Hydrothermal liquefaction (HTL) is a promising process for converting wet biomass and organic wastes into bio-crude oil. It also produces an aqueous product referred to as post-hydrothermal liquefaction wastewater (PHWW) containing up to 40% of the original feedstock carbon, which reduces the overall energy efficiency of the HTL process. This study investigated the feasibility of using anaerobic digestion (AD) to treat PHWW, with the aid of activated carbon. Results showed that successful AD occurred at relatively low concentrations of PHWW (≤ 6.7%), producing a biogas yield of 0.5 ml/mg CODremoved, and ∼53% energy recovery efficiency. Higher concentrations of PHWW (≥13.3%) had an inhibitory effect on the AD process, as indicated by delayed, slower, or no biogas production. Activated carbon was shown to effectively mitigate this inhibitory effect by enhancing biogas production and allowing digestion to proceed at higher PHWW concentrations (up to 33.3%), likely due to sequestering toxic organic compounds. The addition of activated carbon also increased the net energy recovery efficiency of AD with a relatively high concentration of PHWW (33.3%), taking into account the energy for producing activated carbon. These results suggest that AD is a feasible approach to treat PHWW, and to improve the energy efficiency of the HTL processes.

INTRODUCTION

Hydrothermal liquefaction (HTL) is a promising thermochemical process for converting wet organic biosolids into bio-crude oil for fuel or other biochemical products (Toor Rosendahl & Rudolf 2011) with a net positive energy balance. It uses elevated temperatures (200–400 °C) and pressures (10–15 MPa) to convert feedstock organics into four products: (1) bio-crude oil, (2) bio-char solid residue, (3) CO2-rich gas, and (4) an aqueous post-HTL wastewater (PHWW). The HTL process offers several advantages compared to other chemical or thermochemical fuel production processes, such as transesterification or pyrolysis, including the ability to use a wide variety of low-lipid feedstocks (e.g. livestock manure, wastewater sludge, and many types of fast-growing algae), shorter processing times, and a lower energy consumption for feedstock dewatering/drying (Toor Rosendahl & Rudolf 2011).

Despite all the advantages of HTL, several issues have limited its development beyond laboratory scale, one of which is limited energy efficiency. That is, HTL cannot convert all the organics in the feedstock into bio-crude oil, leaving behind an aqueous product with high concentrations of potentially valuable organics and nutrients: up to 40% of the carbon and up to 80% of the nutrients from the feedstock is released into PHWW (Yu et al. 2011). Carbon that partitions to the PHWW represents a reduction in the net energy yield captured into HTL oil. While several previous studies have focused on increasing the oil yield and minimizing carbon release into the aqueous product through HTL process optimization (pressure, temperature, retention time, etc.), relatively little attention has been paid to energy recovery from PHWW. The US Department of Energy has highlighted the importance of carbon recovery in biofuel production as exemplified by their recently initiated CHASE program (Carbon, Hydrogen, and Separation Efficiencies in Bio-Oil Conversion Pathways). This program seeks to develop technologies that make better use of PHWW and other thermochemical process sidestreams to improve the overall carbon efficiency and commercial viability of these bio-oil conversion processes. Although some recent studies have investigated resource recovery from sidestreams of thermochemical conversion processes, the main focus has been on nutrient recovery rather than carbon recovery (Jena et al. 2011; Biller et al. 2012).

Anaerobic digestion (AD) is the most widely used method for carbon/energy recovery from wet organic wastes. Key advantages include minimal dewatering needed prior to AD, low sludge production, and in particular the production of biogas as a renewable fuel. When synergistically integrated with HTL and algal cultivation, the AD of PHWW could help facilitate carbon recovery from the atmosphere and multi-cycle nutrient reuse, which can maximize the bioenergy production from waste streams (this is discussed in detail in the section ‘Integration of anaerobic digestion with the hydrothermal liquefaction process’). However, PHWW has also been reported to be toxic to both mammalian cells (Pham et al. 2013) and microorganisms (Zhou 2010). Many compounds identified in PHWW could potentially be inhibitory or toxic to AD processes, including ammonia and various organic compounds. These inhibitory effects could potentially be mitigated through dilution or the addition of an adsorptive medium, such as activated carbon. The use of activated carbon in AD processes in wastewater treatment systems has been reported to improve process stability, counteract inhibition (caused by toxic compounds), enhance removal efficiencies of recalcitrant and toxic pollutants, and provide immobilization of anaerobic microbes (Çeçen & Aktaş 2011). Therefore, the objective of this study is to investigate the feasibility of converting organics in PHWW to biogas via AD, in order to improve effluent water quality and increase the overall energy efficiency of the HTL process.

MATERIAL AND METHODS

PHWW and anaerobic inoculum

The PHWW used in this study was a mixture of PHWW from HTL of swine manure under different process conditions (pressure and reaction time). Key characteristics of the PHWW are provided in Table 1. Further characterization of water quality parameters for PHWW resulting from swine manure conversions under different HTL operating conditions was reported previously in detail by Appleford (2005). A mesophilic anaerobic inoculum was collected from a full-scale anaerobic digester operating at the Urbana-Champaign Sanitary District, and was used within 2 hours after collection. Characteristics of the anaerobic inoculum are also provided in Table 1.

Table 1

Characteristics of post-hydrothermal liquefaction wastewater (PHWW), supernatant of centrifuged sludge, and anaerobic sludge inoculum

Parameter PHWW Supernatant of centrifuged sludge Anaerobic sludge inoculum 
COD (g/l) 104,060 1,356 18,100 
TANa (g/l) 2,247 740 – 
Total nitrogen (g/l) 5,355 2,220 – 
Total phosphorus (g/l) 1,001 215 – 
pH 5.60 7.75 – 
Solids content (g/l) – – 29,500 
Suspended solids (g/l) – – 28,350 
Parameter PHWW Supernatant of centrifuged sludge Anaerobic sludge inoculum 
COD (g/l) 104,060 1,356 18,100 
TANa (g/l) 2,247 740 – 
Total nitrogen (g/l) 5,355 2,220 – 
Total phosphorus (g/l) 1,001 215 – 
pH 5.60 7.75 – 
Solids content (g/l) – – 29,500 
Suspended solids (g/l) – – 28,350 

aTotal ammonia nitrogen: the sum of ammonia nitrogen and ammonium nitrogen.

Anaerobic batch test design and set-up

Batch anaerobic tests were conducted to investigate (1) the anaerobic digestibility of PHWW and (2) the effect of adsorptive media on the AD of PHWW. Mesophilic batch digestion experiments were carried out at 37 °C in an incubation chamber. As shown in Table 2, reactors were loaded with a range of PHWW concentrations from 3.3% to 66.7%. The PHWW was diluted into the supernatant of the anaerobic inoculum sludge, which provided a growth medium suitable for the anaerobic culture. The supernatant was obtained after centrifuging the inoculum sludge at 4,000 rpm for 20 min. Powdered activated carbon (PAC) (Norit DARCO Ultra 100) was added to four of the reactors. All the reactors were sealed with rubber septa and caps. The headspace of the reactors was purged with N2 gas for 10 minutes to ensure anaerobic conditions. All of the reactors were manually mixed once a day for 30 seconds prior to measuring biogas volume. Control batch reactors containing only inoculum and sludge supernatant were also prepared.

Table 2

Batch experiment set-up for anaerobic digestion of post-hydrothermal liquefaction wastewater (PHWW)

ID PHWW (& of liquid volume) Supernatant of centrifuged sludge (& of liquid volume) Sludge inoculum (& of liquid volume) PAC dosage (g/l) Initial pH 
Control 0% 96.7% 3%  7.92 
Control* 0% 80.0% 20%  8.12 
3.3% 3.3% 93.3% 3%  7.67 
6.7% 6.7% 90.0% 3%  7.41 
13.3%* 13.3% 66.7% 20%  7.43 
26.7%* 26.7% 53.3% 20%  6.86 
33.3% 33.3% 63.3% 3%  6.64 
66.7% 66.7% 30.0% 3%  5.72 
6.7% + PAC 6.7% 90.0% 3% 1.4 7.41 
26.7% + PAC* 26.7% 53.3% 20% 2.4 6.86 
33.3% + PAC 33.3% 63.3% 3% 1.4 6.50 
66.7% + PAC 66.7% 30.0% 3% 1.4 5.82 
ID PHWW (& of liquid volume) Supernatant of centrifuged sludge (& of liquid volume) Sludge inoculum (& of liquid volume) PAC dosage (g/l) Initial pH 
Control 0% 96.7% 3%  7.92 
Control* 0% 80.0% 20%  8.12 
3.3% 3.3% 93.3% 3%  7.67 
6.7% 6.7% 90.0% 3%  7.41 
13.3%* 13.3% 66.7% 20%  7.43 
26.7%* 26.7% 53.3% 20%  6.86 
33.3% 33.3% 63.3% 3%  6.64 
66.7% 66.7% 30.0% 3%  5.72 
6.7% + PAC 6.7% 90.0% 3% 1.4 7.41 
26.7% + PAC* 26.7% 53.3% 20% 2.4 6.86 
33.3% + PAC 33.3% 63.3% 3% 1.4 6.50 
66.7% + PAC 66.7% 30.0% 3% 1.4 5.82 

*These four test conditions were conducted in 120 ml serum bottles with 100 ml operational/liquid volume while all the other test conditions were conducted in 200 ml serum bottles with 150 ml operational/liquid volume. In addition, these four tests had larger amount of inoculum (inoculum was 20% of total operational volume versus 3% in the other tests).

Measurements and calculations

Biogas production was regularly monitored using a water displacement column and flask filled with deionized water acidified to pH 2 using H2SO4. Biogas production was normalized by subtracting biogas production from the control reactors. The methane (CH4) concentration of the biogas in the reactors' headspace was analyzed using a Varian CP-3600 gas chromatograph (GC) equipped with a thermal conductivity detector as described previously (Zhou 2010). Filtered (0.45 μm pore size nylon filter) water samples before and after digestion were analyzed for chemical oxygen demand (COD) by visible light absorbance after dichromate digestion, according to standard methods (Clesceri et al. 1999). Two types of biogas production yields (ml biogas/mg CODadded and ml biogas/mg CODremoved) were calculated, by dividing the cumulative biogas production amount by either the COD added into the reactor or removed from the reactor (the COD removal from the control reactors was also ‘subtracted out’). The COD and NH4+-N concentrations in the influent/effluent during the biogas trials are shown in Table S1 in the supplementary information (available in the online version of this paper).

RESULTS AND DISCUSSION

Biogas production

Figure 1(a) shows cumulative biogas production from the anaerobic batch test with PHWW concentrations ranging from 3.3% to 66.7%. The response of the AD process to different PHWW concentrations was quite variable. At low concentrations of PHWW (3.3% and 6.7%), biogas production started promptly after the test began and steadily increased for about 25 days. At the end of the digestion period (65 days), these two conditions reached a biogas yield of about 0.32 ml/mg CODadded. Biogas samples from these two conditions were collected near the end of the steady biogas production phase (Day 27) and GC analysis showed that high quality biogas was produced with a methane content above 70% (Figure 1(c)). Assuming typical values of methane content to be 65%, this gives us an energy recovery efficiency of around 53% of the theoretical maximum (theoretical maximum conversion values for COD to methane was 0.25 mg CH4/mg CODadded, (0.39 ml CH4/mg CODadded) based on typical reaction stoichiometry (Metcalf & Eddy et al. 2003)).

Under higher PHWW concentrations (13.3% and 26.7%), a biogas production lag phase was observed. As shown in Figure 1(a) and (b), after about 8 days of negligible biogas production in the 13.3% PHWW condition and 35 days in the 26.7% PHWW condition, biogas production rose sharply and resulted in a final biogas yield of 0.35 ml/mg CODadded and 0.11 ml/mg CODadded, respectively, at the end of the 65-day digestion period. Since both of these biogas production curves were still increasing at the end of the test, the maximum biogas production potential of these two conditions is very likely to be higher than the highest value achieved during this study. Biogas samples from these conditions were collected and analyzed for methane content on Day 44, which was the early stage of steady biogas production for both. As shown in Figure 1(c), methane content in the 13.3% PHWW condition was 78%, while methane content in the 26.7% PHWW condition was lower at 58%. However, since the biogas was sampled at a very early stage in the case of the 26.7% PHWW condition, it is quite possible that the methane content would increase to a level similar to the conditions with lower PHWW concentrations.

For the highest PHWW concentrations tested, 33.3% and 66.7%, complete inhibition of the AD process was observed, indicated by almost no biogas production during the entire 65-day digestion period, as well as low methane content (sampled at Day 27), as shown in Figure 1.
Figure 1

Biogas production in batch anaerobic digestion reactors. (a) Cumulative biogas production for a 65-day incubation period containing various concentrations (3.3-66.7%) of post-hydrothermal liquefaction wastewater (PHWW). (b) Lag phase for biogas production. (c) Methane content of biogas sampled during exponential increase phase of biogas production (13.3% and 26.7% tests were sampled at Day 44, all the other tests were sampled at Day 27).

Figure 1

Biogas production in batch anaerobic digestion reactors. (a) Cumulative biogas production for a 65-day incubation period containing various concentrations (3.3-66.7%) of post-hydrothermal liquefaction wastewater (PHWW). (b) Lag phase for biogas production. (c) Methane content of biogas sampled during exponential increase phase of biogas production (13.3% and 26.7% tests were sampled at Day 44, all the other tests were sampled at Day 27).

The batch test results show that PHWW concentrations at or above 13.3% inhibited AD and resulted in an extended biogas production lag phase and eventual failure of the AD process with little to no biogas production. Several factors may have contributed to the observed inhibition. Ammonia is one possibility since the PHWW contained high levels of total ammonia nitrogen (TAN, the summation of ammonia nitrogen and ammonium nitrogen) (3.57 g/l). A wide range of inhibiting TAN concentrations for AD has been reported; 1.7 g/l to 14 g/l TAN has been reported to cause a 50% reduction in methane production (Chen 2008). In this study, reactors with 26.7%, 33.3%, and 66.7% PHWW showed fairly high TAN levels, (∼1.7 g/l, 1.9 g/l and 2.7 g/l NH4+-N, respectively). However, in all of our test conditions, the concentrations of free ammonia (NH3), the main cause of TAN inhibition, were below 100 g/l (calculated based on pH and temperature), which is believed to be acceptable for AD in general (Chen 2008). In addition, given that complete inhibition of biogas production occurred at the low end of reported inhibitory TAN levels (∼1.9 g/l), it would suggest that other significant inhibitory effects were most likely present. Similarly, the significant lag phase observed in the 13.3% and 26.7% PHWW conditions, which corresponded to TAN levels of 1.4 g/l and 1.7 g/l, respectively, suggests that other factors played a significant role in the inhibition of AD.

The low pH of PHWW may be a possible factor contributing to the inhibition of AD of PHWW. A pH range of 6.5 to 8.2 is generally considered suitable for successful AD processes (Speece 1983). Initial pH in the 66.7% PHWW condition was 5.82, which is lower than the recommended pH range. Thus, the complete inhibition observed in this test condition was in all likelihood significantly affected by the low pH. Although the pH in the 33.3% PHWW condition (pH = 6.67) was within the typical range of feasible AD, it is only slightly higher than the recommended minimum of 6.5, and we suspect this lower pH at least partially contributed to the inhibition of AD.

Toxic organic compounds may have contributed significantly to the inhibition observed especially in experimental conditions with high PHWW concentrations. A variety of organic compounds were identified in the PHWW from swine manure conversion, including sugars, isosorbide, indole, 3-amonio-phenol, 2-cyclopenten-1-one, carboxylic acids, ketones, alcohols, various cyclic hydrocarbons, and many nitrogen-containing compounds such as amides, azines and pyrroles (Appleford 2005). Many of these compounds are toxic to AD processes, including phenol and nitrophenols (Borja Alba & Banks 1997), alcohols (Demirer & Speece 1998), carboxylic acids (Stergar Zagorc-Konan & Zgajnar-Gotvanj 2003) and various cyclic compounds such as benzene and nitrobenzene (Bhattacharya Qu & Madura 1996), pyridine and its derivatives (Wu & Huang 1998). However, many factors can influence the toxicity during AD processes, including specific toxicant concentrations (inhibitory concentration levels vary widely for specific toxicants), biomass concentration, toxicant exposure time, cell age, feeding pattern, acclimation, and temperature (Chen 2008). In order to identify the specific toxicants that contributed to inhibition during AD of PHWW in this study, further investigation is needed, including characterization of the organic compounds in PHWW, and a series of toxicology studies for each individual compound as well as mixtures of compounds.

Compared to tissue cells or other microorganisms, anaerobic microbes seem to be relatively tolerant of the mixture of chemical compounds found in PHWW. Zhou (2010) reported that 5% of the same PHWW as was used in this study would completely inhibit algal growth. Other research has found that the organic compounds extracted from 7.5% PHWW from algal biomass conversion resulted in a 50% reduction in mammalian cell growth, and in this case the toxicity of the extracted PHWW only accounted for 10–20% of the toxicity of full strength PHWW (Pham et al. 2013; Pham, personal communication). Under certain concentrations of PHWW, the presence of the relatively long lag phase followed by significant biogas production indicates the ability of inhibited anaerobic microorganisms to adapt and acclimate to the toxic compounds in PHWW. Reversibility of toxic effects is commonly noted during anaerobic processes (Chen 2008). It is expected that long-term adaptation would reduce the lag phase of biogas production and eventually allow AD to proceed more quickly with higher concentrations of PHWW. Even more importantly, anaerobic microbes can break down many toxic compounds including a variety of polycyclic aromatic hydrocarbons and nitrogen heterocyclic compounds, such as phenol, phenol derivatives, picoline, and pyridine (Rabus 1995; Carmona 2009), which have been reported to cause toxicity in PHWW. Therefore, it is likely that the inhibitory effects of PHWW will gradually decrease as the AD process proceeds. This highlights the benefit of AD as a potential detoxification step for PHWW treatment. For instance, cultivating algae in PHWW offers significant potential advantages for nutrient recycling and bioenergy production (discussed in detail in the ‘Integration of anaerobic digestion with the hydrothermal liquefaction process’ section). However, algae are more sensitive to the toxicity of PHWW than are anaerobic microbes. Therefore, AD could be an advantageous pretreatment step to reduce the toxicity of PHWW before sending PHWW for use in algal cultivation.

The effect of activated carbon on the anaerobic digestion of PHWW

One way to enhance AD of PHWW is to use an adsorptive medium to sequester potentially toxic compounds and thereby mitigate potential inhibitory effects. For long-term operations in practical applications, granular activated carbon would likely be preferred because it can be easily retained in the reactor, whereas PAC would generally be lost with effluent and/or sludge removal unless special process/reactor configurations were used, such as a continuous membrane bioreactor (Ng Sun & Fane 2006) or upflow anaerobic sludge blankets. In this study, PAC was used because equilibrium could be achieved faster, and this expedited the experiment. Results from this study showed that activated carbon can indeed accelerate biogas production, reduce the lag phase, and facilitate successful digestion under higher concentrations of PHWW. Figure 2 compares biogas production, methane content, and biogas production lag phase from AD of PHWW with and without the addition of PAC. In conditions with a relatively low concentration of PHWW (6.7%), no obvious inhibition effect was observed in terms of a lag phase prior to biogas production (Figure 2(b)), and the addition of PAC only slightly increased the rate of biogas production (Figure 2(a)). The methane content of the biogas from digestion of 6.7% PHWW was not affected by the addition of PAC; both conditions showed a high methane content of 77% and 78%, respectively, at Day 27 (Figure 2(c)).
Figure 2

Comparison of biogas production and methane content for batch anaerobic digestion tests with and without the addition of powdered activated carbon (PAC). (a) Cumulative biogas production. (b) Lag phase for biogas production. (c) Methane content of biogas produced (26.7% and 26.7% + PAC tests were sampled on Day 44, all other tests were sampled on Day 27).

Figure 2

Comparison of biogas production and methane content for batch anaerobic digestion tests with and without the addition of powdered activated carbon (PAC). (a) Cumulative biogas production. (b) Lag phase for biogas production. (c) Methane content of biogas produced (26.7% and 26.7% + PAC tests were sampled on Day 44, all other tests were sampled on Day 27).

At a higher PHWW concentration of 26.7%, AD without PAC incurred a 35-day lag phase prior to biogas production. However, the addition of PAC resulted in a reduced lag phase of only 23 days. The shortened lag phase is valuable in practical applications especially during reactor start-up or recovery after an upset. Note that we do not anticipate that PAC addition would affect the final methane content of biogas, and it was expected that both the 26.7% and 26.7% + PAC test conditions would eventually achieve similar methane content. The observed difference in methane content of the 26.7% (58% methane content) and 26.7% + PAC (68% methane content) test conditions is likely related to the fact that the sampling point (Day 44) occurred during different stages of biogas production for these two test conditions: Day 44 was still in very early stages of biogas production for the 26.7% condition, but in the case of the 26.7% + PAC condition, steady biogas production was already underway. Methane content during the start-up of AD is normally lower than that during steady biogas production because methanogenesis is the last of several sequential steps in AD. Since the biogas production curves for both the 26.7% and 26.7% + PAC conditions were still increasing at the end of the test, the maximum biogas production potential of these two conditions is very likely to be higher than the highest value achieved during this study, and would be expected to reach the same biogas production levels as other successful tests (∼0.32 ml biogas/mg CODadded).

At an even higher PHWW concentration of 33.3%, almost no biogas production occurred without PAC. The addition of PAC mitigated the complete inhibition of AD, and biogas production commenced after a 34-day lag phase. Methane content in the 33.3% PHWW condition with PAC continued to increase, from 34% on Day 27 to 43% on Day 34 and 64% on Day 47, indicating successful AD. Note that neither the addition of PAC nor the PHWW concentration is expected to have affected the methane content in successful AD trials. Instead, the CH4 content in all successful trials was anticipated to be similar during steady biogas production. Again, the observed differences in methane content between the 33.3% + PAC condition and other successful AD trials (Figure 2C) was probably related to the fact that the sampling point occurred at different stages of biogas production, and methane content tends to be lower during the start-up or early stages. For the highest PHWW concentration condition (66.7%), the addition of PAC was not able to induce successful AD during the 65-day experimental period.

As discussed previously, low pH and toxic organic compounds in the PHWW may have both contributed to the inhibition of AD. However, in determining the mechanism by which the addition of PAC was able to alleviate the inhibiting effects of PHWW, it was found that addition of PAC had a very limited effect on pH and ammonia levels. Less than a 5% change in these two parameters was observed during the first week of testing when almost all of the physical adsorption would have taken place. Thus, the process enhancement resulting from addition of PAC was most likely a result of a reduction in toxic organic compounds in the aqueous phase. Activated carbon is well known for its ability to adsorb a wide range of organic compounds, including toxic compounds identified in PHWW such as indole, alcohols, ketones, various cyclic hydrocarbons, and many nitrogenous organic compounds, (e.g. amides, azines, pyrroles, amino-phenol, and pyridine and its derivatives) (Appleford 2005; Pham et al. 2013). Pham et al. (2013) found that the treatment of PHWW by activated carbon can greatly reduce the cytotoxicity to mammalian cells. In AD of coal gasification wastewater, which shares some similar chemical compounds with PHWW (various aromatic compounds and nitrogen-containing compounds, including pyridine, methyol-, dimethyl-, and ethyl-substituted pyridines), Suidan et al. (1983) reported that the presence of granular activated carbon served to sequester the inhibitory and non-biodegradable components of the wastewater and permit the breakdown of degradable compounds during AD.

There are several possible mechanisms for the reduction of organic toxic compounds in the presented study. The reduction may be a result of pure physical adsorption of toxic compounds. The adsorption process generally requires only minutes to days to achieve equilibrium, depending on the adsorbent chosen, which is significantly faster than the net rates for AD. Thus, PAC serves as a rapid physico-chemical sink for inhibitory or toxic substrates, allowing the microbes to survive within an otherwise adverse environment. It is also possible that simultaneous adsorption and biodegradation of toxic compounds occurred (Aktaş & Çeçen 2007) resulting in continuous bioregeneration of the adsorbent in situ. In this case, the activated carbon serves as a buffer, where the initial adsorption process helps to decrease toxic compound levels, and thereby enhances microbial growth and survival. As the remaining aqueous phase toxic compounds are degraded, adsorbed compounds slowly desorb and are re-released into the aqueous phase. Microbes then have an opportunity to degrade the released compounds without experiencing significant inhibitory effects. Bioregeneration of activated carbon has been reported and discussed in several previous studies, including AD of coal gasification wastewater (Suidan et al. 1983) as well as the aerobic biodegradation of phenolic compounds (Lee & Lim 2005) and aromatic compounds (De Jonge et al. 1996).

The production of activated carbon is energy intensive, and a critical question is whether enhancement of AD through addition of activated carbon resulted in an overall benefit for energy recovery. Therefore, the net energy recovery efficiency was calculated taking into consideration the energy spent on producing activated carbon, both virgin and recycled (Bayer et al. 2005), as shown in Figure 3 (see Table S2 in the supplementary information for detailed calculations, available in the online version of this paper). A negative net energy recovery efficiency indicates that the energy spent on producing activated carbon is more than the energy recovered in the form of methane, and therefore the addition of activated carbon is not favorable. If the efficiency decreases compared to the condition without activated carbon, it indicates that the energy content in the additional CH4 recovered as a result of adding activated carbon is still less than the energy spent on producing activated carbon, and thus is also not favorable from an energy perspective. As can be seen in Figure 3, for the 6.7% PHWW condition, the addition of activated carbon resulted in a decreased energy recovery efficiency. This is because activated carbon did not improve the quantity of biogas produced, according to our experimental result. For the 26.7% conditions, although the addition of activated carbon resulted in decreased energy recovery efficiencies, it also reduced the lag phase for biogas production by 34% (see Figure 2(b)), which can be a valuable feature for practical applications. For the 33.3% PHWW conditions, the advantage of adding activated carbon was very clear: energy recovery efficiencies increased from zero to 40% and 49% with virgin and recycled activated carbon, respectively. For the 66.7% PHWW conditions, no biogas was produced even after the addition of activated carbon, resulting in negative energy recovery efficiencies. Therefore, it can be seen that under certain conditions the addition of activated carbon to AD of PHWW can improve the overall energy recovery efficiency. As discussed earlier, it is also worth mentioning that in situ bioregeneration during AD can greatly decrease the need to periodically replace or thermally regenerate the activated carbon, thus resulting in significant improvements to the net energy recovery efficiency with activated carbon.
Figure 3

Comparison of net energy recovery efficiency for trials without activated carbon, with virgin activated carbon, and with recycled activated carbon.

Figure 3

Comparison of net energy recovery efficiency for trials without activated carbon, with virgin activated carbon, and with recycled activated carbon.

Total organic matter removal

Biogas production in the conditions with 3.3%, 6.7% and 13.3% PHWW, which achieved nearly plateaued levels, were tested for total dissolved organic matter removal. Results showed that 50%, 55%, and 45% of organic matter was removed, respectively, as indicated by COD concentrations. PHWW had been reported to have limited biodegradability in another previous study, although this study used aerobic biodegradation (Zhou 2010). Further research is needed to characterize the residual un-digested compounds and develop effective methods to remove them. Potential methods include adsorption (with activated carbon or resins), ozone oxidation, and/or a second stage biological treatment with aerobic or other specialized microbial cultures.

Based on the rate of cumulative biogas production and total organic removal, calculations indicated that the conditions with 3.3%, 6.7% and 13.3% PHWW had biogas production efficiencies of 0.49 ml, 0.50 ml, and 0.50 ml biogas/mg CODremoved, respectively. Given that the methane content for all of these conditions at a single time point during steady-state biogas production was above 70% (Figure 1(c)), indicating successful and efficient AD, we assume the overall methane production to be around 60–70% of the total biogas produced (Metcalf & Eddy et al. 2003) (methane content of the biogas was not routinely measured throughout this study). In this case, AD of these three conditions resulted in an estimated methane yield of around 0.3–0.35 ml methane/mg CODremoved. This estimated methane yield is fairly good in comparison to AD of other types of high strength industrial wastewater, where methane yield has been reported to generally be in the range of 0.25–0.35 ml methane/mg CODremoved (Ersahin et al. 2011). This comparison highlights that PHWW is a suitable feedstock for AD processes, and can achieve fairly efficient carbon/energy recovery from PHWW.

Integration of anaerobic digestion with the hydrothermal liquefaction process

Anaerobic digestion can provide multiple benefits when synergistically integrated with HTL. First, the overall carbon efficiency and energy production of HTL could be increased by recovering organic carbon from PHWW in the form of methane-rich biogas. The biogas could be used to provide the heat needed for HTL, which would effectively reduce the parasitic energy demand. Successful AD could recover 53% of energy from PHWW (as discussed in the ‘Biogas production’ section), which can increase the overall bioenergy conversion efficiency of HTL (energy content in biofuel products over the energy content in incoming feedstock) from 40–65% (Gai et al. 2014) to as much as 80%. In addition, AD could potentially function as a pretreatment step before sending PHWW for algal cultivation, to facilitate nutrient recycling and recovery. PHWW contains up to 80% of the incoming feedstock nutrients (e.g. nitrogen) (Yu et al. 2011), which provides great potential for nutrient recycling and additional biomass production if used for algae cultivation (Jena et al. 2011; Biller et al. 2012). However, PHWW has been shown to be toxic to algae, most likely because of toxic organic compounds and ammonia nitrogen, etc. (Pham et al. 2013; Zhou et al. 2013). Anaerobic microbes have been widely reported to break down many toxic compounds including a variety of polycyclic aromatic hydrocarbons and nitrogen heterocyclic compounds, such as phenol, phenol derivatives, picoline, and pyridine (Rabus 1995; Carmona 2009), which have been reported to cause toxicity in PHWW. In addition, the present study has shown that ∼50% of the organic content in PHWW could be removed through AD. Therefore, it is expected that AD can serve as a detoxification step for PHWW before recycling it for algal biomass production, which facilitates beneficial nutrient reuse and provides additional bioenergy feedstock for HTL (Zhou et al. 2013).

CONCLUSIONS

This study investigated the feasibility of recovering carbon/energy from PHWW through AD. Results showed that AD was successful when treating appropriate concentrations of PHWW (≤6.7%), which yielded ∼ 0.5 ml biogas/mg CODremoved and ∼53% energy recovery efficiency. Higher PHWW concentrations (≥13.3%) showed an inhibitory effect on AD with delayed, reduced, or even no biogas production. Activated carbon mitigated the inhibition by shortening the lag phase of biogas production and allowing AD to occur at higher concentrations of PHWW (up to 33.3%). The addition of activated carbon was also shown to increase the net energy recovery efficiency for AD of 33.3% PHWW when taking into account the energy spent on producing activated carbon. Thus, AD is a feasible and advantageous process for energy recovery from PHWW.

ACKNOWLEDGEMENTS

This material is based upon work that was supported in part by the National Institute of Food and Agriculture, US Department of Agriculture, under award number 2014-67019-21568 and Dudley Smith Initiative. We would also like to acknowledge the support of the Clean Energy Education Initiative funded by the National Science Foundation and the Graduate College at the University of Illinois at Urbana-Champaign.

REFERENCES

REFERENCES
Aktaş
Ö.
Çeçen
F.
2007
Bioregeneration of activated carbon: A review
.
Int. Biodeterior. Biodegrad.
59
,
257
272
.
Appleford
J. M.
2005
Analyses of the products from the continuous hydrothermal conversion process to produce oil from swine manure. Master's thesis, University of Illinois at Urbana-Champaign, Urbana, IL, USA
.
Bhattacharya
S. K.
Qu
M.
Madura
R. L.
1996
Effects of nitrobenzene and zinc on acetate utilizing methanogens
.
Water Res
.
30
,
3099
3105
.
Biller
P.
Ross
A. B.
Skill
S. C.
Lea-Langton
A.
Balasundaram
B.
Hall
C.
Riley
R.
Llewellyn
C. A.
2012
Nutrient recycling of aqueous phase for microalgae cultivation from the hydrothermal liquefaction process
.
Algal Res
.
1
,
70
76
.
Çeçen
F.
Aktaş
Ö.
2011
Activated Carbon for Water and Wastewater Treatment: Integration of Adsorption and Biological Treatment
.
Wiley-VCH Verlag, Weinheim, Germany
.
Chen
Y.
2008
Inhibition of anaerobic digestion process: A review
.
Bioresour. Technol.
99
,
4044
4064
.
Clesceri
L. S.
Greenberg
A. E.
Andrew
D. E.
1999
Standard Methods for the Examination of Water and Wastewater
.
American Public Health Association/American Water Works Association/Water Environment Federation
,
New York
.
De Jonge
R. J.
Breure
A. M.
Van Andel
J. G.
1996
Bioregeneration of powdered activated carbon (PAC) loaded with aromatic compound
.
Water Res
.
30
,
875
882
.
Ersahin
M. E.
Ozgun
H.
Dereli
R. K.
Ozturk
I.
2011
Anaerobic treatment of industrial effluents: an overview of applications
. In:
Waste Water: Treatment and Reutilization
(
Einschlag
F.
, ed.).
InTech
,
New York
, pp.
3
28
.
Jena
U.
Vaidyanathan
N.
Chinnasamy
S.
Das
K. C.
2011
Evaluation of microalgae cultivation using recovered aqueous co-product from thermochemical liquefaction of algal biomass
.
Bioresour. Technol.
102
,
3380
3387
.
Metcalf & Eddy
,
Tchobanoglous
G.
Burton
F. L.
& Stensel
H. D.
2003
Wastewater Engineering: Treatment and Reuse
,
4th edn
.
McGraw-Hill
,
Boston, USA
.
Ng
C. A.
Sun
D.
Fane
A. G.
2006
Operation of membrane bioreactor with powdered activated carbon addition
.
Separ. Sci. Technol
.
41
,
1447
1466
.
Pham
M.
Schideman
L.
Scott
J.
Rajagopalan
N.
Plewa
M.
2013
Chemical and biological characterization of wastewater generated from hydrothermal liquefaction of Spirulina
.
Environ. Sci. Technol.
47
,
2131
2138
.
Rabus
R.
1995
Anaerobic degradation of ethylbenzene and other aromatic hydrocarbons by new denitrifying bacteria
.
Arch. Microbiol.
163
,
96
103
.
Speece
R. E.
1983
Anaerobic biotechnology for industrial wastewater treatment
.
Environ. Sci. Technol.
17
,
416
427
.
Stergar
V.
Zagorc-Konan
J.
Zgajnar-Gotvanj
A.
2003
Laboratory scale and pilot plant study on treatment of toxic wastewater from the petrochemical industry by UASB reactors
.
Water Sci. Technol
.
48
,
97
102
.
Suidan
M. T.
Strubler
C. E.
Kao
S.
Pfeffer
J. T.
1983
Treatment of coal gasification wastewater with anaerobic filter technology
.
Journal (Water Pollution Control Federation)
55
,
1263
1270
.
Zhou
Y.
2010
Improving algal biofuel production through nutrient recycling and characterization of photosynthetic anomalies in mutant algae species. Master's thesis, University of Illinois at Urbana-Champaign, Urbana, IL, USA
.

Supplementary data