Suitability of fecal sludge from composting toilets as feedstock for carbonization

Most people living in high-income countries have access to sewer-based sanitation, e.g. 95% of households in Germany (DSTATIS 2015). Sewer-based sanitation is based on well-established technologies enabling hygienically safe treatment of wastewater. Thereby its long-living infrastructure is based on high capital expenditure (capex) and requires high operational expenditures (opex), mainly for maintenance and electric energy. Due to their centralized structures most technologies are increasingly inefficient and almost unsuitable for sparsely populated regions, e.g. rural areas. Wastewater is increasingly considered as a resource, but the technological costs for nutrient recovery are high, because of their high dilution (Larsen et al. 2013; Anand & Apul 2014).

On-site sanitation is widespread in many low-income countries, e.g. 65–100% of urban areas in sub-Saharan Africa (Strauss et al. 2000). On-site sanitation has no need for a sewer system and commonly includes on-site collection (e.g. from pit latrines and septic tanks), transport (e.g. in trucks) and off-site treatment of the obtained fecal sludge (FS). However, due to insufficient decentralized sanitation infrastructure, FS often reaches the environment untreated causing environmental pollution as well as severe health problems (Strande et al. 2014).

Composting toilets, being also denominated as dry toilets or bio-toilets (Lopez Zavala & Funamizu 2006), are on-site sanitation facilities that are used as an alternative to centralized sewer-based sanitation systems. Composting toilets are operated waterless and typically sawdust is added for odor reduction and for facilitating a post-treatment by aerobic composting. Composting toilets provide the opportunity of being operated fully decentralized, reduce capex and opex for sanitation infrastructure, save water and may even produce a valuable fertilizer. Thus, composting toilets are most commonly used in rural areas and in areas with water shortages (Anand & Apul 2014).

Fecal sludge from composting toilets (CFS) is typically treated by aerobic composting that allows for the stabilization, sanitization and reuse of the treated solids as fertilizer or soil conditioner (Anand & Apul 2014). However, aerobic composting requires several weeks to provide safely sanitized and fully stabilized compost, which induces the need for high reactor volumes or high demand of treatment space. Depending on the availability of treatment space and time, this can lead to high opex. Also, aerobic composting has to be performed under controlled conditions, e.g. frequent aeration and turning, to maintain process temperature levels leading to hygienically safe sanitation (Niwagaba et al. 2009; Germer et al. 2010; Sossou et al. 2014).

Carbonization, including hydrothermal carbonization (HTC) and pyrolysis, proved to be an effective method for the conversion of different types of biological sludge into high-quality char with a broad range of applications (Libra et al. 2011). Carbonization can provide safe sanitation within minutes or hours due to the high process temperatures (Feachem et al. 1981). During carbonization the feedstock volume is significantly reduced and char is produced. The economical valorization of the char can lead to an opex decrease of the entire sanitation process (Diener et al. 2014). Thus, carbonization can be considered as an alternative valorization technology for CFS, overcoming common challenges of aerobic composting.

Pyrolysis of FS from various sources has been subject to few studies in the past, including raw feces (Ward et al. 2014), solids from the settling chamber of a septic tank (Liu et al. 2014), FS from septic tanks and pit latrines (Gold et al. 2018) and sewage sludge from wastewater treatment plants (Bagreev et al. 2001; Inguanzo et al. 2002; Bridle & Pritchard 2004; Wang et al. 2012; Huang et al. 2017; Zhu et al. 2017). HTC of FS from various sources has also been investigated by several researchers, including primary sewage sludge from a settling tank of a wastewater treatment plant (Danso-Boateng et al. 2013, 2015), FS from a septic tank (McGaughy & Reza 2017) and raw feces and sewage sludge using microwave HTC (Afolabi et al. 2015; Afolabi & Sohail 2017). These studies demonstrated that FS is a promising feedstock for carbonization. FS char exhibits beneficial properties, making it a high-value product that can be used, for example, as solid fuel, for soil enhancement, and for carbon sequestration (Gold et al. 2018). However, the use of CFS as feedstock, for either pyrolysis or HTC, has not been studied, so far.

The application of carbonization for sanitation of FS still faces challenges. HTC requires high pressure reactor and safety equipment, which induces high capex and opex. Also, HTC char has to be dewatered and dried, if used as solid fuel (Afolabi & Sohail 2017). Pyrolysis, conversely, requires a dry input material. Pyrolysis of sewage sludge with ≥50% total solids (TS) in a single reactor assembly was reported (Gerber 2011). However, a sophisticated reactor design using screw or double-screw feeding or conveying is required (Zhu et al. 2017). FS that commonly exhibits TS contents <50% (Strauss et al. 2000; Semiyaga et al. 2017; Strande et al. 2018) has to be mechanically dewatered and dried before pyrolysis, which increases the energy demand of the process. One approach to decrease the high energy demand for pyrolysis is the direct combustion of pyrolysis gas and (non-condensed) liquid to provide process heat for drying and pyrolysis, as demonstrated by several authors (Wang et al. 2012; Liu et al. 2014). However, the feasibility of an energy self-sufficient process operation is limited by the feedstock properties, such as TS content, and the product properties, such as amount and calorific value of the pyrolysis gases and liquids.

The objective of this study was to investigate the suitability of CFS as feedstock for pyrolysis and HTC. We performed chemical analysis on CFS and on its carbonization products that were obtained at standard carbonization conditions (pyrolysis at 500 °C for 1 h and HTC at 200 °C for 4 h). Experimental feedstock and char characteristics were compared to the available literature data of FS. Furthermore, we investigated how the addition of sawdust (SD) to FS (as typical for CFS) influences the energy balance of a model pyrolysis process. By addition of SD, TS and chemical oxygen demand (COD) of the input to the pyrolysis process increases and, in turn, the required energy for drying per kg of feedstock decreases. Thus, the aim was to demonstrate that SD addition can be used as an inexpensive method to achieve energy self-sufficient pyrolysis operation when using FS with low TS as feedstock. As an advantageous side-effect, the addition of SD before sludge drying was also shown to enhance drying performance because of an increased heat exchange capacity and an increased bed volume (Li et al. 2016), which consequently decreases the required drying reactor capacity. It has also been demonstrated that the char surface area increases with the addition of SD before pyrolysis of different types of biological sludge, such as sewage sludge (Huang et al. 2017) or anaerobically digested pig manure (Troy et al. 2012), adding value to the char.

CFS samples were collected from six different composting toilets (Ökolocus/Goldeimer, Germany). Three of these toilets were operated on a household scale with 4–10 frequent users over a period of six months from May to October 2016, in Germany. The other three toilets were operated at music festivals in Germany with an unknown number of users over a period of 5 days in July and August 2016, respectively. One of the sample toilets used direct urine diversion, while the other five toilets were operated with liquid drainage, i.e. urine and feces were mixed in the collection chamber and excess liquid was drained through a hole at the bottom of the container. Softwood sawdust was used as bulking material and toilet paper was added by the users. Each toilet was sampled once after operation. Therefore the toilet content was completely emptied, manually mixed with a shovel in a pile and then equally distributed on a tarpaulin. Five equivolumetric samples of approximately 600 g were randomly taken from different parts of the tarpaulin and merged to a full sample. The samples were stored at −18 °C prior to the experiments.

Pyrolysis and HTC were performed using the three samples from household scale CFS.

For CFS pyrolysis 500 g of each tested sample was dried for 24 h at 105 °C, yielding between 100 and 150 g dried sample. Pyrolysis was performed in a muffle furnace (XCHAMB 1,280 by XERION, Germany) using an aliquot of 70 g of the dried sample in one repetition. The muffle furnace was heated at a rate of 8 °C min⁻¹ from ambient temperature to 500 °C and the temperature was held for 1 h. The pyrolysis furnace was purged with N₂ during the entire operation. Purging started 30 minutes prior to the experiments.

HTC was performed in a 500 mL stirred autoclave (BR-300, Berghof Products + Instruments, Germany using a BTC 3,000 control panel) at 200 °C for 4 h with each sample in one repetition; 375 g of sample was used per experimental run according to the holding capacity of the reactor. All samples with TS concentrations higher than 20% were diluted to 20% TS with deionized water to guarantee proper mixing of the samples by the stirrer (range of TS of the used samples: 6.6–28%). After the HTC reactor cooled down to ambient temperature the produced chars were pressure filtered using a vacuum pump and MN 615 paper filters (12 μm pore size, by Marcherey-Nagel, Germany). The filter cake was subsequently dried at 105 °C for 24 h before analysis.

The tested combinations of temperature and reaction time for pyrolysis and HTC represent average reaction conditions that are commonly used in literature (Libra et al. 2011). In our case, all studies that used fecal sludge as feedstock for pyrolysis or HTC applied either of these reaction conditions (Danso-Boateng et al. 2013, 2015; Liu et al. 2014; Ward et al. 2014; Afolabi et al. 2015; Afolabi & Sohail 2017; Gold et al. 2018), thus, allowing for good comparison of our results with those from literature.

CFS samples and gained char samples were analyzed for TS, ash, C (carbon), H (hydrogen), N (nitrogen), S (sulfur) and higher heating value (HHV). CFS was additionally analyzed for volatile solids (VS).

HHV of CFS and char samples were measured using a calorimeter (6400 Automatic Isoperibol Calorimeter by Parr Instrument GmbH, Germany).

Mass and energy balances for a model pyrolysis process of CFS (Figure 1) were calculated. As feedstock CFS was assumed to be a mixture of FS and SD. For the calculations we used literature data as a basis, taking into account data of FS, sewage sludge (if data on FS were unavailable) and SD.

All performed interpolations can be found in the E-supplementary data of this work in the online version of the paper.

The analyzed CFS exhibited a TS of 21 ± 9%, a VS of 90 ± 3%_TS and an ash content of 9 ± 3%_TS (Table 1). Measured carbon (C) and nitrogen (N) contents of CFS were 48 ± 2%_TS and 2.3 ± 1.1%_TS, respectively, measured COD was 290 ± 135 g kg_wet^–1. These characteristics differ from those of FS, which possess lower TS (<20%), VS (<80%_TS) and COD (<100 g L⁻¹), along with higher ash content (>10%_TS). The measured HHV of CFS was 20.2 ± 2.3 MJ kg^–1, and thus, slightly higher than that of pure feces which are reported to be around 17 MJ kg^–1 (Rose et al. 2015).

We presume that the addition of SD and the solid-liquid separation in CFS are the main reasons for these deviations. SD contributes to the increased TS because of its own high TS (Ronewicz et al. 2017) and liquid separation leads to the additional loss of water. In turn, the high TS of CFS leads to an increase of COD and N concentration per wet mass, due to the increased amount of solids per kg. With SD addition in CFS, VS also increases and ash content decreases compared to FS, because of the high fraction of organic matter contained in SD (ash content SD: 0.1–8%_TS (Libra et al. 2011) compared to ash content of pure feces with 7.5–16%_TS (Rose et al. 2015)). Microbial degradation during settling and storage of the materials in the toilets will also alter the chemical composition, in particular, a decrease of VS and COD and increase of ash content can be expected (Libra et al. 2011). SD addition causes porous bulking of CFS during storage, thus fostering aerobic degradation of the material. Conversely, FS forms a liquid–solid suspension (Strande et al. 2014) and is thus expected to predominantly undergo anaerobic degradation during storage.

Char that was produced by pyrolysis of CFS in this study exhibited an ash content of 24 ± 5%_TS, which is considerably higher than the ash content of FS-derived pyrolysis char (>37%_TS) (Table 2). This is most likely related to the comparably low ash content of CFS (9 ± 3%_TS) and the high ash content of FS (34 ± 21%_TS). The low ash content of CFS char indicates a high organic content, which explains the comparatively high C (64 ± 5%_TS) and N (2.9 ± 1.4%_TS) contents, high HHV (24.3 ± 1.9 MJ kg⁻¹) and low char yield (32.6 ± 2.6%). C and N account for 84.0 ± 0.5%_TS and 3.9 ± 1.9%_TS, respectively, when calculated on an ash-free basis being well in line with values of 80–85%_TS and 4–8%_TS calculated from the literature values of FS chars, respectively (Liu et al. 2014; Ward et al. 2014; Gold et al. 2018). The relatively low yield of CFS char is attributed to SD addition and, hence, the low ash content of the feedstock. This can be expected, as during pyrolysis inorganic compounds remain in the char, while organic compounds are partly volatized (Vreugdenhil & Zwart 2009; Libra et al. 2011). When applying similar pyrolysis conditions to SD, even lower char yields of around 25–30% were shown (Yan et al. 2011; Ronsse et al. 2013), which also supports the aforementioned reasoning. A decrease of char yield with addition of SD to sewage sludge was also demonstrated by Troy et al. (2012) and Huang et al. (2017). The HHVs of the here gained CFS-derived pyrolysis chars (24.3 ± 1.9 MJ kg^–1) are considerably higher than the literature values of FS chars (9–18 MJ kg^–1). This, again, can be related to the high organic content of CFS-derived pyrolysis char (indicated by its low ash content) as well as to the variation of feedstock compositions between FS and CFS. HHV values reported by other authors of, for example, 17.9 ± 0.3 MJ kg^–1 (Ward et al. 2014) and 9.1 ± 1.0 MJ kg^–1 (Gold et al. 2018) are gained from chars with ash contents of 37 and 67%, respectively.

When using HTC on CFS the produced char exhibited an ash content of 22 ± 12%_TS and C and N contents of 57 ± 5 and 2.7 ± 0.8%_TS, respectively. In comparison to the literature the ash content of FS HTC char is slightly higher (25–36%_TS), and C content is lower (38–49%_TS) (Table 2). The HHV of CFS-derived by HTC char was 22.9 ± 2.4 MJ kg^–1 and is, thus, in the range of FS-derived HTC chars (18.5–25.0 MJ kg^–1). The comparatively high HTC yield of CFS char (75.3 ± 6.9%) can be attributed, again, to the addition of SD. Zhang et al. (2017) observed a considerable increase of HTC char yield when increasing the amount of SD in a sewage sludge–sawdust mixture used as feedstock for HTC. The moderate fluctuations of the observed yields from CFS HTC chars may be explained by the different composition of the used FS samples, for instance by different amounts of SD or different period of storage before experiment. Also the unequal TS concentrations of the tested samples may have had an impact on the HTC char yield, as low TS were shown to slightly decrease char yield in HTC of FS (Danso-Boateng et al. 2013).

Figure 2(a) shows the calculated ratio of SD to FS on a wet basis X_SD/FS (Equation (6)) that is required to achieve energy self-sufficient operation at varying FS TS (TS_FS). Without the addition of SD (X_SD/FS = 0), energy self-sufficient operation is achieved at pyrolysis temperatures of ≥500 °C and at TS_FS ≥ 30%. At pyrolysis temperatures of <500 °C TS_FS would have to be even higher to achieve energy self-sufficiency. As FS with a TS > 20% is unlikely to be found in practice (Table 1), energy self-sufficient operation without SD addition is not possible. Liu et al. (2014) calculated energy self-sufficient process operation using FS from the settling chamber of a septic tank. They used a similar process scheme and assumptions for their investigation and calculated that TS_FS must be ≥43% at 550 °C pyrolysis temperature to achieve energy self-sufficient operation. This TS_FS is even higher than the values derived in this study (TS_FS ≥ 30% at 500 and 600 °C) and supports the aforementioned presumption.

With decreasing TS_FS energy self-sufficient operation can be achieved by the addition of SD which increases X_SD/FS (Figure 2(a)). The rate at which X_SD/FS has to increase with decreasing TS_FS varies among the different pyrolysis temperatures, as reflected by the inclinations of the graphs. Above 400 °C pyrolysis temperature has a small influence and above 500 °C its influence is negligible. However, the pyrolysis temperature has a high influence at 300 °C, as a considerably higher X_SD/FS is required than at higher pyrolysis temperatures. These observations can be explained by the decrease of the solid product fraction (i.e. char) with increasing pyrolysis temperature (Figure 2(c)), which in turn causes an increase of the gaseous and liquid product fractions. Thus, more energy for process heating is available from burning of pyrolysis liquids and gases. Above 500 °C no influence of X_SD/FS is observed, corresponding to very little changes in the distribution of the products above these temperatures (Ronsse et al. 2013; Liu et al. 2014). At the lowest investigated TS_FS of 5%, circa 0.55 kg_SD,wetkg_FS,wet⁻¹ have to be added to achieve energy self-sufficient operation regardless of pyrolysis temperature.

In Figure 2(b), X_SD/FS,TS (Equation (7)) required for energy self-sufficient operation was plotted over varying TS_FS to demonstrate the fraction of total solids in the char that are derived from SD and from FS, respectively. The graphs demonstrate that at decreasing TS_FS, X_SD/FS,TS has to increase exponentially. For example, at 5% TS_FS an X_SD/FS of 0.55 (Figure 2(a)) corresponds to an X_SD/FS,TS of around 8.5 (Figure 2(b)). Thus, circa 85% of TS of the char is derived from SD and only circa 15% of the char TS originates from FS. This clearly illustrates that when using FS with low TS as feedstock the gained CFS char will be predominantly SD-derived at energy self-sufficient operating conditions.

As shown in Figure 2(c), Y_CFS′ increases with an increasing TS_FS under energy self-sufficient operation. These observations can be expected and are related to the fact that we assumed Y_CFS′ as the sum of Y_SD and Y_FS (Equation (16)). With increasing TS_FS SD addition has to increase to achieve energy self-sufficient operation as shown in Figure 2(b). Y_SD is generally lower than Y_FS, thus, Y_CFS′ decreases with higher X_SD/FS,TS (Ronsse et al. 2013; Liu et al. 2014; Ward et al. 2014). Figure 2(c) also shows that Y_CFS′ decreases with increasing pyrolysis temperature, which is related to the fact that both Y_SD and Y_FS decrease with increasing pyrolysis temperature.

Figure 2(d) shows HHV_CFS,char′ with increasing TS_FS and at different pyrolysis temperatures under energy self-sufficient operation. At 300 °C the pyrolysis temperature HHV_CFS,char′ increases with increasing TS_FS while it decreases at pyrolysis temperatures ≥400 °C. HHV_CFS,char′ is similar at a TS_FS of circa 20% for all pyrolysis temperatures. This interesting finding can be explained by the fact that the HHV of SD-derived char (HHV_SD,char) increases with increasing pyrolysis temperature, while the HHV of FS-derived char (HHV_FS,char) decreases (Ward et al. 2014; Ronewicz et al. 2017). When TS_FS increases, less SD has to be added to achieve energy self-sufficient operation, thus, HHV_CFS,char′ approximates HHV_FS,char. In turn, when TS_FS decreases, more SD has to be added and HHV_CFS,char′ approximates HHV_SD,char. This finding also indicates that if a high HHV_CFS,char′ is desired and FS with TS_FS < 20% is used, i.e. a significant amount of sawdust has to be added (circa X_SD/FS,TS > 1), a high pyrolysis temperature should be favored, while a low pyrolysis temperature should be favored at TS_FS > 20%.

Other than in industrialized countries, where composting toilets are common for on-site sanitation, in developing countries predominantly pit latrines and septic tanks are used (Strande et al. 2014, 2018). As shown in Table 1, FS from these sources exhibits low TS (<10%). When applying the proposed approach of adding SD to FS prior to pyrolysis, several considerations have to be made. As indicated by our results a relatively high amount of SD has to be added to achieve energy self-sufficient operation, because of the low TS. Thus, the approach is limited by the availability of SD. Optionally, instead of SD, other organic bulking materials such as rice husk or corn husk may be used to increase TS (Libra et al. 2011; Ronsse et al. 2013). However, their applicability needs to be elucidated as these materials may behave differently during pyrolysis. The amount of required SD may also be reduced by an additional mechanical dewatering step prior to SD addition that increases TS. For the approach discussed here, we generally recommend favoring low pyrolysis temperatures (300 °C) when aiming for high yield, e.g. when char should be produced for soil amendment. Conversely, when aiming for char with high HHV, e.g. for fuel production, a high pyrolysis temperature should be applied (600 °C). The latter aspect is contrary to pyrolysis of pure FS that exhibits an opposite trend, i.e. an increased HHV is obtained at lower pyrolysis temperatures (Ward et al. 2014; Ronewicz et al. 2017).

CFS exhibits favorable feedstock characteristics for pyrolysis, although a broader fundament of data, including fecal sludge from different sources, e.g. people with different diets, as well as different bulking materials is needed for further validation. The addition of SD to FS enables energy self-sufficient pyrolysis operation, even when using FS with low TS (≥5%). It also allows the production of pyrolysis char with high HHV and low ash content. HTC of CFS is also suitable, however, due to its high TS a dilution of the feedstock is required to enable proper reactor mixing. In summary, carbonization provides a promising concept for the time-efficient, decentralized and hygienically safe valorization of CFS that can be applied as an alternative to conventional treatment approaches.

The authors thank Ronny Neuenfeldt for technical support and Benjamin Wirth for proofreading. The authors thank Goldeimer gGmbH, Hamburg, Germany, and Ökolocus GmbH, Leipzig, Germany, for providing fecal sludge from composting toilets. The project BioFAVOR was founded by the German Federal Ministry of Education and Research (BMBF) within the scope of the ideas competition ‘Neue Produkte für die Bioökonomie’ (grant number 031B0254). Falk Harnisch acknowledges support by the Federal Ministry of Education and Research (Research Award ‘Next generation biotechnological processes – Biotechnology 2020 +’) and the Helmholtz-Association (Young Investigators Group). This work was supported by the Helmholtz Association within the Research Program Renewable Energies.

The Supplementary Material for this paper is available online at http://dx.doi.org/10.2166/washdev.2019.047.