- Email: [email protected]
High-rate anaerobic degradation of 5 and 6 carbon sugars under thermophilic and mesophilic conditions
Bioresource Technology 101 (2010) 3925–3930
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
High-rate anaerobic degradation of 5 and 6 carbon sugars under thermophilic and mesophilic conditions C. Forbes *, D. Hughes, J. Fox, P. Ryan, E. Colleran Environmental Microbiology Research Unit, Department of Microbiology, National University of Ireland, Galway, Ireland
a r t i c l e
i n f o
Article history: Received 8 September 2009 Received in revised form 5 January 2010 Accepted 9 January 2010 Available online 4 February 2010 Keywords: Thermophilic AD Pre-hydrolysis Pectin Cellulose OFMSW
a b s t r a c t In this research paper, a comparison between thermophilic and mesophilic anaerobic degradation of a variety of the simple sugar components of carbohydrate rich biomass is presented. In order to investigate the degradability of these basic sugars, three synthetic sugar based influents were supplied to two high rate upflow anaerobic hybrid reactors (UAHR) operated at 37 °C (R1) and 55 °C (R2). These influent streams were: D-glucose/sucrose; L-arabinose/D-xylose and L-rhamnose/D-galacturonic acid. The reactors were challenged in terms of influent composition rather than loading rate and were therefore operated at a maximum volumetric loading rate (VLR) of 4.5 gCOD l 1 d 1 during stable reactor performance. It was found that a switch from a D-glucose/sucrose synthetic influent to an influent composed of L-arabinose/Dxylose resulted in failure of the mesophilic reactor while the thermophilic UAHR was able to tolerate the change of sugar influent at an unchanged VLR of 4.5 gCOD l 1 d 1. A subsequent phasing-in approach was used to introduce new sugar influent streams and proved highly successful. The physiology of the biomass was assessed and it was noted that thermophilic anaerobic digestion (AD) involved the formation of acetate and H2, implying the involvement of homoacetogenic bacteria, while mesophilic AD proceeded via the formation of other intermediates. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction In unprocessed plant materials, carbohydrate may account for up to 60–95% of total dry weight. Plant-derived raw materials are utilized widely in the food, beverage, paper and packaging industries and thus constitute a significant portion of the waste from these industries. The waste generated from the food industry ranges from relatively unprocessed food materials (surplus and spoiled or contaminated stock) to the by-products of food processing activities in an industrial context. In recent years, a number of novel approaches for the treatment of these wastes with concomitant energy recycling have been investigated, including the production of bioethanol and biodiesel as well as biodegradable plastics (Callaghan et al., 1999). From both an environmental and economic perspective, however, anaerobic treatment is arguably the best option for remediation of these wastes, simultaneously reducing their pollution potential while producing a value added end product in the form of a renewable biofuel, methane gas. There are limitations, however, to direct anaerobic treatment of carbohydrate rich materials which extend to the organic fraction of
* Corresponding author. Present address: Environmental Research Institute, University College Cork, Lee Road, Ireland. Tel.: +353 214901975; fax: +353 214901932. E-mail address: [email protected] (C. Forbes). 0960-8524/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2010.01.019
municipal solid wastes (OFMSW), food wastes and plant materials used as energy crops. The effective and high-rate anaerobic treatment of solid wastes is impeded by the comparatively lengthy time taken to hydrolyse complex polymers such as cellulose, pectin and hemi-celluloses, which comprise the major components of nonanimal food wastes, to their constituent monomeric sugars (Noike et al., 1985). Additionally, the biodegradability of solid wastes is highly dependent on the composition of cellulose, hemicellulose and lignin which are characteristic to the different components of OFMSW (Hartmann and Ahring, 2006). Successful treatment of these wastes in single stage reactors involves the operation of Continuously Stirred Tank Reactors (CSTRs) with retention times in the region of 10–30 days. A commonly employed method to circumvent these problems involves pre-hydrolysis of these materials to yield more readily degradable short chain polymers and monomers prior to anaerobic treatment. By using pre-treatment strategies, improved hydrolysis of complex polysaccharide containing wastes can be achieved. As a result of this approach, the retention times of waste in anaerobic digesters can be considerably shortened, and higher methane yields can be produced. A number of pre-hydrolysis strategies have been investigated in order to eliminate the initial rate limiting hydrolysis step. Pre-treatment methods include mechanical, chemical, and biological based approaches as well as various combinations of these. Those currently employed have been reviewed extensively elsewhere (Mata-Alvarez et al., 2000;
3926
C. Forbes et al. / Bioresource Technology 101 (2010) 3925–3930
Hartmann and Ahring, 2006). Another major limitation in the use of single stage anaerobic digestion of OFMSW and highly biodegradable food wastes is the rapid acidification of these wastes resulting in pH decrease in the reactor, inhibiting the activity of methanogenic Archaea. These instability factors preclude the use of high-rate anaerobic digesters with retention times measured in hours rather than days (Mata-Alvarez et al., 2000). In the current trial, three distinct synthetic sugar influents were devised in order to assess the anaerobic biodegradability of prehydrolysed carbohydrate wastes. Each influent stream consisted of two simple sugars, which were applied to the UAHRs in tandem and designed to represent the carbohydrate macromolecule components of OFMSW. The first waste stream supplied to the reactors consisted of D-glucose and sucrose, both of which have been demonstrated to be readily degraded anaerobically (Forbes et al., 2009). Most carbohydrates contain D-glucose, often as their sole building block, as in the examples of starch, glycogen, and most importantly, cellulose. Sucrose is a disaccharide molecule, the structure of which inhibits further binding to other saccharide moieties. It is normally extracted from either sugar cane or beet and is commonly used as a sweetener and extensively employed in the baking industry and in food preservation. The second waste stream supplied to the reactors consisted of L-arabinose and D-xylose. Both are pentose monosaccharides and found as part of more complex polymers, such as xylans, hemicelluloses and pectins. The third influent stream was comprised L-rhamnose and D-galacturonic acid, the major monosaccharide components of pectin (Grohmann and Bothast, 1994; Renard et al., 1995). It was anticipated that the reactor sludges would adapt readily to these three synthetic influents as these reactors had previously been employed in a 650 days trial during which both had treated a variety of more complex hydrolysate influents (Forbes et al., 2009).
2. Methods 2.1. Reactor design and operation Two laboratory scale UAHRs, both with an active volume of 3.9 l, were used in the current trial. These reactors had previously been inoculated and operated at 37 °C (R1) and 55 °C (R2) for 670 days as described by Forbes et al. (2009), followed by a period of 200 days of application of D-glucose and sucrose. On day 1 of the current trial, both reactors were supplied with the first trial influent, which consisted of D-glucose and sucrose on an equal COD ratio to a total VLR of 4.5 gCOD l 1 d 1. This was achieved by operating the reactors at a 2 days hydraulic retention time (HRT), with an influent concentration of 9 gCOD l 1. The reactors were maintained on this influent until day 91 when the influent compo-
sition was changed to an equal ratio, on a COD basis, of L-arabinose and D-xylose to a total of 9 gCOD l 1. Feeding with L-arabinose and D-xylose was continued at the same rate until day 103. Due to instability, the feeding was then stopped for 6 days (Table 1; Period III). During this period, the reactors were unfed and only 10 ml of buffer and associated nutrients were added on occasion to allow removal of effluent samples for analysis. L-arabinose and D-xylose were reintroduced on day 110, at a lower combined influent COD concentration of 4.5 g l 1, at a VLR of 2.25 gCOD l 1d 1. On day 126, the L-arabinose and D-xylose influent was replaced by D-glucose and sucrose at a VLR of 4.5 gCOD l 1 d 1 (Period V). This VLR was maintained until day 201. L-arabinose and D-xylose were then phased back into the influent at a combined COD of 3 gCOD l 1 on day 202. The VLR of 4.5 gCOD l 1d 1 was maintained by inclusion of D-glucose and sucrose at an influent COD of 6 gCOD l 1 during the period (Period VI). On day 216, the influent L-arabinose and D-xylose was increased to a combined gCOD l 1 of 6, while correspondingly reducing the D-glucose/sucrose in the influent to 3 gCOD l 1 (Table 1; Period VII). On day 264, D-glucose and sucrose were removed totally from the influent and the combined COD concentration of L-arabinose and D-xylose was increased to 9gCOD l 1 d 1 (Table 1; Period VIII). L-rhamnose and D-galacturonic acid were phased into the reactor influent, beginning on day 342. This was achieved by reducing the combined COD of L-arabinose and D-xylose to 6 gCOD l 1 and introducing L-rhamnose and D-galacturonic acid at 3 gCOD l 1 (Table 1; Period IX). The influent concentration of L-rhamnose and Dgalacturonic was increased to 6 gCOD l 1 on day 363 (Table 1; Period X). L-arabinose and D-xylose were withdrawn from the reactor influents on day 400 with a concomitant increase in the combined influent of L-rhamnose and D-galacturonic acid to 9 gCOD l 1. Sole feeding of the latter two sugars was continued at a HRT of 2 days and a VLR of 4.5 gCOD l 1 d 1 until day 430, at which point the trial concluded. Throughout the duration of the trial, the COD:N:P ratio of all influents was maintained at 1000:5:0.5 by supplementation with NH4Cl and KH2PO4 to the required concentrations. Buffering was carried out by addition of NaHCO3 (12 g l 1) and influent waste streams were supplemented with micronutrients (1 ml l 1), as recommended by Shelton and Tiedje (1984).
2.2. Analytical techniques Samples of reactor influent/effluent were routinely analysed for COD (g l 1), VFAs (mg l 1) and pH as previously described (Forbes et al., 2009). Biogas was sampled for CH4 determination according to Standard Methods (APHA, 2001).
Table 1 Influent regime showing the contribution of each sugar to each influent stream, on a COD basis (gCOD l
).
Phase
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
Start of phase (day of trial)
1 4.5
91 –
103 –
110 –
126 4.5
202 3
216 1.5
264 –
342 –
363 –
400 –
L-Arabinose
4.5 –
– 4.5
– –
– 2.25
4.5 –
3 1.5
1.5 3
– 4.5
– 3
– 1.5
– –
D-Xylose
–
4.5
–
2.25
–
1.5
3
4.5
3
1.5
–
L-Rhamnose
–
–
–
–
–
–
–
–
1.5
3
4.5
D-Galacturonic
–
–
–
–
–
–
–
–
1.5
3
4.5
9 2 4.5
9 2 4.5
0 390* 0
4.5 2 2.25
9 2 4.5
9 2 4.5
9 2 4.5
9 2 4.5
9 2 4.5
9 2 4.5
9 2 4.5
D-Glucose Sucrose
acid Total influent COD (g l HRT (days) VLR (gCOD l 1 d 1) *
1
1
)
Based on 10 mls of buffer per day being supplied to the reactors in order to obtain an effluent sample.
C. Forbes et al. / Bioresource Technology 101 (2010) 3925–3930
3927
2.3. Specific substrate utilisation (SSU) tests SSU tests were performed on days 194, 305 and 430 of the trial, corresponding with times at which the reactors were fully adapted to each synthetic influent. SSU tests assessed the activity of fermentative species in the reactor sludges to degrade the various sugars applied during the trial. Tests were carried out at 37 °C and 55 °C in 60 ml serum vials containing sludge (2–5 gVSS l 1), anaerobic buffer and one of each of the trial sugars (2.5 gCOD l 1) to a total of 30 ml. Samples were removed intermittently from test vials, were centrifuged at 10,000g for 10 min and the supernatant analysed spectrophotometrically using the Dubois method for the quantification of carbohydrates (Dubois et al., 1956). Briefly, one volume of sample was mixed with one volume of 5% phenol solution, and the mixture cooled at 4 °C for 10 min. Five volumes of concentrated H2SO4 were rapidly added to the pre-cooled mixture, with continuous mixing, followed by boiling for 5 min. Samples were then cooled and incubated at room temperature for 30 min prior to determining the absorbance at 490 nm. In vial carbohydrate concentrations were plotted against time to give a degradation profile for each sugar. Specific substrate utilisation rates were calculated by using the period of most rapid carbohydrate degradation from each profile, and expressing this as the rate of carbohydrate depletion per gram of volatile suspended solids (mg carbohydrate g VSS 1 hr 1). 2.4. Specific methanogenic activity (SMA) tests Specific methanogenic activity tests were also performed on days 194, 305 and 450 of the trial. SMA measurements were carried out using the pressure transducer method developed by Colleran and Pistilli (1994). Briefly, the procedure involved the measurement of the biogas pressure increase developing in sealed vials fed with the non-gaseous substrates ethanol (30 mM), propionate (30 mM), butyrate (15 mM) and acetate (30 mM), or of the decrease in vials pressurised to 1 atm pressure with the gaseous substrate H2/CO2 (80:20). Tests were carried out in triplicate at 37 °C and 55 °C in 20 ml serum vials containing sludge (2–5 g VSS l 1), anaerobic buffer (Hungate, 1966) and one of each of the test substrates. Appropriate temperature conversion factors were used to record the final results as ml CH4 gVSS 1 day 1 at standard temperature and pressure (STP).
Fig. 1. Percentage soluble COD removal by R1 and R2 during the trial. Different R1 R2). operational periods are denoted by arrows (Table 1). (
methane production by day 100 (Fig. 1). There was a parallel increase in effluent VFA concentrations (Fig. 2A). By day 102, the mesophilic effluent VFA concentrations were 1440, 3310 and 3490 mg l 1 for acetate, propionate and butyrate, respectively (Fig. 2A). By contrast, the thermophilic effluent VFA concentrations were considerably lower, with levels of 676 and 1709 mg l 1 for acetate and propionate, respectively, and with negligible levels of butyrate (Fig. 2B). These high concentrations of short chain fatty acids resulted in pH levels of lower than 5.5 in both reactors by day 100. It is noteworthy that Kim et al. (2002) defined reactor failure in terms of an effluent discharge pH lower than 5.5.
3. Results and discussion 3.1. Reactor operation Prior to the start of the current study, both reactors had been operated for 900 days at 37 °C (R1) and 55 °C (R2). For the 200 days immediately preceding the trial, both reactors operated in a stable fashion on an influent composed of D-glucose/sucrose. The current trial initially utilized the same substrates as previously at a VLR of 4.5 gCOD l 1 d 1 and a HRT of 2 days for a 90 day period (Table 1). Despite fluctuations it was clear that both reactors could achieve >95% COD removal under the conditions imposed (Fig. 1). On day 91, the influent was changed to L-arabinose and D-xylose (on an equal COD basis), while maintaining the VLR and HRT (Table 1). This resulted in an immediate and dramatic decrease in COD removal efficiencies (Fig. 1). The methane content of the produced biogas also decreased to 14% from the mesophilic reactor and 12% from the thermophilic reactor. Within five HRTs the performance of the thermophilic reactor had deteriorated to 35% COD removal, with methane comprising only 16% of the biogas produced. The response of the mesophilic reactor was even more severe, with total failure in terms of COD removal efficiency and
Fig. 2. Effluent VFA levels (mg l 1) from the mesophilic reactor, R1 (A), and the thermophilic reactor, R2 (B) throughout the trial period. Different operational Acetate propionate butyrate. periods are denoted by arrows (Table 1).
3928
C. Forbes et al. / Bioresource Technology 101 (2010) 3925–3930
The influent supply to both reactors was discontinued on day 103. An immediate reduction in effluent VFA concentrations was noted (Fig. 2), together with a decrease in effluent COD (Fig. 1). This suggested that the rapid acidification following introduction of Larabinose and D-xylose had not caused complete inhibition of sludge syntrophic and archaeal populations. L-arabinose and D-xylose were reintroduced on day 110 (Period IV, Table 1) at a lower COD concentration of 2.25 gCOD l 1 for both sugars. This resulted in a VLR of 2.25 gCOD l 1 d 1 at a 2 days HRT. Although the performance of the mesophilic reactor deteriorated, the thermophilic reactor performance showed greater stability, with 50% COD removal efficiency on day 126 (Fig. 1) and a biogas methane content of 30%. The effluent VFA profiles (Fig. 2) confirm the better performance of the thermophilic reactor during this period. To avoid total failure of the mesophilic reactor, on day 126 the influent to both reactors was changed to the D-glucose/sucrose influent at the previous VLR of 4.5 gCOD l 1 d 1 while maintaining the HRT at 2 days (Table 1). As a result, an immediate improvement was observed in the performance of both reactors. By day 202, both reactors had fully recovered and regularly exhibited >90% COD removal efficiencies. L-arabinose and D-xylose were reintroduced to the reactors on day 202 (influent concentration of 1.5 gCOD l 1 for each sugar) while maintaining a reduced combined D-glucose/sucrose influent of 6 gCOD l 1 (Phase VI, Table 1). This initial phasing in of L-arabinose and D-xylose had no significant effect on the COD removal efficiency of either reactor, with COD removal efficiencies remaining between 92% and 98% in both (Fig. 1). On day 216, the L-arabinose/D-xylose influent concentration was increased, with a corresponding decrease in the D-glucose/sucrose content of the influent (Phase VII, Table 1). This had minimal effect on the performance of both reactors, although a minor perturbation was noted for the thermophilic reactor on day 218 (Fig. 1). This was accompanied by a short-term elevated discharge of acetate and propionate in the R2 effluent (Fig. 2B). Omission of D-gluose/sucrose and an increase in the L-arabinose/D-xylose influent concentrations (4.5 gCOD l 1 d 1 and 2 days HRT) on day 264 were accommodated readily by both reactors and were accompanied by increased biogas methane percentages to 85% from both reactors and no significant effect on effluent VFA (Fig. 2). L-rhamnose and D-galacturonic acid were introduced to both reactors while gradually decreasing the L-arabinose/D-xylose influent concentrations (Periods IX–XI, Table 1). Both reactors adapted to the new influent sugars and, by the end of the trial, the COD removal efficiency was >95% (Fig. 1) and effluent VFA levels from both reactors were very low (Fig. 2). 3.2. Mesophilic versus thermophilic reactor performance During initial introduction of L-arabinose and D-xylose from day 91 onwards (Table 1: Period II), a severe decrease was noted in the performance of both reactors within 1 day. The performance of both reactors, and the pH of the effluents reached levels categorised as ‘reactor failure’ by other authors (Kim et al., 2002). It was notable during this period of instability that the mesophilic reactor was more severely impacted by the sudden change to the new influent stream. In most reports to date, thermophilic anaerobic digestion has been reported as a more sensitive technology and associated with high effluent VFA, particularly propionate, accumulation due to the poor substrate affinities of some organisms (van Lier et al., 1996, 1997). This was not the case in the current study, however, and throughout the trial, effluent propionate discharge was a feature of the mesophilic reactor rather than its thermophilic counterpart. The deteriorated performance of the mesophilic reactor during this time may have been due to the pro-
liferation of acetogens at the expense of methanogens. Acetogenic oxidation of arabinose by the mesophilic Clostridium scatologenes strain SL1 has been shown to produce acetate and butyrate and may account for the elevated levels of these VFAs and decreased methane production during reactor failure (Kusel et al., 2000). During the phasing in of the sugar influents, both the thermophilic and mesophilic reactors performed comparably, in terms of COD removal efficiency and % biogas methane. The phasing in of new influent constituents was a highly successful approach, affording the biomass time to adapt to the new sugars, most likely by the development of new bacterial populations. All six trial sugars were degraded under both thermophilic and mesophilic conditions. The overall average COD removal efficiencies during the final trial period for both reactors were very similar 93.8% for R1 and only slightly better, 96.4%, for R2, with the % CH4 biogas of the two reactors highly comparable at 60.6% for R1 and 61.3% for R2. These findings are in contrast to what has been observed in other studies, which have generally shown thermophilic anaerobic digestion to be a more fastidious and sensitive process than mesophilic digestion. 3.3. Biomass physiology 3.3.1. Specific substrate utilisation (SSU) tests Specific substrate utilisation tests assessed the activity of the various fermentative species found within the reactor sludge. The rate of degradation of each sugar by the sludges during the trial is presented in Table 2. It is evident that, by day 194, the reactor sludges had a high capacity to degrade D-glucose and sucrose, and the degradation rates for these sugars exceeded the rates for the others tested. This is not surprising given that the reactor influent prior to this test was composed solely of these sugar moities. The biomass was again tested on day 305, by which time both reactors had successfully adapted to L-arabinose and D-xylose as substrates. From Table 2 it is clear that the SSU rates upon these sugars were comparable at both 37 °C and 55 °C. SSU values for L-arabinose and D-xylose were similar for those of the other tests sugars for the mesophilic sludge, with the exception of glucose which was almost twofold higher than mesophilic degradation of L-arabinose. By contrast, degradation for L-arabinose and D-xylose exceeded those of all other test sugars for the thermophilic sludge (Table 2). The most significant trend evident from the data obtained on day 430 was that SSU rates for all substrates were markedly higher than in the previous two tests, with rates as high as 38.04 mg gVSS 1 h 1 being recorded for the mesophilic degradation of D-glucose (Table 2). Notably high SSU values were also observed for thermophilic degradation of L-rhamnose and mesophilic degradation of D-glucose, sucrose, and D-galacturonic acid. It is not surprising that a faster rate would be evident in the degradation of Dgalacturonic acid, as this sugar had been applied in the reactor influent in the period preceding the test.
Table 2 Substrate utilisation rates for R1 (37 °C) and R2 (55 °C) sludges against sugar substrates used throughout the trial. Rates expressed in mg gVSS 1 hr 1. Day of test
194
Temperature
37 °C
55 °C
37 °C
55 °C
37 °C
55 °C
D-Glucose
19.9
15.4
9.6
5.8
38.0
18.5
Sucrose
13.9 11.4
22.1 3.2
6.1 5.0
2.8 6.8
36.6 7.4
12.2 8.1
D-Xylose
6.6
5.2
5.7
5.2
10.0
7.4
L-Rhamnose
8.8
6.1
4.9
2.2
14.3
32.5
13.9
13.9
5.2
2.2
29.7
16.0
L-Arabinose
D-Galacturonic
acid
305
430
3929
C. Forbes et al. / Bioresource Technology 101 (2010) 3925–3930 Table 3 Specific methanogenic activity (SMA) profiles of R1 (37 °C) and R2 (55 °C) sludges cultivated throughout the trial. Biomass activity expressed in ml CH4 (STP) g are the mean of triplicates ± standard deviation. 305
1
VSS d
1
. Values
Day of test
194
Substrate
37 °C
55 °C
37 °C
55 °C
37 °C
55 °C
H2CO2 Acetate Propionate Butyrate Ethanol
756.9 ± 46.0 55.3 ± 3.8 38.4 ± 3.4 21.2 ± 2.8 65.2 ± 3.7
284.7 ± 30.5 117.1 ± 9.3 22.1 ± 2.6 165.5 ± 15.6 89.3 ± 8.6
508.8 ± 21.8 109.9 ± 11.7 62.3 ± 3.2 40.2 ± 2.0 88.2 ± 1.5
231.7 ± 32.1 36.8 ± 2.1 4.6 ± 1.6 31.7 ± 1.1 22.2 ± 1.5
137.4 ± 11.2 73.1 ± 11.3 77.1 ± 7.8 9.6 ± 2.5 90.1 ± 3.5
225.9 ± 15.6 53.1 ± 0.9 7.8 ± 1.2 12.1 ± 3.1 14.1 ± 0.8
It was of interest, however, that SSU values for D-glucose and sucrose increased (Table 2; day 430). Since D-glucose and sucrose had not been included in the reactor influent since day 264, the SSU results suggest that fermentative species developed during influent feed Periods of IX–XI (Table 1) were D-glucose and sucrose utilizers with a substrate range for the 6-carbon sugars, L-rhamnose and D-galacturonic acid. Reactor performance and SSU results (Table 2, day 305) suggest that this had not been the case for the pentose sugars, L-arabinose and D-xylose. It is likely that new fermentative species, which were present in the reactor at low levels, had to be developed in order to degrade L-arabinose and D-xylose at the concentrations presented during the trial. From SSU results on day 305, it does not appear that the 6-carbon sugar fermenters were involved in the degradation of 5-carbon sugars. At the three test times, SSU results were higher at 37 °C than at 55 °C for all of the sugars tested, with the exceptions of L-arabinose at 55 °C on day 305 and 430, and sucrose on day 194 (Table 2). 3.3.2. Specific methanogenic activity (SMA) tests The SMA values calculated throughout this trial are presented in Table 3. It is evident that on day 194, following reactor recovery, the mesophilic biomass exhibited markedly high activity on the gaseous substrate, H2/CO2. This is probably a direct result of the preceding poor reactor performance which would likely have increased H2 pressures, and led to a proliferation in H2 utilisers as the reactor recovered. It is evident that as the trial progressed, the activity of the biomass cultivated in the mesophilic reactor increased against the syntrophic substrates, propionate and ethanol. This indicated the improved ability of this reactor sludge to degrade the intermediate organic acids that had accumulated and resulted in failure of this reactor in response to the introduction of Larabinose and D-xylose on day 91. A lower activity on butyrate at 37 °C was observed which may indicate that sugar degradation proceeded through the formation of other fatty acid intermediates. With respect to the activity of the methanogens within the R1 sludge, a marked decrease in hydrogenotrophic activity was observed as the trial progressed. This, together with an increase in the activity of the acetotrophic methanogens suggests a shift from hydrogen based methanogenesis to acetotrophic methanogenesis as the successive influents were phased into R1. By contrast, SMA values for the thermophilic biomass against all test substrates was found to decrease as the trial progressed (Table 3). The only exception to this finding was a marginal increase in the SMA values against acetate and propionate on day 430. The comparatively low SMA values against propionate for the R2 biomass were typical of thermophilic sludges (Ahring, 1994; Ahring et al., 1993). A marked decrease in the activity of the thermophilic biomass on butyrate and ethanol was noted as the trial progressed on the different sugars (Table 3). It is possible that intermediates produced by the different substrate fermenters, (e.g. formic and lactic acids etc.) may not have been detected during the thermophilic trial. Alternatively, sugar substrates may have been degraded directly to acetate by homoacetogenic bacteria growing heterotrophically. The role of
430
acetogens during thermophilic digestion, however, requires further investigation. Although mesophilic acetogenic oxidation of sugars has been previously observed, heterotrophic involvement of acetogens at 55 °C remains unclear (Drake et al., 2006; Winter and Wolfe, 1980). In both reactors, the decreased SMA values against H2/CO2 and acetate (Table 3) may reflect the relatively crude SMA measurement technique. Since the sugar fermentative species are likely to grow rapidly, the contribution of the slower-growing archaeal and syntrophic organisms to the overall VSS is likely to decrease. This apparent reduction in SMA values would be more pronounced for the slower-growing acetotrophs rather than the hydrogenotrophic Archaea. 4. Conclusions The current study demonstrated that thermophilic AD was comparable to mesophilic AD during the treatment of simple sugars. At the beginning of this trial, both processes were comparable on a D-glucose/sucrose influent. The introduction of 5 carbon sugars, however, resulted in near failure of both reactors. It is of interest that the thermophilic reactor recovered more rapidly when Dglucose and sucrose were reintroduced. The subsequent introduction of novel influents at initial low and gradually increasing VLRs was accommodated by both reactors, with marginally better performance by R1 (mesophilic) for the duration of the trial. Given fermentative Bacteria may vary considerably with regard to their sugar fermentative profiles, this study highlighted the importance of gradual introduction of new sugar substrates to both mesophilic and thermophilic reactors. Acknowledgements The receipt of financial support from the Irish Environmental Protection Agency is gratefully acknowledged. References Ahring, B.K., 1994. Status on science and application of thermophilic anaerobic digestion. Water Sci. Technol. 30, 241–249. Ahring, B.K., Schmidt, J.E., Winthernielsen, M., Macario, A.J.L., Demacario, E.C., 1993. Effect of medium composition and sludge removal on the production, composition and architecture of thermophilic (55 °C) acetate-utilizing granules from an upflow anaerobic sludge blanket reactor. Appl. Environ. Microbiol. 59, 2538–2545. APHA, 2001. Standard Methods for the Examination of Water and Wastewaters. Washington DC 20005: American Public Heath Association, American Waterworks Association and Water Environment Federation. Callaghan, F.J., Wase, D.A.J., Thayanithy, K., Forster, C.F., 1999. Co-digestion of waste organic solids: batch studies. Bioresour. Technol. 67, 117–122. Colleran, E., Pistilli, A., 1994. Activity test system for determining the toxicity of xenobiotic chemicals to the methanogenic process. Ann. Microbiol. Enzymol. 44, 1–8. Drake, H.L., Küsel, L., Matthies, C., 2006. Acetogenic prokaryotes. In: The Prokaryotes. Springer, New York, pp. 354–420. Dubois, M., Gilles, K.A., Hamilton, J., Rebers, P.A., Smith, F., 1956. Colorimetric method for determination of sugars and related substances. Anal. Chem. 28, 350–356. Forbes, C., O’Reilly, C., Gilleran, C., McLaughlin, L., Tuohy, M., Colleran, E., 2009. Application of high rate, high temperature anaerobic digestion to fungal
3930
C. Forbes et al. / Bioresource Technology 101 (2010) 3925–3930
thermozyme hydrolysates from carbohydrate wastes. Water Res. 43, 2531– 2539. Grohmann, K., Bothast, R.J., 1994. Pectin-rich residues generated by processing of citrus-fruits, apples, and sugar-beets – enzymatic hydrolysis and biological conversion to value-added products. In: Enzymatic Conversion of Biomass for Fuels Production. Oxford University Press, New York, pp. 372–390. Hartmann, H., Ahring, B.K., 2006. Strategies for the anaerobic digestion of the organic fraction of municipal solid waste: an overview. Water Sci. Technol. 53, 7–22. Hungate, R.E., Smith, W., Clarke, R.T.J., 1966. Suitability of butyl rubber stoppers for closing anaerobic roll culture tubes. J. Bacteriol. 91, 908–909. Kim, M., Ahn, Y.H., Speece, R.E., 2002. Comparative process stability and efficiency of anaerobic digestion; mesophilic vs. thermophilic. Water Res. 36, 4369–4385. Kusel, K., Dorsch, T., Acker, G., Stackebrandt, E., Drake, H.L., 2000. Clostridium scatologenes strain SL1 isolated as an acetogenic bacterium from acidic sediments. Int. J. Syst. Evol. Microbiol. 50, 537–546. Mata-Alvarez, J., Mace, S., Llabres, P., 2000. Anaerobic digestion of organic solid wastes. An overview of research achievements and perspectives. Bioresour. Technol. 74, 3–16.
Noike, T., Endo, G., Chang, J.E., Yaguchi, J.I., Matsumoto, J.I., 1985. Characteristics of carbohydrate degradation and the rate-limiting step in anaerobic digestion. Biotechnol. Bioeng. 27, 1482–1489. Renard, C., Crepeau, M.J., Thibault, J.F., 1995. Structure of the repeating units in the rhamnogalacturonic backbone of apple, beet and citrus pectins. Carb. Res. 275, 155–165. Shelton, D.R., Tiedje, J.M., 1984. General method for determining anaerobic biodegradation potential. Appl. Environ. Microbiol. 47, 850–857. van Lier, J.B., Martin, J.L.S., Lettinga, G., 1996. Effect of temperature on the anaerobic thermophilic conversion of volatile fatty acids by dispersed and granular sludge. Water Res. 30, 199–207. van Lier, J.B., Rebac, S., Lettinga, G., 1997. High-rate anaerobic wastewater treatment under psychrophilic and thermophilic conditions. Water Sci. Technol. 35, 199– 206. Winter, J.U., Wolfe, R.S., 1980. Methane formation from fructose by syntrophic associations of Acetobacterium woodii and different strains of methanogens. Arch. Microbiol. 124, 73–79.
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
High-rate anaerobic degradation of 5 and 6 carbon sugars under thermophilic and mesophilic conditions C. Forbes *, D. Hughes, J. Fox, P. Ryan, E. Colleran Environmental Microbiology Research Unit, Department of Microbiology, National University of Ireland, Galway, Ireland
a r t i c l e
i n f o
Article history: Received 8 September 2009 Received in revised form 5 January 2010 Accepted 9 January 2010 Available online 4 February 2010 Keywords: Thermophilic AD Pre-hydrolysis Pectin Cellulose OFMSW
a b s t r a c t In this research paper, a comparison between thermophilic and mesophilic anaerobic degradation of a variety of the simple sugar components of carbohydrate rich biomass is presented. In order to investigate the degradability of these basic sugars, three synthetic sugar based influents were supplied to two high rate upflow anaerobic hybrid reactors (UAHR) operated at 37 °C (R1) and 55 °C (R2). These influent streams were: D-glucose/sucrose; L-arabinose/D-xylose and L-rhamnose/D-galacturonic acid. The reactors were challenged in terms of influent composition rather than loading rate and were therefore operated at a maximum volumetric loading rate (VLR) of 4.5 gCOD l 1 d 1 during stable reactor performance. It was found that a switch from a D-glucose/sucrose synthetic influent to an influent composed of L-arabinose/Dxylose resulted in failure of the mesophilic reactor while the thermophilic UAHR was able to tolerate the change of sugar influent at an unchanged VLR of 4.5 gCOD l 1 d 1. A subsequent phasing-in approach was used to introduce new sugar influent streams and proved highly successful. The physiology of the biomass was assessed and it was noted that thermophilic anaerobic digestion (AD) involved the formation of acetate and H2, implying the involvement of homoacetogenic bacteria, while mesophilic AD proceeded via the formation of other intermediates. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction In unprocessed plant materials, carbohydrate may account for up to 60–95% of total dry weight. Plant-derived raw materials are utilized widely in the food, beverage, paper and packaging industries and thus constitute a significant portion of the waste from these industries. The waste generated from the food industry ranges from relatively unprocessed food materials (surplus and spoiled or contaminated stock) to the by-products of food processing activities in an industrial context. In recent years, a number of novel approaches for the treatment of these wastes with concomitant energy recycling have been investigated, including the production of bioethanol and biodiesel as well as biodegradable plastics (Callaghan et al., 1999). From both an environmental and economic perspective, however, anaerobic treatment is arguably the best option for remediation of these wastes, simultaneously reducing their pollution potential while producing a value added end product in the form of a renewable biofuel, methane gas. There are limitations, however, to direct anaerobic treatment of carbohydrate rich materials which extend to the organic fraction of
* Corresponding author. Present address: Environmental Research Institute, University College Cork, Lee Road, Ireland. Tel.: +353 214901975; fax: +353 214901932. E-mail address: [email protected] (C. Forbes). 0960-8524/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2010.01.019
municipal solid wastes (OFMSW), food wastes and plant materials used as energy crops. The effective and high-rate anaerobic treatment of solid wastes is impeded by the comparatively lengthy time taken to hydrolyse complex polymers such as cellulose, pectin and hemi-celluloses, which comprise the major components of nonanimal food wastes, to their constituent monomeric sugars (Noike et al., 1985). Additionally, the biodegradability of solid wastes is highly dependent on the composition of cellulose, hemicellulose and lignin which are characteristic to the different components of OFMSW (Hartmann and Ahring, 2006). Successful treatment of these wastes in single stage reactors involves the operation of Continuously Stirred Tank Reactors (CSTRs) with retention times in the region of 10–30 days. A commonly employed method to circumvent these problems involves pre-hydrolysis of these materials to yield more readily degradable short chain polymers and monomers prior to anaerobic treatment. By using pre-treatment strategies, improved hydrolysis of complex polysaccharide containing wastes can be achieved. As a result of this approach, the retention times of waste in anaerobic digesters can be considerably shortened, and higher methane yields can be produced. A number of pre-hydrolysis strategies have been investigated in order to eliminate the initial rate limiting hydrolysis step. Pre-treatment methods include mechanical, chemical, and biological based approaches as well as various combinations of these. Those currently employed have been reviewed extensively elsewhere (Mata-Alvarez et al., 2000;
3926
C. Forbes et al. / Bioresource Technology 101 (2010) 3925–3930
Hartmann and Ahring, 2006). Another major limitation in the use of single stage anaerobic digestion of OFMSW and highly biodegradable food wastes is the rapid acidification of these wastes resulting in pH decrease in the reactor, inhibiting the activity of methanogenic Archaea. These instability factors preclude the use of high-rate anaerobic digesters with retention times measured in hours rather than days (Mata-Alvarez et al., 2000). In the current trial, three distinct synthetic sugar influents were devised in order to assess the anaerobic biodegradability of prehydrolysed carbohydrate wastes. Each influent stream consisted of two simple sugars, which were applied to the UAHRs in tandem and designed to represent the carbohydrate macromolecule components of OFMSW. The first waste stream supplied to the reactors consisted of D-glucose and sucrose, both of which have been demonstrated to be readily degraded anaerobically (Forbes et al., 2009). Most carbohydrates contain D-glucose, often as their sole building block, as in the examples of starch, glycogen, and most importantly, cellulose. Sucrose is a disaccharide molecule, the structure of which inhibits further binding to other saccharide moieties. It is normally extracted from either sugar cane or beet and is commonly used as a sweetener and extensively employed in the baking industry and in food preservation. The second waste stream supplied to the reactors consisted of L-arabinose and D-xylose. Both are pentose monosaccharides and found as part of more complex polymers, such as xylans, hemicelluloses and pectins. The third influent stream was comprised L-rhamnose and D-galacturonic acid, the major monosaccharide components of pectin (Grohmann and Bothast, 1994; Renard et al., 1995). It was anticipated that the reactor sludges would adapt readily to these three synthetic influents as these reactors had previously been employed in a 650 days trial during which both had treated a variety of more complex hydrolysate influents (Forbes et al., 2009).
2. Methods 2.1. Reactor design and operation Two laboratory scale UAHRs, both with an active volume of 3.9 l, were used in the current trial. These reactors had previously been inoculated and operated at 37 °C (R1) and 55 °C (R2) for 670 days as described by Forbes et al. (2009), followed by a period of 200 days of application of D-glucose and sucrose. On day 1 of the current trial, both reactors were supplied with the first trial influent, which consisted of D-glucose and sucrose on an equal COD ratio to a total VLR of 4.5 gCOD l 1 d 1. This was achieved by operating the reactors at a 2 days hydraulic retention time (HRT), with an influent concentration of 9 gCOD l 1. The reactors were maintained on this influent until day 91 when the influent compo-
sition was changed to an equal ratio, on a COD basis, of L-arabinose and D-xylose to a total of 9 gCOD l 1. Feeding with L-arabinose and D-xylose was continued at the same rate until day 103. Due to instability, the feeding was then stopped for 6 days (Table 1; Period III). During this period, the reactors were unfed and only 10 ml of buffer and associated nutrients were added on occasion to allow removal of effluent samples for analysis. L-arabinose and D-xylose were reintroduced on day 110, at a lower combined influent COD concentration of 4.5 g l 1, at a VLR of 2.25 gCOD l 1d 1. On day 126, the L-arabinose and D-xylose influent was replaced by D-glucose and sucrose at a VLR of 4.5 gCOD l 1 d 1 (Period V). This VLR was maintained until day 201. L-arabinose and D-xylose were then phased back into the influent at a combined COD of 3 gCOD l 1 on day 202. The VLR of 4.5 gCOD l 1d 1 was maintained by inclusion of D-glucose and sucrose at an influent COD of 6 gCOD l 1 during the period (Period VI). On day 216, the influent L-arabinose and D-xylose was increased to a combined gCOD l 1 of 6, while correspondingly reducing the D-glucose/sucrose in the influent to 3 gCOD l 1 (Table 1; Period VII). On day 264, D-glucose and sucrose were removed totally from the influent and the combined COD concentration of L-arabinose and D-xylose was increased to 9gCOD l 1 d 1 (Table 1; Period VIII). L-rhamnose and D-galacturonic acid were phased into the reactor influent, beginning on day 342. This was achieved by reducing the combined COD of L-arabinose and D-xylose to 6 gCOD l 1 and introducing L-rhamnose and D-galacturonic acid at 3 gCOD l 1 (Table 1; Period IX). The influent concentration of L-rhamnose and Dgalacturonic was increased to 6 gCOD l 1 on day 363 (Table 1; Period X). L-arabinose and D-xylose were withdrawn from the reactor influents on day 400 with a concomitant increase in the combined influent of L-rhamnose and D-galacturonic acid to 9 gCOD l 1. Sole feeding of the latter two sugars was continued at a HRT of 2 days and a VLR of 4.5 gCOD l 1 d 1 until day 430, at which point the trial concluded. Throughout the duration of the trial, the COD:N:P ratio of all influents was maintained at 1000:5:0.5 by supplementation with NH4Cl and KH2PO4 to the required concentrations. Buffering was carried out by addition of NaHCO3 (12 g l 1) and influent waste streams were supplemented with micronutrients (1 ml l 1), as recommended by Shelton and Tiedje (1984).
2.2. Analytical techniques Samples of reactor influent/effluent were routinely analysed for COD (g l 1), VFAs (mg l 1) and pH as previously described (Forbes et al., 2009). Biogas was sampled for CH4 determination according to Standard Methods (APHA, 2001).
Table 1 Influent regime showing the contribution of each sugar to each influent stream, on a COD basis (gCOD l
).
Phase
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
Start of phase (day of trial)
1 4.5
91 –
103 –
110 –
126 4.5
202 3
216 1.5
264 –
342 –
363 –
400 –
L-Arabinose
4.5 –
– 4.5
– –
– 2.25
4.5 –
3 1.5
1.5 3
– 4.5
– 3
– 1.5
– –
D-Xylose
–
4.5
–
2.25
–
1.5
3
4.5
3
1.5
–
L-Rhamnose
–
–
–
–
–
–
–
–
1.5
3
4.5
D-Galacturonic
–
–
–
–
–
–
–
–
1.5
3
4.5
9 2 4.5
9 2 4.5
0 390* 0
4.5 2 2.25
9 2 4.5
9 2 4.5
9 2 4.5
9 2 4.5
9 2 4.5
9 2 4.5
9 2 4.5
D-Glucose Sucrose
acid Total influent COD (g l HRT (days) VLR (gCOD l 1 d 1) *
1
1
)
Based on 10 mls of buffer per day being supplied to the reactors in order to obtain an effluent sample.
C. Forbes et al. / Bioresource Technology 101 (2010) 3925–3930
3927
2.3. Specific substrate utilisation (SSU) tests SSU tests were performed on days 194, 305 and 430 of the trial, corresponding with times at which the reactors were fully adapted to each synthetic influent. SSU tests assessed the activity of fermentative species in the reactor sludges to degrade the various sugars applied during the trial. Tests were carried out at 37 °C and 55 °C in 60 ml serum vials containing sludge (2–5 gVSS l 1), anaerobic buffer and one of each of the trial sugars (2.5 gCOD l 1) to a total of 30 ml. Samples were removed intermittently from test vials, were centrifuged at 10,000g for 10 min and the supernatant analysed spectrophotometrically using the Dubois method for the quantification of carbohydrates (Dubois et al., 1956). Briefly, one volume of sample was mixed with one volume of 5% phenol solution, and the mixture cooled at 4 °C for 10 min. Five volumes of concentrated H2SO4 were rapidly added to the pre-cooled mixture, with continuous mixing, followed by boiling for 5 min. Samples were then cooled and incubated at room temperature for 30 min prior to determining the absorbance at 490 nm. In vial carbohydrate concentrations were plotted against time to give a degradation profile for each sugar. Specific substrate utilisation rates were calculated by using the period of most rapid carbohydrate degradation from each profile, and expressing this as the rate of carbohydrate depletion per gram of volatile suspended solids (mg carbohydrate g VSS 1 hr 1). 2.4. Specific methanogenic activity (SMA) tests Specific methanogenic activity tests were also performed on days 194, 305 and 450 of the trial. SMA measurements were carried out using the pressure transducer method developed by Colleran and Pistilli (1994). Briefly, the procedure involved the measurement of the biogas pressure increase developing in sealed vials fed with the non-gaseous substrates ethanol (30 mM), propionate (30 mM), butyrate (15 mM) and acetate (30 mM), or of the decrease in vials pressurised to 1 atm pressure with the gaseous substrate H2/CO2 (80:20). Tests were carried out in triplicate at 37 °C and 55 °C in 20 ml serum vials containing sludge (2–5 g VSS l 1), anaerobic buffer (Hungate, 1966) and one of each of the test substrates. Appropriate temperature conversion factors were used to record the final results as ml CH4 gVSS 1 day 1 at standard temperature and pressure (STP).
Fig. 1. Percentage soluble COD removal by R1 and R2 during the trial. Different R1 R2). operational periods are denoted by arrows (Table 1). (
methane production by day 100 (Fig. 1). There was a parallel increase in effluent VFA concentrations (Fig. 2A). By day 102, the mesophilic effluent VFA concentrations were 1440, 3310 and 3490 mg l 1 for acetate, propionate and butyrate, respectively (Fig. 2A). By contrast, the thermophilic effluent VFA concentrations were considerably lower, with levels of 676 and 1709 mg l 1 for acetate and propionate, respectively, and with negligible levels of butyrate (Fig. 2B). These high concentrations of short chain fatty acids resulted in pH levels of lower than 5.5 in both reactors by day 100. It is noteworthy that Kim et al. (2002) defined reactor failure in terms of an effluent discharge pH lower than 5.5.
3. Results and discussion 3.1. Reactor operation Prior to the start of the current study, both reactors had been operated for 900 days at 37 °C (R1) and 55 °C (R2). For the 200 days immediately preceding the trial, both reactors operated in a stable fashion on an influent composed of D-glucose/sucrose. The current trial initially utilized the same substrates as previously at a VLR of 4.5 gCOD l 1 d 1 and a HRT of 2 days for a 90 day period (Table 1). Despite fluctuations it was clear that both reactors could achieve >95% COD removal under the conditions imposed (Fig. 1). On day 91, the influent was changed to L-arabinose and D-xylose (on an equal COD basis), while maintaining the VLR and HRT (Table 1). This resulted in an immediate and dramatic decrease in COD removal efficiencies (Fig. 1). The methane content of the produced biogas also decreased to 14% from the mesophilic reactor and 12% from the thermophilic reactor. Within five HRTs the performance of the thermophilic reactor had deteriorated to 35% COD removal, with methane comprising only 16% of the biogas produced. The response of the mesophilic reactor was even more severe, with total failure in terms of COD removal efficiency and
Fig. 2. Effluent VFA levels (mg l 1) from the mesophilic reactor, R1 (A), and the thermophilic reactor, R2 (B) throughout the trial period. Different operational Acetate propionate butyrate. periods are denoted by arrows (Table 1).
3928
C. Forbes et al. / Bioresource Technology 101 (2010) 3925–3930
The influent supply to both reactors was discontinued on day 103. An immediate reduction in effluent VFA concentrations was noted (Fig. 2), together with a decrease in effluent COD (Fig. 1). This suggested that the rapid acidification following introduction of Larabinose and D-xylose had not caused complete inhibition of sludge syntrophic and archaeal populations. L-arabinose and D-xylose were reintroduced on day 110 (Period IV, Table 1) at a lower COD concentration of 2.25 gCOD l 1 for both sugars. This resulted in a VLR of 2.25 gCOD l 1 d 1 at a 2 days HRT. Although the performance of the mesophilic reactor deteriorated, the thermophilic reactor performance showed greater stability, with 50% COD removal efficiency on day 126 (Fig. 1) and a biogas methane content of 30%. The effluent VFA profiles (Fig. 2) confirm the better performance of the thermophilic reactor during this period. To avoid total failure of the mesophilic reactor, on day 126 the influent to both reactors was changed to the D-glucose/sucrose influent at the previous VLR of 4.5 gCOD l 1 d 1 while maintaining the HRT at 2 days (Table 1). As a result, an immediate improvement was observed in the performance of both reactors. By day 202, both reactors had fully recovered and regularly exhibited >90% COD removal efficiencies. L-arabinose and D-xylose were reintroduced to the reactors on day 202 (influent concentration of 1.5 gCOD l 1 for each sugar) while maintaining a reduced combined D-glucose/sucrose influent of 6 gCOD l 1 (Phase VI, Table 1). This initial phasing in of L-arabinose and D-xylose had no significant effect on the COD removal efficiency of either reactor, with COD removal efficiencies remaining between 92% and 98% in both (Fig. 1). On day 216, the L-arabinose/D-xylose influent concentration was increased, with a corresponding decrease in the D-glucose/sucrose content of the influent (Phase VII, Table 1). This had minimal effect on the performance of both reactors, although a minor perturbation was noted for the thermophilic reactor on day 218 (Fig. 1). This was accompanied by a short-term elevated discharge of acetate and propionate in the R2 effluent (Fig. 2B). Omission of D-gluose/sucrose and an increase in the L-arabinose/D-xylose influent concentrations (4.5 gCOD l 1 d 1 and 2 days HRT) on day 264 were accommodated readily by both reactors and were accompanied by increased biogas methane percentages to 85% from both reactors and no significant effect on effluent VFA (Fig. 2). L-rhamnose and D-galacturonic acid were introduced to both reactors while gradually decreasing the L-arabinose/D-xylose influent concentrations (Periods IX–XI, Table 1). Both reactors adapted to the new influent sugars and, by the end of the trial, the COD removal efficiency was >95% (Fig. 1) and effluent VFA levels from both reactors were very low (Fig. 2). 3.2. Mesophilic versus thermophilic reactor performance During initial introduction of L-arabinose and D-xylose from day 91 onwards (Table 1: Period II), a severe decrease was noted in the performance of both reactors within 1 day. The performance of both reactors, and the pH of the effluents reached levels categorised as ‘reactor failure’ by other authors (Kim et al., 2002). It was notable during this period of instability that the mesophilic reactor was more severely impacted by the sudden change to the new influent stream. In most reports to date, thermophilic anaerobic digestion has been reported as a more sensitive technology and associated with high effluent VFA, particularly propionate, accumulation due to the poor substrate affinities of some organisms (van Lier et al., 1996, 1997). This was not the case in the current study, however, and throughout the trial, effluent propionate discharge was a feature of the mesophilic reactor rather than its thermophilic counterpart. The deteriorated performance of the mesophilic reactor during this time may have been due to the pro-
liferation of acetogens at the expense of methanogens. Acetogenic oxidation of arabinose by the mesophilic Clostridium scatologenes strain SL1 has been shown to produce acetate and butyrate and may account for the elevated levels of these VFAs and decreased methane production during reactor failure (Kusel et al., 2000). During the phasing in of the sugar influents, both the thermophilic and mesophilic reactors performed comparably, in terms of COD removal efficiency and % biogas methane. The phasing in of new influent constituents was a highly successful approach, affording the biomass time to adapt to the new sugars, most likely by the development of new bacterial populations. All six trial sugars were degraded under both thermophilic and mesophilic conditions. The overall average COD removal efficiencies during the final trial period for both reactors were very similar 93.8% for R1 and only slightly better, 96.4%, for R2, with the % CH4 biogas of the two reactors highly comparable at 60.6% for R1 and 61.3% for R2. These findings are in contrast to what has been observed in other studies, which have generally shown thermophilic anaerobic digestion to be a more fastidious and sensitive process than mesophilic digestion. 3.3. Biomass physiology 3.3.1. Specific substrate utilisation (SSU) tests Specific substrate utilisation tests assessed the activity of the various fermentative species found within the reactor sludge. The rate of degradation of each sugar by the sludges during the trial is presented in Table 2. It is evident that, by day 194, the reactor sludges had a high capacity to degrade D-glucose and sucrose, and the degradation rates for these sugars exceeded the rates for the others tested. This is not surprising given that the reactor influent prior to this test was composed solely of these sugar moities. The biomass was again tested on day 305, by which time both reactors had successfully adapted to L-arabinose and D-xylose as substrates. From Table 2 it is clear that the SSU rates upon these sugars were comparable at both 37 °C and 55 °C. SSU values for L-arabinose and D-xylose were similar for those of the other tests sugars for the mesophilic sludge, with the exception of glucose which was almost twofold higher than mesophilic degradation of L-arabinose. By contrast, degradation for L-arabinose and D-xylose exceeded those of all other test sugars for the thermophilic sludge (Table 2). The most significant trend evident from the data obtained on day 430 was that SSU rates for all substrates were markedly higher than in the previous two tests, with rates as high as 38.04 mg gVSS 1 h 1 being recorded for the mesophilic degradation of D-glucose (Table 2). Notably high SSU values were also observed for thermophilic degradation of L-rhamnose and mesophilic degradation of D-glucose, sucrose, and D-galacturonic acid. It is not surprising that a faster rate would be evident in the degradation of Dgalacturonic acid, as this sugar had been applied in the reactor influent in the period preceding the test.
Table 2 Substrate utilisation rates for R1 (37 °C) and R2 (55 °C) sludges against sugar substrates used throughout the trial. Rates expressed in mg gVSS 1 hr 1. Day of test
194
Temperature
37 °C
55 °C
37 °C
55 °C
37 °C
55 °C
D-Glucose
19.9
15.4
9.6
5.8
38.0
18.5
Sucrose
13.9 11.4
22.1 3.2
6.1 5.0
2.8 6.8
36.6 7.4
12.2 8.1
D-Xylose
6.6
5.2
5.7
5.2
10.0
7.4
L-Rhamnose
8.8
6.1
4.9
2.2
14.3
32.5
13.9
13.9
5.2
2.2
29.7
16.0
L-Arabinose
D-Galacturonic
acid
305
430
3929
C. Forbes et al. / Bioresource Technology 101 (2010) 3925–3930 Table 3 Specific methanogenic activity (SMA) profiles of R1 (37 °C) and R2 (55 °C) sludges cultivated throughout the trial. Biomass activity expressed in ml CH4 (STP) g are the mean of triplicates ± standard deviation. 305
1
VSS d
1
. Values
Day of test
194
Substrate
37 °C
55 °C
37 °C
55 °C
37 °C
55 °C
H2CO2 Acetate Propionate Butyrate Ethanol
756.9 ± 46.0 55.3 ± 3.8 38.4 ± 3.4 21.2 ± 2.8 65.2 ± 3.7
284.7 ± 30.5 117.1 ± 9.3 22.1 ± 2.6 165.5 ± 15.6 89.3 ± 8.6
508.8 ± 21.8 109.9 ± 11.7 62.3 ± 3.2 40.2 ± 2.0 88.2 ± 1.5
231.7 ± 32.1 36.8 ± 2.1 4.6 ± 1.6 31.7 ± 1.1 22.2 ± 1.5
137.4 ± 11.2 73.1 ± 11.3 77.1 ± 7.8 9.6 ± 2.5 90.1 ± 3.5
225.9 ± 15.6 53.1 ± 0.9 7.8 ± 1.2 12.1 ± 3.1 14.1 ± 0.8
It was of interest, however, that SSU values for D-glucose and sucrose increased (Table 2; day 430). Since D-glucose and sucrose had not been included in the reactor influent since day 264, the SSU results suggest that fermentative species developed during influent feed Periods of IX–XI (Table 1) were D-glucose and sucrose utilizers with a substrate range for the 6-carbon sugars, L-rhamnose and D-galacturonic acid. Reactor performance and SSU results (Table 2, day 305) suggest that this had not been the case for the pentose sugars, L-arabinose and D-xylose. It is likely that new fermentative species, which were present in the reactor at low levels, had to be developed in order to degrade L-arabinose and D-xylose at the concentrations presented during the trial. From SSU results on day 305, it does not appear that the 6-carbon sugar fermenters were involved in the degradation of 5-carbon sugars. At the three test times, SSU results were higher at 37 °C than at 55 °C for all of the sugars tested, with the exceptions of L-arabinose at 55 °C on day 305 and 430, and sucrose on day 194 (Table 2). 3.3.2. Specific methanogenic activity (SMA) tests The SMA values calculated throughout this trial are presented in Table 3. It is evident that on day 194, following reactor recovery, the mesophilic biomass exhibited markedly high activity on the gaseous substrate, H2/CO2. This is probably a direct result of the preceding poor reactor performance which would likely have increased H2 pressures, and led to a proliferation in H2 utilisers as the reactor recovered. It is evident that as the trial progressed, the activity of the biomass cultivated in the mesophilic reactor increased against the syntrophic substrates, propionate and ethanol. This indicated the improved ability of this reactor sludge to degrade the intermediate organic acids that had accumulated and resulted in failure of this reactor in response to the introduction of Larabinose and D-xylose on day 91. A lower activity on butyrate at 37 °C was observed which may indicate that sugar degradation proceeded through the formation of other fatty acid intermediates. With respect to the activity of the methanogens within the R1 sludge, a marked decrease in hydrogenotrophic activity was observed as the trial progressed. This, together with an increase in the activity of the acetotrophic methanogens suggests a shift from hydrogen based methanogenesis to acetotrophic methanogenesis as the successive influents were phased into R1. By contrast, SMA values for the thermophilic biomass against all test substrates was found to decrease as the trial progressed (Table 3). The only exception to this finding was a marginal increase in the SMA values against acetate and propionate on day 430. The comparatively low SMA values against propionate for the R2 biomass were typical of thermophilic sludges (Ahring, 1994; Ahring et al., 1993). A marked decrease in the activity of the thermophilic biomass on butyrate and ethanol was noted as the trial progressed on the different sugars (Table 3). It is possible that intermediates produced by the different substrate fermenters, (e.g. formic and lactic acids etc.) may not have been detected during the thermophilic trial. Alternatively, sugar substrates may have been degraded directly to acetate by homoacetogenic bacteria growing heterotrophically. The role of
430
acetogens during thermophilic digestion, however, requires further investigation. Although mesophilic acetogenic oxidation of sugars has been previously observed, heterotrophic involvement of acetogens at 55 °C remains unclear (Drake et al., 2006; Winter and Wolfe, 1980). In both reactors, the decreased SMA values against H2/CO2 and acetate (Table 3) may reflect the relatively crude SMA measurement technique. Since the sugar fermentative species are likely to grow rapidly, the contribution of the slower-growing archaeal and syntrophic organisms to the overall VSS is likely to decrease. This apparent reduction in SMA values would be more pronounced for the slower-growing acetotrophs rather than the hydrogenotrophic Archaea. 4. Conclusions The current study demonstrated that thermophilic AD was comparable to mesophilic AD during the treatment of simple sugars. At the beginning of this trial, both processes were comparable on a D-glucose/sucrose influent. The introduction of 5 carbon sugars, however, resulted in near failure of both reactors. It is of interest that the thermophilic reactor recovered more rapidly when Dglucose and sucrose were reintroduced. The subsequent introduction of novel influents at initial low and gradually increasing VLRs was accommodated by both reactors, with marginally better performance by R1 (mesophilic) for the duration of the trial. Given fermentative Bacteria may vary considerably with regard to their sugar fermentative profiles, this study highlighted the importance of gradual introduction of new sugar substrates to both mesophilic and thermophilic reactors. Acknowledgements The receipt of financial support from the Irish Environmental Protection Agency is gratefully acknowledged. References Ahring, B.K., 1994. Status on science and application of thermophilic anaerobic digestion. Water Sci. Technol. 30, 241–249. Ahring, B.K., Schmidt, J.E., Winthernielsen, M., Macario, A.J.L., Demacario, E.C., 1993. Effect of medium composition and sludge removal on the production, composition and architecture of thermophilic (55 °C) acetate-utilizing granules from an upflow anaerobic sludge blanket reactor. Appl. Environ. Microbiol. 59, 2538–2545. APHA, 2001. Standard Methods for the Examination of Water and Wastewaters. Washington DC 20005: American Public Heath Association, American Waterworks Association and Water Environment Federation. Callaghan, F.J., Wase, D.A.J., Thayanithy, K., Forster, C.F., 1999. Co-digestion of waste organic solids: batch studies. Bioresour. Technol. 67, 117–122. Colleran, E., Pistilli, A., 1994. Activity test system for determining the toxicity of xenobiotic chemicals to the methanogenic process. Ann. Microbiol. Enzymol. 44, 1–8. Drake, H.L., Küsel, L., Matthies, C., 2006. Acetogenic prokaryotes. In: The Prokaryotes. Springer, New York, pp. 354–420. Dubois, M., Gilles, K.A., Hamilton, J., Rebers, P.A., Smith, F., 1956. Colorimetric method for determination of sugars and related substances. Anal. Chem. 28, 350–356. Forbes, C., O’Reilly, C., Gilleran, C., McLaughlin, L., Tuohy, M., Colleran, E., 2009. Application of high rate, high temperature anaerobic digestion to fungal
3930
C. Forbes et al. / Bioresource Technology 101 (2010) 3925–3930
thermozyme hydrolysates from carbohydrate wastes. Water Res. 43, 2531– 2539. Grohmann, K., Bothast, R.J., 1994. Pectin-rich residues generated by processing of citrus-fruits, apples, and sugar-beets – enzymatic hydrolysis and biological conversion to value-added products. In: Enzymatic Conversion of Biomass for Fuels Production. Oxford University Press, New York, pp. 372–390. Hartmann, H., Ahring, B.K., 2006. Strategies for the anaerobic digestion of the organic fraction of municipal solid waste: an overview. Water Sci. Technol. 53, 7–22. Hungate, R.E., Smith, W., Clarke, R.T.J., 1966. Suitability of butyl rubber stoppers for closing anaerobic roll culture tubes. J. Bacteriol. 91, 908–909. Kim, M., Ahn, Y.H., Speece, R.E., 2002. Comparative process stability and efficiency of anaerobic digestion; mesophilic vs. thermophilic. Water Res. 36, 4369–4385. Kusel, K., Dorsch, T., Acker, G., Stackebrandt, E., Drake, H.L., 2000. Clostridium scatologenes strain SL1 isolated as an acetogenic bacterium from acidic sediments. Int. J. Syst. Evol. Microbiol. 50, 537–546. Mata-Alvarez, J., Mace, S., Llabres, P., 2000. Anaerobic digestion of organic solid wastes. An overview of research achievements and perspectives. Bioresour. Technol. 74, 3–16.
Noike, T., Endo, G., Chang, J.E., Yaguchi, J.I., Matsumoto, J.I., 1985. Characteristics of carbohydrate degradation and the rate-limiting step in anaerobic digestion. Biotechnol. Bioeng. 27, 1482–1489. Renard, C., Crepeau, M.J., Thibault, J.F., 1995. Structure of the repeating units in the rhamnogalacturonic backbone of apple, beet and citrus pectins. Carb. Res. 275, 155–165. Shelton, D.R., Tiedje, J.M., 1984. General method for determining anaerobic biodegradation potential. Appl. Environ. Microbiol. 47, 850–857. van Lier, J.B., Martin, J.L.S., Lettinga, G., 1996. Effect of temperature on the anaerobic thermophilic conversion of volatile fatty acids by dispersed and granular sludge. Water Res. 30, 199–207. van Lier, J.B., Rebac, S., Lettinga, G., 1997. High-rate anaerobic wastewater treatment under psychrophilic and thermophilic conditions. Water Sci. Technol. 35, 199– 206. Winter, J.U., Wolfe, R.S., 1980. Methane formation from fructose by syntrophic associations of Acetobacterium woodii and different strains of methanogens. Arch. Microbiol. 124, 73–79.