Abstract

Anaerobic digestion (AD) is a proven technology for energy production from the stabilization and reduction of sewage waste. The AD and impact of ultrasonic pretreatment of four waste activated sludges (WASs) from conventional and three non-conventional municipal wastewater treatment plants were investigated. WAS from a conventional activated sludge (CAS) system, a rotating biological contactor (RBC), a lagoon, and a nitrifying moving-bed biofilm reactor (MBBR) were pretreated with ultrasonic energies of 800–6,550 kJ/kg total solids to illustrate the impact of sludge type and ultrasonic pretreatment on biogas production (BGP), solubilization, and digestion kinetics. The greatest increase in BGP over the control of pretreated sludge did not coincide consistently with greater sonication energy but occurred within a solubilization range of 2.9–7.4% degree of disintegration and are as follows: 5% ± 3 biogas increase for CAS, 12% ± 9 for lagoon, 15% ± 2 for nitrifying MBBR, and 20% ± 2 for RBC. The effect of sonication on digestion kinetics was inconclusive with the application of modified Gompertz, reaction curve, and first-order models to biogas production. These results illustrate the unique response of differing sludges to the same levels of sonication energies.

INTRODUCTION

Biological treatment is the conventional means of treating municipal wastewaters. However, the conversion of soluble substrates into biomass results in a stream of waste activated sludge (WAS) that requires further management. Anaerobic digestion (AD) is an important sludge management tool that is used to degrade biomass, reduce waste volume, and stabilize the organic, putrescible content of the sludge (Appels et al. 2008a). As a by-product, AD produces a renewable, energy-rich biogas (BG) which can be used to reduce the energy requirements of the wastewater treatment processes of municipal plants (Shen et al. 2015).

AD is a multistage process involving hydrolysis, acidogenesis, acetogenesis, and methanogenesis (Pavlostathis & Giraldo-Gomez 1991; Appels et al. 2008a). A cost intensive detraction to AD is its long hydraulic retention time (HRT) relative to aerobic digestion of sludge, resulting in larger footprints and higher capital costs. The rate limiting phase for the AD of sludge generated from municipal wastewater treatment is hydrolysis (Pavlostathis & Giraldo-Gomez 1991; Barber 2005). Digestion can be enhanced through the application of various pretreatment strategies designed to disintegrate the recalcitrant cell material of WAS resulting in a shorter hydrolysis phase and a smaller fraction of undigested material in the effluent (Bougrier et al. 2006; Carrère et al. 2010; Alqaralleh et al. 2015; Cano et al. 2015). This reduces the HRT requirements and increases biogas production (BGP).

Ultrasound is a promising pretreatment technology that has been demonstrated at full-scale to yield 3–10 kW of energy from increased methane production for every kilowatt of applied ultrasonic power (Barber 2005; Xie et al. 2007; Cano et al. 2015). The hydromechanical sheer stresses induced by cavitation disrupt sludge flocs and rupture microbial cell walls, thus releasing intracellular material and soluble organic matter (Wang et al. 2005). Sonicated sludge can also indirectly improve AD by enhancing the buffering capacity of anaerobic phases sensitive to acid accumulation through increased system alkalinity and formation of bicarbonate. Methanogenic biomass can also be improved by 45–140% with sonication densities of 0.18–0.52 W/mL (Mao & Show 2007). Ultrasonic pretreatment has been shown to unfavourably reduce the dewaterability of pre-digested sludge, yet enhance the dewaterability after AD, thereby reducing the cost of residual solids treatment (Şahinkaya & Sevimli 2013).

The investigation of ultrasonic pretreatment on sludges collected from different municipal wastewater treatment plants (WWTPs) has resulted in a wide range of reported performance parameters such as optimal energies, solubilizations, and biogas increases (BGI) (Appels et al. 2008b; Cano et al. 2015). The current literature is lacking with respect to ultrasonic pretreatment research on sludges generated by lagoons and attached growth, biofilm technologies. The different structure of the biomaterial and, in particular, the extracellular polymeric substances (EPS) component of biofilm sludge, as compared to conventional suspended systems, may have an impact on the efficacy of sonication. For example, nutritional, environmental, and operational conditions have been shown to have an impact on the EPS matrix of a rotating biological contactor (RBC) and activated sludges (Martín-Cereceda et al. 2001).

In this work, we investigate the impact of sonication and AD on a conventional sludge and three alternative sludges that are not conventionally digested anaerobically. In particular, the impact of various energy levels of sonication on BGP and solubilization were quantified through soluble chemical oxygen demand (sCOD) analysis and bioassays. Linear and non-linear regression models based on cumulative BGP were applied to compare and interpret digestion results. These investigations provide necessary information for the application of ultrasonic pretreatment to alternative sludges to determine potential viability and BG response of the AD of varying sludges.

MATERIALS AND METHODS

Sludge sources

Sludge was collected from four municipal WWTPs operating suspended and attached growth, biofilm biological treatment technologies. The inoculum and four waste sludges were characterized for total chemical oxygen demand (tCOD), sCOD, total solids (TS), and volatile solids (VS) (Table 1). The first source of sludge for this study was thickened WAS (TWAS) that was collected from the Robert O. Pickard Environmental Centre (ROPEC) conventional activated sludge (CAS) facility, located in Ottawa, Canada. The plant is designed to treat an average of 545,000 m3/day with a solid retention time (SRT) of 5–7 days. ROPEC was not operated to achieve nitrification. The bacteriological seed for AD (inoculum) was the same for all biogas tests during the study. Inoculum was collected from the mesophilic anaerobic digestors of the ROPEC facility operating at a 48/52% mixture of primary sludge and TWAS with an SRT of 20 days. The second source of sludge in this study was harvested from the Water Pollution Control Plant in Wendover, Canada and was sludge produced by a biofilm treatment system. This facility conducts secondary, biological treatment through the use of three RBCs with a maximum capacity of 1,260 m3/day of municipal sewage. As the system is an attached growth technology, the SRT values are unknown. The sludge of the RBC technology is produced through the erosion, abrasion, and potential sloughing off excess biofilm mass from the rotating contactors. The third source of sludge was collected from the municipal WWTP at Masson-Angers, Canada that consists of four aerated lagoons in series to treat a combined maximum annual volume of 230,000 m3 without achieving nitrification. Settled waste sludge was harvested from the fourth lagoon in the multilagoon system (Delatolla & Babartusi 2005). Sludge removal records from the lagoons indicate that the age of the harvested sludge was 5 years. The fourth source of sludge was a temporary, post-carbon-removal, nitrifying moving-bed biofilm reactor (MBBR) pilot system that was fed the final effluent from the Masson-Angers lagoon (Young et al. 2016). As the MBBR pilot is an attached growth technology, the SRT of the system is unknown. Sludge of the nitrifying MBBR technology is produced through the erosion of biofilm carriers that are kept in constant motion in the MBBR basins (Karizmeh et al. 2014; Forrest et al. 2016).

Table 1

Characteristics of inoculum and harvested sludges

Source tCOD (mg/L) sCOD (mg/L) TS (%) VS (%) VS/TS (%) 
Inoculum 21,200 ± 400 11,300 ± 100 1.48 ± 0.01 0.88 ± 0.01 59.5 ± 0.7 
CAS 73,200 ± 1,100 15,400 ± 300 6.67 ± 0.02 4.43 ± 0.03 66.4 ± 0.5 
Lagoon 24,100 ± 1,200 1,200 ± 200 3.64 ± 0.05 1.51 ± 0.03 42.0 ± 1.0 
RBC 25,200 ± 200 420 ± 10 3.02 ± 0.02 1.70 ± 0.01 56.3 ± 0.5 
MBBR 20,800 ± 2,200 515 ± 3 2.18 ± 0.02 1.53 ± 0.01 70.2 ± 0.8 
Source tCOD (mg/L) sCOD (mg/L) TS (%) VS (%) VS/TS (%) 
Inoculum 21,200 ± 400 11,300 ± 100 1.48 ± 0.01 0.88 ± 0.01 59.5 ± 0.7 
CAS 73,200 ± 1,100 15,400 ± 300 6.67 ± 0.02 4.43 ± 0.03 66.4 ± 0.5 
Lagoon 24,100 ± 1,200 1,200 ± 200 3.64 ± 0.05 1.51 ± 0.03 42.0 ± 1.0 
RBC 25,200 ± 200 420 ± 10 3.02 ± 0.02 1.70 ± 0.01 56.3 ± 0.5 
MBBR 20,800 ± 2,200 515 ± 3 2.18 ± 0.02 1.53 ± 0.01 70.2 ± 0.8 

Values are sample mean ± standard deviation (n ≥ 3).

Sonication

RBC, lagoon, and MBBR sludge samples were gravity settled, centrifuged, and diluted with supernatant to an initial concentration of 6.5% TS in order to replicate the solids concentration of the TWAS samples collected from the CAS facility. All samples, including the TWAS, were further diluted to 4.5% TS with a buffer/micronutrient medium (Table S1, available with the online version of this paper) to maintain pH and ensure anaerobic growth would not be limited by a lack of trace nutrients (Kennedy & Droste 1985). Sonication was performed with a 450 Branson Digital Sonifier (Emerson Industrial, Connecticut), with a probe diameter of 13 mm, operating at 20 kHz and peak capability of 400 W. Sonication power (ES, in kJ/kg) was quantified with specific energy (Equation (1)): 
formula
(1)
where P is the power (J), t is the duration of sonication (sec), v is the sample volume (L), and TSo is the initial total solids (g/L).

Samples were sonicated in 200 mL batches for 1, 2, 5, and 10 min which correspond to specific energies of 800 ± 40, 1,550 ± 130, 3,770 ± 300, and 6,550 ± 530 kJ/kg TS. These values bracket the range of sonication power as reported by Bougrier et al. (2005) which defines low power (ES < 1,000 kJ/kg TS) as the level at which disintegration is limited to floc disruption and high power (1,000 kJ/kg TS < ES < 7,000 kJ/kg TS) as the level at which cell lysis occurs. Beyond ES = 7,000 kJ/kg TS, there is little reported benefit to increasing BG production with the rate of solubilisation being reported to decrease (Bougrier et al. 2005). pH stability due to sonication was verified for each sludge. Samples were sonicated without temperature control to mimic conditions of full-scale applications. Temperature increases were uniform for all sludge types resulting in an 8 °C increase for 1 min of sonication, 14 °C increase for 2 min, 32 °C increase for 5 min, and 50 °C increase for 10 min.

Solubilization was quantified through the degree of sludge disintegration (DD) calculated as the ratio of sCOD increase after sonication to the total possible sCOD increase (Equation (2)). Sonication did not affect tCOD, and thus tCOD is consistent for each sludge type after sonication. 
formula
(2)
where sCOD is soluble COD, sCOD0 is the initial soluble COD of untreated sludge, and tCOD is total COD.

Bioassays

The bioassay tests measured BGP according to the procedure by Owen et al. (1979) as single stage mesophilic (35 °C) assays on sludges of 4.5% TS to test the effect of sonication on BGP. As mentioned above, the samples were diluted to 4.5% TS with a buffer/micronutrient medium, to maintain pH and ensure anaerobic growth would not be limited by a lack of trace nutrients (Kennedy & Droste 1985). CAS samples were digested in 500 mL reactor vessels containing 300 mL substrate and 60 mL inoculum. Due to sample volume limitations, all other assays were conducted at a 1:10 scale to the CAS assays. RBC, lagoon, and MBBR tests were conducted in replicates of 4 in 50 mL reaction vessels containing 30 mL substrate and 6 mL inoculum. Bottles were purged with N2 gas for 2 min, closed with butyl rubber stoppers, and sealed with an aluminium crimp. Samples were incubated and shaken at 35 °C in a Psycrotherm controlled environment incubator shaker (New Brunswick Scientific Co. Inc., USA). BGP was sampled daily and measured by manometer.

CAS sample assays were conducted in triplicate at 35 °C utilizing the AMPTS II (Bioprocess Control, Sweden). Quantity of produced BG was automatically logged every hour by the system.

Analytical methods

TS and VS were measured as per standard method 2540 (APHA 2012). Samples for tCOD analysis were homogenized and measured using HACH method 10212. To mitigate the potential interference of filamentous bacteria causing bias during filtration, sCOD was separated by centrifuging sludge samples at 8,000 g for 20 min and measured using HACH method 8000.

Data analysis

Three non-linear models for the estimation of performance parameters were compared to empirical data for BGP (Table 2). The following models were shown by Donoso-Bravo et al. (2010) as being effective at modelling production parameters in batch systems. Experimental ultimate biogas production (Bo), maximum production rate (Rm), and lag time (λ) parameters were determined as described by Lay et al. (1996). Cumulative biogas yield, B (mL/g VS) and time of digestion t (h) are independent variables. The modified Gompertz (GM) Equation (3) has been used successfully to model biogases in multiple AD systems. The transference function or Reaction Curve (RC) model (Equation (4)) is based on control principles by considering the process as a system receiving inputs and generating outputs to predict maximum gas production. A first-order (FO) kinetic model (Equation (5)) based on substrate degradation is used to find the coefficient of the limiting rate (kH), which, for AD, is assumed to be hydrolysis. Non-linear optimization and statistical analysis were performed. Comparisons of data were conducted through t-tests, one-way analysis of variance (ANOVA), and Tukey's honest significant difference (HSD) range distribution, with p-value less than 0.05 indicating significance.

Table 2

Models for the determination of BGP parameters

Model Equation  Reference 
GM  (3) Lay et al. (1996)  
RC  (4) Redzwan & Banks (2004)  
FO  (5) Pavlostathis & Giraldo-Gomez (1991)  
Model Equation  Reference 
GM  (3) Lay et al. (1996)  
RC  (4) Redzwan & Banks (2004)  
FO  (5) Pavlostathis & Giraldo-Gomez (1991)  

RESULTS AND DISCUSSION

Sludge particulate solubilization

Sonication pretreatment solubilizes particulate and cellular material to shorten the limiting hydrolysis phase of AD, thus potentially increasing the overall rate of treatment and potentially the extent of stabilization. The degree of sludge disintegration can be used to determine the extent of solubilization caused by pretreatment. The present study demonstrates that solubilisation increased with sonication time for all sludge samples analysed in this study (Figure 1). A one-way ANOVA comparing the differing sludge types and DD was conducted for 1 and 10 min levels of sonication. The differences in sludge type had a significant effect on mean DD at 1 min [F(3, 50) = 14.05, p < 0.0001] and at 10 min [F(3, 50) = 171.9, p < 0.0001]. Post hoc comparison using the Tukey's HSD test indicates that there is only a significant difference between 1 min RBC sludge (DDm = 1.36%, SD = 0.32) and the three other samples of 1 min MBBR (DDm = 3.76%, SD = 0.97), 1 min lagoon (DDm = 3.77%, SD = 0.27), and 1 min CAS (DDm = 4.60%, SD = 1.70). There is no significant difference in 1 min DD between MBBR, lagoon, and CAS at this minimum level of sonication. At the highest applied level of sonication (10 min), results indicate that DD of 10 min CAS (DDm = 19.39%, SD = 2.47), 10 min lagoon (DDm = 14.39%, SD = 0.54), 10 min RBC (DDm = 5.92%, SD = 0.34), and 10 min MBBR (DDm = 27.07%, SD = 2.57) are all significantly different from each other. This suggests that sludge source can be a cause of significant variation in disintegration; the effect becoming more pronounced at higher levels of sonication. This is contrary to previous research that simply modeled DD linearly by the single variable of sonication time alone (Zhang et al. 2007).

Figure 1

Solubilization and yield (YsCOD) vs. sonication times for all sludges.

Figure 1

Solubilization and yield (YsCOD) vs. sonication times for all sludges.

In this study, BG yield was measured as BG produced per mass of sCOD consumed during digestion. As DD increased and more sCOD was available for digestion, the yield decreased (Figure 1). A comparison of yield (YsCOD) at 0 and 1 min sonication for each sludge type was conducted by the Student's t-test to determine the impact of low-powered sonication. There was a significant YsCOD decrease (p < 0.0001) between 0 and 1 min across all sludge types: CAS (0 min YsCOD = 1.54, SD = 0.11; 1 min YsCOD = 0.98, SD = 0.09), lagoon (0 min YsCOD = 8.36, SD = 2.56; 1 min YsCOD = 1.64, SD = 0.34), RBC (0 min YsCOD = 61.64, SD = 7.81; 1 min YsCOD = 6.18, SD = 0.60), and MBBR (0 min YsCOD = 1.59, SD = 0.04; 1 min YsCOD = 1.17, SD = 0.05). The decrease in biogas yield continued as the degree of sonication was increased to 10 min. In particular, the RBC sludge showed the greatest effect as biogas yield decreased 10- and 60-fold from 62 to 6.2 and from 62 to 1.2 for 1 and 10 min sonication times.

The statistically significant difference in YsCOD at various sonication intensities may be caused by either the release of recalcitrant COD or by the inhibition to BGP. Similar effects have been noted in high temperature thermalizations greater than 170 °C (Carrère et al. 2010; Kim & Lee 2012). The cause in thermal pretreatments is thought to be the caramelization or burning of substrates and the conversion of carbohydrates and amino acids through Maillard reactions to melanoidins that are toxic or impossible to degrade (Şahinkaya & Sevimli 2013). Although the bulk temperature of 1 min sonicated samples increased by only 8 °C, sonication can still cause high temperature effects in the sludge through the extreme local conditions of cavitation where the bubbles can have temperatures up to 5000 K (Flint & Suslick 1991; Tiehm et al. 2001). It was previously reported that similar inhibition to BGP was caused by high power sonication thought to arise from melanoidin formation or release of inhibitory long chain fatty acids from bacterial cell membranes (Appels et al. 2008b; Kim & Lee 2012). However, all four sludges used in this study show a significant decrease in YsCOD when pretreated with the lowest limit of ultrasonic power (1 min, ES = 800 kJ/kg TS), which is lower than the power threshold of 1,000 kJ/kg TS required for cell lysis (Bougrier et al. 2005). Inhibition due to sonication is therefore more likely to be caused by melanoidin formation from high temperature cavitation rather than the release of inhibitory compounds from cell membranes.

BGP and effect of sonication

Four waste sludges from various sources and of differing natures were digested under similar conditions with similar initial solids concentrations. The effect of sonication on cumulative BGP normalized per mass of sludge (mL BG/g sludge), digested for 0, 1, 2, 5, and 10 min of sonication, is illustrated in Figure 2. Cumulative BGP normalized per mass of VS is also illustrated in Figure 2 to compare to conventional normalization methods. This study investigates the BG potential of various sludge types, of which differing VS fractions are inherent characteristics. Hence, the following discussion focuses on the BG produced normalized by sludge mass. It should be noted, however, that the BGP presented normalized per mass of sludge and normalized per mass of VS demonstrate similar responses to sonication. By mass, the ultimate BGP was statistically greater than the control for all sonicated samples. In the subsequent section, modelling is conducted with BGP data normalized by mass VS.

Figure 2

Cumulative BGP for all sludge types and sonication times normalized by mass of sludge and by mass of VS.

Figure 2

Cumulative BGP for all sludge types and sonication times normalized by mass of sludge and by mass of VS.

For BG produced from CAS sludge, there was not a significant effect of sonication on BGP with a 95% confidence (p < 0.05) for the five conditions [F(4, 10) = 2.623, p = 0.0985]; however, a t-test performed on 0 min and 1 min BGP at a lower confidence of 90% determined there is a significant difference [t(3) = 2.967, p = 0.0592] of ultimate BGP between the control, 0 min (BGPm = 7.6, SD = 0.2), and 1 min (BGPm = 8.00, SD = 0.09) sonication levels representing a 5% ± 3 increase in BGP over the control. While the DD increased with increased sonication for CAS (Figure 1), BGP from 1 min sonicated samples did not differ significantly (p > 0.1) from 2, 5, and 10 min sonicated samples and concomitantly the biogas yield (YsCOD) based on soluble COD decreased.

Sludge samples from the CAS system yielded lower increases in BGP with sonication in comparison to the BGP increases of the alternative sludges studied here. It should be noted that the CAS plant operated with a low SRT and as such the sludge is likely to contain less cell debris and less recalcitrant material than CAS systems operating at longer SRTs. The small differences in BGI from sonication suggest that the pretreatment was less effective on a more readily digestible CAS sludge or energy levels applied in this study were less than the threshold required to exceed floc disintegration and enter a cell lysis phase. This implies that the reaction of CAS sludge to sonication is unique with respect to the reactions of the alternative sludges to the same sonication energy levels.

An ANOVA conducted on the ultimate BGP of lagoon sludges indicated that sonication had a significant effect for all five applied sonication energy levels [F(4, 15) = 5.994, p = 0.0044]. 1 min of sonication produced the greatest final quantity of BG (BGPm = 1.326, SD = 0.003) for all lagoon sludges, increasing BGP over the control by 12% ± 9. A subsequent post hoc comparison using Tukey's HSD test found BGP from 1 min sonicated sludge was significantly greater (p < 0.05) than the control, 0 min (BGPm = 1.2, SD = 0.1), and 10 min (BGPm = 1.19, SD = 0.02) ultimate BGP, yet not significantly different (p > 0.05) than 2 min (BGPm = 1.28, SD = 0.04) and 5 min (BGPm = 1.29, SD = 0.03). The overall low ultimate BGP of 1.19–1.32 mL/g sludge suggests that a majority of the biodegradable organics were digested over the 5 years the sludge has accumulated at the benthic zone of the lagoon. As such, it is likely that the tCOD was refractory and not susceptible to sonication pretreatment.

An ANOVA conducted on ultimate BGP of RBC sludges indicated that sonication had a significant effect on ultimate BGP for this sludge type [F(4, 15) = 49.75, p < 0.0001]. In this case, 10 min of sonication produced the greatest final volume of BG (BGPm = 3.12, SD = 0.01) for RBC sludges, increasing BGP over the control by 20% ± 2. A post hoc comparison found BGP from 10 min sonicated sludge significantly greater (p < 0.05) than 0 min (BGPm = 2.59, SD = 0.04), 1 min (BGPm = 2.80, SD = 0.06), 2 min (BGPm = 2.87, SD = 0.04), and 5 min (BGPm = 2.98, SD = 0.09). This is the only case within the four sludge types tested where increased sonication energy correlated directly with increased BGP. Sonicating RBC sludge for 10 min exhibited the greatest increase in biogas, 20% ± 2 compared to the other sludges (Figure 3). The fact that biofilm sludge has a high component of extracellular polysaccharide material the sonication pretreatment may have a very positive impact on solubilizing the EPS and making it more readily available for digestion.

Figure 3

Biogas increase vs. solubilization as the degree of disintegration for all sludge types. Shaded area represents the region of maximum biogas increases for four sludge types tested.

Figure 3

Biogas increase vs. solubilization as the degree of disintegration for all sludge types. Shaded area represents the region of maximum biogas increases for four sludge types tested.

An ANOVA conducted on ultimate BGP of MBBR sludges indicated that sonication again had a significant effect for all levels of sonication [F(4, 10) = 19.33, p = 0.0001]. The greatest increase of BGP production over the control, 15% ± 2, occurred for the 2 min sonicated sludge (BGPm = 5.49, SD = 0.2), which was significantly greater (p < 0.05) than 0 min (BGPm = 4.79, SD = 0.06) and 10 min (BGPm = 5.17, SD = 0.07), yet not significantly different (p > 0.05) from 1 min (BGPm = 5.30, SD = 0.04) and 5 min (BGPm = 5.4, SD = 0.2) BGP. This similarity in the response of the MBBR sludge to the RBC sludge indicates that sonication may have a greater effect on sludge generated by biofilm technologies.

The effects of sonication on solubilization of sludge and increase of BGP appear to vary according to the sludge source tested. Bougrier et al. (2005) reported the effective sonication energy ranges of floc disruption and cell lysis based on municipal secondary WAS using high-load aeration floatation-thickening to 1.85% TS. Other researchers using different secondary sludge sources found similar phases of floc disruption and cell lysis, yet at differing energy levels (Zhang et al. 2008). The present study shows that sludges react uniquely to ultrasound based on their source and treatment technology. This is likely due to the varying composition of the sludge, with the results demonstrating that biofilm sludge and hence high EPS content sludge are more responsive to ultrasound treatment; as demonstrated by the DD and BGP of the RBC and MBBR sludge. The ultrasound treatment in this study appears to effectively disintegrate and disrupt the EPS and embedded cells in biofilm sludge, thus increasing readily available substrate for digestion and ultimately BGP. The level of solubilisation due to sonication and peak BGP vary according to sludge type. However, peak BGP of all sludge types occurred when the DD was within a small range of 2.9–7.4% (Figure 3). The sonication energy level that achieved the peak BGP was unique for each different sludge type. Each sludge had a unique value of sonication energy for the point of diminishing returns where DD may increase, but BGP is inhibited. CAS sludge yielded lower increases in BGP after sonication in comparison to the BGP increases of the alternative sludges studied here. The small differences in increased BGP after sonication for CAS sludge suggest that the energy levels used were less than the threshold required to exceed floc disintegration and enter a cell lysis phase. This would make the reaction of CAS sludge to sonication unique with respect to the reactions of the alternative sludges to the same sonication energy levels. The alternative sludges also reacted uniquely to sonication in terms of BGP increases. These results could explain the large range of literature values for the increased BGP from sonication pretreatment of 6.3–53% (Table 3). Hence, variation in reported results of AD of sonicated WAS may be due to variations in treatment style, initial sCOD, sludge age, and influent concentrations that are unique to each source sludge studied (Wang et al. 2005). This could explain the wide variation in BGP increases.

Table 3

Comparison of reported results for AD of sonicated sludge types

Sludge type Ultrasonic conditions
 
AD conditions
 
Findings Source 
Frequency (kHz) Energy level Duration Scale Temperature Residence time (days) Biogas (BGI) and/or methane (MI) increase 
Industrial WAS 20 0.33 W/mL 20 min Batch Mesophilic 40 104% MI Chu et al. (2002)  
290% MI 
Municipal WAS 41 – 150 min Semi-continuous Mesophilic 41.6% BGI Tiehm et al. (2001)  
20 – 30 min Batch Mesophilic 33 23% BGI Onyeche et al. (2002)  
27 200–300 W/L 2.5–30 min Batch Mesophilic 19 10–20% BGI Grönroos et al. (2005)  
20 7,000 kJ/kg TS Batch Mesophilic 16 40% BGI Bougrier et al. (2005)  
31 10 W/cm² 90 sec Pilot Mesophilic 16% BGI Nickel & Neis (2007)  
20 1 W/mL 1 min Batch Mesophilic – 5.6% MI 6.3% BGI Şahinkaya & Sevimli (2013)  
20 800 kJ/kg TS Batch Mesophilic 30 5% BGI Current study 
TWAS 20 0.52 W/mL 1 min Batch Mesophilic 16 53% BGI Mao & Show (2007)  
25 1,020 W/L – Batch – 40% BGI Appels et al. (2008b)  
1/3 Primary 2/3 WAS 20 13.7 W/cm² 1.5 sec Full – 30 45% BGI Xie et al. (2007)  
53% Primary 47% WAS 31 – 96 sec Batch Mesophilic 28 30% BGI Tiehm et al. (1997)  
75% Primary 25% WAS 20 11,000 kJ/kg TS Semi-continuous Mesophilic 20 31% BGI Benabdallah El-Hadj et al. (2007)  
Thermophilic 15 16% BGI 
Not specified – – – 14 Full – 12–69 20–50% BGI Barber (2005)  
20 10.8 kW/kg TS Pilot – 20 42% BGI Pérez-Elvira et al. (2009)  
Lagoon 20 800 kJ/kg TS Batch Mesophilic 30 12% BGI Current study 
RBC 20 6,550 kJ/kg TS Batch Mesophilic 30 20% BGI 
MBBR 20 1,550 kJ/kg TS Batch Mesophilic 30 15% BGI 
Sludge type Ultrasonic conditions
 
AD conditions
 
Findings Source 
Frequency (kHz) Energy level Duration Scale Temperature Residence time (days) Biogas (BGI) and/or methane (MI) increase 
Industrial WAS 20 0.33 W/mL 20 min Batch Mesophilic 40 104% MI Chu et al. (2002)  
290% MI 
Municipal WAS 41 – 150 min Semi-continuous Mesophilic 41.6% BGI Tiehm et al. (2001)  
20 – 30 min Batch Mesophilic 33 23% BGI Onyeche et al. (2002)  
27 200–300 W/L 2.5–30 min Batch Mesophilic 19 10–20% BGI Grönroos et al. (2005)  
20 7,000 kJ/kg TS Batch Mesophilic 16 40% BGI Bougrier et al. (2005)  
31 10 W/cm² 90 sec Pilot Mesophilic 16% BGI Nickel & Neis (2007)  
20 1 W/mL 1 min Batch Mesophilic – 5.6% MI 6.3% BGI Şahinkaya & Sevimli (2013)  
20 800 kJ/kg TS Batch Mesophilic 30 5% BGI Current study 
TWAS 20 0.52 W/mL 1 min Batch Mesophilic 16 53% BGI Mao & Show (2007)  
25 1,020 W/L – Batch – 40% BGI Appels et al. (2008b)  
1/3 Primary 2/3 WAS 20 13.7 W/cm² 1.5 sec Full – 30 45% BGI Xie et al. (2007)  
53% Primary 47% WAS 31 – 96 sec Batch Mesophilic 28 30% BGI Tiehm et al. (1997)  
75% Primary 25% WAS 20 11,000 kJ/kg TS Semi-continuous Mesophilic 20 31% BGI Benabdallah El-Hadj et al. (2007)  
Thermophilic 15 16% BGI 
Not specified – – – 14 Full – 12–69 20–50% BGI Barber (2005)  
20 10.8 kW/kg TS Pilot – 20 42% BGI Pérez-Elvira et al. (2009)  
Lagoon 20 800 kJ/kg TS Batch Mesophilic 30 12% BGI Current study 
RBC 20 6,550 kJ/kg TS Batch Mesophilic 30 20% BGI 
MBBR 20 1,550 kJ/kg TS Batch Mesophilic 30 15% BGI 

Modelling of BGP results

Three kinetic models previously used for methane production to describe the AD process and critical digestion performance parameters were tested against BGP normalized by mass of VS (Figure 2) (Donoso-Bravo et al. 2010). Overall, the models showed strong correlation (r2 = 0.920–0.999) with the data and are deemed useful for the accurate determination and comparison of design parameters. As an example, Figure 4 illustrates the results of the non-linear regression using the three models for 1 min sonication times. The complete results can be found in Table S2 (available with the online version of this paper). Non-linear regression results for 0, 2, 5, and 10 min sonications can be found in Figures S1–S4 (available online). The GM model had overall stronger fit (r2 ≥ 0.961) and could accommodate a larger data set that had significant lag times better when compared to RC (r2: 0.920–0.995) and FO (r2: 0.917–0.992), but tended to underestimate the maximum rate of BGP (slope) for the curve as noted by Donoso-Bravo et al. (2010).

Figure 4

GM, RC, and FO models fit to mean cumulative BGP for 1 min of sonication for CAS (a)–(c), lagoon (d)–(f), RBC (g)–(i), and MBBR (j)–(l). Standard deviation of BGP is not included for CAS.

Figure 4

GM, RC, and FO models fit to mean cumulative BGP for 1 min of sonication for CAS (a)–(c), lagoon (d)–(f), RBC (g)–(i), and MBBR (j)–(l). Standard deviation of BGP is not included for CAS.

A one-way ANOVA was conducted on each model parameter to compare the effects of sonication on Bo, Rm, Lag, and kH for 0, 1, 2, 5, and 10 min of sonication pretreatment. All 32 tests showed a significant effect of sonication on the individual design parameters at the p < 0.05 level. There is no case in which the control sludge without sonication pretreatment has the greatest modeled maximum (Rm) or overall (kH) kinetic rate (Table S2). There is no case in which the highest BGP occurs at the same sonication level as the highest maximum (Rm) or overall (kH) rate.

Ultimate BGP has already been discussed in the previous section. Since there are two functions that can model the Rm parameter, a comparison of the standard deviation of residuals (Sy.x) for GM and RC models was conducted over the same, truncated data set as used by the RC model to determine which model deviated the least from experimental values and could thus more accurately portray Rm. For CAS sludge, the more accurate parameter is derived from the GM model while RBC, Lagoon, and MBBR Rm are better predicted by the RC model. A comparison of Rm and kH modelled parameters with DD were found to be significantly different (p < 0.05) based on separate ANOVA tests; yet, there is no clear effect of DD on Rm or kH, which is concurrent with Donoso-Bravo et al.'s (2010) findings. It is possible that the increased initial sCOD generated from sonication pretreatment has no effect on the maximum rate of digestion or the overall apparent hydrolysis coefficient while still significantly being able to impact the ultimate BG yield.

During lag time before active digestion, the inoculum's hydrolytic microbes require time to adjust enzyme production to accommodate the available substrates. Since inoculum was not acclimatized to the substrates before digestion, a significant lag phase was expected, but is usually neglected in kinetic analysis due to highly variable and uncertain lag phase length in batch experiments (Lay et al. 1996). Experimental lag was determined by extrapolating a line from the point of Rm to the axis with slope of Rm (Lay et al. 1996). However, CAS data show apparent diauxic characteristics where multi-phasic BGP is evident, separated by multiple, short duration plateaus in the profile (Figure 2 – CAS), making the aforementioned method unreliable and inconsistent since the maximum BGP rate could occur in different phases resulting in a non-sensical value for lag. In this case, a better separation between lag and growth phases was determined by the time when BGP first exceeded 2 mL/day. The BGP varies in the lag phase yet, when it reaches 2 mL/day, it does not decrease again. Using this estimation for the time when lag phase ends matches with the graphical data (Figure 2) and will be used to delineate between lag phase and the start of the active digestion phase, which ends when BGP again decreases below 2 mL/day.

The hourly resolution of CAS BGP data illuminated multiple production phases separated by mini-plateaus representing diauxic performance (Figure 2 – CAS). The CAS control sludge without pretreatment has four active phases of BGP separated by short lag times in 345 ± 2 h of active digestion and an FO modelled hydrolysis coefficient of 3.90·10−3 h−1. The 1, 2, and 5 min sonicated sludge digestions also have four phases of BGP. Their apparent hydrolysis coefficient did not differ significantly with that of the control. However, the 10 min sonicated sludge has three phases of BGP. When compared to the control, it results in a significantly shorter [t(4) = 37.74, p < 0.0001] active digestion time of 272 ±2 h and a faster hydrolysis coefficient of 4.60·10−3 h−1. This indicates that sonication pretreatment is effective at homogenizing the most recalcitrant substrates that compose the fourth BGP peak into ones that can be digested with other more preferable, more easily digestible substrates. This effectively shortens active digestion time through the removal of an additional lag phase.

Overall, excepting 1, 2, and 10 min sonication pretreatment of MBBR sludge, sonication increases the maximum rate of digestion over non-sonicated sludge in the other 13 cases, but not in a clearly definable pattern. Ideal sonication energy may be unique for each sludge type and dependent on system requirements. Designing a system for both the greatest increase in kinetics and greatest BGP may not be possible.

CONCLUSIONS

In this work, sonication pretreatment was proved to significantly increase tCOD biogas yield in mesophilic batch assays of the four sludge types tested. It had a greater impact on increasing BGP from biofilm type sludge (15–20% for MBBR and RBC) as compared to conventional, suspended growth technologies (5–12% for CAS and Lagoon). The different sludge types tested responded uniquely to the same levels of sonication energies. An optimal specific energy for the greatest production of BG was not found that coincided for all sludges. Instead, optimal specific energy was unique for each sludge, but the peak BGP for all sludges occurred within a small solubilization range of 2.9–7.4% DD. Sonication pretreatment exhibited significant BGP inhibition relative to sCOD, even at the lowest applied energy levels of 800 kJ/kg TS. In most cases, there was no significant difference (p < 0.05) in increased BGP between low (1 min, 800 kJ/kg TS) and high (10 min, 6,550 kJ/kg TS) energy levels. The use of automated BG logging revealed diauxic growth patterns in the BGP of CAS WAS. The duration of the active phase of BGP decreased significantly in the AD of sludge sonicated for 10 min (6,550 kJ/kg TS) where the number of active phases of BGP was reduced from 4 to 3.

Three models were used to fit experimental data and determine ultimate BGP, maximum rate of digestion, lag time, and rate of hydrolysis coefficient. GM, RC, and FO models show strong correlations with sonicated waste sludge BGP. The GM model was useful for fitting experimental data with significant lag time. Rm and kH parameters were successfully determined with RC and FO models when data sets were truncated to remove lag time. Sonication had no clear effect on Rm or kH, which were shown to be poor indicators of the effect of sonication pretreatment on digestion kinetics. Overall, ultrasound impacted BGP, disintegration and digestion rates in ways that could not be easily correlated with the basic characterizations and models employed in this study.

REFERENCES

REFERENCES
Alqaralleh
R.
,
Delatolla
R.
&
Kennedy
K.
2015
Anaerobic digestion of simulated-organic fraction of municipal solid waste: effect of alkaline pretreatment
.
International Journal of Environment and Waste Management
16
(
2
),
166
185
.
https://doi.org/10.1504/IJEWM.2015.071292
.
APHA, WEF, AWWA
2012
Standard Methods for the Examination of Water and Wastewater
,
22nd edn
.
APHA, WEF, AWWA
,
Washington, DC
.
Appels
L.
,
Baeyens
J.
,
Degrève
J.
&
Dewil
R.
2008a
Principles and potential of the anaerobic digestion of waste-activated sludge
.
Progress in Energy and Combustion Science
34
(
6
),
755
781
.
https://doi.org/10.1016/j.pecs.2008.06.002
.
Appels
L.
,
Dewil
R.
,
Baeyens
J.
&
Degrève
J.
2008b
Ultrasonically enhanced anaerobic digestion of waste activated sludge
.
International Journal of Sustainable Engineering
1
(
2
),
94
104
.
https://doi.org/10.1080/19397030802243319
.
Barber
W. P.
2005
The effects of ultrasound on sludge digestion
.
Water and Environment Journal
19
(
1
),
2
7
.
https://doi.org/10.1111/j.1747-6593.2005.tb00542.x
.
Benabdallah El-Hadj
T.
,
Dosta
J.
,
Márquez-Serrano
R.
,
Mata-Álvarez
J.
,
El-hadj
T. B.
,
Dosta
J.
&
Ma
R.
2007
Effect of ultrasound pretreatment in mesophilic and thermophilic anaerobic digestion with emphasis on naphthalene and pyrene removal
.
Water Research
41
(
1
),
87
94
.
https://doi.org/10.1016/j.watres.2006.08.002
.
Bougrier
C.
,
Carrère
H.
&
Delgenès
J. P.
2005
Solubilisation of waste-activated sludge by ultrasonic treatment
.
Chemical Engineering Journal
106
(
2
),
163
169
.
https://doi.org/10.1016/j.cej.2004.11.013
.
Bougrier
C.
,
Albasi
C.
,
Delgenès
J. P.
&
Carrère
H.
2006
Effect of ultrasonic, thermal and ozone pre-treatments on waste activated sludge solubilisation and anaerobic biodegradability
.
Chemical Engineering and Processing: Process Intensification
45
(
8
),
711
718
.
https://doi.org/10.1016/j.cep.2006.02.005
.
Cano
R.
,
Pérez-Elvira
S. I.
&
Fdz-Polanco
F.
2015
Energy feasibility study of sludge pretreatments: a review
.
Applied Energy
149
,
176
185
.
https://doi.org/10.1016/j.apenergy.2015.03.132
.
Carrère
H.
,
Dumas
C.
,
Battimelli
A.
,
Batstone
D. J.
,
Delgenès
J. P.
,
Steyer
J. P.
&
Ferrer
I.
2010
Pretreatment methods to improve sludge anaerobic degradability: a review
.
Journal of Hazardous Materials
183
(
1–3
),
1
15
.
https://doi.org/10.1016/j.jhazmat.2010.06.129
.
Chu
C. P.
,
Lee
D. J.
,
Chang
B. V.
,
You
C. S.
&
Tay
J. H.
2002
‘Weak’ ultrasonic pre-treatment on anaerobic digestion of flocculated activated biosolids
.
Water Research
36
(
11
),
2681
2688
.
https://doi.org/10.1016/S0043-1354(01)00515-2
.
Delatolla
R.
&
Babartusi
S.
2005
Parameters affecting hydraulic behavior of aerated lagoons
.
Journal of Environmental Engineering
131
(
10
),
1404
1413
.
Donoso-Bravo
A.
,
Pérez-Elvira
S. I.
&
Fdz-Polanco
F.
2010
Application of simplified models for anaerobic biodegradability tests. Evaluation of pre-treatment processes
.
Chemical Engineering Journal
160
(
2
),
607
614
.
https://doi.org/10.1016/j.cej.2010.03.082
.
Flint
E. B.
&
Suslick
K. S.
1991
The temperature of cavitation
.
Science
253
(
5026
),
1397
1399
.
https://doi.org/10.1126/science.253.5026.1397
.
Grönroos
A.
,
Kyllönen
H.
,
Korpijärvi
K.
,
Pirkonen
P.
,
Paavola
T.
,
Jokela
J.
&
Rintala
J.
2005
Ultrasound assisted method to increase soluble chemical oxygen demand (SCOD) of sewage sludge for digestion
.
Ultrasonics Sonochemistry
12
(
1–2) (Special Issue
),
115
120
.
https://doi.org/10.1016/j.ultsonch.2004.05.012
.
Karizmeh
M.
,
Delatolla
R.
&
Narbaitz
R. M.
2014
Investigation of settleability of biologically produced solids and biofilm morphology in moving bed bioreactors (MBBRs)
.
Bioprocess and Biosystems Engineering
37
(
9
),
1839
1848
.
Kennedy
K. J.
&
Droste
R. L.
1985
Startup of anaerobic downflow stationary fixed film (DSFF) reactors
.
Biotechnology and Bioengineering
27
(
8
),
1152
1165
.
https://doi.org/10.1002/bit.260270810
.
Kim
D. J.
&
Lee
J.
2012
Ultrasonic sludge disintegration for enhanced methane production in anaerobic digestion: effects of sludge hydrolysis efficiency and hydraulic retention time
.
Bioprocess and Biosystems Engineering
35
(
1–2
),
289
296
.
https://doi.org/10.1007/s00449-011-0588-x
.
Lay
J.-J.
,
Li
Y.-Y.
&
Noike
T.
1996
Effect of moisture content and chemical nature on methane fermentation characteristics of municipal solid wastes
.
Journal of Environmental Engineering Division
552
(
VII-1
),
101
108
.
https://doi.org/10.2208/jscej.1996.552_101
.
Mao
T.
&
Show
K.-Y.
2007
Influence of ultrasonication on anaerobic bioconversion of sludge
.
Water Environment Research: A Research Publication of the Water Environment Federation
79
(
4
),
436
441
.
https://doi.org/10.2175/106143006X123049
.
Martín-Cereceda
M.
,
Jorand
F.
,
Guinea
A.
&
Block
J. C.
2001
Characterization of extracellular polymeric substances in rotating biological contractors and activated sludge flocs
.
Environmental Technology
22
(
8
),
951
959
.
https://doi.org/10.1080/09593332208618231
.
Nickel
K.
&
Neis
U.
2007
Ultrasonic disintegration of biosolids for improved biodegradation
.
Ultrasonics Sonochemistry
14
(
4
),
450
455
.
https://doi.org/10.1016/j.ultsonch.2006.10.012
.
Onyeche
T. I.
,
Schläfer
O.
,
Bormann
H.
,
Schröder
C.
&
Sievers
M.
2002
Ultrasonic cell disruption of stabilised sludge with subsequent anaerobic digestion
.
Ultrasonics
40
(
1–8
),
31
35
.
https://doi.org/10.1016/S0041-624X(02)00087-2
.
Owen
W. F.
,
Stuckey
D. C.
,
Healy
J. B.
,
Young
L. Y.
&
McCarty
P. L.
1979
Bioassay for monitoring biochemical methane potential and anaerobic toxicity
.
Water Research
13
(
6
),
485
492
.
https://doi.org/10.1016/0043-1354(79)90043-5
.
Pavlostathis
S. G.
&
Giraldo-Gomez
E.
1991
Kinetics of anaerobic treatment – a critical review
.
CRC Critical Reviews in Environmental Control
21
(
January 2013
),
411
490
.
https://doi.org/10.1080/10643389109388424
.
Pérez-Elvira
S.
,
Fdz-Polanco
M.
,
Plaza
F. I.
,
Garralón
G.
&
Fdz-Polanco
F.
2009
Ultrasound pre-treatment for anaerobic digestion improvement
.
Water Science and Technology
60
(
6
),
1525
1532
.
https://doi.org/10.2166/wst.2009.484
.
Redzwan
G.
&
Banks
C.
2004
The use of a specific function to estimate maximum methane production in a batch-fed anaerobic reactor
.
Journal of Chemical Technology and Biotechnology
79
(
10
),
1174
1178
.
https://doi.org/10.1002/jctb.1107
.
Şahinkaya
S.
&
Sevimli
M. F.
2013
Sono-thermal pre-treatment of waste activated sludge before anaerobic digestion
.
Ultrasonics Sonochemistry
20
(
1
),
587
594
.
https://doi.org/10.1016/j.ultsonch.2012.07.006
.
Shen
Y.
,
Linville
J. L.
,
Urgun-demirtas
M.
,
Mintz
M. M.
&
Snyder
S. W.
2015
An overview of biogas production and utilization at full-scale wastewater treatment plants (WWTPs) in the United States: challenges and opportunities towards energy-neutral WWTPs
.
Renewable and Sustainable Energy Reviews
50
,
346
362
.
https://doi.org/10.1016/j.rser.2015.04.129
.
Tiehm
A.
,
Nickel
K.
&
Neis
U.
1997
The use of ultrasound to accelerate the anaerobic digestion of sewage sludge
.
Water Science and Technology
11
,
121
128
.
https://doi.org/10.1016/S0273-1223(97)00676-8
.
Tiehm
A.
,
Nickel
K.
,
Zellhorn
M.
,
Neis
U.
&
Tiehm
A.
2001
Ultrasonic waste activated sludge disintegration for improving anaerobic stabilization
.
Water Research
35
(
8
),
2003
2009
.
https://doi.org/10.1016/S0043-1354(00)00468-1
.
Wang
F.
,
Wang
Y.
&
Ji
M.
2005
Mechanisms and kinetics models for ultrasonic waste activated sludge disintegration
.
Journal of Hazardous Materials
123
(
1–3
),
145
150
.
https://doi.org/10.1016/j.jhazmat.2005.03.033
.
Xie
R.
,
Xing
Y.
,
Ghani
Y. A.
,
Ooi
K.
&
Ng
S.
2007
Full-scale demonstration of an ultrasonic disintegration technology in enhancing anaerobic digestion of mixed primary and thickened secondary sewage sludge
.
Journal of Environmental Engineering and Science
6
(
5
),
533
541
.
https://doi.org/10.1139/S07-013
.
Young
B.
,
Delatolla
R.
,
Ren
B.
,
Kennedy
K.
,
Laflamme
E.
&
Stintzi
A.
2016
Pilot-scale tertiary MBBR nitrification at 1 °C: characterization of ammonia removal rate, solids settleability and biofilm characteristics
.
Environmental Technology
3330
(
May
),
1
9
.
https://doi.org/10.1080/09593330.2016.1143037
.
Zhang
P.
,
Zhang
G.
&
Wang
W.
2007
Ultrasonic treatment of biological sludge: floc disintegration, cell lysis and inactivation
.
Bioresource Technology
98
(
1
),
207
210
.
https://doi.org/10.1016/j.biortech.2005.12.002
.
Zhang
G.
,
Zhang
P.
,
Gao
J.
&
Chen
Y.
2008
Using acoustic cavitation to improve the bio-activity of activated sludge
.
Bioresource Technology
99
(
5
),
1497
1502
.
https://doi.org/10.1016/j.biortech.2007.01.050
.

Supplementary data