Effective pathogen removal by low temperature thermal pre-treatment and anaerobic digestion for Class A biosolids production from sewage sludge

Pathogen inactivation during anaerobic digestion is a requirement for ensuring the safety and sustainability of biosolids reuse. The pathogens and parasites requirements for Class A biosolids are defined in the Mexican Standard NOM-004-SEMARNAT-2002 (SEMARNAT 2003) as thermotolerant coliforms, lower than 1,000 Most Probable Number (MPN)/gTS (TS, total solids); Salmonella, <3 MPN/gTS; and viable helminth eggs (HE) lower than 1 HE/gTS. These specifications are similar to U.S. standards except for HE (1 HE/4gTS in the EPA rule) and for not setting a limit for enteric viruses. Also, the Mexican standard includes a three level classification (A, B and C) with emphasis on HE content (Rubio-Loza & Noyola 2010).

Several authors have emphasized the importance of pre-treatments for enhancing sludge hygienization over other effects, especially when pathogen control is required for a proper final disposal (Mocé-Llivina et al. 2003; Skiadas et al. 2005; Arthurson 2008; Sahlström et al. 2008; Popat et al. 2010; Viau et al. 2011; Astals et al. 2012). Sludge pre-treatment may be accomplished by different methods, such as mechanical, chemical, thermal and biological (Montusiewicz et al. 2010).

Among these options, thermal pre-treatment can achieve efficient microbial pathogens removal and may have lower costs if produced biogas is used for thermal energy supply. This kind of pre-treatment has been studied using a wide range of temperatures, from 60 to 270 °C under different exposure times. The most common treatment temperatures are between 60 and 180 °C; those under 100 °C are considered as low temperature thermal treatments (e.g. pasteurization) and have the advantage of reduced energy requirements compared with higher temperature treatments. On the other hand, temperatures over 200 °C have been found to be responsible for refractory compound formation (Climent et al. 2007).

The decay of microorganisms in full-scale thermal treatment systems is influenced by operational factors and sludge characteristics, notably the heat transfer in large masses of sludge. Microorganisms may also be protected from heat and other detrimental factors by imbedding into the sludge solids (Mocé-Llivina et al. 2003). Moreover, at low temperatures (<100 °C), treatment time plays a more important role than treatment temperature (Lang & Smith 2008). At present, there is a growing concern about pathogen survival in low numbers after exposure to these processes and subsequently a re-growth to hazardous levels once in favorable environmental conditions (Iranpour et al. 2005; Higgins et al. 2007; Arthurson 2008; Dentel et al. 2008; Jiang et al. 2013).

Heat resistance of microorganisms is mostly influenced by the length of time and temperature of treatment, but environmental factors such as pH, water activity (a_w), composition of media, and so on, may drastically influence their heat resistance (Esteban et al. 2013). Also, the cooling rate after the treatment could have an important role during pathogen elimination, mainly because a mild and long cooling phase may not be detrimental to the survival of microorganisms. The latter could lead to re-growing of pathogens at more favorable conditions (Gupte et al. 2003; Iranpour et al. 2005; Higgins et al. 2007; Lang & Smith 2008; Esteban et al. 2013). A better understanding of survival conditions of pathogens during the heating and cooling phases of low-temperature pre-treatment is required in order to prevent regrowth/reactivation of this kind of microorganisms.

In a previous work, Rubio-Loza & Noyola (2010) found that a short thermophilic (55 °C) pre-treatment stage (3 days retention time) in an acidogenic reactor upstream from a mesophilic methanogenic reactor, was an excellent measure for improving pathogen removal. In order to reduce the size of the pre-treatment reactor, this work evaluated a low temperature thermal pre-treatment step (0.5 and one hour retention time) for pathogen inactivation in sewage sludge before entering the anaerobic digester. The inactivation patterns of thermotolerant coliforms and Salmonella spp. at different temperatures were obtained as well as under different cooling rate conditions, aiming to achieve the highest removal of these pathogen indicators with no possibility of reactivation or re-growth during the subsequent mesophilic anaerobic digestion phase.

The mixture of primary sludge (PS) and waste activated sludge (WAS) was obtained from a small (3 m³ d⁻¹) wastewater treatment plant at the Institute of Engineering of the National Autonomous University of Mexico, Mexico City. The sample was prepared considering weight percentage of 60% PS and 40% WAS. Total solids in PS were 60 to 70 g L⁻¹ (volatile fraction 60 to 70%) and 25 to 35 g L⁻¹ in thickened secondary sludge (volatile fraction 75 to 85%). The combined sludge had a final mean concentration of 37.9 ± 5.6 gTS L⁻¹, 30.3 ± 6.1 gVS (volatile solids) L⁻¹, 7.6 ± 2.9 gFS (fixed solids) L⁻¹, and was stored in a refrigerator at 4 °C. The sample was renewed every three weeks. Thermotolerant coliforms and Salmonella spp. content of mixed sludge was 6.6 ± 0.6 and 7.7 ± 0.6 log MPN g⁻¹ TS, respectively.

The thermal pre-treatments were applied in a thermal hydrolyzing device. The device is a continuous stirred batch reactor of 1,000 mL with a heating system and a fan located at the bottom for temperature control and cooling purposes; both plugged to a digital electronic controller with a thermocouple and a timer. An aluminum jacket constituted the heating system with 200 W resistance. The timer was activated when the fluid inside the device reached the desired temperature, so, treatment duration was defined as the elapsed time between this moment and when the sample started to be cooled.

A volume of 800 mL of the sludge mixture at a temperature of 4–6 °C was introduced in to the thermal-hydrolyzing device. The time before the device reached the preset temperatures corresponds to the non-isothermal heat treatment. This period was about 20 to 35 min, and during this time samples were taken for the determination of indicator microorganisms.

Once the preset temperature was reached, a sample (50 mL) was taken (called T₀), followed by subsequent samples at different times. At the end of the preset treatment time, the selected cooling treatment was applied immediately in order to quench the sludge and stop the thermal inactivation. Three replicates were measured for each sample.

The survival patterns were generated with the Weibull equation (Van Boekel 2002) and compared with an empirical distribution function which predicts an initial sigmoidal shoulder and softer negative slope on the tail (Peleg 1999). The non-linear regression, the coefficients of determination (r²) and the root mean-square error (RMSE) were adjusted and calculated using Microsoft Excel 2010.

At the end of the isothermal heat treatment, the sludge samples were cooled at three different cooling conditions: air-flow cooling, immersion on ice-bath and immersion on ice-bath with 10% sodium chloride. The purpose of these experiments was to evaluate the cooling rate impact over the final counting and re-growth of indicator microorganisms. The final temperature values reached for each cooling treatment were 25 °C for air-flow and ice bath cooling, and 20, 10 and 0 °C for ice-bath with salt cooling.

Two mesophilic digesters, described in detail in Rubio-Loza & Noyola (2010), were maintained in a constant-temperature room (35 ± 3 °C). The hydraulic retention times for each digester were 10 and 13 days. Feeding with 70 °C, 60 min pretreated sludge was made on a daily batch basis. Total, volatile and fixed solids, pH and alkalinity were determined in the raw and the effluent of the digesters according to standard methods (APHA/AWWA/WEF 1999). The alkalinity ratio was determined with partial alkalinity (pH endpoint 5.75) and total alkalinity (pH end point 4.30) following Jenkins et al. (1983). Daily gas production was quantified by water displacement using an inverted graduated cylinder, and biogas composition was measured using a gas chromatograph (Fisher Scientific 1200 Gas Partitioner with TCD) with a Porapak Q column and helium as a carrier gas.

For each treatment and cooling combination, one batch (800 mL) of resulting pretreated sludge was incubated in sterile Erlenmeyer flask of 1,000 ml capacity fitted with a magnetic stirrer and closed with a rubber stopper. Headspace was purged with nitrogen gas to assure anaerobic conditions. The flasks were incubated at 35 ± 1 °C, with constant mechanical stirring. The minimum incubation period was 36 h and a maximum of 72 h.

Quantification of thermotolerant coliforms, Salmonella spp., and HE was conducted in accordance with the methodologies indicated by the NOM-004-SEMARNAT-2002 (SEMARNAT 2003) (based on standard methods (APHA/AWWA/WEF 1999); Sections 9221B, 9221E and 9260B), using the most probable number technique by multiple tubes in a series of three.

The first approach was to determine the pathogen indicators content in sludge samples treated at different temperatures (50, 60, 70 °C) at three different treatment times (30, 60 and 90 min). These results are shown in Figure 1, where almost complete destruction of microbial indicators can be seen on every treatment time at 70 °C, but not at 60 or 50 °C. Based on these results, sludge pre-treatment at 70 °C for 60 min was chosen for feeding the anaerobic digesters in order to achieve the removal of organic matter from the already hygienized sludge. However, both digesters (10 and 13 days of HRT) produced a digested sludge with increased pathogen numbers, reaching levels over five and three log MPN g⁻¹ TS for thermotolerant coliforms and Salmonella, respectively. HE did not follow the same pattern as they were absent in the treated sludge, demonstrating that they were completely inactivated at the pre-treatment step. Total counts for pretreated sludge used for feeding were 2.8 × 10¹ MPN gTS⁻¹ for thermotolerant coliforms and below the detection limit for Salmonella spp.

With the previous results, more detailed monitoring of the treatment temperature was applied. Also, the treatment at 50 °C was discarded, and higher temperature (80 °C) was added. The average temperature behavior during the heating phase at 60, 70 and 80 °C was recorded at different times. The slopes for these treatments were 2.3 ± 0.2 °C min⁻¹ for a final target temperature of 60 °C and 2.6 ± 0.2 °C min⁻¹ for 70 and 80 °C. The time needed to reach each preset treatment temperature was on average 30 ± 3 min.

The experimental survival data for thermotolerant coliforms during the non-isothermal heating phase are depicted in Figures 2(a) and 2(b). A reduction of two log units for the 70 °C heating ramp, and almost four log units in the case of 80 °C were observed. On the contrary, there were no changes in the number of thermotolerant coliforms at the end of the 60 °C non-isothermal heating. The same happened for the 60 and 70 °C heating ramps for Salmonella spp. before the start of the isothermal phase. An interesting result is that survival data for Salmonella spp. for the 80 °C heating ramp remained unchanged until the last five minutes, when a sudden decrease of Salmonella spp. counts appeared (data not shown).

Figures 2(c) and 2(d) shows the thermotolerant coliform and Salmonella spp. survival curves together with the expected inactivation derived from the isothermal heat resistance data using Weibull and empirical models. Both models predicted similar lethality values, although the empirical model predicted higher survival in all cases and was closer to the experimental data (Table 1). It should be noticed that the contribution of the non-isothermal heating phase on inactivation was larger at higher temperatures, limiting their reduction once in the isothermal heating phase of such cases.

Results show that after thermal pre-treatment (non-isothermal and isothermal heating phases) at 60, 70 or 80 °C for 60 min, counts of thermotolerant coliforms and Salmonella spp. decreased between 6 and 7 log units and 7 and 8 log units, respectively. Moreover, during the isothermal pre-treatment at 80 °C, Salmonella spp. was already under detection limits, as a result of the effectiveness of the previous non-isothermal heating phase.

At this point, it was clear that pre-treatment at 60 °C had an unsatisfactory performance. As a result, it was discarded, and the following experiments were performed only at 70 and 80 °C for 60 min, where the best pathogen removal results were obtained. Figure 3 presents the temperature behavior of the three different cooling operations applied (air-flow cooling, immersion in ice-bath and immersion in ice-bath with 10% sodium chloride). The initial cooling rates were 1.2, 2 and 4 °C min⁻¹, respectively. Only the fastest after the heat pre-treatments at 70 and 80 °C was able to achieve inactivation of thermotolerant coliforms and Salmonella spp. without evidence of reactivation or re-growth within – at least – 48 h under anaerobic conditions. On the other hand, if the cooling rate is slow (2 °C min⁻¹) or there is no thermal shock (1.2 °C min⁻¹), both heating treatments resulted in reactivation and re-growth of thermotolerant coliforms and Salmonella spp. in a period not exceeding 24 h. Table 2 summarizes the data obtained in 12 h intervals.

The great diversity of microbial composition of sewage sludge is one of the key issues that makes it very difficult to determine reliable conditions for its sanitization. Van Boekel (2002) mentioned the important role of population heterogeneity of the same microorganisms, or biological variation, when explaining the nonlinear survival curves behavior. More recently, Van Derlinden et al. (2011) confirmed the existence of Escherichia coli subpopulations, which showed different responses triggered by heat stress.

At this point, the Weibull model fitted less accurately to the survival data, because it does not consider thermal adaptation of microbial subpopulations. On the contrary, the empirical model includes parameters that allowed a better prediction of the survival behavior at the end of the heating phase (Table 1). When the curve presents an upward concavity, it may be attributed to the fact that the most sensitive cells are inactivated first, and the remaining population becomes progressively sturdier. If the semi-logarithmic survival curve exhibits a downward concavity, that may be explained by accumulated damage on the remaining population which makes the surviving cells progressively more susceptible to the lethal treatment (Bermúdez-Aguirre & Corradini 2012).

Apparently, thermal processes may induce a ‘viable but non-culturable’ (VBNC) condition, from which some bacteria may later recover. An initial rapid decline of the population can be observed when certain microorganisms enter the VBNC state. The stressed cells remain in this status until they adapt their physiology; meanwhile, the more resilient and culturable microorganisms showed a slower decay or may continue as a transition cells, entering and leaving the VBNC state, with no apparent variation in the microbial counts (Gupte et al. 2003). This behavior may be found in the experimental data presented in Figures 2(c) and 2(d).

Such reactivation behavior has been demonstrated after high-speed centrifugal dewatering of thermophilically digested biosolids in both Escherichia coli and thermotolerant enterococci (Iranpour et al. 2005; Higgins et al. 2007; Qi et al. 2007; Viau & Peccia 2009). The microorganisms that enter the VBNC state often do so after exposure to environmental stress such as nutrient or substrate deprivation, presence of metals, chlorine, salinity, low and high temperatures, as some authors have found (Higgins et al. 2007). Once the environmental stress has passed, some of these microorganisms may reactivate, sometimes in larger quantities than before the presence of the stress conditions.

This issue is explained by Iranpour et al. (2005) in the case of thermophilically digested biosolids which were exposed to a slow drop in temperature during post-digestion processing. In that case, they solved the re-growth problem by insulating and electrical heat-tracing the post-digestion steps (screening, centrifuge dewatering, storage and transport).

Traditionally, thermotolerant coliforms have been used as an indicator of the presence of enteric pathogens in biosolids (Higgins et al. 2007). However, it could be possible for some subpopulations to remain viable even though microbial indicators seem to be killed as was observed at the end of the thermal pre-treatment. When a sudden decrease of temperature is used (membrane and cell integrity compromised), complete removal and inactivation of entire microbial populations may be obtained. Studies about the effect of cooling rates after thermal treatments on E. coli vegetative cells and Bacillus sporothermodurans IC4 spores (Conesa et al. 2009; Esteban et al. 2013) describe contradictory effects. However, none of them applied a severe cooling step to obtain complete removal of their microbial cell tested; suggesting that differential resistance of several subpopulations could be avoided in this way.

Apparently, pathogen removal evaluation using culture based enumeration of indicator organisms such as thermotolerant coliforms or Salmonella, may underestimate the actual number of viable organisms present in the sample, especially for thermal treatments (Jiang et al. 2013; Fu et al. 2014). As a result, the digestion processes traditionally used to meet certain regulations based on these indicators may not achieve the expected goals.

A 60 min thermal treatment of sludge at 60, 70 and 80 °C, followed by gradual, ambient cooling allowed reactivation and re-growth of thermotolerant coliforms and Salmonella spp. within the first 24 h of a mesophilic anaerobic digestion phase. Consequently, these conditions are not effective for a thermal pre-treatment as a part of a sludge stabilization process and production of Class A biosolids.

More effective thermal treatment (60 min at 70 °C and 80 °C followed by fast cooling in ice-bath with sodium chloride) achieved a decrease of between 6 and 7 log units of thermotolerant coliforms and between 7 and 8 log units of Salmonella spp., with no reactivation or re-growth during the following 48 h under anaerobic mesophilic conditions.

Thermal treatments should be evaluated with additional methods in order to identify the VBNC state of some microorganisms and incorporate them to biosolids regulations.

The first author would like to thank DGAPA-UNAM (Dirección General de Asuntos de Personal Académico) for a postdoctoral fellowship. The authors wish to thank the laboratory work carried out by Evelyn Martínez (anaerobic digesters) and Mauricio Magos (regrowth experiments). The technical assistance of Danai Montalvan-Sorrosa, Margarita Cisneros and Roberto Briones is greatly appreciated.