Major factors influencing bacterial leaching of heavy metals (Cu and Zn) from anaerobic sludge

Environmen tal Pollution 85 (1994) 175-184

MAJOR FACTORS INFLUENCING BACTERIAL LEACHING OF HEAVY METALS (Cu A N D Zn) FROM ANAEROBIC SLUDGE D. Couillard, M. Chartier & G. Mercier Institut National de la Recherche Scientifique (INRS-Eau), UniversitOdu Quebec, 2700 rue Einstein, Sainte-Foy, Quebec, Canada, G1 V 4C7 (Received 19 November 1992; accepted 22 April 1993) tion of heavy metals in the sludges (Oliver & Cosgrove, 1974; Cheng et al., 1975; Stoveland et aL, 1979; Nelson, 1986; Stephenson & Lester, 1987a,b) often makes them improper to land spreading. Heavy metals can accumulate in plants and they can be transferred to the animal kingdom through the food chain (Chaney, 1973). In Canada and the US, more than 50% of sewage sludges exceed the governmental norms for at least one of the metals (Lue Hing et al., 1980; Wozniak & Huang, 1982; Wong & Henry, 1984; St. Yves & Beaulieu, 1988). Metal removal by way of solubilization is thus an attractive solution. Chemical solubilization of metals has been attempted on many occasions (Jenkins et aL, 1981; Wozniak and Huang, 1982; Kilt et al., 1983). However, this process proved to be too costly because of its large acid requirements (Sch6nborn & Hartmann, 1978; Wong & Henry, 1984; Tyagi & Couillard, 1987; Tyagi et aL, 1988). Moreover, this process is often found to be ineffective to solubilize copper and lead from anaerobically digested sludges. On the other hand, the biological solubilization of metals is a process both simple and inexpensive. It is executed by thiobacilli, principally Thiobacillus ferrooxidans, a chemo-autotrophic bacteria living on the oxidation of reduced sulphur compounds and/or ferrous iron. The solubilization mechanism can be direct or indirect (Lundgren & Silver, 1980; Hutchins et al., 1986; Torma, 1986). In the direct mechanism, metal sulphides (MS) are oxidized into sulphates by the bacteria (cf. eqn (1)):

Abstract

Anaerobically digested sewage sludges were treated for heavy metal removal through a biological solubilization process called bacterial leaching (bioleaching). The solubilization of copper and zinc from these sludges is described in this study: using continuously stirred tank reactors with and without sludge recycling at different mean hydraulic residence times (1, 2, 3 and 4 days). Significant linear equations were established for the solubilization of zinc and copper according to relevant parameters: oxygen reduction potential ( ORP), p H and residence time (t). Zinc solubilization was related to the residence time with a r2 (explained variance) of 0.82. Considering only t = 2 and 3 days explained variance of 0.31 and 0-24 were found between zinc solubilization as a function of ORP and p H indicating a minor importance of those two factors for this metal in the range of p H and ORP experimented. Cu solubilization was weakly correlated to mean hydraulic residence time (r z -- 0.48), while it was highly correlated to ORP (r 2 = 0.80) and p H (r 2 = 0.62) considering only t of 2 and 3 days in the case of p H and ORP. The ORP dependence of Cu solubilization has been clearly demonstrated in this study. In addition to this, the importance of the substrate concentration for Cu solubilization has been confirmed. The hypothesis of a biological solubilization of Cu by the indirect mechanism has been supported. The results permit, under optimum conditions, the drawing of linear equations which will allow prediction of metal solubilization efficiencies from the parameters p H ( Cu), ORP ( Cu) and residence time ( Cu and Zn), during the treatment. The linear regressions will be a useful tool for routine operation of the process.

bacteria

MS + 202

) MSO4

(1)

In the indirect mechanism, ferric irons (Fe 3+) generated by the oxidation of ferrous ions (Fe 2+) by T. ferrooxidans (cfi eqn (2)) react in a purely chemical way with metal sulphides to produce Fe 2+ again (cfi Eqn (3)). This Fe 2+ can be oxidized by 7". ferrooxidans once or many times in a cyclic process. Sulphur produced in this process will then be oxidized into sulphuric acid by the thiobacilli (cfi Eqn (4)), which helps to lower the pH.

Keywords: sewage sludge, heavy metals bioleaching, Thiobacillusferrooxidans, pH, ORP, Cu, Zn. INTRODUCTION The spreading of sewage sludges on agricultural lands appears more than ever to be the most attractive and inexpensive disposal solution (EPA, 1984; Davis, 1987; Nicholson, 1988). Moreover, these sludges have good fertilizing values (Davis, 1987). However, the accumulaEnviron. Pollut. 0269-7491/94/$07.00 © 1994 Elsevier Science Limited, England. Printed in Great Britain

2Fe2+ + i O 2 + 2H +T.fe . . . . . idans > 2Fe 3+ + H20

(2)

MS + 2Fe 3+ --->M 2* + 2Fe 2+ + S°

(3)

2S ° + 302 + 2H20 175

Thiobacilli

> 2H2SO4

(4)

176

D. Couillard, M. Chartier, G. Mercier

Pyrites and ferrous sulphate can also be used as substrates by the bacteria. Finally, once made soluble, the metals will be collected from the supernatant after a solid-liquid separation process, leaving a sludge suitable for land spreading. In their studies on heavy metals present in the sludge of an anaerobic digestor, Theis and Hayes (1980) drafted graphs linking soluble and insoluble forms of heavy metals to pH and oxygen reduction potential (ORP). They found that temperature, pH and aeration are of great importance for the effectiveness of metal extraction from sludges with acid. According to their work, the effectiveness of metal solubilization mainly depends on the chemical environment. The forms in which metals appear in sludges are related to pH, ORP, the presence of a chelating agent and the ligands concentration (Gould & Genetelli, 1978; Hayes et al., 1980). Wozniak and Huang (1982) studied the variables modifying the acidification efficiency for metal removal from sludges. They found that it was a function of pH, total solids concentration, types of metal and acidification time. Work on the biological solubilization of metals from sewage sludges is recent (Wong & Henry, 1984; Mercier, 1988; Tyagi et al., 1988; Couillard & Mercier, 1990, 1991a, b; Couillard & Chattier, 1991; Couillard et al., 1991). Wong and Henry (1984) noticed from bacterial leaching in batch experiments that aeration affected the solubilization of Cu, Cd, Zn and Ni, but did not affect that of Pb. According to their work, metal solubilization is also affected by temperature and pH. Tyagi et al. (1988) found that bacterial leaching of metals in batch reactors reaches higher rates as total solids concentration and pH values decrease. When the process will be applied in a wastewater treatment plant, metal analysis will be too long to perform to be an effective control parameter. If solubilization efficiency can be linked to pH and ORP the operators can adjust pumping or substrate addition hour by hour to keep pH and ORP in the optimum zone. Operation of bacterial leaching on a daily basis would be easier if simple relations between solubilization and residence time or pH or ORP are established. These linear regressions then serve as control parameters in real operation of the process. The objectives of this study are to identify and quantify the relations between the parameters pH, ORP or mean hydraulic residence time (t) and metal solubilization efficiencies during treatment of anaerobically digested sewage sludge. This would permit the development of a practicle tool for the control of the process. The comparative efficiency of the reactors used in this study has been the subject of a previous publication (Couillard & Mercier, 1990). METHODS The reactors used in this study, a 2.5 litre continuously stirred tank reactor (CSTR) and a 3.0 litre CSTR with recycling (CSTRWR), were run simultaneously as

described by Mercier (1988) and Couillard and Mercier (1990). They are made of polycarbonate and all parts coming in contact with the sludge were made of plastic or glass. The temperature was held constant in both reactors at 30 + I°C by heated water circulating through the bath containing the reactors. Both reactors were supplied with pre-aerated and pre-acidified sludge, pH in the feeding sludge was maintained at 4 by a pH controller and 5% v/v sulphuric acid. ORP was maintained between 50 and 250 mV with controlled aeration. Five masterflex pumps operating according to a programmed Chroncontrol microprocessor provided the supply of the sludge throughout the different compartments of each reactor. In the CSTRWR, a recycling rate (~) of 50% was maintained. The reactors were run at mean hydraulic residence times (t) of 1, 2, 3 and 4 days. In all cases the concentration of the substrate, FeSO4.7H20, was 1 g litre ~. Mean hydraulic residence time (t) is defined as the reactor volume (V) divided by the flow rate (Q) into the reactor (excluding recycled flow) (Grady & Lim, 1980; Ramalho, 1983). Thus, mean hydraulic residences times are equal for CSTR and CSTRWR. But solid retention time is higher in the C S T R W R because 50% of the inflow, enriched in solids by decantation, is returned in the CSTRWR. In fact V in a CSTR: t = - - = 0c = St F

(5)

V VXv in a CSTRWR: t = - - ~ S, F FwX~r

(6)

(Grady and Lira, 1980) in this CSTRWR: 0c ~ S,

(7)

where 0c is the thiobacilli retention time, V is the volume of reactor (litres), Xv is the concentration of volatile solids in the reactor (g litre ~), Fw is the flow wasted from settler (litres day '), Xvr is the concentration of volatile solids in the recycled flow (g litre ') and S, is the volatile solid retention time. Equations (5) and (6) are usually applied in wastewater treatment, where the concentration of volatile solids is assumed to be the active biomass (Grady & Lim, 1980). In the present case, the concentration of volatile solids is not the active population of thiobacilli as organic matter is present in great quantity. Unfortunately, the counting of Thiobacillus ferrooxidans in sludge is not yet technically possible and 0~ cannot be determined in the CSTRWR. The cell retention time in this C S T R W R is not necessarily equal to the volatile solids retention time as it is not known if the bacteria is mainly attached to the solids or if they are free in the solution. If thiobacilli are mainly free in the solution, 0¢ could be smaller than St and approach the value of t. In the CSTR cell retention time is equal to solid retention time as it is assumed to be perfectly mixed and there is no recirculation. Usually, the C S T R W R is

Bacterial leaching of heavy metals from anaerobic sludge

177

Table 1. Total heavy metals and total solids content of the anaerobic sludge used for each mean hydraulic residence time

Mean hydraulic residence time in reactor (days)

Cu (mg litre ~1)

Zn (mg litre 1)

Ni (mg litre 1)

Cd (mg litre -1)

Pb (mg litre 1)

1 2 3u 4

26-9 28.2 25.3 27.2

21.0 22.8 20.3 25.8

1.00 0.91 0-81 0-84

0.10 0-13 0-12 0.10

6.8 5-6 5.3 6.4

Cr (mg litre 1.52 1.38 1.26 1.37

l)

Total solids (%) 3-2 + 0.1 2.9 _+0.1 3.0 + 0.1 2.9 + 0-1

"For both runs (1 g litre ~ and 3 g litre i FeSO4.7H20).

more efficient than a CSTR for biological process (Grady & Lim, 1980; Ramalho, 1983) but for the bacterial leaching of metals from sewage sludge the opposite happened. The CSTR was more efficient than the C S T R W R when 3 g FeSO4.7H20 litre I was used (Couillard & Mercier, 1990). The reactors were considered stable after 3 × t, the time required to obtain reasonably stable values of pH and ORP, especially if the operation is in the o p t i m u m range of operation (t, concentration of substrate) of the process. Experimental results support this assumption (Couillard & Mercier, 1990). Only the values (pH, ORP, metal solubilization) of stable systems for each mean hydraulic residence time (t = 1, 2, 3 and 4 days with 1 g FeSO4.7H20 litre l) were used for calculation, for example from day 9 to day 12 for t = 3 days. An additional run, with 3 g litre t substrate concentration was experimented at t = 3 days (also stabilized by running during 3 × t) (Couillard & Mercier, 1990). The bacterial strain Thiobacillusferrooxidans (ATCC 19859) was pre-acclimated to grow on anaerobic sludges to which the substrate FeSO4.7H20 was added. The anaerobic sludges were collected at the outflow of the anaerobic digestor at Ville de Deux-Montagnes (Qudbec) wastewater treatment plant, a 10 600 inhabitant town without any major industry. Table 1 gives the total solids content and the metal composition of the sludges used at each t. pH, ORP, and dissolved 02 percentage values were taken daily for each t, except for t -- 1 day at which they were taken twice daily. O R P and pH were determined in the reactors and in the recycled sludge with a Fisher Accumet p H meter (model 805 MP). The O R P electrode reading was regularly corrected with the help of K o d a k Ektachrome 217 quinhydron at pH 4 and 7. Sludge samples (20 ml) for metal analysis were also taken daily, centrifuged at 17 000 × g for 10 min and the supernatant was stored at 4°C with 0.2 ml HNO3. Metal concentration of the supernatant in each reactor were determined according to A P H A (1985) instructions with a flame atomic absorption spectrophotometer (Varian AA-575). Cu and Zn in the sludges were analysed after 10 ml sludge were digested with 10 ml HNO3, 5 ml H F and fumed with 7 ml HC104 (APHA, 1985). Finally, volatile and inert solids were determined according to A P H A (1985) method 209 F and dissolved 02 concentration was measured with a YSI 54 oxymeter.

R E S U L T S AND D I S C U S S I O N The activity of thiobacilli causes a p H drop by H2SO 4 production (Sch6nborn & Hartmann, 1978; Wong & Henry, 1983; Torma, 1986; Tyagi et aL, 1988). The thiobacilli also increase the O R P (Wong and Henry, 1983; Tyagi et al., 1988; Couillard & Mercier, 1990) by decreasing the free electron concentrations. This O R P decrease results from the oxidation of Fe 2+ or of sulphide to sulphate. The sulphide is also a good source of energy for the thiobacilli (Kelly, 1982). The effect of thiobacilli on the decrease of p H and the increase of O R P has been clearly demonstrated during batch treatment of aerobic sewage sludge (Couillard et al., 1991). Figures 1-6 show the linear regressions found for the solubilization of each of the metals according to corresponding pH, O R P and mean hydraulic residence time. The confidence interval on the slope and the prediction interval is also shown on these figures. After many attempts, it was found that only the values of t -- 2 and t -- 3 days could be used to show the importance of pH and ORP. This is probably due to the major dependence of solubilization towards t which dissimulate others relations. So for the regressions relating metal solubilization to pH or ORP, the only values considered were those for t -- 2 and 3 days. Table 2 gives the regression equations of these relations as well as their explained variance (r 2) and their significance at the 95% confidence level according to the Student's t-test. The confidence interval on the slope, at the 95% confidence level, and on the intercept were also calculated. In all cases, Zn and Cu solubilization related significantly (in a statistical sense) to parameters ORP, pH and t (of Table 2). Simple linear regressions have been chosen because they give simple equations which permit easy prediction for routine operation. It also permits to find a threshold value which is a value of p H or O R P at which the first zero solubilization value is statistically observed. Figure 1 shows Zn solubilization as a function of the O R P while Fig. 2 relates it to p H and Fig. 3 to the residence time. Residence time was the main factor explaining zinc solubilization with a r 2 = 0.82. Zinc solubilization was residence time dependent for all p H and O R P values studied. F r o m Fig. 3, it appears that a t of 3 or 4 days is necessary to maximize Zn solubilization when 1 g litr~ 1 of FeSO4.7H20 is used. A threshold

178

D. Couillard, M . Chartier, G. M e r c i e r 100

I

I

,

I

I

~

,

I

I

[

,

,

,

,

I

I

,

I

I

8 0

N

¢o

•

.

,

.

.

•

•

.

•

:~

~

,~.

•

-

•

°--

~3

N

°-.--

0

(/1

. . . . . . . . . . .

.....

.Q 6 0

-N

• • o

-N o

"

"

°

~. "

~

•

"

"

~

¢

.

~

.......

"

"

~

-

"

.~

~

~

. . ")~. •

~

.

~

,

"

"

"

"

"

"

"

7 "

"

_

"

--

°

• • • •

40

I

J

2 0 0

i

I

J

5 3 0 0

[

i

I

I

I

4 0 0

I

I

I

I

I

5 0 0

ESO0

ORP (mV) Fig. 1. Linear regression for Zn solubilization as a function of ORP for mean hydraulic residence time of 2 and 3 days (dotted lines are interval on the slope (narrowest interval) and prediction interval, both at the 95% confidence level).

value of p H 7.2 is calculated from the equation in Table 2, which is far from the p H = 5-5 threshold calculated by Theis and Hayes (1980) and from the one obtained by A d a m s and Sanders (1984) (pH = 5-8). In fact, the weak r 2 ( c f Table 2) for p H comes from the importance of the time factor which makes the determination of the actual p H threshold inaccurate. In fact, it can be concluded that in the range of pH studied the significance of this relation is minor and pH iO0

l'

does not influence much solubilization so the threshold value of 7.2 is probably far from reality• The ORP related equation (Table 2) shows a - 1299 mV threshold for zinc. This value is very distant from that of 150 mV predicted by Theis and Hayes (1980) at pH = 5. It can be explained by the fact that the importance of O R P appears to be quite low for zinc with an r 2 of 0.31. The time factor being too important it is impossible to get accurate value for threshold value of ORP. On the

1

f

'...

I

l

"Y¢~

"

1

I

[

1

I

I

[

,

8 0

tO

--

'" - ) 4 " ~

"

"

"

"

,

--

°_

~

D

~

~

~

,.....

-

N

0

6 O

O'3

• • - ,

t

4 0 2

t

1

J

J

I

I

I

~4

I

i""

[

2 . 5

3

~

l

~

I

I

1 3 . 5

I

I 4

pH

Fig. 2. Linear regression for Zn solubilization as a function of pH for mean hydraulic residence time of 2 and 3 days (dotted lines are interval on the slope (narrowest interval) and prediction interval, both at the 95% confidence level).

Bacterial leaching of heavy metals from anaerobic sludge .100

¢O

i 79

I

8 O

j

.4.-

,"

•

.•

.'

N

c~ 0 (/)

tl

6o

~

..." '" ,.

F

-•

..

4 0

0

.1

2

'=1

,4

5

Mean hydraulic residence flme (d) Fig. 3. Linear regression for Zn solubilization as a function of t (dotted lines are interval on the slope (narrowest interval) and prediction interval, both at the 95% confidence level). Figures 4, 5 and 6 present copper solubilization as a function o f O R P , p H and mean hydraulic residence time, respectively O R P is the major factor correlating with Cu solubilization with an explained variance of 0,80 (<[i Table 2). T o obtain more than 50% Cu solubilization, the O R P must be higher than =480 mV with the specified conditions (see the M e t h o d s section). Linear regression o f Cu solubilization versus p H has an explained variance o f 0-62. A p H lower than 2-9 gives

whole, zinc was more easily made soluble because p H and O R P values in these trials were far from minimal solubilization thresholds (pH = 5-5 to 5-8; O R P = 150 mV). Thus, t comes out as the only major factor influencing Zn solubilization (in the range studied) with an explained variance (r e, see Table 2) o f 0.82. Three or 4 days are necessary to obtain g o o d results. This fact is interesting and will help in the control o f the bacterial leaching•

~o

4-0 N

2

0 (.f'l)r--

.o

'

7i -

*

•

~OO

~

..

|

"

s,4

.

.

."

,'

300

'

I

.. '

)4

'~34 . •" ~..

.

'

'

/

.~ .

"

"' * . ~

~

.

~00

500

600

ORP (mv) Fig. 4. Linear regression tbr Cu solubilization as a function of ORP for mean hydraulic residence time of 2 and 3 days (dotted lines are interval on the slope (narrowest interval) and prediction inler~al, both at the 95% contidence lcvclL

D. Couillard, M. Chartier, G. Mercier

180 BO

.~.1

I

.I

I,.•

I

I

.,•..

SO

[

I

..

co

.m .4,.-

C~ N 0_ ,_

4o

J3 0 U3

~

"~

.

.

20

•

',

' . ,

':\":, I

1

I

2

[

"

1

2 . 5

3

r.

[,

I

3,5

4

pH Fig. 5, Linear regression for Cu solubilization as a function of pH for mean hydraulic residence time of 2 and 3 days (dotted lines are interval on the slope (narrowest interval) and prediction interval both at the 95% confidence level)•

the present case those are practically interpolations within the range of experimental value. Zinc threshold values were obtained by extrapolation, which is not accurate• The fact that the experiments took place within these ranges could explain the predominance of those parameters over the residence time, which shows an r 2 of 0.48. In fact, Theis and Hayes (1980) calculated that Cu was made soluble at an O R P value higher than about 250 mV and a p H value lower than 4.5 for

50°/,, Cu solubilization and more. Note that only t = 2 or 3 days have been considered for the two former regressions for the reasons explained above. In the experiments, the values of pH were in the range of 2.82-3.43 while those for the O R P were between 215 and 548 inV. As the threshold value (0% solubilization) of pH (3.52, cf Fig. 5) and O R P for copper (221 mV, ~f Fig. 4) are included in this range, the determination of these values is much more accurate than for zinc. In ,

50

q

° °

m

, • • " "

•

r-

~

~°

,

- "

40 m

° ° ° "

m

,

• ° • • "

.)~

o " "

° "

-~

° ° ° ° "w

cO 30

• ~.

N

.

-

-

-

•

°

°

.

°

"

o

-

°

.

-

o

°

--

°~

o

20

J I

10 - -

.

I O

[

m

°

I

Jl,,llll t

l 2

Ill,lllll 3

1 4

I 5

Mean hydraulic residence tlme (d) Fig. 6. Linear regression for Cu solubilization as a function of t (dotted lines are interval on the slope (narrowest interval) and prediction interval, both at the 95% confidence level).

Bacterial leaching of heavy metals from anaerobic sludge

181

Table 2. Zn and Cu solubilization related to parameters ORP, pH and mean hydraulic residence time (t)

Metal

Linear regression

n

Explained variance (rz)

Statistical significance (T value)

Confidence intervals on slope (95%)

Confidence intervals intercept (95%)

Zn

y solubilization = 51.96 + 0-04 (ORP) y solubilization -- 118.22 - 16.31 (pH) y solubilization = 46.04 + 9.30 (t)"

30 30 44

0.31 0.24 0-82

sign. sign. sign.

[0.003; 0.072] [-27-38; -5.24] [9.56; 12.70]

[42.60; 61.29] [85.93;150.50] [38.66; 47.35]

Cu

y solubilization -- -39.75 + 0.18 (ORP) y solubilization -- 271.39 - 77.01 (pH) y solubilization = 18.04 + 5.30 (t)"

29 29 40

0.80 0.62 0-48

sign• sign• sign.

[0•14; 0 . 2 1 ] [ - 100•76; -53-44] [3.12; 6.31]

[-54.60; -24.89] [198-05;344.72] [13.79; 22-84]

" t, mean hydraulic residence time.

10 3 M Cu in anaerobically digested sludges. The results in Fig. 4 show a zero Cu solubilization at an O R P value of 221 inV. This threshold value is close to the theoretical 250 mV threshold found by Theis and Hayes (1980). For p H (cf. Fig. 5), the results indicate that a value higher than 3.5 (3.43-3.72 in the interval of slope and intercept) prevents Cu solubilization. This threshold differs from the one found by Theis and Hayes (1980) (4-5) and Adams and Sanders (1984) (4-5), the latter with thickened and frozen raw sludges. F r o m these results it is possible to conclude that Cu behaves really in a different way than Zn. Cu being p H and O R P dependent it will limit more the efficiency of the process than zinc. In fact, the control of the process could be based only on copper solubilization, because when copper was solubilized zinc was also solubilized. The O R P (Cu) solubilization as shown by these results, could account for the weak solubilization rates

obtained with chemical processes for anaerobic sludge (Hayes et al., 1980; Jenkins et al., 1981). In chemical processes, only p H is regulated, leaving O R P at values too low to permit Cu solubilization. The O R P factor could also explain the better metal solubilization efficiencies obtained by Hayes et al. (1980) from aerobic sludges, in comparison to anaerobic sludges. The O R P values in aerobically digested sludges are generally above 400 mV (Ahlberg & Boyko, 1972). In the biological leaching of metals from sewage sludge, it is clearly demonstrated that O R P increase as p H decrease (t = 2 and 3 days) (see Fig. 7). An explained variance of 0.76 implies that these parameters are strongly correlated. This is logical as p H and O R P are the result of the biological activity of the thiobacilli. This study does not pretend to determine the chemical form or speciation of metals in sewage sludge, this

BOO

,

,

I

550

500

>

,~

450

-

E 400

13_ 350

0 :300

250

200 2

2.5

3

3.5

,4

pH Fig. 7. Linear regression for ORP as a function of pH (n = 27 and r 2 = 0.76) for mean hydraulic residence time of 2 and 3 days (dotted lines are interval on the slope (narrowest interval), and prediction interval, both at the 95% confidence level)•

182

D. Couillard, M. Chartier, G. Mercier Table 3. Reactor effectiveness using 1 g litre i or 3 g FeSO4"TH20 litre -I for t = 3 days

Cu

Zn

CSTR

Mean solubilization efficiency (%) Solubilization rate increase (+) or decrease ( - ) (%)

CSTRWR

CSTR

CSTRWR

1 g litre 1

3 g litre i

1 g litre ~

3 g litre i

1 g litre i

3 g litre i

1 g litre i

3 g litre i

41.8

62-2

36.9

50.1

72.2

77-4

70.8

64.1

+20.4

+13.2

being a field of research in itself (Lake et al., 1984; Legret et al., 1987), But from these results it is clear, that since copper is solubilised to a limit of 65% (Fig. 5) and zinc 80°/,, (Fig. 2) the intracellular fraction can be no more than 20% of the zinc and 35% of the copper. Intracellular fraction is theoretically not available unless the cells are destroyed. In order to determine if the solubilization of Cu and Zn depends on the concentration of the substrate, an experiment has been led at t = 3 days (stabilized by running for more than 3 x t) with two different substrate concentrations: 1 g litre ~ and 3 g litre ~ FeSO4-7H20. The results appear in Table 3. According to Lundgren and Silver (1980), metals being made soluble by the indirect mechanism are very O R P dependent. As solubilization is strongly influenced by O R P it can be proposed that Cu is oxidized by the bacteria following this mechanism. Moreover, in the indirect mechanism, the concentration of the substrate added should influence metal solubilization since the substrate is the source of iron. In fact, a shortage of Fe 2+ means less Fe 3+ produced, which, in turn, slows down the chemical transformation of sulphides into sulphates, given that Fe 3+ is present in a lesser quantity. Thus, a metal oxidized by the indirect mechanism will be dependent on the concentration of FeSO4.7H20. The authors observed this influence of substrate concentration on Cu solubilization. The Cu solubilization efficiencies were higher both in the C S T R and in the C S T R W R with 3 g litre ~ of substrate instead of 1 g litre ~, The increase was from 41.8% to 62.2% in the C S T R and from 36.9% to 50.1% in the C S T R W R . For Zn, the solubilization efficiencies were slightly higher with 3 g litre l substrate in the C S T R (+ 5-2%) and slightly lower in the C S T R W R (-6.7%), These results seem to show that Cu solubilization is substrate concentration and O R P dependent whereas nothing similar could be concluded in the case of Zn, a difference of about 5% not being assumed significant. Moreover, the effect for Zn is positive in one case (CSTR) and negative in the other case (CSTRWR). The hypothesis according to which Cu would more likely be solubilized by the indirect mechanism (FeSO4.7H,O and O R P dependent) is thus supported as demonstrated previously

+5.2

-6.7

(Couillard & Mercier, 1990). This tendency has not been discerned for Zn. Finally, this demonstrates the importance of the substrate concentration in the event of optimizing Cu solubilization efficiencies in the operation of a reactor. It is important to understand that direct demonstration of the type of mechanism is not an easy task. First, the indirect mechanism is a cyclic process and one molecule of iron can solubilize one or more molecules of metal sulphides so the knowledge of the quantity of Fe 3+ formed does not necessarily determine the quantity of metal sulphides actually oxidized by Fe 3÷. In fact, a direct mechanism is probably also operating and even if concentrations of S2 , S O t , Fe 2+ and Fe 3+ are determined, it would be impossible to differentiate between sulphide directly and indirectly oxidized. The indirect demonstration used in this experiment is one of the few ways to demonstrate the action of the indirect mechanism. The interest of these linear regressions is to predict Zn and Cu solubilization by the evaluation of easily determined parameters: pH, O R P and residence time. This demonstrates a certain constancy in the biological process of metal solubilization, even in a matrix as complex as sewage sludges. Linear regressions allows the operators of the process (in a real situation) to anticipate solubilization efficiencies with measurement of pH and ORP, thus optimizing the functioning of a reactor for routine operation in terms of costs and effectiveness. CONCLUSION This research has shown the relative importance of pH, O R P and hydraulic residence time on Cu and Zn solubilization. It also stressed the importance of other factors such as substrate concentration. The weak Cu solubilization rates obtained with 1 g litre ~ substrate shed light on the necessity to carry out additional research with higher substrate concentrations in order to improve solubilization. Cu solubilization was highly dependent on ORP. This tends to support the hypothesis of Cu solubilization by the indirect mechanism, p H ranks second with regards to its influence in the case of Cu, residence time

Bacterial leaching o f heavy metals f r o m anaerobic sludge

c o m i n g last. Zn solubilization is mainly explained by the residence time (r 2 = 0.82), leaving O R P and p H far behind within ranges prevailing in the functioning o f the reactor in biological solubilization processes. F o r the reasons mentioned, copper solubilization will be the limiting factor which will permit the control o f the process with simple linear regressions in a wastewater treatment plant.

ACKNOWLEDGEMENTS This research was made possible t h r o u g h grants f r o m Natural Sciences and Engineering Research Council o f C a n a d a (grant OGP0003711), the MinistOre de l'l~ducation o f the Province o f QuObec (grant F C A R EQ-3029) and the Centre Qu6b6cois de Valorisation de la Biomasse, Qu6bec, Canada.

REFERENCES Adams, T. McM. & Sanders, J. R. (1984). The effects of pH on the release to solution of zinc, copper and nickel from metal loaded sewage sludges. Environ. Pollut., B8, 85 99. Ahlberg, N. R. & Boyko, B. L. (1972) Evaluation and design of aerobic digesters. J. o f Water Pollut. Control Fed., 44(4), 634-43. APHA (1985). Standard Methods for the Examination o f Waste and Wastewater (16th edn). American Public Health Association, Washington, DC, USA. Chaney, R. L. (1973). Crop and Food Chain Effect o f Toxic Elements in Sludges and Effluents on Land. National Association of State University and Land, Grant Colleges, Washington, DC, USA, pp. 129-40. Cheng, M. H., Patterson, J. W. & Minear, R. A. (1975). Heavy metals uptake by activated sludge. J. Water Pollut. Control Fed., 47(2), 362 76. Couillard, D. & Chattier, M. (1991). Biological removal of heavy metals from aerobic sludge in batch reactors. J. Biotechnol., 20, 163 80. Couillard, D. & Mercier, G. (1990). Bacterial leaching of heavy metals from sewage sludge Bioreactors comparison. Environ. Pollut., 66, 237 52. Couillard, D. & Mercier, G. (1991a). Optimum residence time (in CSTR and airlift reactor) for bacterial leaching of metals from anaerobic sewage sludge. Water Res., 25(2), 211 18. Couillard, D. & Mercier, G. (1991b) ProcOd6 de solubilisation biologique des m6taux dans les boues ana6robies d'Opuration: Filtrabilit6, neutralisation et teneurs en N et P des boues traitOes. Can. J. Chem. Engng, 69, 779-87. Couillard, D, Chartier, M. & Mercier, G. (1991). Bacterial leaching of heavy metals from aerobic sludge. Bioresource Technol., 36, 293 302. Davis, R. D. (1987). Use of sewage sludge on land in the United Kingdom. Water Sci. Technol., 19(8), 1-8. EPA (1984). Use and Disposal of Municipal Wastewater Sludge (EPA 625/10-84-003). u S Environmental Protection Agency, Cincinnati, OH 45268, USA. Gould, M. & Genetelli, E. J. (1978) Heavy metal complexation behaviour in anaerobically digested sludges. Water Res., 12, 505-12. Grady Jr, C. P. L. & Lira, H. C. (1980). Biological Wastewater Treatment. Marcel Dekker, New York, USA. Hayes, T. D., Jewell, W. J. & Kabrick, R. M. (1980) Heavy metal removal from sludges using combined biological/

183

chemical treatment. Proc. 34th ConJerence on Industrial Waste, Ann Arbor Science Publishers, Ann Arbor, MI, USA, pp. 529-43. Hutchins, S. R., Davidson, M. S., Brierley, J. A. & Brierley, C. L. (1986). Microorganisms in reclamation of metals. Ann. Rev. Microbiol., 40, 311-36. Jenkins, R. L., Scheybeler, B. J., Smith, M. L., Baird, R., Lo, M. P. & Haug, R. T. (1981). Metals removal and recovery from municipal sludge. J. Water Pollut. Control. Fed., 53, 25 32. Kelly, D. P. (1982). Biochemistry of the chemolithotrophic oxidation of inorganic sulphur. In Trans. Royal Society of London, B298, 499 528. Kiff, R. J., Cheung, Y. H. & Brown, S. (1983). Heavy metal removal from sewage sludges, factors governing detoxification process efficiency. In Proc. o f the 3th ConJerence on Heavy Metals in the Environment. CEP Consultants, Edinburgh, UK, pp. 401-4. Lake, D. L., Kirk, P. W. W. & Lester, J. M. (1984). Fractionation, characterization and speciation of heavy metals in sewage sludge and sludge-amended soils: A review. J. Environ. Qual., 13(2), 175 83. Legret, M., Diver, L. & Marchandise, P. (1987). Mobilit6 et extraction des m6taux lourds associ6s aux boues des stations d'6puration. Water Res., 21(5), 541 7. Lue Hing, C., Zeng, D. R., Sawyer, B., Guth, E. & Whitebloom, S. (1980). Industrial waste pretreatment and EPA cadmium limitations. J. Water Pollut. Control. Fed., 52, 2538-51. Lundgren, D. G. & Silver, M. (1980). Ore leaching by bacteria. Ann. Rev. Microbiol., 34, 263 83. Mercier, G. (1988). L'extraction biologique des m6taux lourds des boues ana6robies d'6puration. Master theses, INRS-Eau, University of Qu6bec, Qu6bec, Canada. Nelson, P. O. (1986) Studies of the uptake of heavy metals by activated sludge. In Proc. International Symposium on Metal Speciations, Separations and Recover)'. Industrial Waste Elimination Research Center, Lewis Publishers, Chelsea, Michigan, USA, pp. VIII-69-VIII-76. Nicholson, J. P. (1988). What are we doing with sludge? Pollut. Engng., 20(3), 14. Oliver, B. G. & Cosgrove, E. G. (1974). The efficiency of heavy metal removal by a conventional activated sludge treatment plant. Water Res., 8, 869 74. Ramalho, R. S. (1983). Introduction to Wastewater Treatment Processes, (2nd edn). Academic Press, New York, USA, p. 296. Sch/Snborn, W. & Hartmann, H. (1978). Bacterial leaching of metals from sewage sludge. Eur. J. Appl. Microbiol. Biotechnol., 5, 305 13. Stephenson, T. & Lester, J. N. (1987a). Heavy metal removal during the activated sludge process. I. Extent of soluble and insoluble metal removal mechanisms. Sci. Tot. Environ., 63, 199 214. Stephenson, T. & Lester, J. N. (1987b). Heavy metal removal during the activated sludge process. II. Insoluble metal removal mechanisms. Sci. Tot. Environ., 63, 215-30. Stoveland, S., Astruc, M., Lester, J. N. & Perry, R. (1979). The balance of heavy metals through a sewage treatment works. II. Chromium, nickel and zinc. Sci. Tot. Envir., 12, 25-34. St. Yves, A. & Beaulieu, R. (1988). Caract6risation des boues de 34 stations d'6puration des eaux us6es municipales (jan. f6v. 1988). Minist6re de l'Environnement du Qu6bec, Direction G6n6rale de l'Assainissement des Eaux, Direction de l'Assainissement Agricole, no. 262. Theis, T. L. & Hayes, T. D. (1980). Chemistry o f Wastewater Technology, ed. A. J. Rubin. Ann Arbor Science Publishers, Ann Arbor, MI, USA, pp. 403-419.

184

D. Couillard, M. Chartier, G. Mercier

Torma, A. E. (1986). Biohydrometallurgy as an emerging technology. Biotechnol. Bioengng Syrup., 16, 49-63. Tyagi, R. D. & Couillard, D. (1987). Bacterial leaching of metal from digested sewage sludge. Process Biochem., 22, 114-17. Tyagi, R. D., Couillard, D. & Tran, F. (1988). Heavy metal removal from anaerobically digested sludge by chemical and microbiological methods. Environ. Pollut., 50, 295-316.

Wong, L. & Henry, J. G. (1983). Bacterial leaching of heavy metals from anaerobically digested sewage sludge. Water Pollut. Res. J. Can., 18, 151-62. Wong, L. & Henry, J. G. (1984). Decontaminating biological sludge for agricultural use. Water Sci. Technol., 17, 575-86. Wozniak, D. J. & Huang, J. Y. C. (1982). Variables affecting metals removal from sludge. J. Water Pollut. Control Fed., 54, 1574~80.