Catalyzed and non-catalyzed APLs were produced by pyrolysis of commercially available, dried biosolids at 800°C. The biosolids were a mix of anaerobically digested primary sludge and raw waste activated sludge (Milorganite) from the Jones Island Water Resource Recovery Facility (Milwaukee, Wisconsin) as described elsewhere (Liu et al., 2017). Air stripping was utilized to reduce the NH₃-N concentration in some APL samples before anaerobic digestion. For this, 30 mL of catalyzed or non-catalyzed APL was aerated (2 L/min, 1 atm, 20°C) for 9 h to air-strip NH₃-N.

ATA tests were performed to measure the effect of APLs on the rate of methane production from acetate (Owen et al., 1979; Speece, 2008). For ATA analysis, each of six serum bottles was fed a different APL concentration (Table 1) and all received 10 g/L calcium acetate with basal nutrient media (Speece, 2008). Serum bottles were inoculated with biomass taken from a lab-scale anaerobic digester fed non-fat dry milk and basal nutrients media. NH₃-N loadings in bottles receiving air-stripped APL were significantly lower than those in bottles receiving non-air-stripped APL (Table 1). Therefore, toxicity from constituents other than NH₃-N was identified using results from bottles receiving air-stripped APL. In addition, preliminary ATA tests were conducted using ammonium chloride to determine the response of the inoculum to NH₃-N.

Biogas volume was measured daily using a 100 mL wetted-barrel glass syringe by inserting a needle through serum bottle septa. Biogas methane concentration was measured by gas chromatography (GC System 7890A, Agilent Technologies, Irving, TX, USA) with a thermal conductivity detector GC-TCD. VFA concentrations were measured by gas chromatography (GC System 7890A, Agilent Technologies, Irving, TX, USA) using a flame ionization detector (GC-FID). Total solids (TS), volatile solids (VS), total suspended solids (TSS), volatile suspended solids (VSS), and NH₃-N concentrations were measured using standard methods (American Public Health Association, 1998). The pH was measured using a probe and meter (Orion 4 Star, Thermo, Waltham, MA, USA). Statistical analyses including average and standard deviation calculations were performed using Microsoft Excel 2015.

APL constituent concentrations were measured using two different gas chromatography-mass spectrophotometry (GC-MS) methods and two different GC-FID methods. The first GC-MS method employed a DB wax column (Agilent Technologies), whereas the second employed an HP-5MS column (Agilent, Technologies). The GC-MS DB wax column analysis was performed using a GC-MS system (HP-6890 GC with a 5972A mass selective detector) and column with 30 m × 0.53 mm ID × 1.0 μm film thickness (DB-Wax column, Agilent Technologies). The injector employed a split ratio of 50:1 and the injection volume was 1 μL. The GC oven was programmed with an initial temperature of 50°C for 1 min, temperature ramp of 15°C/min and a final temperature of 220°C and final hold time of 40 min. The flow rate was constant at 2.0 mL/min of the helium carrier gas. Both catalyzed and non-catalyzed APLs were diluted with water at ratios of 1:18.7 and 1:45.6, respectively. Also, samples were acidified to pH = 1 using concentrated HCl solution. The second GC-MS analysis was performed using a GC-MS system (HP-7890A GC with a 5975A mass selective detector) fitted with an HP-5MS column with 30 m × 0.25 mm ID × 1.0 μm film thickness (HP5-MS column, Agilent Technologies). The injection volume was 0.5 μL with a split ratio of 10:1. The oven was programmed with an initial temperature of 50°C for 1 min, temperature ramp of 10°C/min and a final temperature of 250°C and final hold time of 4 min. The flow rate was constant at 1.2 mL/min of the helium carrier gas. Both catalyzed and non-catalyzed APLs were undiluted.

The first GC-FID analysis was performed using a GC (GC System 7890B, Agilent Technologies, Irving, TX, USA) that employed a DB FFAP column with 30 m × 0.25 mm ID × 0.25 μm film thickness. The operating system used was OpenLAB CDS—Acquisition Build version 2.1.0.433. The carrier gas was helium with a constant flow rate of 18 mL/min. The helium make-up was 32 mL/min, hydrogen flow at 35 mL/min with an air flow of 300 mL/min. The oven initial temperature was 95°C for 2 min, ramped at 10°C/min and a final temperature of 200°C at 40°C/min with 5 min hold time. A sample volume of 5 μL was injected with a split ratio of 2:1.

The second GC-FID method (GC-FID, Bruker Corporation, Bruker Daltonics, Inc., USA) utilized a 1701 capillary column as described elsewhere (Rover et al., 2015). Quantification of organic compounds by GC-FID was conducted using standard curves for each analyte as described previously (Choi et al., 2014) and concentrations were measured as the mass of analyte per mass of APL (% w/w, raw basis).

Autocatalysis during pyrolysis reduced the APL COD from >200 gCOD/L for non-catalyzed APL to 32 ± 1.2 gCOD/L for catalyzed APL (~85% COD reduction) (Table 2). The catalysis effect on NH₃-N concentration varied, with non-catalyzed APL NH₃-N concentration of 63 ± 0.1 g/L and catalyzed APL NH₃-N concentration from 32 to 72 g/L. High NH₃-N concentration in the liquid product obtained from sewage sludge pyrolysis was highlighted previously (Park et al., 2008; Fonts et al., 2009). Concentrations of VFAs in APLs were below detection after autocatalytic pyrolysis (Table 3B). In contrast, non-catalyzed APL contained approximately 24 g/L of acetic acid (Table 3A). This is consistent with previous studies which have found high acetic acid concentration in APL (Torri and Fabbri, 2014). Acetic acid is an easily degradable substrate in anaerobic digestion; therefore, the presence of acetic acid indicates that APL could potentially serve as a co-digestate. The majority of organic compounds detected by GC-MS were nitrogen-containing compounds in both non-catalyzed and catalyzed APLs (Tables 3A,B). The highest percent peak area measured by GC-MS HP-5MS column analysis was observed for pyrrole in both non-catalyzed and catalyzed APLs. Acetic acid concentration measured by GC-MS DB wax column analysis was three times higher in non-catalyzed compared to catalyzed APL. However, peak areas of identifiable compounds detected by both GC-MS methods only contributed to a fraction of the total peak area observed, whereas the majority of peaks were unidentified (Tables 3A,B).

Nearly all the compounds identified by GC-FID in both APLs were aromatic hydrocarbons (Tables 3A,B). Compounds detected by GC-FID in non-catalyzed APL included phenols, xylene, ethylbenzene, cresol and acetophenone. Among these compounds, phenol, m-cresol, p-cresol, and 2,4-dimethylphenol contributed to a significant fraction of the detected substances (1.39% by weight) and are known to inhibit anaerobic processes at IC₅₀ concentrations of 2,100, 890, 91, and 71 mg/L, respectively (Fedorak and Hrudey, 1984; Blum and Speece, 1991). Phenolic compounds damage the cell membrane which adversely affects membrane proteins and cell wall permeability (Madigou et al., 2016). A recent study has summarized that selection of key microbial populations would help improve the removal of refractory pollutants including phenol and p-cresol (Franchi et al., 2018).

In addition, ethylbenzene and m-xylene are not known to be degraded under strictly anaerobic conditions and, therefore, do not serve as a substrate for methane production (Edwards and Grbic-Galic, 1994). Catalyzed APL contained fewer GC-FID-identified organic constituents compared to non-catalyzed APL (Table 3B). Only 4 organic compounds were detected by GC-FID in the catalyzed APL; 3,5-dimethoxy-4-hydroxybenzaldehyde, 2,5-dimethoxybenzyl alcohol, ethylbenzene, and styrene, whereas in the non-catalyzed APL, 23 compounds were identified.

Both catalyzed and non-catalyzed APLs contained high concentrations of NH₃-N that could inhibit anaerobic digestion (Table 2). NH₃-N concentrations higher than ~3–5 g/L NH₃-N at neutral pH and 35°C exert toxicity and decrease unacclimated methanogenic activity (Speece, 1996; Liu and Sung, 2002). Air stripping decreased NH₃-N concentration by approximately 80% in both non-catalyzed and catalyzed APL (Table 2). Air stripping did not significantly change the COD concentration in non-catalyzed APL; however, it reduced the catalyzed APL COD concentration by 30%, indicating catalyzed APL contained a higher fraction of volatile organic carbon than non-catalyzed APL. The pH of both catalyzed and non-catalyzed APLs significantly decreased after aeration, ostensibly due to NH₃-N removal (Table 2).

GC-FID analyses of catalyzed and non-catalyzed APLs after air stripping were conducted (Tables 4A,B). Before and after air stripping, essentially the same compounds were present in the non-catalyzed APL. In catalyzed APL after air stripping, however, only 2 compounds were detected; 3,5-dimethoxy-4-hydroxybenzaldehyde and 2,5-dimethoxybenzylalcohol. Therefore, air stripping of catalyzed APL removed the volatile organic chemicals, ethylbenzene, and styrene. The IC₅₀ values of ethylbenzene and styrene to unacclimated methanogenic biomass are each approximately 150 mg/L (Blum and Speece, 1991; Araya et al., 2000). Removing these potentially inhibitory chemicals from APL in addition to NH₃-N by air stripping may have resulted in reduced APL toxicity.

ATA testing was performed on both non-air-stripped, catalyzed and non-catalyzed APL samples as well as on air-stripped catalyzed APL. Since the non-catalyzed APL NH₃-N concentration was significantly below potentially inhibitory concentrations in ATA tests, no ATA testing was not performed on air-stripped, non-catalyzed APL (Table 1).

Non-catalyzed APL significantly inhibited methane production at high concentrations of 2.3 and 4 gCOD/L (Figure 1A). However, at lower concentrations of 0.06 and 0.30 gCOD/L, no significant inhibition was observed, and the initial methane production rates were similar to those of the control system that received no APL. This shows that APL digestion may be feasible at low organic loading rates, but toxicity inhibits the process at high loading rates. The inhibition could be attributed to the presence of compounds such as phenols that are known to adversely affect unacclimated methane-producing cultures at high concentrations (e.g., IC₅₀ for phenol is 2,100 mg/L) (Hübner and Mumme, 2015). Others have also suggested that methane production inhibition could be due to APL constituents including phenols, oxygenated and nitrogen-containing organics (Torri and Fabbri, 2014). In addition, high concentration of phenols in unacclimated anaerobic digesters impedes AD processes resulting in decreased biogas production (Madigou et al., 2016). Metabolic inhibition and modification of community composition are phenomena that negatively affect phenol degradation efficiency (Rosenkranz et al., 2013).

For catalyzed APL, high concentrations of 2.3 and 4.0 gCOD/L, caused total cessation of methane production (Figure 1B). It was originally assumed that the catalyzed APL would be less toxic than non-catalyzed APL since it contained fewer organic compounds; however, the opposite was observed, and catalyzed APL was more inhibitory than the non-catalyzed APL. The methane production rates for 0.3 and 0.6 gCOD/L APL doses were also lower than those of the controls. The lowest COD loading (0.06 gCOD/L) did not inhibit methane production rate.

Methane production data were analyzed to determine the concentration of APL that inhibited methane production rate by 50% (i.e., the IC₅₀ value). The IC₅₀ values were 0.3 and 2.3 gCOD/L for catalyzed and non-catalyzed APLs, respectively (Figure 1C). The highest APL COD loadings that demonstrated < 10% inhibition were approximately 0.5 and 0.1 g COD/L for non-catalyzed and catalyzed APL, respectively.

In order to determine the toxicity of NH₃-N to the seed biomass, an ATA analysis was performed with NH₃-N concentration range of 0, 3, 6, and 9 g/L (Figure 2A). High NH₃-N concentrations of 6 and 9 g/L NH₃-N resulted in high methane production inhibition, whereas a lower concentration of 3 g/L NH₃-N, resulted in low, but observable inhibition. The NH₃-N IC₅₀ value was 2 g/L NH₃-N for the seed biomass (Figure 2B). NH₃-N inhibition can be due to multiple factors including a change in the intracellular pH and inhibition of specific enzyme reactions (Rajagopal et al., 2013). It has also been proposed that NH₃ is a cause of inhibition because it can permeate the cellular membrane and lead to protein imbalance and/or potassium deficiency (Speece, 1996; Yenigün and Demirel, 2013).

ATA testing was performed on catalyzed APL after NH₃-N was removed by air stripping. After NH₃-N was stripped, the NH₃-N concentration in all the ATA bottles was <1 g/L and should not have been significantly inhibitory (Table 1). Therefore, any methane production rate inhibition was ostensibly a result of APL constituents other than NH₃-N. ATA results using catalyzed APL after air stripping show that aeration did reduce toxicity at the 0.6 gCOD/L dose, but inhibition was still observed at higher concentrations (Figure 3A). Even after NH₃-N was stripped, methane production essentially ceased at 2.3 and 4 gCOD/L APL doses. NH₃-N concentrations in these two systems were approximately 0.5 and 0.9 g/L, respectively, which were significantly lower than IC₅₀ value of 2 g/L. Therefore, NH₃-N toxicity was not the main cause of inhibition and other APL constituents inhibited methane production.

Dose response curves for methane production rate vs. COD and NH₃-N concentrations were plotted and IC₅₀ values were calculated (Figures 3B,C). After air stripping, biomass showed higher tolerance to the APL and the IC₅₀ value increased by approximately 300% (Figure 3B). Under both air-stripped and non-air-stripped conditions, the NH₃-N concentrations were significantly less than the IC₅₀ value, but methane production inhibition was still observed (Figure 3C). Therefore, NH₃-N was not the main or only inhibitory constituent. Inhibitory organic constituents in the APL exerted significant toxicity to the unacclimated biomass.

In summary, higher inhibition was observed using catalyzed vs. non-catalyzed APL despite the catalyzed APL having a lower COD concentration and fewer aromatic compounds. Energy recovery from APL may require pretreatment processes to decrease toxicity of the APL, prior to anaerobic digestion. Utilizing acclimated anaerobic biomass as an inoculum should also be considered to reduce inhibition and improve co-digestion. The APL for this work was produced at 800°C. It may be that APL produced from biosolids at other temperatures has different anaerobic toxicity properties. In addition, the APL produced for this work was derived from a mix of digested primary and raw waste activated sludges; the mix had a relatively high organic nitrogen content. APL derived from other traditional pyrolysis feedstock with lower nitrogen content, such as corn stover or wood, produces APL with lower total nitrogen and probably has different toxicity effects in anaerobic digesters.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.