentration was about 2750 mg/liter, and the COD removal efficiency was higher than 90% throughout all the dilution rates selection in the present study. The value of the Sludge Volume Index (SVI) was the range of 37.0 to 58.5 and provided good settleability of sludge. The sludge yield was 0.53 for the system of X_R = 5000 mg/liter and 0.57 for the system of X_R = 10,000 mg/liter. The carbohydrate and the protein content of the cells were 10.1–21.6% and 35.6–50.6%, respectively. For predicting the reactor biological solid and effluent COD of the CMOAS system by using the Ramanathan-Gaudy model, two sets of values for the biological kinetic constants should be considered since it provided the best fit of predicted values of the observed values. In the present study, μ_m = 0.4 hr^?1, k_s = 92 mg/liter for 1/3 ? D ? 1, and μ_m = 0.05 hr^?1, k_s = 11.1 mg/liter for 1/9 ? D < 1/3 were used to calculate the predicted values of reactor biological solid and effluent filtrate COD.     

Response to various shock loadings of the constant recycle-sludge-concentration activated-sludge process

The stability of the model of a completely mixed activated-sludge process holding the recycle sludge concentration, X_R, as a system constant subjected to pH, temperature, potassium cyanide, and phenol shock loading was investigated. Soft-drink bottling wastewater was used and maintained at 1000 mg/liter chemical oxygen demand (COD). The hydraulic ratio and recycle sludge concentration were maintained at 0.3 and 7000 mg/liter, respectively. An initial dilution rate of ¼ hr^?1 was maintained for pH and temperature shock loading, with ¼ and ? hr^?1 for KCN shock loading and ¼, ?, and \documentclass{article}\pagestyle{empty}\begin{document}$\frac{1}{16}$\end{document} hr^?1 for phenol shock loading. It was found that the present system could handle pH shock loading as low as 4.0 and as high as 10.4 without any serious disruption of biological solid concentration and filtrate COD. At pH 4.0 shock loading, filamentous organisms were predominant. Temperature shock loading could be handled from 23 to 36°C without any leakage of effluent filtrate COD. At 46°C temperature shock, a 14 hr period was required to recuperate to the new steady state and provided only 85% of COD removal efficiency. For KCN (50 mg/liger) and phenol (85 mg/liter) shock loading, the dilution rates should be lower than \documentclass{article}\pagestyle{empty}\begin{document}$\frac{1}{16}$\end{document} hr^?1 in order to shorten the transient period and improve the effluent quality. Biological kinetic constants included cell yield value, maximum growth rate, and the saturation constant, which was varied with the qualitative shock applied.     

Biological kinetic behaviors and operational performances in aerated and oxygenated systems

Biological kinetic behaviors of the oxygenated and aerated activated sludge process were studied and compared in both once-through and constant sludge recycle systems. The models derived by Herbert, Elsworth, and Telling [J. Gen. Microbiol., 14 , 601 (1956)] and Ramanathan and Gaudy [Biotechnol. Bioeng., 11 , 207 (1969)] were used for the studies of once-through and constant sludge recycle systems, respectively. Soft drink waste water was used for the growth limiting substrate. Temperature was controlled within 30 ± 2°C. The influent substrate concentration was maintained at 1,000 mg/liter. The experiments were conducted at various dilution rates (from \documentclass{article}\pagestyle{empty}\begin{document}$ \frac{1}{9} $\end{document} to 1/1.0 hr^?1), and recycle solids concentration values (from 5,000 to 10,000 mg/liter), with hydraulic recycle ratio, α, at 0.3. Biological kinetic constants were evaluated and compared. It was found that these constants were different for the aerated and oxygenated systems within a certain range of dilution rates studied. The critical dilution rates for diluting out effluent chemical oxygen demand (COD) occurred at 0.1 and 0.2 hr^?1in the once-through operation, and 0.2 and 0.4 hr^?1in the sludge recycle operation for aerated and oxygenated systems, respectively. Observed sludge yield values and specific growth rate were varied with the type of aeration and with and without constant sludge recycle concentration applied. Sludge carbohydrates and proteins content in the oxygenation system (cell recycle) were 10.1–21.6% and 35.6–52.2%. Sludge volume index in the air and oxygenation systems varied from 41.4 to 354 and 31.9 to 58.5, respectively.     

Effect of Fluoride on Nitrification of a Concentrated Industrial Waste

The potential for biological nitrification of an industrial waste containing 4,000 mg of ammonia N (NH₄⁺-N) and 10,000 mg of fluoride per liter was investigated. Ammonium sulfate and sodium fluoride were tested in various combinations of 100 to 2,000 mg of NH₄⁺-N per liter and 0 to 5,000 mg of F⁻ per liter in suspended-growth stirred-tank reactors containing enriched cultures of nitrifying bacteria from a municipal sewage treatment plant. The stirred-tank reactors were fed once per day at a constant hydraulic retention period and cell retention time of 10 days. Temperature was 23°C, and pH was 7.0 to 7.5. Clarified secondary effluent was used to make up feeds and to provide minor nutrients. Steady-state data, confirmed by mass balances, were obtained after five to six retention periods. In the absence of fluoride, nitrification efficiency was near 100% for up to 500 mg of NH₄⁺-N per liter. The influence of fluoride was studied at a low ammonia concentration (100 mg/liter) and exerted no significant effect on nitrification at concentrations of up to 200 mg/liter. Maximum effect of fluoride was reached at 800 mg of F⁻ per liter, and no greater inhibition was observed for up to 5,000 mg of F⁻ per liter. At the highest concentrations studied, ion pairing of ammonium and fluoride may exert a significant effect on kinetic coefficients. Kinetic analyses showed maximum specific substrate removal rates (q_max) of NH₄⁺-N to be about 2.3 mg of N per mg of volatile suspended solids per day in the absence of fluoride and 0.85 mg of N per mg of volatile suspended solids per day in the presence of fluoride. The form of inhibition due to the presence of fluoride was shown to be not competitive, conforming to a mixed inhibition model.     

Production of Candida utilis protein from peat extracts

Sphagnum peat extracts or hydrolysates have been obtained and used as a culture medium for the production of Candida utilis biomass as single cell proteins. Acid hydrolysis of ground peat (4–60 mesh) in an autoclave operated under a set of conditions for acid strength (0.3-1.5 (v/v) H₂SO₄), holding time (1–4 hr), temperature (100–165°C), and weight ratio of dry peat to solution (3.3–16.7 g dry peat/100 g solution) yielded carbohydrate-rich extracts of different concentrations (1–34g/liter). The best yield (mg total carbohydrate/g dry peat) was obtained for a holding time of I hr and a temperature of 152°C. Low peat concentratio (4.1 g dry peat/100 g solution)resulted in high yield(280mg total carbohydrate/gdry peat) with a corresponding low carbohydrate content in hydrolysate (13 g/liter), while a lower yield with a higher carbohydrate content (34 g/liter)in hydrolysate were found when increasing peat concentration (16.7 g dry peat/100 g solution). Shake-fladk experiments using peat hydrolysates as the culture medium together with NH₄OH (～4.8 g/liter) and K₂HPO₄(5 g/liter) as nitrogen and phosphate supplement, respectively, gave a maximum biomass concentration of 7.5 g/liter after 60 hr at 30°C and 200rpm. Batch cultivation in a fermentor under controlled conditions for aeration (4.2 liter/min), agitation (500rpm), temperature (30°C), and pH (5.0) produced a maximum biomass of 10 g/liter after 20 hr with a specific growth rate of 0.13 hr^?1. For the continuous cultivation, a maximal biomass productivity of 1.24 g/gliter-he was obtained at a dilution rate of 0.125 hr ^?1. Monod's equation's equation has been used for the estimation of the coefficients μ_Max, K_s, and Y. It was found that the yield coefficient Y is not constant during the progress of batch cultivation.     

Biokinetic parameters and behavior of <Emphasis Type="Italic">Aeromonas hydrophila</Emphasis> during anaerobic growth

The biokinetics of glucose metabolism were evaluated in Aeromonas hydrophila during growth in an anaerobic biosystem. After approx 34 h growth, A. hydrophila metabolized 5,000 mg glucose l⁻¹ into the end-products ethanol, acetate, succinate and formate. The maximum growth rate, μ _m, half saturation coefficients, K _s, microbial yield coefficient, Y, cell mass decay rate coefficient, k _d, and substrate inhibition coefficient, K _si were 0.25 ± 0.03 h⁻¹, 118 ± 31 mg glucose l⁻¹, 0.12 μg DNA mg glucose⁻¹, 0.01 h⁻¹, and 3,108 ± 1,152 mg glucose l⁻¹, respectively. These data were used to predict the performance of a continuous growth system with an influent glucose concentration of 5,000 mg l⁻¹. Results of the analysis suggest that A. hydrophila will metabolize glucose at greater than 95% efficiency when hydraulic retention times (HRTs) exceed 7 h, whereas the culture is at risk of washing out at an HRT of 6.7 h.     

Effect of high substrate concentration and cell feedback on kinetic behavior of heterogeneous populations in completely mixed systems

Growth kinetics of heterogeneous populations of sewage origin were studied in completely mixed reactors of the once-through type at a high concentration of incoming substrate, 3000 mg/l glucose, and in systems employing cell feedback or sludge recycle at an incoming substrate concentration of 1000 mg/1 glucose. The recycle flow rate employed was 25% of the incoming feed flow, and the concentration of cells in the recycle was maintained as closely as possible at 150% of the cell concentration in the reactor. Studies were made at various dilution rates. Throughout these studies, batch experiments using cells grown at the various dilution rates were made to determine k_s and μ_m values. As in previous studios using heterogeneous populations, the relationship between specific growth rates μ and substrate concentration S was represented better by the Monod equation than by any other which was tested. The growth “constants” μ_m, k_s, and Y were found to fall in the same general range as those determined in previous studies in once-through systems operated at 1000 mg/l glucose. It was observed that cell recycle, even at the relatively low concentration factor employed in these studies, greatly enhanced the flocculating and settling characteristics of the cells.     

Denitrification by Alcaligenes denitrificans in a closed rotating biological contactor

Biological denitrification using a pure culture of Alcaligenes denitrificans was investigated in a closed rotating biological contactor, which operated with a hydraulic retention time of 2 h, a carbon/nitrogen ratio of 2:1, with a dissolved O₂ concentration below 6 mg l^–1 and under three different phosphate concentrations. Alcaligenes denitrificans was not repressed by O₂ limitation and the removal of nitrate was about 30% more efficient at the intermediate phosphate concentration (20 mg P l^–1).     

The application of constant recycle solids concentration in activated sludge process.

The applicability of the model derived by Ramanathan and Gaudy (Biotechnol. Bioeng., 11, 207, (1969)) for completely mixed activated sludge treatment holding the recycle solids concentration as a system constant was investigated using an actual industrial organic wastewater. Short-term experiments were conducted at various dilution rates (1/8, 1/6, 1/4, 1/2, 1/1.5 hr-1) for two recycle solids concentration values (5000 and 7000 mg/liter). The influent substrate concentration was maintained at 1000 mg/liter COD and the hydraulic recycle ratio- alpha, was kept at 0.3. It was found that for bottling plant (Pepsi Cola) wastewaters, a steady state with respect to reactor biological solids and effluent COD, at different dilution rates, could be attained, lending experimental evidence to the assumption that a steady state could be reached in developing the model and also affecting the applicability of the model in industrial organic wastewater. The reactor biological solids and effluent COD calculated from the model closely agreed with the observed values at dilution rates lower than 0.5 hr-1. Operation at dilution rates higher than 0.5 hr-1 will washout the biological solids from the reactor and the recycle substrate concentration will be apparent if the concentration of XR were not increased.     

Biomethanation: Anaerobic fermentation of CO2, H2 and CO to methane

Studies to examine the microbial fermentation of coal gasification products (CO₂, H₂ and CO) to methane have been done with a mixed culture of anaerobic bacteria selected from an anaerobic sewage digestor. The specific rate of methane production at 37°C reached 25 mmol/g cell hr. The stoichiometry for methane production was 4 mmol H₂/mol CO₂. Cell recycle was used to increase the cell concentration from 2.5 to 8.3 g/liter; the volumetric rate of methane production ran from 1.3 to 4 liter/liter hr. The biogasification was also examined at elevated pressure (450 psi) and temperature to facilitate interfacing with a coal gasifier. At 60°C, the specific rate of methane production reached 50 mmol/g cell hr. Carbon monoxide utilization by the mixed culture of anaerobes and by a Rhodopseudomonas species was examined. Both cultures are able to carry out the shift conversion of CO and water to CO₂ and hydrogen.     

Modelling of seasonal variation in nitrogen retention and in-lake concentration: A four-year mass balance study in 16 shallow Danish lakes

The mass balance for total nitrogen (N) was studied over a four-year period in 16 shallow mainly eutrophic 1st order Danish lakes. Water was sampled in the main inlet of each lake 18–26 times annually, and from the outlets and the lake 19 times annually. Water was also sampled from minor inlets, although less frequently. N input and output were calculated using daily data on discharge (Q), the latter being obtained either from the Q/H relationship based on automatic recordings of water level (H) for the main in- and outlet, or by means of Q/Q relationships for the minor inlets. Annual mean N retention in the lakes ranged from 47 to 234 mg N m^–2 d^–1, and was particularly high in lakes with high N loading. Annual percentage retention (N _ret –y%) ranged from 11 to 72%. Non-linear regression analysis revealed that hydraulic retention time and mean depth accounted for 75% of the variation in annual mean N _ret –y% and, in combination with inlet N concentration, accounted for 84% of the variation in the in-lake N concentration. N _ret % varied according to season, being higher in the second and third quarter than in the first and fourth quarter (median 18–19%). A simple model was developed for predicting monthly nitrogen retention (N _ret –m) on the basis of external N loading, the lake water pool of nitrogen N _pool , hydraulic loading and lake water temperature. Calibration of only two parameters on data from the randomly selected 8 out of 16 lakes rendered the model capable of accurately simulating seasonal dynamics of the in-lake N concentration and N _ret –m in all 16 lakes. We conclude that with regard to shallow, eutrophic lakes with a relatively low hydraulic retention time, it is now possible to determine not only annual mean nitrogen retention, but also the seasonal variation in N _ret–m . Prediction of seasonal variation in N loading of downstream N-limited coastal areas is thereby rendered much more reliable.     

Steady-state model for activated sludge with constant recycle sludge concentration

In a previous report it was concluded that steady-state operation of completely mixed reactors for growth of heterogeneous microbial populations, i.e., activated sludge processes, was extremely difficult to attain if maintenance of a constant sludge recycle ratio, c, was required, and equations were devised in which the concentration of cells in the recycle, x_R, rather than the recycle ratio, was constant. In this report the equations are developed and computational analysis shows the effect on substrate and cell concentrations in the reactor of operational variables such as inflowing feed concentration, hydraulic recycle ratio, recycle sludge concentration, dilution rate, and the biological “constants” μ_m, k_s, and Y. The stabilizing effect of operating with constant x_R on the dilute-out pattern is shown.     

Practical design equations for trickling-filter process

The concept of solid retention time (SRT) was applied in the trickling-filter process. A rational model of the trickling-filter process employing activated-sludge-process operational parameters was presented. The design equation was developed as follows; 1/SRT = [(S₀ ? S_n)/X ]·(F/V)·Y ? k_d, where SRT is the sludge retention time, S₀ is the influent substrate concentration; S_n is the effluent substrate concentration; X is the total cell mass retained per unit filter volume; V is the total volume of the filter; F is the influent flow rate; Y is the cell yield, and k_d is the cell decay rate. A laboratory-scale trickling-filter pilot plant treating synthetic sucrose waste-water was studied to verify the present design equation. The solid retention time was evaluated from the total slime mass (active and inactive) retained and the sludge wasted daily. It was found that the present design equation could be applied for designing the trickling-filter process by the application of SRT employed in the activated sludge process. Also, the SRT could be related to the hydraulic loading and influent substrate concentration for a given filter medium. The variation of SRT by the hydraulic loading at constant organic loading was observed and could be expressed by the mechanistic model. When SRT was maintained more than 12 days, it provided the highest five-day biological oxygen demand (BOD₅) removal, minimum sludge production, and lowest sludge volume index (SVI) value. The present model does include both microbial growth kinetic concepts, which can be more practical and meaningful for the design of a trickling filter.     

Membrane bioreactor for drinking water denitrification

The aim of this study is to evaluate the performance of a membrane bioreactor with cell recycle to be used for drinking water denitrification, when operated with a high nitrate load (up to 7.68?kgNO₃ ^?/m³?day) and low hydraulic retention time (down to 0.625?h). Nitrate and nitrite were always completely removed for all the operational conditions used. The effluent's nitrite concentration kept below 0.1?mg NO₂ ^?/l with exception of a short period, during the reactor start-up, when it accumulates. The performance of the membrane bioreactor was also evaluated using a groundwater containing 148?mg NO₃ ^?/l. Nitrate and nitrite concentration in the effluent were below the recommended values for drinking water when the reactor was controlled at pH 7.0. The membrane flux decreases during operation as a consequence of membrane fouling. The flux decrease was more severe during operation with synthetic medium than with contaminated groundwater due to the existence of molecular complexes in the synthetic broth. A backshock technique was used to reduce the surface fouling of the membrane. Combining this technique with the use of a reserve asymmetric structured membrane it was found that the membrane flux remains nearly unchanged.     

Pentachlorophenol biodegradation kinetics of an oligotrophic fluidized-bed enrichment culture

A fluidized-bed reactor (FBR) was used to enrich an aerobic chlorophenol-degrading microbial culture. Long-term continuous-flow operation with low effluent concentrations selected oligotrophic microorganisms producing good-quality effluent for pentachlorophenol(PCP)-contaminated water. PCP biodegradation kinetics was studied using this FBR enrichment culture. The results from FBR batch experiments were modeled using a modified Haldane equation, which resulted in the following kinetic constants: q _max = 0.41 mg PCP mg protein⁻¹ day⁻¹, K _S = 16 μg l⁻¹, K _i = 5.3 mg l⁻¹, and n = 3.5. These results show that the culture has a high affinity for PCP but is also inhibited by relatively low PCP concentrations (above 1.1 mg PCP l⁻¹). This enrichment culture was maintained over 1 year of continuous-flow operation with PCP as the sole source of carbon and energy. During continuous-flow operation, effluent concentrations below 2 μg l⁻¹ were achieved at 268 min hydraulic retention time (t _HR) and 2.5 mg PCP l⁻¹ feed concentration. An increase in loading rate by decreasing t _HR did not significantly deteriorate the effluent quality until a t _HR decrease from 30 min to 21 min resulted in process failure. Recovery from process failure was slow. Decreasing the feed PCP concentration and increasing t _HR resulted in an improved process recovery. Received: 10 October 1996 / Received revision: 21 January 1997 / Accepted: 24 January 1997     

Sublethal Cytotoxic Effects of Mercuric Chloride on the Ciliate Tetrahymena pyriformis*

SYNOPSIS. The behavior and ultrastructure of Tetrahymena pyriformis was assessed after exposure to dosages of 8 and 16% of the lethal concentration of HgCl² (TLm 96 hr). The lower dosage caused no abnormal changes in cell motility, activity of the water explusion vesicles, or cell shape; the higher dosage caused deleterious changes in these parameters. The higher sublethal HgCl² concentration (0.50 mg/liter) elicited damage of several cell structures. This damage persisted and accumulated with time up to 24 hr. At the lower HgCl² dosage (0.25 mg liter) there were extensive changes after 1-hr exposure involving primarily mitochondria; however, all major changes were repaired after 24 hr of constant exposure to the HgCl², indicating adaptation to the toxicant. Based solely on cytotoxic evidence an attempt is made to apply the findings defining what constitutes a “safe'’concentration of HgCl₂ in the cell's environment.     

Thermodynamic Analysis of Energy Transfer in Acidogenic Cultures

A global thermodynamic analysis, normally used for pure cultures, has been performed for steady‐state data sets from acidogenic mixed cultures. This analysis is a combination of two different thermodynamic approaches, based on tabulated standard Gibbs energy of formation, global stoichiometry and medium compositions. It takes into account the energy transfer efficiency, ?, together with the Gibbs free energy dissipation, ΔG_o, analysis of the different data. The objective is to describe these systems thermodynamically without any heat measurement. The results show that ? is influenced by environmental conditions, where increasing hydraulic retention time increases its value all cases. The pH effect on ? is related to metabolic shifts and osmoregulation. Within the environmental conditions analyzed, ? ranges from 0.23 for a hydraulic retention time of 20 h and pH 4, to 0.42 for a hydraulic retention time of 8 h and a pH ranging from 7–8.5. The estimated values of ΔG_o are comparable to standard Gibbs energy of dissipation reported in the literature. For the data sets analyzed, ΔG_o ranges from –1210 kJ/mol_x, corresponding to a stirring velocity of 300 rpm, pH 6 and a hydraulic retention time of 6 h, to –20744 kJ/mol_x for pH 4 and a hydraulic retention time of 20 h. For average conclusions, the combined approach based on standard Gibbs energy of formation and global stoichiometry, used in this thermodynamic analysis, allows for the estimation of Gibbs energy dissipation values from the extracellular medium compositions in acidogenic mixed cultures. Such estimated values are comparable to the standard Gibbs energy dissipation values reported in the literature. It is demonstrated that ? is affected by the environmental conditions, i.e., stirring velocity, hydraulic retention time and pH. However, a relationship that relates this parameter to environmental conditions was not found and will be the focus of further research.     

Treatment of cyanide-containing wastewater from the food industry in a laboratory-scale fixed-bed methanogenic reactor

During the process of producing cassava starch from Manihot esculenta roots, large amounts of cyanoglycosides were released, which rapidly decayed to CN⁻ following enzymatic hydrolysis. Depending on the varying cyanoglycoside content of the cassava varieties, the cyanide concentration in the wastewater was as high as 200 mg/l. To simulate anaerobic stabilization, a wastewater with a chemical oxygen demand (COD) of about 20 g/l was prepared from cassava roots and was fermented in a fixed-bed methanogenic reactor. The start-up phase for a 99% degradation of low concentrations of cyanide (10 mg/l) required about 6 months. After establishment of the biofilm, a cyanide concentration of up to 150 mg CN⁻/l in the fresh wastewater was degraded during anaerobic treatment at a hydraulic retention time of 3 days. All nitrogen from the degraded cyanide was converted to organic nitrogen by the biomass of the effluent. The cyanide-degrading biocoenosis of the anaerobic reactor could tolerate shock concentrations of cyanide up to 240 mg CN⁻/l for a short time. Up to 5 mmol/l NH₄Cl (i.e. 70 mg N/l = 265 mg NH₄Cl/l) in the fresh wastewater did not affect cyanide degradation. The bleaching agent sulphite, however, had a negative effect on COD and cyanide removal. For anaerobic treatment, the maximum COD space loading was 12 g l⁻¹ day⁻¹, equivalent to a hydraulic retention time of 1.8 days. The COD removal efficiency was around 90%. The maximum permanent cyanide space loading was 50 mg CN⁻ l⁻¹ day⁻¹, with tolerable shock loadings up to 75 mg CN⁻ l⁻¹day⁻¹. Under steady-state conditions, the cyanide concentration of the effluent was lower than 0.5 mg/l. Received: 15 August 1997 / Received revision: 10 October 1997 / Accepted: 14 October 1997