Continuous anaerobic treatment of autoxidized bark extracts in laboratory-scale columns

Continuous Anaerobic Treatment of Autoxidized Bark Extracts in Labora tory-Scale Columns

J. Field, M. J. H. Leyendeckers, R. Sierra-Alvarez, and G. Lettinga Department of Water Pollution Control, Wageningen Agricultural University, Wageningen, The Netherlands

L. H. A. Habets Paques B. V, Balk, The Netherlands

Received February 28, 199OMccepted June 22 1990

Debarking wastewaters of the forest industry contain high concentrations of tannins that are inhibitory to methane bacteria. The tannins can be polymerized to nontoxic col- ored compounds by the applications of an autoxidation pretreatment, enabling the anaerobic treatment of easily biodegradable components in the wastewater. The continu- ous anaerobic treatment of untreated and autoxidized pine bark extract was studied in laboratory-scale columns packed with a granular sludge bed. The autoxidation dou- bled the conversion efficiency of bark extract COD to methane (from 19 to 40%). After 5 months of operation, anaerobic treatment of the autoxidized extracts was feasible at high influent concentrations (14 g COD/L) and loading rates (26 g biodegradable COD/L * d ) with 98% elimination of the biodegradable fraction. The detoxification pretreat- ment polymerized the toxic tannins to poorly biodegradable high molecular weight tannins and humic compounds which were not eliminated during anaerobic treatment. Although the original tannins of the untreated extract were elimi- nated by 60%, they were not biodegraded to volatile fatty acids and methane but instead were transformed to pheno- lic degradation intermediates (phenol, p-cresol, 3-phenyl- propionate, and carboxycyclohexane). Therefore, the autoxidation pretreatment did not decrease the content of readily biodegradable substrates which accounted for 53% of the extract COD. The recalcitrant COD expected in the effluents of reactors treating autoxidized debarking waste- water can be effectively separated by calcium precipitation prior to anaerobic treatment.

INTRODUCTION

Previous studies indicate that the debarking wastewa- ters of the forest industry are troublesome to treat by the anaerobic methods.’-3 The problems observed were due to their methanogenic toxicity. Aqueous extracts of bark were studied to determine the inhibiting com- pounds of debarking wastewaters. The tannins, which account for 30-60% of the chemical oxygen demand

* Present address: Department of Chemical Engineering, Fac- ulty of Sciences, +tonornous University of Barcelona, 08193 Bel- laterra (Barcelona), Spain.

(COD) and 70-90% of the UV absorbance, were shown to cause most of the methanogenic toxicity observed in coniferous bark extract^.^ The toxicity is postulated to result from hydrogen bonding interactions between tan- nins and functional proteins of bacteria? Autoxidative treatments can be applied to detoxify the bark ex- tracts6 The detoxification effect is due to the oxidative polymerization of the toxic low molecular weight (MW) tannins to nontoxic high MW tannins and humic-like nontannic compounds.’ These high MW compounds are unable to penetrate to the sensitive sites (i.e., func- tional proteins) of methane Therefore, oxi- dative polymerization reactions can potentially be applied as a pretreatment, enabling the anaerobic treat- ment of biodegradable substrates in highly toxic debark- ing waste streams.

The previous work,627 however, is based on laboratory batch experiments conducted over relatively short time periods (-2 weeks). The purpose of this study was to evaluate the role of autoxidative detoxification meth- ods in improving the long term anaerobic treatment of coniferous bark extract. Since the autoxidation pre- treatment might change the biodegradability of the tan- nins and other phenols present in the debarking wastewater, an additional objective was to determine if the detoxification method lowers the content of sub- strate available for methane production.

MATERIALS AND METHODS

Bark Extracts

Throughout this study paper filtered aqueous extracts of bark were prepared by adding 60°C tap water to ground air dried bark of scot’s pine (Pinus sylvestris) and shak- ing for 3 h with N2 gas in the head space. Additionally, 250 mg/L of ascorbic acid was added to the extract to prevent premature autoxidation. The pine bark extracts

Biotechnology and Bioengineering, Vol. 37, Pp. 247255 (1991) 0 1991 John Wiley i3 Sons, Inc. CCC 0006-3592/91/030247-010$0400

were prepared to different concentrations by using dif- ferent amounts of air dried bark which ranged from 18 to 144 g/L. On the average, the tannins accounted for 36.5 25.4% of the extract COD and 77.8 27.7% of the extract UV absorbance. The average yield of tannins by the extraction procedure was 5.0 21.3% of the bark on a dry weight/dry weight basis.

Analytical Methods

The COD (the micromethod with dichromate) and volatile suspended solids (VSS) were determined ac- cording to standard methods? The UV absorbance of the extracts and samples were based on the absorbance of 215 nm in a 1-cm quartz cuvette as described previ- o ~ s l y . ~ The color of the extracts was based on the vis- ible absorbance measured at 440 nm.6 The tannins were measured according to the polyvinylpyrrolidone (PVP) method described in detail by Field et aL4 In summary, this method is based on the disappearance of COD and UV absorbance when the extract is shaken together with 14.3 g/L PVP for 1 h. The HPLC chromatographic procedure and the low MW tannin determination based on the HPLC results were described in previous publi- cation~?-~ The VFA analysis and the determination of phenol, p-cresol, and carboxycylohexane were based on the gas chromatographic procedure described by Field et aL9 The analysis of trans-cinnamic acid, 3-phenyl- propionic acid, and phenylacetic acid was based on a similar gas chromatographic procedure using an oven temperature of 190°C instead of 130°C.

Column Experiments

The continuous experiments were conducted in small glass columns which contained 0.15 L of liquid volume as illustrated in Figure 1. These were placed in a con- stant temperature room of 30 22°C. The influents of the columns were prepared from the bark extracts by adding required nutrients of bacterial growth as de- scribed in Field et aL9 and by adding alkalinity in the form of NaHC03 at approximately 1 g/g biodegradable COD (-0.5 g/g extract COD). The influents which were prepared for 3-7-day intervals of operation were placed in a refrigerator. The head space of the influent containers were filled with Nz gas, and as they were pumped empty, Nz gas was provided from a gas bag to replenish the void volume.

The methane gas production was measured in mari- otte flasks of 10 L volume filled with 5% NaOH to scrub out the CO2 from the biogas.

In this study two separate experiments were con- ducted with the columns. The first experiments consisted of three columns treating (1) untreated (unoxi- dized); (2) tannin-free (PVP treated); and (3) autoxi- dized pine bark extract. These columns were each seeded with 20 g VSS/L reactor with a granular sludge originally cultivated on potato processing wastewater

u 5 8/

t T 1 &

I

Figure 1. Design of the 0.15-L columns used in the continuous experiments of this study: (1) influent; (2) glass beads; (3) granular sludge bed; (4) screen; (5) gas separator; (6) biogas to mariotte flask; (7) effluent water lock; (8) syphon break; (9) effluent sam- pling point; (10) effluent discharge.

but obtained after several months of storage in an anaerobic lagoon. After seeding, each of the columns were operated under identical conditions with a pH 7 neutralized VFA stock solution (C2:C3:C4 = 24:34:41% COD) supplied at 4.2 g COD/L and an av- erage loading of 12 g COD/L - d for 1 month followed by an additional 2 weeks with 8.4 g COD/L and an aver- age loading of 22 g COD/L . d, at which time the COD elimination efficiency was 95% or greater. After this 6-week adaption period to the VFA substrate, the column experiment was started (day 0 of the experi- ment) by changing to the bark extract influents.

The second experiment consisted of one column treating calcium precipitated autoxidized pine bark ex- tract. This column was seeded with 31 g VSS/L reactor with the sludge obtained from column 3 (of the first ex- periment). The experiment was started (day 0) directly with the bark extract influent.

The biodegradability of the bark extracts during the column experiments was evaluated with the following parameters: M = the COD-based methane yield as a percentage of the influent COD; VFA = VFA remain- ing in the effluent as a percentage of the influent COD; ECOD = the percentage of COD elimination based on the filtered effluent sample; Cells = percentage of con- version of influent COD to cells (estimated from ECOD minus M ) ; A = percentage of conversion of influent COD to acidified products (the sum of M and VFA); BD = percentage of biodegradability of the influent COD under the operational conditions (sum of A and Cells); Y = specific cell yield in COD per unit of COD biodegraded; ECOD,, = the percentage of elimination of the biodegradable COD, =[1 - (VFA/BD)] x 100; MBD = the COD-based methane yield as a percentage of the biodegradable COD in the influent.

Methanogenic Activity Assays

The activity test was conducted for the granular sludge recovered from column experiments. The sludge

248 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 37, FEBRUARY 1991

(1.3 g VSS/L) was fed 4 g COD/L neutralized VFA sub- strate containing acetate, propionate, and butyrate in a ratio of 1 : 1 : 1 dry wt or 24: 34: 41 COD basis. Nutrients were added as described previously for the column in- fluents. The assay was conducted in two feedings. At the end of the first feeding, the old VFA media was replaced with new VFA media. Methane gas produc- tion was measured in 1.0-L mariotte flasks filled with 5% NaOH.

Anaerobic Biodegradability Test

The biodegradability of untreated bark extracts or tannin-free extracts was assayed with sludge recovered at the end of the column experiments. Alkalinity and nutrients were added in accordance with the additions described previously for the column influents. Sludge and extract concentrations utilized are described in the results.

Bark Extract Pretreat ment s

Autoxidation of the bark extracts was conducted by raising the pH of the bark extract to pH 11.5 (requiring 0.7-1 g NaOH/L), and aerating with a porous aeration stone. The aeration rate applied was high (approxi- mately 30 v/v per hour) to ensure that the supply of air was not limited. In the first continuous experiment an aeration period of 16 h was utilized. After autoxidation, the extract pH was readjusted to 6.5 with HC1. The av- erage decrease in the tannin content by the autoxidation performed on 19 extracts used to prepare influents in this study was 50.8 &10.9% Autoxidation had no signif- icant effect on the COD concentration of the extract.

Tannin-free extracts were prepared by PVP treatment of bark extracts which was conducted in accordance with the PVP determination of tannins, i.e., utilizing approximately 14.3 g PVP/L of bark extract. The tan- nin-free extracts contained on the average 63.5% of the COD and 22.2% of the UV absorbing matter compared to untreated extracts.

In the second continuous experiment, calcium treat- ment of autoxidized bark extracts was conducted in or- der to precipitate high MW polymerized products. In a previous study, the ability of calcium to precipitate the high MW autoxidation products was observed." The extract was aerated for 1 h at a starting pH of 11.5, and then 500 mg/L of Ca2+ (as CaC12) was added, at which time the pH was approximately 9. The precipitate was allowed to settle for 2 h and was then decanted through a paper filter. The treated extract was adjusted to pH 6.5 with HCl. The 1-h autoxidation treatment re- moved 4.0 k3.856, 16.3 +3.2%, and 43.5 +11.3% of the extract COD, UV, and tannins, respectively. The com- bined autoxidation and calcium treatment removed 29.6 +2.1%, 72.3 +3.6%, and 86.0 43.2% of the ex- tract COD, UV, and tannins, respectively.

RESULTS

Effect of Autoxidation on Extract Characteristics

The changes in the pine bark extract characteristics as a function of the autoxidation pretreatment time are illustrated in Figure 2 for a typical experiment. The fig- ure illustrates that the autoxidative polymerization caused coloration of the extract associated with a de- crease in the tannin content and an even larger decrease of the low MW tannin content. Most of the changes occurred after only 1 h of autoxidation, when the con- centration of the toxic low MW tannins was reduced by 75%. The polymerization reactions caused a small de- crease in the total UV absorbance (only 20%) and no decrease in the COD.

Continuous Treatment of Extract

Treatment Performance

The autoxidation of the bark extracts increased the re- actor performance. The column fed with the untreated extract provided lower levels of COD elimination and had higher effluent volatile fatty acid (VFA) concentra- tions compared to the column fed with autoxidized ex- tract (Table I). This was clearly due to the inhibition of methane production caused by the tannins in the un- treated extract. The experiment averaged only 19% conversion of the COD to methane with the column fed untreated extract; whereas twice the yield of meth- ane (40%) was obtained by autoxidizing the extract (Table 11). The low yield of methane from the untreated extract resulted in the incomplete utilization of VFA, which in turn was responsible for an incomplete elimi- nation of biodegradable COD. In contrast, the au- toxidized extract fed column provided the maximum obtainable methane yield possible, since a nearly com- plete elimination of the biodegradable substrate oc-

100% A - 4

k 40

20

1 ' 0 0.2 0.4 0.6 0.8 1.0 1.2

Ari tos ida t io i i 'Time LOG( l + h )

Figure 2. Changes in pine bark extract characteristics as a func- tion of the autoxidation time (using a starting pH of 11.5). Extract UV 215nrn (A) as a percentage of the unoxidized extract UV 215 nm, 1281x, 1 cm; tannin concentration (W) expressed as a per- centage of the unoxidized extract tannin concentration, 2088 mg COD/L; low MWtannin concentrution (0) expressed as a percentage of the unoxidized tannin concentration; extract color (A) measured as VIS 440 nm and expressed as a percentage of the maximum VIS 440-nm formation during autoxidation, 5.21x, 1 cm. The COD concentration of the extract was 6000 mg/L.

FIELD ET AL.: ANAEROBIC TREATMENT OF AUTOXIDIZED BARK EXTRACTS 249

Table I. The period averaged loading parameters and reactor performance during the continuous anaerobic treatment of pine bark extracts.

Influent concentration (g COD/L) HRT (h) COD loading (g COD/L . d) Substrate loading (g CODBD/L . d)

untrt auto PVP untrt auto PVP untrt auto PVP untrt auto PVP

6.5 7.1 4.3 10.6 10.3 10.3 14.8 16.7 10.0 6.0 6.8 6.1 B 17-21 10.2 9.9 7.8 6.4 5.9 5.7 41.4 44.2 34.7 16.4 16.9 20.5 C 22-56 3.9 3.8 2.8 6.7 4.8 5.6 16.4 20.1 13.1 9.2 10.4 9.1 D 57-75 4.8 4.7 3.4 5.7 5.5 7.1 20.1 20.4 12.5 12.0 11.3 9.5 E 76-93 5.9 5.7 4.2 5.1 4.8 5.2 32.1 30.8 22.7 17.0 15.8 18.3 F 94-119 10.1 9.5 6.6 6.2 6.8 7.8 42.1 43.3 23.5 24.4 25.0 19.7 G 120-133 8.2 14.0 4.0 7.0 14.8 7.2 34.6 26.9 19.8 16.6 14.1 16.8 H 134-155 12.9 13.9 9.8 7.0 8.4 7.3 46.8 41.3 33.3 25.9 26.0 30.0

Period Days

A 0-16

curred (Table 11). The pine bark tannins were highly detoxified by the autoxidation. Table I shows that the application of high organic loading rates of approxi- mately 41 g COD/L . d (26 g biodegradable COD/L . d) and high influent concentrations (14 g COD/L) was fea- sible by the end of the experiment with 98% removal of the biodegradable COD.

The preparation of tannin-free bark extract increased the reactor performance to the same extent as the autoxidation treatment. Although the COD elimina- tion and conversion of COD to methane was higher (Table I), this was only due to the fact that the poorly biodegradable tannin COD fraction was removed from the influent by the PVP treatment. The quantity of methane produced and COD eliminated was the same, and both columns provided the same percentage of biodegradable substrate elimination (Table 11).

Although strong inhibition was evident in the column fed untreated extract, some adaptation to the toxic tan- nins had occurred. In periods B and C, from 42 to 67% of the biodegradable COD was present as VFA in the reactor effluent. The VFA was composed of about 70% acetate and 30% propionate (on a COD basis), indicating that in this initial period both acetate and

propionate degradations were inhibited. As the reactor performance improved in periods D to H, the aceto- clastic methanogens were no longer severely inhibited. However, the effluent VFA still accounted for 21-26% of the biodegradable COD and the VFA was composed mostly of propionate (from 70 to 90% of the VFA on a COD basis). Therefore, the toxicity of propionate deg- radation was persistent throughout the entire course of the experiment.

Sludge A ctivit y

The sludge obtained from each of the reactors at the end of the continuous experiment was assayed for specific methanogenic activity (Table 111). The specific methanogenic activity of the tannin-free extract sludge was the highest. The untreated extract sludge had 80% less activity based on the results from the first VFA feeding. This indicates the inhibitory effect of the bark tannins on the biomass activity. Table IV indicates that particularly propionate degradation was highly inhibited since almost no propionate was utilized by this sludge even after 2 weeks. The increase in activity that did oc- cur in the second VFA feeding can be attributed to the

Table 11. Average organic loading, treatment efficiency, and biodegradation of pine bark extract during the continuous experiments.

Experimental averageb

Parametera

COD load CODBD load ECOD M VFA Cells A BD

ECODBD MBD Y

Units

g COD/L . d g CODBD/L . d

% infl. COD % infl. COD % infl. COD % infl. COD % infl. COD % infl. COD

% infl. CODBD % infl. CODBO G C O D 4 g CODBD

untrt

29.1 15.5

37.4 18.8 15.8 18.4 34.7 53.1

69.9 35.4 0.33

auto PVP

29.1 20.1 15.4 15.3

51.8 74.3 39.5 56.5

1.4 2.4 12.3 16.4 40.7 59.0 52.9 76.3

97.3 97.0 74.7 74.0 0.22 0.21

auto + Ca

16.2 11.5

66.2 47.8 4.8

18.4 52.6 71.1

93.1 67.2 0.26

a Parameters defined in Materials and Methods. Abbreviations: untrt, untreated extract; auto, autoxidized (pH 11.5, 16 h) extract; PVP,

tannin-free extract (i.e., treated with PVP); auto + Ca, autoxidized extract (pH 11.5, 1 h) followed by calcium precipitation.

250 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 37. FEBRUARY 1991

Table I. (continued)

ECOD (% infl. COD) A4 (% infl. COD) Effluent VFA (mg COD/L) pH (median value)

Period Days untrt auto PVP untrt auto PVP

A 0-16 B 17-21 C 22-56 D 57-75 E 7 6 9 3 F 94-119 G 120-133 H 134-155

29.9 39.3 60.1 14.9 41.2 60.4 32.9 52.3 66.8 45.3 54.0 73.7 38.1 50.3 76.6 43.4 53.7 79.2 38.0 49.8 83.1 42.0 60.0 85.9

25.6 34.4 51.1 4.6 28.6 35.5

13.0 38.4 52.7 21.4 41.2 56.9 15.3 35.9 46.2 22.6 43.7 70.1 26.5 49.6 68.7 20.3 38.8 57.5

untrt auto PVP

724 93 40 2987 218 520 936 20 100 613 26 62 791 35 123

1385 325 127 799 380 47

1759 193 155

untrt auto PVP

6.8 7.0 7.2 6.3 7.4 7.3 7.2 7.2 7.2 7.3 7.3 7.2 7.1 7.3 7.1 1.1 7.1 6.9 6.9 7.2 6.8 6.8 6.9 6.8

(PVP treated).

Table 111. Sludge concentration and sludge activity at the end of the continuous experiments fed pine bark extract.

Sludge activityb (g COD/g VSS . d) Sludge concentration

in reactor Sludge" (g VSS/L) First Second

untrt auto PVP

70 68 60

0.111 0.350 0.370 0.561 0.559 0.797

a The experiment was started with 20 g VSS/L granular sludge, having an activity of 0.62 g COD/g VSS . d. Abbreviations: untrt, sludge from column fed untreated extract; auto, sludge from column fed autoxidized extract; PVP, sludge from column fed tannin-free (PVP treated) extract.

First and second VFA feedings.

onset of butyrate degradation (Table IV). The autoxi- dized extract sludge had a high activity, but it was some- what lower than that of the tannin-free extract sludge. One possible explanation for this is that the autoxida- tion pretreatment was less extensive with increasingly concentrated extracts. Thus, when high influent con- centrations were utilized at the end of the continuous experiment, some inhibitory effects could have been imparted if the extracts were not completely detoxified.

Biodegradability of Extracts

The conversion of influent COD to potential substrates for methane production can be determined from the sum of methane produced and VFA left unutilized (i.e., acidification). The experimental averaged acidification of the untreated pine bark extract was 35% (Table 11). The average acidification of 41% from the autoxidized extract indicates that the detoxification treatment did not lower but rather increased the potential substrate for methane. The acidification of the tannin-free extract averaged 59% of the influent COD, but if expressed as a percentage of the whole extract COD (before PVP treatment), it was equal to 38%.

Biodegradability indicates the potential COD elimina- tion by additionally accounting for the COD converted to cells. The experiment averaged biodegradability of the extract COD was the same (53%) for the untreated and autoxidized extracts (Table 11). These were the

Table IV. Acetate (CZ), propionate (C3), and butyrate (C4) concen- trations in the media at the end of the first and second feedings of the sludge activity test conducted for the sludge recovered at the end of the continuous pine bark extract experimentP

End first feeding End second feeding (mg COD/L) (mg COD&)

Sludge CZ c3 c4 c2 c3 c4 ~~ ~ ~ ~~

untrt 171 1389 1537 134 1436 99 auto 243 252 28 25 33 0 PVP 94 147 11 32 59 0

Abbreviations: untrt, column fed untreated extract; auto, column fed autoxidized extract; PVP, column fed tannin-free extract

FIELD ET AL.: ANAEROBIC TREATMENT OF AUTOXIDIZED BARK EXTRACTS 251

a Each feeding lasted 7 days, the starting concentration of Cz, C3, and C4 at each feeding was 979, 1387, and 1673 mg COD/L. Abbre- viations are defined in the footnote of Table 111.

same since the cell yield was distinctly higher in the column treating untreated extract. The abnormally high cell yield in this column may have resulted from adsorp- tion of some COD components on the sludge, which in turn is accounted for as cells based on the method in which the cell production is estimated. The biodegrad- ability of the tannin-free extract averaged 76%; how- ever, the high value is merely an artifact of removing the poorly biodegradable tannins by PVP treatment. If expressed as a percentage of the whole extract COD (i-e., prior to PVP treatment), the tannin-free extract biodegradability was only 49%. Since this value is only slightly less than the biodegradability found for the tan- nin containing extract, the tannin fraction is thus not an important source of biodegradable matter under the conditions prevailing during continuous anaerobic treatment.

The acidification and biodegradability of the extract by the tannin-free and autoxidized extract fed columns increased during the course of the experiment (Fig. 3). In contrast, the acidification of the untreated extract remained largely unchanged, and a lower increase in the biodegradability was observed. These results in- dicate that significant adaptation to the acidification and biodegradability of certain components in the bark extracts occurred in the uninhibited columns, while a similar level of adaptation was not evident in the inhib- ited column.

The reduced acidification of untreated compared to detoxified extracts might have resulted from the incom-

A B C D E F G H Time Period

Biodeeradation' r 0 Cells = VFA EZ3 Methane Acidification - c t

Figure 3. Period averaged acidification and biodegradation of the influent COD during the continuous anaerobic treatment of unoxi- dized, autoxidized, and PVP treated pine bark extracts: acidifica- tion = percentage of conversion of the COD supplied to VFA and CHI; biodegradation = acidification + cells.

plete biodegradation of phenolic degradation intermedi- ates. Their degradation appeared to be disturbed when high VFA concentrations were present in the reactor as a result of inhibited methane production. Figure 4 illustrates that these intermediates (phenol, p-cresol, and carbox ycyclohexane) were linearly related to the logarithm of the effluent VFA concentration during the continuous column experiments. On one occasion (day 126) other intermediates of phenolic compound degradation were investigated; these included truns- cinnamate, phenylacetate, and 3-phenylpropionate. Of the three, only 3-phenylpropionate was found (116- 137 mg COD/L) in the columns which had significant

LOG(VFA mg C O D / L ) 0 Col 1 n Col 2 0 Col 3

Figure 4. Sum of the effluent phenolic intermediate concentra- tion (phenol,p-cresol, and carboxycyclohexane) as a function of the effluent VFA concentration during the continuous anaerobic treat- ment of untreated (Col l ) , tannin-free (Col 2), and autoxidized (Col 3) extracts. The results were obtained between day 94 and day 155 of the experiment (Rz = 0.72).

levels of effluent VFA (600-800 mg COD/L) on the day of sampling.

During periods C to F of the continuous experiment, the elimination of UV absorbance and tannins by an- aerobic treatment was measured (Fig. 5). About half of the UV absorbance and 60% of the tannins were elimi- nated from the untreated extract. Very low levels of UV absorbance were eliminated from the autoxidized ex- tract which was expected assuming that the autoxidized tannin products whould have an increased recalcitrance. The elimination of UV absorbance which did occur cor- responded to the quantity of UV absorbance units re- moved from the tannin-free extract. No elimination of the autoxidized tannins was evident. The HPLC chro- matograms of the influents and effluents, sampled on day 106, illustrate the partial elimination and biotrans- formation of the UV absorbing compounds in the un- treated extracts by anaerobic treatment (Fig. 6). In contrast, the high MW autoxidation products were not altered by anaerobic treatment. Although a few of the UV absorbing compounds were degraded in the autoxi- dized extract, these were the nontannic phenolic com- pounds that were not altered by the autoxidation.

The elimination of tannins from the untreated ex- tracts must have resulted from biological transforma- tions to phenolic intermediates of lower UV absorbance or to compounds not having UV absorbance. In any case, their elimination did not coincide with any in- crease in the acidification or biodegradation of the ex- tract beyond that observed in the autoxidized extract.

At the end of the experiment, sludge from the column experiments was recovered to study the long term anaerobic biodegradability of bark extract in batch assays. The sludge recovered from the column fed, tannin-free extract was able to biodegrade (Fig. 7) more bark extract COD than the sludge obtained from the column fed, untreated extract. The same trend between the sludges was also observed with respect to the UV elimination (Fig. 8) and tannin elimination (Fig. 9).

- 160 E UNTR'I' AUTO PVP

U

i n f l e f f l infl e f f l i n f l e f f l n tannin fraction

tannin-free fraction total U V

Figure 5. Average influent (infl) and effluent (effl) UV 215 nm absorbance of the tannin and tannin-free fractions. The UV mea- surements were conducted in periods C to F of the continuous experiments fed untreated (UNTRT), autoxidized (AUTO), and tannin-free (PVP) pine bark extracts.

252 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 37, FEBRUARY 1991

unt rea ted ; ext rac t i

4 8 1 2 1 6 2 0 2 4 2 6 32 3 6 4 0

tannin- f ree C N a ext rac t - ? I

A- . . . .

- 4 8 I 2 1 6 2 0 2 4 2 8 3 2 3 6 4 0 - 2 ext rac t

autoxidized

0 47 N

4 8 1.?1620242832 3 6 4 0 Retention Time (minu tes )

Figure 6. The HPLC chromatograms of the influents (-) and the effluents (... .....), sampled on day 106, of the continuous anaerobic treatment of untreated, autoxidized, and tannin-free pine bark extracts.

100 7,

60

40

80

60

40

20

0 10 20 30 0

D ay Figure 7. Biodegradability of untreated and tannin-free pine bark extract during batch anaerobic digestion by sludge recovered at the end of the continuous experiments fed untreated (sludge 1) and tannin-free bark (sludge 2) extracts. Sludge 1 and 2 concentra- tions were 1.25 and 1.09 g VSS/L, respectively. The untreated and tannin-free extract concentrations were 3.59 and 2.38 g COD/L at the start of the assay. Untreated extract fed to sludge 1 (+); tannin-free extract fed to sludge 1 (...Ei...); untreated extract fed to sludge 2 (4); tannin-free extract fed to sludge 2 (.. .O...) .

Therefore, adapting the sludge to phenolic degradation in the absence of tannins was better for the degrada- tion of the phenolic compounds in the whole bark ex- tract (including tannins). This was probably due to the higher methanogenic activity of sludge cultivated on the tannin-free fraction, which served to maintain a lower VFA concentration in the media and thereby favor the degradation of phenolic compounds.

Based on the long term batch digestion experiments with the sludge cultivated on the tannin-free fraction of the extract, the ultimate COD biodegradability of the untreated bark extract was 85% after 30 days. The

0 10 20 30 40 50 "

Days Figure 8. The UV absorbance (215 nm) of the assay media during batch anaerobic digestion of untreated and tannin-free pine bark extracts by sludge recovered at the end of the continuous experi- ments fed untreated (sludge 1) and tannin-free (sludge 2) extracts. Details of this experiment are described in Figure 7. The initial UV absorbance of the untreated and tannin-free extracts were 101 and 281x, 1 cm, respectively. Untreated extract fed to sludge 1 (+); tannin-free extract fed to sludge 1 (...tT!...); untreated ex- tract fed to sludge 2 (+); tannin-free extract fed to sludge 2 ( . . . Q . . . ) .

I 0 0

c 2 c 8 0 c z $ 2 60

" c 2 40 - 0 c c - 0

6\"

4

4

" 20

0- 0 10 20 30 40 50

Days

Figure 9. Tannin concentration based on COD and UV (215 nm) measurements of the assay media during batch anaerobic digestion of untreated pine bark extracts by sludge recovered at the end of the continuous experiments fed untreated (sludge 1) and tannin- free (sludge 2) extracts. Details of this experiment are described in Figure 7. The initial tannin concetnration was 1.36 g COD/L or 75 lx, 1 cm, 215 nm absorbance units. Tannin COD sludge 1 (+); tannin UV sludge 1 (...El...); tannin COD sludge 2 (+); tannin UV sludge 2 (...@...).

elimination of extract UV and tannins reached 80 and 75%, respectively. The untreated pine bark extract con- tains more biodegradable substrate than the tannin-free extract (Fig. 7). This would indicate that at least some of the unoxidized tannins are potentially biodegradable to CH4, VFA, and cells if they have a sufficiently long residence time in anaerobic conditions.

Continuous Treatment of Precipitated Extracts

Treatment Performance and Biodegradability

The calcium precipitated autoxidized extract was treated with a higher COD elimination efficiency and acidification and was biodegraded to a greater extent (Tables I1 and V) than the autoxidized extract. This was due to the precipitation of poorly degradable COD prior to the anaerobic treatment. However, the parameters

FIELD ET AL.: ANAEROBIC TREATMENT OF AUTOXIDIZED BARK EXTRACTS 253

Table V. Period averaged loading parameters and reactor performance during the continuous anaerobic treatment of calcium precipitated autoxidized pine bark extracts.

Influent concentration COD loading Substrate loading Period Days (g COWL) HRT (h) (g COD/L . d) (g CODBD/L ' d)

A 0-23 3.7 5.3 17.9 12.2 B 24-54 2.8 6.8 12.5 9.3 c 55-67 2.6 2.8 22.9 16.4

were comparable to the those of the column fed autoxi- dized extract if they are expressed as a percentage of the autoxidized extract COD before calcium precipita- tion. Therefore, calcium pretreatments of autoxidized extracts can be applied to physically remove some of the poorly biodegradable COD without decreasing the COD which is potentially convertible by biological treatment. The combined precipitation and anaerobic treatment of the autoxidized extract reduced the COD by 79% (27 and 52% of the original COD by Ca2+ and anaerobic treatment, respectively). Thus approximately 56% of the residual COD in anaerobically treated autoxidized bark extract could be eliminated by precipitation.

DISCUSSION

Effect of Autoxidation on Treatment Efficiency

The debarking wastewaters of the forest industry are troublesome for anaerobic treatment due to their high content of tannins inhibitory to methane production. An autoxidation pretreatment can be applied to poly- merize the tannins to nontoxic colored compounds. In this study, the anaerobic treatment of aqueous pine bark extracts was improved by the autoxidation pre- treatment. During anaerobic treatment of unoxidized extracts, incomplete conversion of the biodegradable substrates to methane was observed and high concen- trations of VFA were present in the effluent. Anaerobic treatment of autoxidized extracts provided twice as much methane production. The decreased content of inhibitory compounds in the autoxidized extracts was evident from the low effluent VFA concentrations dur- ing the continuous experiment as well as from the high activity of the sludge recovered at the end of the experi- ment. The autoxidation pretreatment permitted the anaerobic treatment of pine bark extracts at high influ- ent concentrations (14 g COD/L) and loading rates of 41 g COD/L * d (26 g biodegradable COD/L - d) with 98% removal of the biodegradable COD.

The column treating pine bark extract which was not detoxified acclimatized considerably to the toxicity of the tannins. Initially, almost no methanogenic activity was evident; however, during the course of the experi- ment the methane production increased. Despite the acclimization, the methanogenic activity never reached the level observed in the uninhibited columns. Since the VFA that accumulated in the effluent was pre- dominantly propionic acid, we can conclude that the acetoclastic methanogens acclimatized to the toxicity,

whereas the toxicity to trophic groups involved directly or indirectly in propionic acid degradation (i.e., either the acetogenic or hydrogenotrophic bacteria) was per- sistent. This is in agreement with the data of Blum et al.," which indicate that the methanogenic toxicity of phenolic compounds is generally more severe for pro- pionic as compared to acetic acid degradation.

Effect of Autoxidation on Anaerobic Nodegradability

The unoxidized tannins, which are responsible for 37% of the extract COD, were eliminated 60% by continuous anaerobic treatment; whereas, the autoxidized products of these tannins were not eliminated at all. The high MW products of autoxidation reactions increase the recalcitrance of the tannin fraction. Poor anaerobic biodegradability of high MW phenolic compounds has frequently been observed in studies with lignin and peat."-" However, the formation of high MW tannin and humic products by autoxidation did not correspond to a decrease in the whole extract biodegradability. The extract biodegradability during anaerobic treatment was 53% for both unoxidized and autoxidized influents. Al- though the tannins of the unoxidized extract were clearly less recalcitrant than the polymerized products, their elimination was due to their partial degradation to phenolic and nonphenolic intermediates as opposed to complete fermentation to VFA and CHI. The inhibition of the methanogenic consortia caused VFA from the acidification of the biodegradable fraction to accumu- late. The high VFA concentration (and presumably also elevated H2) provide conditions which are known to be thermodynamic ally unfavorable for further biodegrada- tion of the phenolic intermediates.I6 Several studies have previously demonstrated that simple phenolic com- pounds or their breakdown intermediates are poorly de- graded if high concentrations of VFA or H2 are present in the media.9,'6-'8 Therefore, methanogenic and aceto- genic bacteria play a vital role in the anaerobic degrada- tion of phenolic compounds by maintaining low VFA and H2 concentrations in the media.

The recalcitrant COD expected in the effluents of anaerobic processes treating autoxidized bark waste- water can be effectively reduced by calcium precipita- tion prior to anaerobic treatment. The precipitation is also effective in reducing the color of the autoxidized extracts."

The authors wish to express their gratitude to PAQUES B.V. (Balk, The Netherlands) for its financial support.

254 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 37, FEBRUARY 1991

Table V. (continued)

ECOD Effluent VFA PH (% infl. COD) M (% infl. COD) (mg C O W ) (median value)

60.3 39.8 322 6.9 70.0 51.1 60 7.1 67.2 53.7 120 6.9

References

1. S. Rekunen, 0. Kallio. T. Nystroem, and 0. Oivanen, Wat. Sci. Technol., 17, 133 (1985).

2. S. Rekunen, PIRA, Paper and Board Division Key-Event, Cost Effective Treatment of Papermill Effluents Using Anaerobic Tech- nologies, Leatherhead, England, January 14-15, 1986.

3. P. K. Latola, Wat. Sci. Technol., 17, 233 (1985). 4. J. A. Field, M. J. H. Leyendeckers, R. Sierra-Alvarez, G. Let-

tinga, and L. H. A. Habets, Wat. Sci. Technol., 20,(1), 219 (1988). 5. J. A. Field, S. Kortekaas, and G. Lettinga, Biol. Wst., 29, 241

(1989). 6. J.A. Field, G. Lettinga, and L. H.A. Habets, 1 Chem. Tech.

Biotechnol. 49, 15 (1990). 7. J. A. Field, G. Lettinga, and L. H.A. Habets, 1 Ferment. Bio-

eng., 69, 148 (1990). 8. APHA, Standard Methods for the Examination of Water and

Wastewater, 16th ed. (American Public Health Association, American Water Works Association and Water Pollution Con- trol Federation, Washington, DC, 1985).

9. J.A. Field, G. Lettinga, and M. Geurts, Biol. Wast., 21, 37

10. J.A. Field, R. Sierra-Alverez, G. Lettinga, and L. H.A. Ha-

11. J.W. Blum, R. Hergenroeder, G. F. Parkin, and R. E. Speece,

12. W. F. Owen, D.C. Stuckey, J. B. Healy, Jr., L.Y. Young, and

13. J. G. Zeikus, A. L. Wellstein, and T. K. Kirk, FEMS Microbiol.

14. R. Benner and R. E. Hodson, Appl. Environ. Microbiol., 50,971

15. P. J. Colberg and L.Y. Young, Appl. Environ. Microbiol., 49, 345

16. P.M. Fedorak, D. J. Roberts, and S. E. Hrudey, War. Res., 20,

17. A. G. Lane, 1 Chem. Tech. Biotechnol., 30, 345 (1980). 18. T. Kobayashi, T. Hashinaga, E. Mikami, and T. Suzuki, Wat.

(1987).

bets, 1 Chem. Tech. Biotechnol. 49, 35 (1990).

1 Wat. Pollut. Ctrl. Fed., 58(2), 122 (1986).

P. L. McCarty, Wat. Res., 13, 485 (1979).

Lett., 15, 193 (1982).

(1985).

(1985).

1315 (1986).

Sci. Technol., 21, 55 (1989).

FIELD ET AL.: ANAEROBIC TREATMENT OF AUTOXIDEED BARK EXTRACTS 255