Evaluation of the energetic potential of sewage sludge by characterization of its organic composition

An important focus of the ‘energy and water nexus’ is the determination and evaluation of the chemically bound energy that is contained in wastewater in the form of organic substances. In conventional wastewater treatment plants (WWTP), organics that are not converted to gas are removed from the wastewater as sewage sludge. Depending on the local boundary conditions, between 50 and 60% of the organic substance in the untreated wastewater is removed via the primary (PS) and surplus (SS)/excess sludge (Svardal & Kroiss 2011; Schaum 2016). The task is to deploy the organic substances energetically.

The anaerobic transformation of the organic substance to biogas enables the generation of electricity and heat in a combined heat and power plant. By incineration/gasification/pyrolysis of sewage sludge thermal energy is generated which can be used, e.g. to dry sewage sludge or to produce electricity in turbines (Schaum & Lux 2011). For the future, also a direct energy conversion into electricity via microbiological fuel cells seems to be feasible; finally, energy storage is possible, in the form of chemically bound energy, in order to enable flexible energy conversion to biogas or heat/steam (Schaum 2016).

The heating value characterizes the energy contained in a substance as the energy that is released by complete combustion. Consequently, the heating value makes it possible to evaluate the energy potential of sewage sludge. Primarily because of the complexity of analyzing the heating value in bomb calorimeters at WWTPs (in general no bomb calorimeter available and sample preparation, i.e. drying and grinding, is required), organic substances contained in the sewage sludge are, in practice, almost exclusively characterized via the total volatile solids (TVS). Therefore, it is not possible to make a direct statement about the quantitative energy potential of the sludge. Correlations between TVS and the heating value have been published (e.g. Baten 1996; Schmelz 2013); however, an evaluation of the results with reference to a co-fermentation (addition of readily available biomass, such as food waste, to the digestion) is lacking. The heating value of a substance can be derived stoichiometrically from the chemical oxygen demand (COD) and also from an elemental analysis. Even though there are a few relevant publications (e.g. Wouda & Boley 1989; Shizas & Bagley 2004), differentiated correlations for sewage sludge treatment are also unavailable. Because the COD is the classical parameter for wastewater and sewage sludge treatment that is analyzed in WWTPs, calculating the heating value via the COD is relevant.

The evaluation of the energy potential becomes particularly important in terms of the thermic disposal/utilization pathways for sludge, such as, e.g. combustion, gasification and pyrolysis. In this research, TVS, the elemental composition, and the COD will be correlated with the heating value in order to characterize the organic substances energetically.

Based on the results that are available in the literature, the fundamental theoretical principles, together with a survey of the current state of the field, will be presented first. These results will then be validated on the basis of the results of measurements performed on a variety of sludge samples taken from large-scale WWTPs.

The higher heating value (HHV, or gross calorific value) is the quotient of the amount of heat that is released when a substance is combusted completely and the mass of the sample, whereby the following boundary conditions apply (DIN 2000a):

• combustion occurs at a constant volume,

• the fuel temperature prior to combustion and that of the combustion products is 25°C,

• water that is present in the fuel and water that is formed by the combustion of hydrogenated compounds in the fuel are available in the liquid phase following combustion,

• the combustion products from carbon and sulfur are present as carbon dioxide and sulfur dioxide in the gas phase, and

• oxidation of nitrogen does not occur.

The lower heating value (LHV, or net calorific value) differs from the HHV of the water that is present in the fuel, and the water that is formed by the combustion of hydrogenated compounds in the fuel are available in the gas phase at 25°C following combustion (DIN 2000a). That means LHV and HHV are different because the condensation enthalpy is taken into account differently: HHV takes it into account; LHV does not.

Because combustion occurs at a constant volume, only the reaction energy is measured. By contrast, at a constant pressure, the volume work that is performed is also measured. It corresponds to the reaction enthalpy (DIN 2000a). With solid fuels, the difference is only 20–30 J/g and, thus, within the range of measurement uncertainty of the heating value analysis. In general, this effect can, therefore, be neglected (Brandt 1999).

with carbon (C), hydrogen (H), sulfur (S), nitrogen (N), oxygen (O) in mass-%; w = water in mass-%; a = ash in mass-%; k_n = factors for the respective percentage of each substance in the total HHV.

Various methods for the determination of the factors are available, mainly based on statistical approaches (Boie 1957; Friedl et al. 2005). Thereby, the accuracy and validity depend on the investigated fuel. Table 1 shows the relevant factors that are derived from Dulong, Boie, and Channiwala.

For pure substances (arginine, glucose, propionic acid), Shizas & Bagley (2004) determined that there was a strong relationship between the heating value calculated from the enthalpy and the measured HHV (deviation: 1.2–4.5%). These authors also analyzed the digested sludge (DS) from the North Toronto WWTP and determined HHV of 15.9 (PS), 12.4 (SS/excess sludge) and 12.7 MJ/kg TS (DS); the LHV was not determined. However, contradictory observations have also been reported: Heidrich et al. (2011) analyzed wastewater samples from two treatment plants in Great Britain and postulated that a correlation between the heating value and COD does not exist, mainly because not all substances are included via the COD (Janicke 1983).

Based on investigations from the 1930s, Fair and Geyer developed first approaches to the correlation between the HHV and the TVS of sewage sludge, taking into account boundary conditions from wastewater technology (Niemitz 1965; Fair et al. 1967; Kempa 1970; Vesilind 1979). Based on these findings, several investigations on sewage sludge regarding the correlation between HHV or LHV and TVS (TVS in [%]) have been undertaken; see equations below. It should be noted that the respective boundary conditions vary to some extent (number and types of sewage sludge samples, etc.).

Note: Investigation of various types of sewage sludge; PS, DS from Germany and the USA, carbon-containing sludge, and PS enriched with ash.

Note: 122 sewage sludge samples

Note: 36 sewage sludge samples from 13 WWTP from the USA

Figure 1 shows a diagram of the approaches for estimating the HHV or LHV via the TVS. By assuming TVS = 50% and calculating the LHV according to these approaches, the estimated range of the LHV is between 10.5 and 12.5 MJ/kg TS, with a mean of 11.4 ± 1.0 MJ/kg TS. The estimation of the HHV reported by Niemitz (1965) shows a distinct deviation in the gradient, compared to the other approaches. There are several possible causes, e.g. chemical analysis, sampling, etc.

Sewage sludge samples from several municipal WWTPs (30,000 to 1.8 million PE, population equivalents) were investigated, as well as several pure substances (glucose, cellulose, albumin, casein, stearic and palmitic acid), in order to classify the results.

Random samples (sample volume approximately 5 L) of different types of sewage sludge (PS, SS/excess sludge, and DS) were collected, whereby not all of the WWTPs were equipped with all treatment stages (e.g. pre-treatment, digestion). Pure substances with a laboratory purity p.a. (‘pro analysi’, analytical grade) are used. The samples were dried at 105°C and ground using a pebble mill. Subsequently, the following analyses were performed:

• TVS, after heating the TS at 550°C, according to DIN (2000b).

• COD, according to DIN (1986), via potassium dichromate.

• Elemental analysis (CHN), carried out by the Departments of Mechanical Engineering, Institute of Energy Systems and Technology and Material Sciences (TU Darmstadt). The percentage of oxygen was calculated under the assumption that the sum of carbon, nitrogen, hydrogen, sulfur, oxygen, and the residue of TVS/ash is 100%. Following aqua regia digestion (DIN 2001), sulfur was analyzed by inductively coupled plasma optical emission spectrometry (ICP-OES; Perkin Elmer – Optima 3200 DV) (DIN 2009).

• The determination of the HHV was performed with a bomb calorimeter C 200 (IKA), according to DIN (2000a). Benzoic acid was used for calibration. Because only dried samples were analyzed, the water content (w) is 0%. Depending on the type of sample, the portions weighed between 200 mg (PS) and 500 mg (SS/DS).

Table 2 shows the composition of the analyzed types of sewage sludge, sorted according to source (PS, SS/excess, DS). Because of the different process stages, there is a gradient in the carbon content, analytically described via the TVS, COD, and elemental carbon (C) for the respective sewage sludge. Due to the conversion of organic matter to methane and carbon dioxide during digestion, the concentration of organic matter is lower in DS, compared to PS and SS.

Figure 2 shows the carbon content and the COD versus the TVS and demonstrates that, with increasing TVS, the carbon content, as well as the COD, increase linearly. The linear relationship also signifies that the specific energy density of the investigated sewage sludge (gradient of the straight line in Figure 2), defined as COD/TVS, does not differ. Independent of the type of sewage sludge, the COD/TVS ratio was determined to be 1.46–1.56 g COD/g TVS; cf. (‘conferre’, compare) Table 3. In contrast, the specific energy density of fats can reach a ratio of 2.9 g COD/g TVS.

With increasing COD, the heating values increase linearly, cf. Figure 3. One has to distinguish between HHV/LHV on the one hand, and the different methods of determination on the other. Thereby, the basic parameter is the standardized determination of the lower heating value (LHV_DIN) according to DIN via the HHV.

All calculation methods yield a correlation coefficient between 0.92 and 0.98, with respect to LHV/HHV and COD. In Table 4, the various calculation methods are compared. The HHV is approximately 8% higher than the respective LHV_DIN; similar results are found in the literature (Brandt 1999). Compared to the other calculation approaches, elemental composition results in greater deviations. This is due to the factors determined by Boie (1957) and Channiwala (WEF 2009), cf. Table 1, which are based on coal or biomass.

The calculated LHV_COD shows a slight deviation of approximately 9% compared to LHV_DIN; cf. Table 4. The COD was determined following the method of oxidation via potassium dichromate; this means there is a difference between these oxidation processes and the processes of combustion (HHV). One has to consider that, with potassium dichromate, organic compounds might not be completely determined. For example, the analytically determined COD of dimethyl sulfone (component of animal and plant organisms) is approximately 6 g/kg TS, whereas the COD should be approximately 1,360 g/kg TS, according to the stoichiometric calculation (Janicke 1983). LHV_DIN is 17.3 MJ/kg TS. This exemplarily illustrates the differences in analytical methods, as well as possible deviations in the correlation of the LHV. However, this is a very specific example; with sewage sludge, in particular, there is a high correlation between the measurement data and the results of the stoichiometric calculations (Janicke 1983; Zeig et al. 2012).

In Figure 4(a), the correlation between LHV_DIN and TVS for the investigated sewage sludge samples is depicted; data from Wouda & Boley (1989) are also included. It is evident that the result is comparable to the approaches of Baten (1996) and Schmelz (2013). The validity of the determined linear correlations mainly depends on the specific energy density (COD/TVS; cf. Table 3) that must be comparable for the compounds under investigation (sewage sludge). In Figure 4(a), this relationship is obvious. In addition, Figure 4(a) exemplarily shows the LHV of pure substances categorized as fats, proteins, and carbohydrates. Due to the different energy densities (COD/TVS; cf. Table 3) – compared to sewage sludge – the correlation line determined for sewage sludge is not valid for the chosen pure substances. With approximately 36 MJ/kg TS, the LHV for the investigated fats lie above the correlation line, while the LHV for carbohydrates (approximately 14 MJ/kg TS) lie below the line that, with a TVS of 100%, would correspond to approximately 21 MJ/kg TS. These data are relevant for the operation of sewage sludge treatment plants when considering the integration of co-substrates, e.g. food waste and sludge from skimming tanks. In Figure 4(b), the correlation between LHV_DIN and COD for the investigated sewage sludge samples is depicted; data from Wouda & Boley (1989) are also included. The theoretical derivation of the COD from the LHV_DIN (cf. Equation (5)) is in good accordance with the analyzed correlation. In addition, the comparison with the analyzed pure substances shows that the results are largely independent of the kind of analyzed substance (in contrast to TVS). One prerequisite, however, is the oxidizability of the organic matter via potassium dichromate.

The measured values of the LHV_DIN and COD can be evaluated by a theoretical COD balance in the WWTP. Partial flows of the wastewater (effluent, gas, sludge, etc.) can be calculated, based on the PE specific COD load of 120 g COD/(PE d) of the influent to a municipal WWTP, where the treatment consists of a primary settling tank, the activated sludge process, and digestion (Cornel et al. 2011; Svardal & Kroiss 2011). This balancing enables the determination of PE specific COD loads of the various sludge flows (PS, SS, and DS).

By including the PE specific sludge loads, COD concentrations in the sewage sludge types are determined, from which the LHV_COD can be calculated, via Equation (5); cf. Table 5. Furthermore, Table 5 shows a high correlation between these theoretically determined values and the measured data for the investigated sewage sludge samples. Within the framework of the balancing determination, the effect of the polymer used for sludge conditioning on the COD and, correspondingly, on the heating value, was investigated. Assuming that 10 g of the active substance in the polymer per kg TS was added for conditioning, the percentage of COD in the DS is between 0.8 and 3.6%, for a COD concentration of the polymer between 830 and 3,610 g COD/kg active polymer substance (depending on the form in which the polymer was added; whether as a powder, dispersion, or emulsion) (Schaum 2016).

However, the balancing shows that there is a high sensitivity of the PE specific sewage sludge loads. In the literature, a wide fluctuation range is reported, whereby higher specific sewage sludge loads, in particular, e.g. for DS, of 50–60 g TS/(PE d) have been established, e.g. by Imhoff et al. (2009). Experience with WWTP operations, together with the presented comparison of the COD balance, shows that one can expect an annual mean value for specific sewage sludge loads, e.g. for DS, of approximately 38 g TS/(PE d). Here, one has to consider that the sewage sludge loads, according to DWA (2014), and as depicted in Table 5, are based on theoretical estimations that have to be validated with measurement data.

The HHV/LHV are decisive factors for thermal sewage sludge disposal (combustion/gasification/pyrolysis) because they represent the energy potential. The key factor is the organic matter in the sewage sludge that is characterized by various parameters, e.g. TVS, COD, and C. Via these characteristics, it is possible to determine the HHV/LHV:

• Calculating the LHV via the COD provides a very good method for estimating the LHV. Thereby, the calculation is relatively independent of the organic composition, on the condition that the oxidizability via potassium dichromate is given. Furthermore, the COD is a classical parameter in wastewater and sewage sludge treatment, determinable with relatively simple methods. The measurement of the COD in sludge can be performed directly at the WWTP, as part of the routine analyses. This makes it possible to produce the basic data required for energetic balancing of the sludge treatment. Similarly, existing analyses can be used subsequently to calculate the heating value. In contrast, measuring the HHV with a bomb calorimeter is much more complex and is not performed at WWTP. This why there are relatively few published measurement data for HHV.

• In contrast, the correlation of the HHV or LHV via TVS is of limited use. One prerequisite here is a constant specific energy density (COD/TVS), as is given only by sewage sludge. The great advantage is the simple analytical determination of TVS.

• When calculating the HHV or LHV via the elemental composition, one has to consider the validity of the coefficients for sewage sludge, because most approaches have been developed for using coal. Depending on the organic composition, it might be necessary to adjust the coefficients, e.g. when applying co-substrates. From the analytical point of view, the determination of the elemental composition is relatively complex compared to the other calculation approaches, and is comparable with the direct determination by bomb calorimeter.

Theoretical COD balancing can be used to validate the measurement data when determining the organic matter in PS, SS/excess, and DS. With this approach, characteristics relevant for the energy potential in sewage sludge treatment, that means COD and HHV/LHV, are derived.

The research project is funded by the Fritz and Margot Faudi Foundation (project number 81).