- Email: [email protected]
Optimisation of sulphate reduction in a methanol-fed thermophilic bioreactor
Water Research 36 (2002) 1825–1833
Optimisation of sulphate reduction in a methanol-fed thermophilic bioreactor Jan Weijmaa,*, Eelco A.A. Botsa, Gabrielle Tandlingera, Alfons J.M. Stamsb, Look W. Hulshoff Pola, Gatze Lettingaa a
b
Sub-department Environmental Technology, Bomenweg 2, Wageningen University, 6700 EV Wageningen, The Netherlands Laboratory of Microbiology, Wageningen University, Hesselink van Suchtelenweg 4, 6703 CT, Wageningen, The Netherlands Received 11 April 2000; received in revised form 28 June 2001; accepted 27 August 2001
Abstract Several methods were tested to optimise sulphate reduction and minimise methane formation in thermophilic (651) expanded granular sludge bed reactors fed with a medium containing sulphate and methanol. Lowering the pH from 7.5 to 6.75 resulted in a rapid decrease of methane formation and a concomitant increase in sulphate reduction. The inhibition of methane formation was irreversible on the short-term. Lowering the COD/SO24 ratio (COD: chemical oxygen demand) from 6 to 0.34 (g/g) rapidly favoured sulphate reduction over methanogenesis. Continuous addition of 2 g L 1 2-bromoethanesulphonate was ineffective as complete inhibition of methanogenesis was obtained only for two days. Inhibition of methanogens by sulphide at pH 7.5 was only effective when the total sulphide concentration was above 1200 mg S L 1. For practical applications, a relatively short exposure to a slightly acidic pH in combination with operating the reactor at a volumetric methanol-COD loading rate close to the maximum volumetric sulphide-COD formation rate. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Competition; Methanogenesis; Methanol; Thermophilic; EGSB; Sulphate reduction
1. Introduction The principle of biotechnological desulphurisation of flue-gases is that sulfur oxyanions like sulphate and sulfite are reduced under anaerobic conditions to hydrogen sulphide, which is subsequently partially oxidised under aerobic conditions to elemental sulfur by colourless sulfur bacteria [1]. Under thermophilic (651C) conditions, methanol is an attractive electron donor and carbon source for the anaerobic step of this process [2]. However, a strong competition between sulphate reducing bacteria (SRB) and methanogenic archaea (MA) may occur for a considerable period of time (up to 5 months) when unadapted sludge is used as inoculum for the anaerobic reactor. Formation of *Corresponding author. E-mail address: [email protected] (J. Weijma).
methane is undesirable as it adversely affects the process economics due to the inefficient use of methanol. Here we describe methods to minimise methane formation in thermophilic anaerobic reactors fed with methanol and sulphate, so that SRB do not have to compete with MA for methanol or products from methanol catabolism like H2 and acetate. One method to minimise methanogenesis is the addition of agents that selectively inhibit growth of methanogens. A widely used inhibiting agent of methanogens is 2-bromoethanesulphonate (BES), which functions as a competitive inhibitor of the methyl-CoM reductase complex (CoM), which is a key enzyme of methanogens [3]. Chloroform is another inhibitor of methanogens, but this compound is less selective as it may also inhibit SRB [4]. Competition between SRB and MA can also be steered by making use of a differential response of their growth rates to changes in environmental condition [5].
0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 3 - 1 3 5 4 ( 0 1 ) 0 0 3 9 0 - 6
1826
J. Weijma et al. / Water Research 36 (2002) 1825–1833
The tolerance of SRB and MA towards the toxic unionised hydrogen sulphide can, for example, be quite different [6]. Free hydrogen sulphide is considered the most toxic form of sulphide [7] presumably because it can easily diffuse through the lipid cell membrane into the bacterial cytoplasm, where it can react with cell components. SRB and MA may also have a different pH-optimum for growth on common substrates. As the speciation of the weak acid hydrogen sulphide into H2S, HS , and S2 is affected by the pH, the effect of a change in pH on the growth of SRB and MA may be the result of a change in the concentration of free hydrogen sulphide. We previously found that at a temperature of 651C and pH 7.5, SRB gradually outcompete MA for methanol in continuously operated anaerobic EGSB reactors in which the total sulphide was produced up to a concentration of 1100 mg S L 1 (total sulphide: the sum of H2S, HS and S2 ), corresponding to a free hydrogen sulphide concentration of about 120 mg S L 1 [2]. It was not clarified if inhibition by hydrogen sulphide played a role in the competition between MA and SRB. Expanded granular sludge bed (EGSB) reactors were used in this study for continuous experiments. EGSB reactors provide good mixing between anaerobic sludge and medium, as a result of the imposed high (>2 m h 1) upflow liquid velocity [8]. EGSB reactors are particularly suitable for processes in which biogas induced mixing is low as in sulphidogenic processes [9].
The objective of the research described in this paper is to find the proper conditions to selectively inhibit or minimise methane formation in thermophilic anaerobic reactors fed with methanol and sulphate, thereby making reducing equivalents available for sulphate reduction.
2. Materials and methods 2.1. Reactor EGSB reactors with a volume of 4 L were used for continuous experiments. The experimental set-up is shown in Fig. 1. A detailed description of the reactor and mineral medium composition of the influent were presented previously [2]. The upward liquid flow was 3 m h 1 in all experiments. The EGSB reactors were kept at a temperature of 651C. Automatic pH control was applied, adding 0.1 N NaOH or 0.1 N HCl when necessary. The pH was controlled at a value of 7.5, unless stated otherwise. A stock solution containing 5 M methanol (Labscan Ltd., Dublin, Ireland) was pumped into the influent with a Gilson Minipuls3 Peristaltic pump at a flow rate necessary to obtain the desired methanol load. Yeast extract (Life Technologies, Paisly, Scotland) was
Fig. 1. EGSB reactor.
J. Weijma et al. / Water Research 36 (2002) 1825–1833
dissolved in the methanol stock solution to give a concentration of 20 mg L 1 in the influent. In EGSB-I experiments were carried out to assess the effect of BES and pH variations on methane formation. EGSB-I was inoculated with 2.5 L (150 g VSS) elutriated sludge from a pilot plant for biological sulphate reduction of scrubber liquid from a flue gas scrubber of a coal-fired power plant (Amercentrale, Geertruidenberg, The Netherlands). This sludge is further referred to as Amer sludge. Amer sludge was kindly provided by Paques Natural Solutions (Balk, The Netherlands), and was developed at 551C on a medium containing ethanol (accounting for about 95% of the chemical oxygen demand or COD), methanol (accounting for about 5% of the COD) and sulphate/sulphite. EGSB-I was started at an hydraulic retention time (HRT) of 13 h, an organic loading rate (OLR) of 2.2 g COD L 1 day 1 and a sulphate loading rate (SLR) of 3.4 g SO24 L 1 day 1. BES (sodium salt of 2-bromoethanesulfonic acid, Acros Organics, New Jersey, USA) was added to the influent from day 9 to 18 at a concentration of 2 g L 1. In a control experiment, a separate EGSB-reactor was inoculated with the same seed material and started at the same operating conditions as EGSB-I, but without BES-addition. The pH in reactor EGSB-I was lowered from 7.5 to 7.15 on day 45, and further reduced to 6.75 on day 55. The pH was reset at 7.5 on day 83. EGSB-II was used for assessment of the effect of the sulphide concentration and the COD/SO24 ratio on methane and sulphide formation. EGSB-II was inoculated with 500 mL (30 g VSS) Amer sludge and 500 mL (25 g VSS) Amer sludge that had been pre-adapted to methanol and sulphate (COD/SO24 ratio of 0.67 g/g) for 3 months at pH 7.5 and a temperature of 651C. EGSB-II was operated at a HRT of 4 h. During the first week, the sulphate concentration in the influent amounted to 0.50 g L 1, resulting in an SLR of 2.6 g SO24 L 1 day 1 while the OLR was kept at a value of 8 g COD L 1 day 1. On day 8 the OLR was increased to 15 g COD L 1 day 1 in order to prevent substrate limitation. From day 21 to day 54, Na2S (Merck, Darmstadt, Germany) was added to the influent from a concentrated stock solution. By adjusting the flow of the sulphide stock solution the sulphide concentration in the reactor could be controlled. The sodium concentration in the reactor was kept at 125710 mM during Na2Saddition by adjusting the NaCl concentration of the influent. At day 62 the SLR was increased from 2.6 to 19.8 g SO24 L 1 day 1, while the OLR was kept at 15.4 g COD L 1 day 1. At day 68 the OLR was decreased to 9.2 g COD L 1 day 1, and further decreased to 6.6 g COD L 1 day 1 at day 72. At day 78, the OLR was reset at 15.4 g COD L 1 day 1. The inoculation material that was used in batch assays for determination of sulphide toxicity (see below) was obtained from reactor EGSB-III which had been
1827
inoculated with the same sludges as used for EGSB-II: 220 mL (13 g VSS) elutriated Amer sludge and 200 mL (10 g VSS) Amer sludge that already had been adapted to methanol and sulphate (COD/SO24 ratio of 0.67 g/g) for 3 months in an EGSB reactor. Operating conditions for this reactor were the same as for EGSB-II except that an OLR of 8.5 g COD L 1 day 1 and a SLR of 13 g SO24 L 1 day 1 were applied. Sludge was taken from this reactor at day 7 for assessment of the effect of sulphide on the specific methanogenic activity. The volumetric sulphidogenic, methanogenic and acetogenic COD-conversion rates in the EGSB on day 7 were 4.3, 1.2 and 0.2 g COD L 1 day 1, respectively. 2.2. Activity assay Batch activity assays were carried out in 117-mL vials for determination of the specific methanogenic activity at various sulphide concentrations. The vials contained 50mL medium and a N2/CO2 gas phase of 67 mL. The mineral medium used for activity assays had a similar composition as the reactor influent and was described elsewhere [2]. However, Na2SO4 was omitted from the medium. Methanol and yeast extract were added from concentrated stock solutions, to give initial concentrations of 1.0 g COD L 1 and 20 mg L 1, respectively. Prior to inoculation, sodium sulphide was added from a concentrated stock solution that was neutralised with HCl. Total sulphide concentrations ranged from 30 to 2500 mg S L 1. Depending on the applied Na2S concentration, the NaClconcentration in the medium was adjusted in order to keep the Na+ concentration at the same value in all bottles. The pH of the medium was adjusted, if necessary, to a value of 7.4–7.6 by addition of a few drops of HCl or NaOH (0.1 N). Calculation of the specific methanogenic activity was conducted as previously [2]. 2.3. Analyses VSS was done according to Dutch standard methods. Methanol and acetate were analyzed with gas chromatography using a ionisation detector. Sulphate and bromide were analysed by HPLC. Sulphide was analysed with a colorimetric method. A detailed description of the analytical procedures can be found elsewhere [2]. The H2S concentration was calculated from the measured total sulphide concentration and pH using a pKa -value for H2S of 6.6 at 651C [10].
3. Results 3.1. Effect of BES addition The effect of addition of the specific methanogenic inhibitor BES on methanogenesis from methanol was
1828
J. Weijma et al. / Water Research 36 (2002) 1825–1833
studied in EGSB-I during day 0–23. Results are shown in Fig. 2a. One week after start-up the methane and sulphide formation amounted to 1.7 and 0.4 g COD L 1 day 1, respectively. In the control reactor, similar levels were reached a few days later, which was probably due to instable pH-control from day 8 to day 10 (Fig. 2b). After day 13 the methane and sulphide formation in the control reactor remained stable. Methanogenesis completely ceased within two days after addition of 2 g L 1 BES to the influent of EGSB-I from day 9 onwards. Despite the continuous addition of BES, already at day 11 some methane was formed again and subsequently the methane formation recovered to a maximum value of 0.9 g COD L 1 day 1 at day 15. Detection of bromide in the effluent from day 9 to 18 indicated breakdown of BES. The sulphide and acetate formation strongly increased by BES addition to maximum values of 2.6 and 1.3 g COD L 1 day 1, respectively, on day 13. The organic and sulphate loading rates were temporarily increased to 4.8 g COD L 1 day 1 (day 12–14) and 6.8 gSO24 L 1 day 1 (day 12–18), respectively, on day 12 to prevent methanol and sulphate limitation. The acetate formation decreased from day 13 onwards and on day 18 acetate
(a)
2.5
(1)
-1
(gCOD.L .day )
(2)
2.0
-1
COD-conversion rate
3.0
1.5 1.0 0.5 0.0 0
10 15 Time (days)
20
25
5
10 15 Time (days)
20
25
(b)
-1
(gCOD.L .day )
1.5
-1
COD-conversion rate
2.0
5
1.0 0.5 0.0 0
Fig. 2. Volumetric sulphidogenic (K), methanogenic (J), and acetogenic (m) COD-conversion rates in EGSB-I from day 0 to day 23 (a) and in the control reactor (b). Both reactors were started at an HRT of 13 h, an OLR of 2.2 g COD L 1 day 1 and a SLR of 3.4 g SO24 L 1 day 1. From day 12 to day 14 the OLR was increased to 4.8 g COD L 1 day 1 and from day 12 to day 18 the SLR was increased to 6.8 g SO24 L 1 day 1. Arrows; (1) start BES (2 g L 1) addition; (2) termination BES addition.
could no longer be detected in the effluent, while the sulphide formation had decreased to 1.4 g COD L 1 day 1. Termination of BES addition on day 18 had no immediate effect on methane, acetate and sulphide formation. In a second experiment with the same seed sludge and using an identical experimental system, a pulse of chloroform was injected in the reactor to give a concentration of 10 mM. Surprisingly, chloroform selectively inhibited sulphate reduction in this experiment and therefore it seemed to be an ineffective agent for inhibition of methanogenesis (data not shown). 3.2. Effect of pH The effect of pH variations in the range 6.75–7.50 on methanol conversion by SRB and MA was studied in EGSB-I during day 45–88. From day 45–54 the pH was lowered from 7.5 to 7.15. In this period methanol was limiting as indicated by the low effluent methanol concentration of less than 48 mg COD L 1. The results in Fig. 3 reveal that the methane formation decreased 25% within one day after the pH drop to 7.15. From the measured total sulphide concentration and pH it can be calculated that concomitantly the free H2S-concentration increased from 24 to 58 mg S L 1. The subsequent adjustment of the pH to 6.75 at day 55 resulted within 2 days in an additional 38% decrease of the methane formation. The free H2S-concentration increased from 63 to 104 mg S L 1. The overall inhibition of the methane formation from day 45–57 amounted to 76%. In case the inhibition merely would be caused by the increased free H2S concentration, a 50% inhibition value of 79 mg S L 1 can be estimated from these data. Following the quick drop in the methane formation from day 55–57 it further decreased slowly from 0.3 g COD L 1 day 1 at day 57 to 0.05 g COD L 1 day 1 at day 82. The methane formation did not recover when the pH was reset to 7.5 from day 83–87. During the period of lowered pH (day 45–81) the sulphide formation gradually increased from 0.9 to 2.0 g COD L 1 day 1. Following the increase of the pH at day 82 the sulphide formation further increased to 2.4 g COD L 1 day 1 at the end of the experiment on day 87, corresponding to a sulphate elimination efficiency of 90%. 3.3. Effect of sulphide Batch experiments were conducted to elucidate the effect of sulphide on the specific methanogenic activity of the seed sludge of EGSB-II. The sludge was reactivated on methanol and sulphate for 7 days in EGSB-III. The specific methanogenic activity on methanol in the absence of sulphate decreased linearly with the total sulphide concentration in the range of 200–1600 mg
1829
J. Weijma et al. / Water Research 36 (2002) 1825–1833
3.0 (1)
(2)
(3)
(5)
(4)
COD-conversion rate (gCOD.L-1.day-1)
2.5 2.0 1.5 1.0 0.5 0.0 0
10
20
30
40
50
60
70
809
0
Time (days) Fig. 3. Volumetric sulphidogenic (K), methanogenic (J), and acetogenic (m) COD-conversion rates in EGSB-I. Arrows: (1) start BES (2 g L 1) addition; (2) termination BES addition; (3) pH reduced from 7.5 to 7.15; (4) pH reduced from 7.15 to 6.75; (5) pH increased from 6.75 to 7.5.
S L 1 at pH 7.5 (Fig. 4), with 50% inhibition occurring at a total sulphide concentration of 980 mg S L 1. Complementary to the batch experiments the effect of sulphide on methanogenesis from methanol was investigated in EGSB-II in the period day 0–61. This reactor was fed with an influent containing methanol and only 0.5 g L 1 sulphate. Sulphate was almost completely reduced in EGSB-II from day 3 onwards as revealed by the low effluent sulphate concentration of 0.05 g L 1. Other results are shown in Fig. 5. It was calculated that sulphate reduction accounted for consumption of a minor part (0.34 gMeOH–COD L 1) of the methanol load. For the remainder of the methanol, no substrate competition occurred between methanogens and sulphate reducers. This enabled us to assess the effect of external sulphide addition on the methane formation independently from substrate competition effects. Sulphide addition was started at day 21 when the methane formation had been relatively stable for a week at 5 g COD L 1 day 1. Inhibition of the methane formation was not observed in the period from day 22–25 at a total sulphide concentration of 600–700 mg S L 1. On the contrary, the methane formation increased slightly from 5.2 to 6.1 g COD L 1 day 1. Only at a total sulphide concentration of about 1250 mg S L 1 the methane formation steadily decreased from 5.6 to 4.9 g COD L 1 day 1 between day 27 and 29. The methane formation further decreased at approximately the same rate to a final value of 4.0 g COD L 1 day 1 at the imposed total sulphide concentration of 1550–1800 mg S L 1 between day 31 and 34. The total sulphide concentration was decreased to 1250 mg S L 1 on day 35 and to 800–1000 mg S L 1 between day 36 and 52. Under these conditions, the methane formation initially (day 36–41) remained relatively stable but from day 41
Fig. 4. Specific methanogenic activity of sludge as function of total sulphide. Sludge was sampled from EGSB-III on day 7. The line represents linear regression.
onwards it steadily recovered until a value of 6.3 g COD L 1 day 1 at day 52. At day 54 the external addition of sulphide was terminated, leading to a drop of the total sulphide concentration to 160 mg L 1 resulting from sulphate reduction alone. At this sulphide level the methane formation increased to 8.4 g COD L 1 day 1 at day 61. 3.4. Effect of COD/SO24 ratio From day 62 to 88 the effect of the COD/SO24 ratio on methanol utilisation was studied in EGSB-II. Results are shown in Table 1 and Fig. 6. At day 62 the sulphate concentration in the influent of EGSB-II was increased from 0.5 to 3.8 g L 1 and accordingly the COD/SO24 ratio dropped from 5.9 to 0.78 (g/g). As shown in Table 1, the effluent sulphate concentration amounted to
1830
4000
8
3500
7
3000
6
2500
5 2000
4 3
1500
2
1000
1
500
-1
4500
9
(mgS.L )
10
Tota l Sulfide
Volumetric methanogenic_ COD_ conversion rate (gCOD.L 1.day 1)
J. Weijma et al. / Water Research 36 (2002) 1825–1833
0
0 0
10
20
30
40
50
60
70
Time (days)
Fig. 5. Experimental results obtained in EGSB-II, operated at an OLR of 15.4 g COD L 1 day 1 and a SLR of 2.6 g SO24 L 1 day 1. Volumetric methanogenic COD-conversion rate (J) and total sulphide concentration (F). Arrows indicate start and termination of addition of sodium sulphide from a concentrated stock solution between day 21 and day 54.
2.2 g L 1 from day 62 to 67 (Period II), showing that sulphate was not limiting the growth of SRB. The shift of the COD/SO24 ratio instantaneously affected the electron flow: the average sulphide formation increased from 2.2 to 6.4 g COD L 1 day 1 from Period I (day 59– 61) to Period II, whereas the average methane formation decreased from 8.4 to 5.4 g COD L 1 day 1. Competition for reducing equivalents between sulphate reducing bacteria and methanogens was further stimulated by reducing the COD/SO24 ratio to 0.46 at day 68 by decreasing the OLR. This had little effect on the sulphide formation, but the methane formation immediately further dropped to 3.5 g COD L 1 day 1 in Period III (day 68–71). Only when the OLR was decreased to 6.6 g COD L 1 day 1 (COD/SO24 ratio 0.30) in Period IV (day 72–77) the sulphide formation became affected as it decreased from 7.1 to 5.0 g COD L 1 day 1 from Period III–IV. The methane formation further decreased to 1.6 g COD L 1 day 1 in these periods. Methanol was limiting in Period IV, as indicated by the low methanol concentration in the reactor of 0.03 g COD L 1. In order to assess whether methane formation could be recovered, the OLR was increased to 15.4 g COD L 1 day 1 at day 78, restoring the methanol overloading conditions from this day onwards (Period V). The methane formation recovered only to 3.1 g COD L 1 day 1, which is a 40% reduction compared to Period II. By contrast, the sulphide formation fully recovered. Thus, by imposing methanol-limiting conditions for 10 days to the sludge, a partial irreversible and selective decrease of the methane formation was achieved. The net acetate formation remained relatively low compared to the sulphide and methane formation throughout the operation of EGSB-II as it amounted
to a maximum of only 0.7 COD L 1 day 1, and the acetate formation became even less (0.2 g COD L 1 day 1) in the period from day 72 to 77, when methanol-limiting conditions prevailed (Fig. 6). Surprisingly, acetate formation did not recover in Period V at methanol overloading conditions.
4. Discussion The results show that a relatively short exposure to a slightly acidic pH and keeping the methanol sludge loading rate close to the maximum specific methanol degradation rate of the sulphidogenic biomass is effective for a rapid development of sulphidogenic biomass with low methanogenic activity. It may be speculated that methanogenesis recovers on the long term due to adaptation of the methanogenic biomass when using any method for its suppression. However, as we described in a previous paper [2] SRB almost completely outcompeted MA for methanol in long-term (up to 150 days) experiments at pH 7.5 using identical reactors and seed sludge, with highly similar organic and sulphate loading rates. Consequently, it is likely that methane formation will remain low once sulphate reduction predominates under methanol-limiting conditions. The experiment carried out in EGSB-I revealed that BES at a concentration of 2 g L 1 does not durably inhibit methanogenesis. Although temporarily methane formation ceased completely following continuous BES addition, it already started to recover within 3 days. This possibly is due to degradation of BES, as indicated by the appearance of bromide in the effluent or to adaptation of the methanogens to BES. The temporary increase in both the sulphide and acetate formation after the BES addition maybe due to interactions between
1831
1
day 0.7 0.670.0 0.570.0 0.270.1 0.170.1
1
VAC g COD L 1
day 8.470.1 5.470.3 3.570.2 1.670.8 3.170.5
1
VMC g COD L 1
day 0.05 2.270.1 F F F
2.270.0 6.470.7 7.170.2 5.070.9 6.370.8
1
VSC g COD L Sulphateeff gL 1
5.9 0.78 0.46 0.34 0.93
0.37 0.4070.01 0.1370.01 0.0370.01 0.4770.16
1
MeOHeff g COD L COD/SO24 g COD/g SO24 day
1
2.670.0 19.870.0 19.870.0 19.670.0 16.470.6
1
SLR g SO24 L day
1
59–61 62–67 68–71 72–77 78–88 I II III IV V
15.470.1 15.470.0 9.270.1 6.670.0 15.470.2
1
OLR g COD L Day Period
Table 1 Performance of EGSB-II under sulfate limiting and methanol limiting conditions. MeOHeff, sulfateeff; methanol, sulfate concentration in effluent. VSC, VMC, VAC: volumetric sulfidogenic, methanogenic and (net) acetogenic COD-conversion rates, respectively. F; not determined
J. Weijma et al. / Water Research 36 (2002) 1825–1833
SRB, MA and AB on substrate level but may also partly be due to degradation of BES. The results obtained with EGSB-I also revealed that chloroform inhibits SRB instead of MA. Recently, also Scholten [4] found inhibition of SRB by chloroform in freshwater sediment. Decreasing the pH from 7.5 to 7.15 and subsequently to 6.75 apparently is effective for a rapid and selective inhibition of methanogenesis in favour of sulphate reduction to sulphide. As a change in pH around neutral values strongly affects the free H2S concentration, this inhibition may be caused by the increased free H2S concentration, viz. from 24 to 104 mg S L 1, rather than from the drop of the pH. On the other hand, results from EGSB-II showed that methane formation remained unaffected at H2S concentrations ranging from 90 to 110 mg S L 1. Based on this observation, the reduced methane formation following the pH drop more likely is the result of inhibition by the low pH rather than of hydrogen sulphide inhibition. Minami et al. [11] found that sulphate reduction becomes stimulated between pH 6.2 and 6.8, while in pH range 7.0–7.5 methanogenesis prevailed under thermophilic (531C) conditions in a packed-bed reactor fed with methanol and sulphate. These contrasting results show that even relatively small changes in the experimental system may have a strong impact on the anaerobic degradation of methanol. Returning to a pH of 7.5 did not result in a recovery of methane formation within a period of 5 days. Such a period is short compared to the long solid retention times typical for high-rate anaerobic reactors and it may be speculated that methanogenesis recovers on the long term. However, this is most unlikely under the methanol-limiting conditions as prevailing from day 20 onwards, because SRB are known to outcompete MA under the applied conditions [2]. In case methanol is not limiting, methanogenesis may recover on the long term. This aspect requires further investigation. The pH drop of 7.5–6.75 for accomplishing a 76% inhibition of methane formation requires only a low amount of acid. Therefore, temporarily imposing slightly acidic pHvalues might represent a practical and feasible method for selective and durable suppression of methanogenesis in a full-scale process. Under sulphate limiting conditions, a 35% decrease of the methane formation was observed in EGSB-II over a 10-day period at a total sulphide concentration ranging from 1200 to as high as 1700 mg S L 1 at pH 7.5. Subsequent lowering of the total sulphide concentration to values between 800 and 1000 mg S L 1 resulted in a gradual and complete recovery of the methane formation within 2 weeks. These results show that high (1200–1700 mg S L 1) total sulphide concentrations must be maintained for several weeks for a considerable decrease of methanogenic activity. In reactors seeded with sludge exerting a low specific sulphidogenic activity such high sulphide concentrations can only be achieved
1832
J. Weijma et al. / Water Research 36 (2002) 1825–1833 12 volumetric COD-conversion rate (gCOD.L-1.day-1)
1
2 3
4
10 8 6 4 2 0 0
20
40
60
80
100
Time (days)
Fig. 6. Volumetric sulphidogenic (K), methanogenic (J), and acetogenic (m) COD-conversions in EGSB-II. EGSB-II was operated at a HRT of 4 h and was started at an OLR of 15.4 g COD L 1 day 1 and a SLR of 2.6 g SO24 L 1 day 1. Arrows: (1) SLR increased to 19.8 g SO24 L 1 day 1 at day 62; (2) OLR decreased to 9.2 g COD L 1 day 1 at day 68; (3) OLR decreased to 6.6 g COD L 1 day 1 at day 72; (4) OLR increased to 15.4 g COD L 1 day 1 at day 78. The dotted line represents the calculated sulphide production that results from the reduction of sulphate to sulphide in the period of external sulphide addition.
by applying very long hydraulic retention times for prolonged periods, which is not very practical. It still remains unclear why methanogenesis increases in the continuous reactor EGSB-II at a total sulphide concentration ranging from 800 to 1000 mg S L 1, while in batch reactors inoculated with the seed sludge of EGSBII a 40–50% inhibition of the specific methanogenic activity was observed in the same concentration range. It can be speculated that already a more sulphide-tolerant methanogenic biomass had developed in EGSB-II compared to the methanogens present in the seed sludge that was used for batch assays. However, it looks improbable that a more sulphide-tolerant methanogenic biomass already developed before the sludge was exposed to high sulphide concentrations in EGSB-II. It is furthermore unlikely that the observed contrasting results from batch and continuous reactors exclusively can be attributed to differences in hydraulic mixing conditions in both systems or to small differences in the applied environmental conditions. In any case, results of batch assays did not have any predictive value for the effect of sulphide in the continuous system. Upon the transition from sulphate limiting (COD/ SO24 ratio of 5.9) to non-sulphate limiting conditions (COD/SO24 ratio of 0.34), an 80% decrease of methane formation was found in EGSB-II. It is relevant to note that the effluent methanol and acetate concentration remained at a value of 0.4–0.5 and 0.1 g COD L 1, respectively, following the first lowering of the COD/ SO24 ratio from 5.9 to 0.78. Therefore, the decreased methane formation following at the lower COD/SO24 ratio cannot be attributed to kinetic limitations of methylotrophic and acetotrophic MA. However, not
methanol or acetate but H2/CO2 may represent the main methanogenic precursor in thermophilic anaerobic sludge cultivated on methanol, as suggested by Weijma [12]. Therefore, the reduced rate of methanogenesis following the first decrease of the COD/SO24 ratio possibly is due to kinetic limitation of hydrogenotrophic methanogens. Subsequent stepwise drops of the COD/SO24 ratio to 0.46 and 0.34 resulted in low effluent methanol and acetate concentrations. Therefore, it cannot be excluded that the observed lower methane formation at these COD/SO24 ratios is due to kinetic limitation of methylotrophic or acetotrophic methanogens. The results furthermore show that methanogenesis can irreversibly but only partially be suppressed by temporarily (EGSB-II, Period IV) applying a low COD/SO24 ratio between two periods (Period III and V) of a higher COD/ SO24 ratio. The results in Table 1 and Fig. 6 show that even at the higher COD/SO24 ratio of 0.93 in Period V, methanogenesis is lower compared to Period III. Theoreticallly, a higher COD/SO24 ratio favours methanogenesis over sulphate reduction, which is confirmed by the results from Periods II and III. Therefore, the degree of inhibition methanogenesis may have been even higher, had the same COD/SO24 ratio been applied in Period V as in Period III but certainly not lower. Most probably, the decay rate of methanogens is increased due to substrate starvation at a lower COD/SO24 ratio. The sulphide formation only decreased when the OLR was lower than the maximum volumetric sulphidogenic CODconversion rate as assessed under overloading conditions. This shows that SRB are superior over MA in the competition for available reducing equivalents. Accordingly, the major part of the reducing equivalents derived
J. Weijma et al. / Water Research 36 (2002) 1825–1833
from methanol only becomes available for the MA when utilisation of these electron equivalents by SRB is limited by the amount of sulphate or by the absence of a sufficient sulphate reducing capacity of the reactor. Thus (temporary) lowering of the COD/SO24 ratio provides a tool to lower the amount of methanol consumed per mol sulphate reduced. However, the sulphate removal efficiency decreases at lower COD/SO24 ratio but temporarily this may well be acceptable, for instance during start-up or when methanogenesis recovers due to process upsets. It is worth noting that lowering the COD/SO24 ratio of the influent to values below 0.67 or lowering the reactor pH from slightly alkaline (7.5–8) to neutral or slightly acidic (6.5–7) values also suppressed methanogenesis in mesophilic high-rate reactors fed with sulphate and a mixture of volatile fatty acids (VFA’s, i.e. acetic, propionic and butyric acid) [13]. The stimulating effect of lowering the COD/SO24 ratio likely can be attributed to the generally more favourable growth kinetic properties of SRB over MA at low concentrations of common substrates like hydrogen and acetate [14]. The similar pHeffect seems coincidental because MA in general do not favour a higher pH compared to SRB. Although the effect of pH might be indirect as it affects the speciation of sulphide, also for sensitivity of SRB and MA towards inhibitory sulphide, no general trend can be recognised from literature data [15].
5. Conclusions *
*
*
Methane formation can be substantially reduced by imposing a slightly acidic pH to thermophilic reactors fed with sulphate and methanol, or by operating the reactor at an organic loading rate close to the maximum volumetric sulphidogenic CODconversion rate. 2-BES is ineffective as durable inhibitor of methanogenesis from methanol at 651C. Total sulphide at concentrations of up to 1000 mg S L 1 does not inhibit thermophilic methanogenesis from methanol. Therefore, such high total sulphide concentrations does not favour sulphate reduction.
Acknowledgements This research was supported by the Netherlands Technology Foundation STW, applied science division of NWO and the technology program of the Ministry of Economic affairs of the Netherlands. The authors thank C.J.N. Buisman and H. Dijkman from Paques Natural
1833
Solutions, Balk, and the Netherlands for valuable discussions.
References [1] Janssen AJH, Sleyster R, van der Kaa C, Jochemsen A, Bontsema J, Lettinga G. Biological sulphide oxidation in a fed-batch reactor. Biotechnol Bioeng 1995;47: 327–33. [2] Weijma J, Stams AJM, Hulshoff Pol LW, Lettinga G. Thermophilic sulphate reduction and methanogenesis with methanol in a high rate anaerobic reactor. Biotechnol Bioeng 2000;67:354–63. [3] Oremland RS, Capone DG. Use of ‘‘specific’’ inhibitors in biogeochemistry and microbial ecology. Adv Microbiol Ecol 1988;10:285–383. [4] Scholten JCM. The effect of sulphate and nitrate on the methane formation by methanogenic archaea in a freshwater sediments. Ph.D. thesis. The Netherlands: Wageningen Agricultural University, 1999. [5] Hao JH, Chen JM, Huang L, Buglass RL. Sulphatereducing bacteria. Crit Rev Environ Sci Technol 1996;26(1):155–87. [6] Maillacheruvu KM, Parkin GF. Kinetics of growth, substrate utilization and sulphide toxicity for propionate, acetate, and hydrogen utilizers in anaerobic systems. Water Environ Res 1996;68(2):1099–106. [7] Reis MAM, Lemos PC, Almeida JS, Carrondo MJT. Evidence for the intrinsic toxicity of H2S to sulphate reducing bacteria. Appl Microbiol Biotechnol 1992;36: 145–7. [8] Lettinga G. Anaerobic digestion and wastewater treatment systems. Antonie van Leeuwenhoek 1995;67:3–28. [9] Dries J, de Smul A, Goethals L, Grootaerd H, Verstraete W. High rate biological treatment of sulphate-rich wastewater in an acetate fed EGSB-reactor. Biodegradation 1998;9(2):103–11. [10] Sill!en LG, Martell AE. Stability constants of metal–ion complexes. Burlington House: The Chemical Society, 1964. [11] Minami K, Tanimoto Y, Tasaki M, Ogawa S, Okamura K. Influence of pH on methane and sulphide production from methanol. J Ferm Technol 1988;66(1):117–21. [12] Weijma, J. Methanol as electron donor for thermophilic sulphate and sulfite reduction. Ph.D thesis. The Netherlands: Wageningen University, 2000. [13] Omil F, Lens P, Hulshoff Pol LW, Lettinga G. Effect of upward velocity and sulphide concentration on volatile fatty acid degradation in a sulphidogenic granular sludge reactor. Process Biochem 1996;31:699–710. [14] Oude-Elferink SJWH, Visser A, Hulshoff Pol LW, Stams AJM. Sulphate reduction in methanogenic bioreactors. FEMS Microbiol Rev 1994;15:119–36. [15] Colleran E, Finnegan S, Lens P. Anaerobic treatment of sulphate-containing waste streams. Antonie van Leeuwenhoek 1995;67:29–46.
Optimisation of sulphate reduction in a methanol-fed thermophilic bioreactor Jan Weijmaa,*, Eelco A.A. Botsa, Gabrielle Tandlingera, Alfons J.M. Stamsb, Look W. Hulshoff Pola, Gatze Lettingaa a
b
Sub-department Environmental Technology, Bomenweg 2, Wageningen University, 6700 EV Wageningen, The Netherlands Laboratory of Microbiology, Wageningen University, Hesselink van Suchtelenweg 4, 6703 CT, Wageningen, The Netherlands Received 11 April 2000; received in revised form 28 June 2001; accepted 27 August 2001
Abstract Several methods were tested to optimise sulphate reduction and minimise methane formation in thermophilic (651) expanded granular sludge bed reactors fed with a medium containing sulphate and methanol. Lowering the pH from 7.5 to 6.75 resulted in a rapid decrease of methane formation and a concomitant increase in sulphate reduction. The inhibition of methane formation was irreversible on the short-term. Lowering the COD/SO24 ratio (COD: chemical oxygen demand) from 6 to 0.34 (g/g) rapidly favoured sulphate reduction over methanogenesis. Continuous addition of 2 g L 1 2-bromoethanesulphonate was ineffective as complete inhibition of methanogenesis was obtained only for two days. Inhibition of methanogens by sulphide at pH 7.5 was only effective when the total sulphide concentration was above 1200 mg S L 1. For practical applications, a relatively short exposure to a slightly acidic pH in combination with operating the reactor at a volumetric methanol-COD loading rate close to the maximum volumetric sulphide-COD formation rate. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Competition; Methanogenesis; Methanol; Thermophilic; EGSB; Sulphate reduction
1. Introduction The principle of biotechnological desulphurisation of flue-gases is that sulfur oxyanions like sulphate and sulfite are reduced under anaerobic conditions to hydrogen sulphide, which is subsequently partially oxidised under aerobic conditions to elemental sulfur by colourless sulfur bacteria [1]. Under thermophilic (651C) conditions, methanol is an attractive electron donor and carbon source for the anaerobic step of this process [2]. However, a strong competition between sulphate reducing bacteria (SRB) and methanogenic archaea (MA) may occur for a considerable period of time (up to 5 months) when unadapted sludge is used as inoculum for the anaerobic reactor. Formation of *Corresponding author. E-mail address: [email protected] (J. Weijma).
methane is undesirable as it adversely affects the process economics due to the inefficient use of methanol. Here we describe methods to minimise methane formation in thermophilic anaerobic reactors fed with methanol and sulphate, so that SRB do not have to compete with MA for methanol or products from methanol catabolism like H2 and acetate. One method to minimise methanogenesis is the addition of agents that selectively inhibit growth of methanogens. A widely used inhibiting agent of methanogens is 2-bromoethanesulphonate (BES), which functions as a competitive inhibitor of the methyl-CoM reductase complex (CoM), which is a key enzyme of methanogens [3]. Chloroform is another inhibitor of methanogens, but this compound is less selective as it may also inhibit SRB [4]. Competition between SRB and MA can also be steered by making use of a differential response of their growth rates to changes in environmental condition [5].
0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 3 - 1 3 5 4 ( 0 1 ) 0 0 3 9 0 - 6
1826
J. Weijma et al. / Water Research 36 (2002) 1825–1833
The tolerance of SRB and MA towards the toxic unionised hydrogen sulphide can, for example, be quite different [6]. Free hydrogen sulphide is considered the most toxic form of sulphide [7] presumably because it can easily diffuse through the lipid cell membrane into the bacterial cytoplasm, where it can react with cell components. SRB and MA may also have a different pH-optimum for growth on common substrates. As the speciation of the weak acid hydrogen sulphide into H2S, HS , and S2 is affected by the pH, the effect of a change in pH on the growth of SRB and MA may be the result of a change in the concentration of free hydrogen sulphide. We previously found that at a temperature of 651C and pH 7.5, SRB gradually outcompete MA for methanol in continuously operated anaerobic EGSB reactors in which the total sulphide was produced up to a concentration of 1100 mg S L 1 (total sulphide: the sum of H2S, HS and S2 ), corresponding to a free hydrogen sulphide concentration of about 120 mg S L 1 [2]. It was not clarified if inhibition by hydrogen sulphide played a role in the competition between MA and SRB. Expanded granular sludge bed (EGSB) reactors were used in this study for continuous experiments. EGSB reactors provide good mixing between anaerobic sludge and medium, as a result of the imposed high (>2 m h 1) upflow liquid velocity [8]. EGSB reactors are particularly suitable for processes in which biogas induced mixing is low as in sulphidogenic processes [9].
The objective of the research described in this paper is to find the proper conditions to selectively inhibit or minimise methane formation in thermophilic anaerobic reactors fed with methanol and sulphate, thereby making reducing equivalents available for sulphate reduction.
2. Materials and methods 2.1. Reactor EGSB reactors with a volume of 4 L were used for continuous experiments. The experimental set-up is shown in Fig. 1. A detailed description of the reactor and mineral medium composition of the influent were presented previously [2]. The upward liquid flow was 3 m h 1 in all experiments. The EGSB reactors were kept at a temperature of 651C. Automatic pH control was applied, adding 0.1 N NaOH or 0.1 N HCl when necessary. The pH was controlled at a value of 7.5, unless stated otherwise. A stock solution containing 5 M methanol (Labscan Ltd., Dublin, Ireland) was pumped into the influent with a Gilson Minipuls3 Peristaltic pump at a flow rate necessary to obtain the desired methanol load. Yeast extract (Life Technologies, Paisly, Scotland) was
Fig. 1. EGSB reactor.
J. Weijma et al. / Water Research 36 (2002) 1825–1833
dissolved in the methanol stock solution to give a concentration of 20 mg L 1 in the influent. In EGSB-I experiments were carried out to assess the effect of BES and pH variations on methane formation. EGSB-I was inoculated with 2.5 L (150 g VSS) elutriated sludge from a pilot plant for biological sulphate reduction of scrubber liquid from a flue gas scrubber of a coal-fired power plant (Amercentrale, Geertruidenberg, The Netherlands). This sludge is further referred to as Amer sludge. Amer sludge was kindly provided by Paques Natural Solutions (Balk, The Netherlands), and was developed at 551C on a medium containing ethanol (accounting for about 95% of the chemical oxygen demand or COD), methanol (accounting for about 5% of the COD) and sulphate/sulphite. EGSB-I was started at an hydraulic retention time (HRT) of 13 h, an organic loading rate (OLR) of 2.2 g COD L 1 day 1 and a sulphate loading rate (SLR) of 3.4 g SO24 L 1 day 1. BES (sodium salt of 2-bromoethanesulfonic acid, Acros Organics, New Jersey, USA) was added to the influent from day 9 to 18 at a concentration of 2 g L 1. In a control experiment, a separate EGSB-reactor was inoculated with the same seed material and started at the same operating conditions as EGSB-I, but without BES-addition. The pH in reactor EGSB-I was lowered from 7.5 to 7.15 on day 45, and further reduced to 6.75 on day 55. The pH was reset at 7.5 on day 83. EGSB-II was used for assessment of the effect of the sulphide concentration and the COD/SO24 ratio on methane and sulphide formation. EGSB-II was inoculated with 500 mL (30 g VSS) Amer sludge and 500 mL (25 g VSS) Amer sludge that had been pre-adapted to methanol and sulphate (COD/SO24 ratio of 0.67 g/g) for 3 months at pH 7.5 and a temperature of 651C. EGSB-II was operated at a HRT of 4 h. During the first week, the sulphate concentration in the influent amounted to 0.50 g L 1, resulting in an SLR of 2.6 g SO24 L 1 day 1 while the OLR was kept at a value of 8 g COD L 1 day 1. On day 8 the OLR was increased to 15 g COD L 1 day 1 in order to prevent substrate limitation. From day 21 to day 54, Na2S (Merck, Darmstadt, Germany) was added to the influent from a concentrated stock solution. By adjusting the flow of the sulphide stock solution the sulphide concentration in the reactor could be controlled. The sodium concentration in the reactor was kept at 125710 mM during Na2Saddition by adjusting the NaCl concentration of the influent. At day 62 the SLR was increased from 2.6 to 19.8 g SO24 L 1 day 1, while the OLR was kept at 15.4 g COD L 1 day 1. At day 68 the OLR was decreased to 9.2 g COD L 1 day 1, and further decreased to 6.6 g COD L 1 day 1 at day 72. At day 78, the OLR was reset at 15.4 g COD L 1 day 1. The inoculation material that was used in batch assays for determination of sulphide toxicity (see below) was obtained from reactor EGSB-III which had been
1827
inoculated with the same sludges as used for EGSB-II: 220 mL (13 g VSS) elutriated Amer sludge and 200 mL (10 g VSS) Amer sludge that already had been adapted to methanol and sulphate (COD/SO24 ratio of 0.67 g/g) for 3 months in an EGSB reactor. Operating conditions for this reactor were the same as for EGSB-II except that an OLR of 8.5 g COD L 1 day 1 and a SLR of 13 g SO24 L 1 day 1 were applied. Sludge was taken from this reactor at day 7 for assessment of the effect of sulphide on the specific methanogenic activity. The volumetric sulphidogenic, methanogenic and acetogenic COD-conversion rates in the EGSB on day 7 were 4.3, 1.2 and 0.2 g COD L 1 day 1, respectively. 2.2. Activity assay Batch activity assays were carried out in 117-mL vials for determination of the specific methanogenic activity at various sulphide concentrations. The vials contained 50mL medium and a N2/CO2 gas phase of 67 mL. The mineral medium used for activity assays had a similar composition as the reactor influent and was described elsewhere [2]. However, Na2SO4 was omitted from the medium. Methanol and yeast extract were added from concentrated stock solutions, to give initial concentrations of 1.0 g COD L 1 and 20 mg L 1, respectively. Prior to inoculation, sodium sulphide was added from a concentrated stock solution that was neutralised with HCl. Total sulphide concentrations ranged from 30 to 2500 mg S L 1. Depending on the applied Na2S concentration, the NaClconcentration in the medium was adjusted in order to keep the Na+ concentration at the same value in all bottles. The pH of the medium was adjusted, if necessary, to a value of 7.4–7.6 by addition of a few drops of HCl or NaOH (0.1 N). Calculation of the specific methanogenic activity was conducted as previously [2]. 2.3. Analyses VSS was done according to Dutch standard methods. Methanol and acetate were analyzed with gas chromatography using a ionisation detector. Sulphate and bromide were analysed by HPLC. Sulphide was analysed with a colorimetric method. A detailed description of the analytical procedures can be found elsewhere [2]. The H2S concentration was calculated from the measured total sulphide concentration and pH using a pKa -value for H2S of 6.6 at 651C [10].
3. Results 3.1. Effect of BES addition The effect of addition of the specific methanogenic inhibitor BES on methanogenesis from methanol was
1828
J. Weijma et al. / Water Research 36 (2002) 1825–1833
studied in EGSB-I during day 0–23. Results are shown in Fig. 2a. One week after start-up the methane and sulphide formation amounted to 1.7 and 0.4 g COD L 1 day 1, respectively. In the control reactor, similar levels were reached a few days later, which was probably due to instable pH-control from day 8 to day 10 (Fig. 2b). After day 13 the methane and sulphide formation in the control reactor remained stable. Methanogenesis completely ceased within two days after addition of 2 g L 1 BES to the influent of EGSB-I from day 9 onwards. Despite the continuous addition of BES, already at day 11 some methane was formed again and subsequently the methane formation recovered to a maximum value of 0.9 g COD L 1 day 1 at day 15. Detection of bromide in the effluent from day 9 to 18 indicated breakdown of BES. The sulphide and acetate formation strongly increased by BES addition to maximum values of 2.6 and 1.3 g COD L 1 day 1, respectively, on day 13. The organic and sulphate loading rates were temporarily increased to 4.8 g COD L 1 day 1 (day 12–14) and 6.8 gSO24 L 1 day 1 (day 12–18), respectively, on day 12 to prevent methanol and sulphate limitation. The acetate formation decreased from day 13 onwards and on day 18 acetate
(a)
2.5
(1)
-1
(gCOD.L .day )
(2)
2.0
-1
COD-conversion rate
3.0
1.5 1.0 0.5 0.0 0
10 15 Time (days)
20
25
5
10 15 Time (days)
20
25
(b)
-1
(gCOD.L .day )
1.5
-1
COD-conversion rate
2.0
5
1.0 0.5 0.0 0
Fig. 2. Volumetric sulphidogenic (K), methanogenic (J), and acetogenic (m) COD-conversion rates in EGSB-I from day 0 to day 23 (a) and in the control reactor (b). Both reactors were started at an HRT of 13 h, an OLR of 2.2 g COD L 1 day 1 and a SLR of 3.4 g SO24 L 1 day 1. From day 12 to day 14 the OLR was increased to 4.8 g COD L 1 day 1 and from day 12 to day 18 the SLR was increased to 6.8 g SO24 L 1 day 1. Arrows; (1) start BES (2 g L 1) addition; (2) termination BES addition.
could no longer be detected in the effluent, while the sulphide formation had decreased to 1.4 g COD L 1 day 1. Termination of BES addition on day 18 had no immediate effect on methane, acetate and sulphide formation. In a second experiment with the same seed sludge and using an identical experimental system, a pulse of chloroform was injected in the reactor to give a concentration of 10 mM. Surprisingly, chloroform selectively inhibited sulphate reduction in this experiment and therefore it seemed to be an ineffective agent for inhibition of methanogenesis (data not shown). 3.2. Effect of pH The effect of pH variations in the range 6.75–7.50 on methanol conversion by SRB and MA was studied in EGSB-I during day 45–88. From day 45–54 the pH was lowered from 7.5 to 7.15. In this period methanol was limiting as indicated by the low effluent methanol concentration of less than 48 mg COD L 1. The results in Fig. 3 reveal that the methane formation decreased 25% within one day after the pH drop to 7.15. From the measured total sulphide concentration and pH it can be calculated that concomitantly the free H2S-concentration increased from 24 to 58 mg S L 1. The subsequent adjustment of the pH to 6.75 at day 55 resulted within 2 days in an additional 38% decrease of the methane formation. The free H2S-concentration increased from 63 to 104 mg S L 1. The overall inhibition of the methane formation from day 45–57 amounted to 76%. In case the inhibition merely would be caused by the increased free H2S concentration, a 50% inhibition value of 79 mg S L 1 can be estimated from these data. Following the quick drop in the methane formation from day 55–57 it further decreased slowly from 0.3 g COD L 1 day 1 at day 57 to 0.05 g COD L 1 day 1 at day 82. The methane formation did not recover when the pH was reset to 7.5 from day 83–87. During the period of lowered pH (day 45–81) the sulphide formation gradually increased from 0.9 to 2.0 g COD L 1 day 1. Following the increase of the pH at day 82 the sulphide formation further increased to 2.4 g COD L 1 day 1 at the end of the experiment on day 87, corresponding to a sulphate elimination efficiency of 90%. 3.3. Effect of sulphide Batch experiments were conducted to elucidate the effect of sulphide on the specific methanogenic activity of the seed sludge of EGSB-II. The sludge was reactivated on methanol and sulphate for 7 days in EGSB-III. The specific methanogenic activity on methanol in the absence of sulphate decreased linearly with the total sulphide concentration in the range of 200–1600 mg
1829
J. Weijma et al. / Water Research 36 (2002) 1825–1833
3.0 (1)
(2)
(3)
(5)
(4)
COD-conversion rate (gCOD.L-1.day-1)
2.5 2.0 1.5 1.0 0.5 0.0 0
10
20
30
40
50
60
70
809
0
Time (days) Fig. 3. Volumetric sulphidogenic (K), methanogenic (J), and acetogenic (m) COD-conversion rates in EGSB-I. Arrows: (1) start BES (2 g L 1) addition; (2) termination BES addition; (3) pH reduced from 7.5 to 7.15; (4) pH reduced from 7.15 to 6.75; (5) pH increased from 6.75 to 7.5.
S L 1 at pH 7.5 (Fig. 4), with 50% inhibition occurring at a total sulphide concentration of 980 mg S L 1. Complementary to the batch experiments the effect of sulphide on methanogenesis from methanol was investigated in EGSB-II in the period day 0–61. This reactor was fed with an influent containing methanol and only 0.5 g L 1 sulphate. Sulphate was almost completely reduced in EGSB-II from day 3 onwards as revealed by the low effluent sulphate concentration of 0.05 g L 1. Other results are shown in Fig. 5. It was calculated that sulphate reduction accounted for consumption of a minor part (0.34 gMeOH–COD L 1) of the methanol load. For the remainder of the methanol, no substrate competition occurred between methanogens and sulphate reducers. This enabled us to assess the effect of external sulphide addition on the methane formation independently from substrate competition effects. Sulphide addition was started at day 21 when the methane formation had been relatively stable for a week at 5 g COD L 1 day 1. Inhibition of the methane formation was not observed in the period from day 22–25 at a total sulphide concentration of 600–700 mg S L 1. On the contrary, the methane formation increased slightly from 5.2 to 6.1 g COD L 1 day 1. Only at a total sulphide concentration of about 1250 mg S L 1 the methane formation steadily decreased from 5.6 to 4.9 g COD L 1 day 1 between day 27 and 29. The methane formation further decreased at approximately the same rate to a final value of 4.0 g COD L 1 day 1 at the imposed total sulphide concentration of 1550–1800 mg S L 1 between day 31 and 34. The total sulphide concentration was decreased to 1250 mg S L 1 on day 35 and to 800–1000 mg S L 1 between day 36 and 52. Under these conditions, the methane formation initially (day 36–41) remained relatively stable but from day 41
Fig. 4. Specific methanogenic activity of sludge as function of total sulphide. Sludge was sampled from EGSB-III on day 7. The line represents linear regression.
onwards it steadily recovered until a value of 6.3 g COD L 1 day 1 at day 52. At day 54 the external addition of sulphide was terminated, leading to a drop of the total sulphide concentration to 160 mg L 1 resulting from sulphate reduction alone. At this sulphide level the methane formation increased to 8.4 g COD L 1 day 1 at day 61. 3.4. Effect of COD/SO24 ratio From day 62 to 88 the effect of the COD/SO24 ratio on methanol utilisation was studied in EGSB-II. Results are shown in Table 1 and Fig. 6. At day 62 the sulphate concentration in the influent of EGSB-II was increased from 0.5 to 3.8 g L 1 and accordingly the COD/SO24 ratio dropped from 5.9 to 0.78 (g/g). As shown in Table 1, the effluent sulphate concentration amounted to
1830
4000
8
3500
7
3000
6
2500
5 2000
4 3
1500
2
1000
1
500
-1
4500
9
(mgS.L )
10
Tota l Sulfide
Volumetric methanogenic_ COD_ conversion rate (gCOD.L 1.day 1)
J. Weijma et al. / Water Research 36 (2002) 1825–1833
0
0 0
10
20
30
40
50
60
70
Time (days)
Fig. 5. Experimental results obtained in EGSB-II, operated at an OLR of 15.4 g COD L 1 day 1 and a SLR of 2.6 g SO24 L 1 day 1. Volumetric methanogenic COD-conversion rate (J) and total sulphide concentration (F). Arrows indicate start and termination of addition of sodium sulphide from a concentrated stock solution between day 21 and day 54.
2.2 g L 1 from day 62 to 67 (Period II), showing that sulphate was not limiting the growth of SRB. The shift of the COD/SO24 ratio instantaneously affected the electron flow: the average sulphide formation increased from 2.2 to 6.4 g COD L 1 day 1 from Period I (day 59– 61) to Period II, whereas the average methane formation decreased from 8.4 to 5.4 g COD L 1 day 1. Competition for reducing equivalents between sulphate reducing bacteria and methanogens was further stimulated by reducing the COD/SO24 ratio to 0.46 at day 68 by decreasing the OLR. This had little effect on the sulphide formation, but the methane formation immediately further dropped to 3.5 g COD L 1 day 1 in Period III (day 68–71). Only when the OLR was decreased to 6.6 g COD L 1 day 1 (COD/SO24 ratio 0.30) in Period IV (day 72–77) the sulphide formation became affected as it decreased from 7.1 to 5.0 g COD L 1 day 1 from Period III–IV. The methane formation further decreased to 1.6 g COD L 1 day 1 in these periods. Methanol was limiting in Period IV, as indicated by the low methanol concentration in the reactor of 0.03 g COD L 1. In order to assess whether methane formation could be recovered, the OLR was increased to 15.4 g COD L 1 day 1 at day 78, restoring the methanol overloading conditions from this day onwards (Period V). The methane formation recovered only to 3.1 g COD L 1 day 1, which is a 40% reduction compared to Period II. By contrast, the sulphide formation fully recovered. Thus, by imposing methanol-limiting conditions for 10 days to the sludge, a partial irreversible and selective decrease of the methane formation was achieved. The net acetate formation remained relatively low compared to the sulphide and methane formation throughout the operation of EGSB-II as it amounted
to a maximum of only 0.7 COD L 1 day 1, and the acetate formation became even less (0.2 g COD L 1 day 1) in the period from day 72 to 77, when methanol-limiting conditions prevailed (Fig. 6). Surprisingly, acetate formation did not recover in Period V at methanol overloading conditions.
4. Discussion The results show that a relatively short exposure to a slightly acidic pH and keeping the methanol sludge loading rate close to the maximum specific methanol degradation rate of the sulphidogenic biomass is effective for a rapid development of sulphidogenic biomass with low methanogenic activity. It may be speculated that methanogenesis recovers on the long term due to adaptation of the methanogenic biomass when using any method for its suppression. However, as we described in a previous paper [2] SRB almost completely outcompeted MA for methanol in long-term (up to 150 days) experiments at pH 7.5 using identical reactors and seed sludge, with highly similar organic and sulphate loading rates. Consequently, it is likely that methane formation will remain low once sulphate reduction predominates under methanol-limiting conditions. The experiment carried out in EGSB-I revealed that BES at a concentration of 2 g L 1 does not durably inhibit methanogenesis. Although temporarily methane formation ceased completely following continuous BES addition, it already started to recover within 3 days. This possibly is due to degradation of BES, as indicated by the appearance of bromide in the effluent or to adaptation of the methanogens to BES. The temporary increase in both the sulphide and acetate formation after the BES addition maybe due to interactions between
1831
1
day 0.7 0.670.0 0.570.0 0.270.1 0.170.1
1
VAC g COD L 1
day 8.470.1 5.470.3 3.570.2 1.670.8 3.170.5
1
VMC g COD L 1
day 0.05 2.270.1 F F F
2.270.0 6.470.7 7.170.2 5.070.9 6.370.8
1
VSC g COD L Sulphateeff gL 1
5.9 0.78 0.46 0.34 0.93
0.37 0.4070.01 0.1370.01 0.0370.01 0.4770.16
1
MeOHeff g COD L COD/SO24 g COD/g SO24 day
1
2.670.0 19.870.0 19.870.0 19.670.0 16.470.6
1
SLR g SO24 L day
1
59–61 62–67 68–71 72–77 78–88 I II III IV V
15.470.1 15.470.0 9.270.1 6.670.0 15.470.2
1
OLR g COD L Day Period
Table 1 Performance of EGSB-II under sulfate limiting and methanol limiting conditions. MeOHeff, sulfateeff; methanol, sulfate concentration in effluent. VSC, VMC, VAC: volumetric sulfidogenic, methanogenic and (net) acetogenic COD-conversion rates, respectively. F; not determined
J. Weijma et al. / Water Research 36 (2002) 1825–1833
SRB, MA and AB on substrate level but may also partly be due to degradation of BES. The results obtained with EGSB-I also revealed that chloroform inhibits SRB instead of MA. Recently, also Scholten [4] found inhibition of SRB by chloroform in freshwater sediment. Decreasing the pH from 7.5 to 7.15 and subsequently to 6.75 apparently is effective for a rapid and selective inhibition of methanogenesis in favour of sulphate reduction to sulphide. As a change in pH around neutral values strongly affects the free H2S concentration, this inhibition may be caused by the increased free H2S concentration, viz. from 24 to 104 mg S L 1, rather than from the drop of the pH. On the other hand, results from EGSB-II showed that methane formation remained unaffected at H2S concentrations ranging from 90 to 110 mg S L 1. Based on this observation, the reduced methane formation following the pH drop more likely is the result of inhibition by the low pH rather than of hydrogen sulphide inhibition. Minami et al. [11] found that sulphate reduction becomes stimulated between pH 6.2 and 6.8, while in pH range 7.0–7.5 methanogenesis prevailed under thermophilic (531C) conditions in a packed-bed reactor fed with methanol and sulphate. These contrasting results show that even relatively small changes in the experimental system may have a strong impact on the anaerobic degradation of methanol. Returning to a pH of 7.5 did not result in a recovery of methane formation within a period of 5 days. Such a period is short compared to the long solid retention times typical for high-rate anaerobic reactors and it may be speculated that methanogenesis recovers on the long term. However, this is most unlikely under the methanol-limiting conditions as prevailing from day 20 onwards, because SRB are known to outcompete MA under the applied conditions [2]. In case methanol is not limiting, methanogenesis may recover on the long term. This aspect requires further investigation. The pH drop of 7.5–6.75 for accomplishing a 76% inhibition of methane formation requires only a low amount of acid. Therefore, temporarily imposing slightly acidic pHvalues might represent a practical and feasible method for selective and durable suppression of methanogenesis in a full-scale process. Under sulphate limiting conditions, a 35% decrease of the methane formation was observed in EGSB-II over a 10-day period at a total sulphide concentration ranging from 1200 to as high as 1700 mg S L 1 at pH 7.5. Subsequent lowering of the total sulphide concentration to values between 800 and 1000 mg S L 1 resulted in a gradual and complete recovery of the methane formation within 2 weeks. These results show that high (1200–1700 mg S L 1) total sulphide concentrations must be maintained for several weeks for a considerable decrease of methanogenic activity. In reactors seeded with sludge exerting a low specific sulphidogenic activity such high sulphide concentrations can only be achieved
1832
J. Weijma et al. / Water Research 36 (2002) 1825–1833 12 volumetric COD-conversion rate (gCOD.L-1.day-1)
1
2 3
4
10 8 6 4 2 0 0
20
40
60
80
100
Time (days)
Fig. 6. Volumetric sulphidogenic (K), methanogenic (J), and acetogenic (m) COD-conversions in EGSB-II. EGSB-II was operated at a HRT of 4 h and was started at an OLR of 15.4 g COD L 1 day 1 and a SLR of 2.6 g SO24 L 1 day 1. Arrows: (1) SLR increased to 19.8 g SO24 L 1 day 1 at day 62; (2) OLR decreased to 9.2 g COD L 1 day 1 at day 68; (3) OLR decreased to 6.6 g COD L 1 day 1 at day 72; (4) OLR increased to 15.4 g COD L 1 day 1 at day 78. The dotted line represents the calculated sulphide production that results from the reduction of sulphate to sulphide in the period of external sulphide addition.
by applying very long hydraulic retention times for prolonged periods, which is not very practical. It still remains unclear why methanogenesis increases in the continuous reactor EGSB-II at a total sulphide concentration ranging from 800 to 1000 mg S L 1, while in batch reactors inoculated with the seed sludge of EGSBII a 40–50% inhibition of the specific methanogenic activity was observed in the same concentration range. It can be speculated that already a more sulphide-tolerant methanogenic biomass had developed in EGSB-II compared to the methanogens present in the seed sludge that was used for batch assays. However, it looks improbable that a more sulphide-tolerant methanogenic biomass already developed before the sludge was exposed to high sulphide concentrations in EGSB-II. It is furthermore unlikely that the observed contrasting results from batch and continuous reactors exclusively can be attributed to differences in hydraulic mixing conditions in both systems or to small differences in the applied environmental conditions. In any case, results of batch assays did not have any predictive value for the effect of sulphide in the continuous system. Upon the transition from sulphate limiting (COD/ SO24 ratio of 5.9) to non-sulphate limiting conditions (COD/SO24 ratio of 0.34), an 80% decrease of methane formation was found in EGSB-II. It is relevant to note that the effluent methanol and acetate concentration remained at a value of 0.4–0.5 and 0.1 g COD L 1, respectively, following the first lowering of the COD/ SO24 ratio from 5.9 to 0.78. Therefore, the decreased methane formation following at the lower COD/SO24 ratio cannot be attributed to kinetic limitations of methylotrophic and acetotrophic MA. However, not
methanol or acetate but H2/CO2 may represent the main methanogenic precursor in thermophilic anaerobic sludge cultivated on methanol, as suggested by Weijma [12]. Therefore, the reduced rate of methanogenesis following the first decrease of the COD/SO24 ratio possibly is due to kinetic limitation of hydrogenotrophic methanogens. Subsequent stepwise drops of the COD/SO24 ratio to 0.46 and 0.34 resulted in low effluent methanol and acetate concentrations. Therefore, it cannot be excluded that the observed lower methane formation at these COD/SO24 ratios is due to kinetic limitation of methylotrophic or acetotrophic methanogens. The results furthermore show that methanogenesis can irreversibly but only partially be suppressed by temporarily (EGSB-II, Period IV) applying a low COD/SO24 ratio between two periods (Period III and V) of a higher COD/ SO24 ratio. The results in Table 1 and Fig. 6 show that even at the higher COD/SO24 ratio of 0.93 in Period V, methanogenesis is lower compared to Period III. Theoreticallly, a higher COD/SO24 ratio favours methanogenesis over sulphate reduction, which is confirmed by the results from Periods II and III. Therefore, the degree of inhibition methanogenesis may have been even higher, had the same COD/SO24 ratio been applied in Period V as in Period III but certainly not lower. Most probably, the decay rate of methanogens is increased due to substrate starvation at a lower COD/SO24 ratio. The sulphide formation only decreased when the OLR was lower than the maximum volumetric sulphidogenic CODconversion rate as assessed under overloading conditions. This shows that SRB are superior over MA in the competition for available reducing equivalents. Accordingly, the major part of the reducing equivalents derived
J. Weijma et al. / Water Research 36 (2002) 1825–1833
from methanol only becomes available for the MA when utilisation of these electron equivalents by SRB is limited by the amount of sulphate or by the absence of a sufficient sulphate reducing capacity of the reactor. Thus (temporary) lowering of the COD/SO24 ratio provides a tool to lower the amount of methanol consumed per mol sulphate reduced. However, the sulphate removal efficiency decreases at lower COD/SO24 ratio but temporarily this may well be acceptable, for instance during start-up or when methanogenesis recovers due to process upsets. It is worth noting that lowering the COD/SO24 ratio of the influent to values below 0.67 or lowering the reactor pH from slightly alkaline (7.5–8) to neutral or slightly acidic (6.5–7) values also suppressed methanogenesis in mesophilic high-rate reactors fed with sulphate and a mixture of volatile fatty acids (VFA’s, i.e. acetic, propionic and butyric acid) [13]. The stimulating effect of lowering the COD/SO24 ratio likely can be attributed to the generally more favourable growth kinetic properties of SRB over MA at low concentrations of common substrates like hydrogen and acetate [14]. The similar pHeffect seems coincidental because MA in general do not favour a higher pH compared to SRB. Although the effect of pH might be indirect as it affects the speciation of sulphide, also for sensitivity of SRB and MA towards inhibitory sulphide, no general trend can be recognised from literature data [15].
5. Conclusions *
*
*
Methane formation can be substantially reduced by imposing a slightly acidic pH to thermophilic reactors fed with sulphate and methanol, or by operating the reactor at an organic loading rate close to the maximum volumetric sulphidogenic CODconversion rate. 2-BES is ineffective as durable inhibitor of methanogenesis from methanol at 651C. Total sulphide at concentrations of up to 1000 mg S L 1 does not inhibit thermophilic methanogenesis from methanol. Therefore, such high total sulphide concentrations does not favour sulphate reduction.
Acknowledgements This research was supported by the Netherlands Technology Foundation STW, applied science division of NWO and the technology program of the Ministry of Economic affairs of the Netherlands. The authors thank C.J.N. Buisman and H. Dijkman from Paques Natural
1833
Solutions, Balk, and the Netherlands for valuable discussions.
References [1] Janssen AJH, Sleyster R, van der Kaa C, Jochemsen A, Bontsema J, Lettinga G. Biological sulphide oxidation in a fed-batch reactor. Biotechnol Bioeng 1995;47: 327–33. [2] Weijma J, Stams AJM, Hulshoff Pol LW, Lettinga G. Thermophilic sulphate reduction and methanogenesis with methanol in a high rate anaerobic reactor. Biotechnol Bioeng 2000;67:354–63. [3] Oremland RS, Capone DG. Use of ‘‘specific’’ inhibitors in biogeochemistry and microbial ecology. Adv Microbiol Ecol 1988;10:285–383. [4] Scholten JCM. The effect of sulphate and nitrate on the methane formation by methanogenic archaea in a freshwater sediments. Ph.D. thesis. The Netherlands: Wageningen Agricultural University, 1999. [5] Hao JH, Chen JM, Huang L, Buglass RL. Sulphatereducing bacteria. Crit Rev Environ Sci Technol 1996;26(1):155–87. [6] Maillacheruvu KM, Parkin GF. Kinetics of growth, substrate utilization and sulphide toxicity for propionate, acetate, and hydrogen utilizers in anaerobic systems. Water Environ Res 1996;68(2):1099–106. [7] Reis MAM, Lemos PC, Almeida JS, Carrondo MJT. Evidence for the intrinsic toxicity of H2S to sulphate reducing bacteria. Appl Microbiol Biotechnol 1992;36: 145–7. [8] Lettinga G. Anaerobic digestion and wastewater treatment systems. Antonie van Leeuwenhoek 1995;67:3–28. [9] Dries J, de Smul A, Goethals L, Grootaerd H, Verstraete W. High rate biological treatment of sulphate-rich wastewater in an acetate fed EGSB-reactor. Biodegradation 1998;9(2):103–11. [10] Sill!en LG, Martell AE. Stability constants of metal–ion complexes. Burlington House: The Chemical Society, 1964. [11] Minami K, Tanimoto Y, Tasaki M, Ogawa S, Okamura K. Influence of pH on methane and sulphide production from methanol. J Ferm Technol 1988;66(1):117–21. [12] Weijma, J. Methanol as electron donor for thermophilic sulphate and sulfite reduction. Ph.D thesis. The Netherlands: Wageningen University, 2000. [13] Omil F, Lens P, Hulshoff Pol LW, Lettinga G. Effect of upward velocity and sulphide concentration on volatile fatty acid degradation in a sulphidogenic granular sludge reactor. Process Biochem 1996;31:699–710. [14] Oude-Elferink SJWH, Visser A, Hulshoff Pol LW, Stams AJM. Sulphate reduction in methanogenic bioreactors. FEMS Microbiol Rev 1994;15:119–36. [15] Colleran E, Finnegan S, Lens P. Anaerobic treatment of sulphate-containing waste streams. Antonie van Leeuwenhoek 1995;67:29–46.