Investigation of Thermal Decomposition and Gases Release from Pre-Drying Municipal Solid Waste (PMSW) Via Pyrolysis Technology

The pre-drying municipal solid waste (PMSW) was collected from dumping sites. The composition of PMSW consisted of food waste, polymeric plastics, paper, cloth pieces, and rubber pieces. The materials contained food (food residue, fruit peels, and vegetable trimmings), plastics (disposable bottles used for water, soft drink, juices, disposal food boxes, and soft drinks, food spoon, and forks, polythene service bags, biscuits wrappers, chips wrappers, straws, binding sticking tape), paper (cardboard boxes, marketing/advertisement broachers, product booklets, cigarette packet, stickers on cardboard, tissue paper, disposal plates), biomass (tree and plant leaves, disposal chopsticks, walnut shells, yard trimming), and rubber (tire pieces, broken chapel pieces, and tube residues). Initially, the collected materials were air-dried for 3 days and then shredded into smaller pieces. Then, a mixture of individuals was prepared with a mass ratio of 48:29:13:7:3 for food waste, polymeric plastics, paper, cloth pieces, and rubber pieces. To achieve optimal results, the pulverized materials were homogenized and sieved to average particle size to 1 mm and then kept in an oven for 12 hours at 105°C. The ready-mixed sample was then put in a desiccator until further experiments and/or their characterization.

The proximate analysis encompassed the contents of water volatiles and ash present in the PMSW. The elemental (chemical) composition (wt. %) was carried out by using an elemental analyzer (model 2400), by which the quantities of elemental carbon, hydrogen, sulfur, and nitrogen were examined. The oxygen and fixed carbon were computed by the difference method. The results of the proximate and ultimate analysis are shown in Table 1.

A thermogravimetric analyzer (TGA) coupled with a mass spectrometer (MS) was used for pyrolysis PMSW. The TG-MS experimental setup is shown in Fig. 1.

Figure 1

The aim was to investigate the weight loss and the gas species evolved during the process, therefore the small weight of about 8 mg was used in all the experiments for the optimal results [26, 58]. The lower heating rates were selected to assist the release of the compounds with the lower molecular weight. To minimize the systematic errors and to establish the baseline, one blank test was conducted prior to the experiments at the heating rates 5 and 15°C/min. The metered sample was loaded in a crucible in TGA, and then each sample was kept under an isothermal condition at 30°C for an hour. Afterwards, the targeted samples were pyrolyzed from room temperature to 900°C at smooth heating rates of 5 and 15°C/min. Argon (Ar) with a flow rate of 500 mL min⁻¹ was supplied. The heated capillary collected the mixture of releasing gases from TGA and conveyed it to MS to analyze and define the arrangement of the released components. In this manner, evolved gas species can be easily identified online. To attain effective results, each experiment was repeated to ensure reproducibility.

MS ionized the volatile species evolved during the thermal breakdown of sample interfaced with TGA. The heated capillary related to TGA supplied identical ions in the form of signals to the MS. Therefore, those received ions were identified and distinguished based on their mass-charge (m/z) ratios. The m/z ratios for hydrogen, methane, steam carbon monoxide, ethane, Argon (Ar), and carbon dioxide are detailed in Table 2. Ar (m/z = 40) was set as a degree to assure the consistency of the released gas components. All the mathematical calculations to determine the gas generation rate were done with documented signals applying the method reported through many studies [42, 43]. The unidentified signals were normalized by the equation below.

Normalized signals for key molecule fragements (1) "i"=ICi*500ICAr*Wt.pre−drying {"\rm{i}"} = {{{\rm{I}}{{\rm{C}}_{\rm{i}}}*500} \over {{\rm{I}}{{\rm{C}}_{{\rm{Ar}}}}*{\rm{Wt}}{._{{\rm{pre}} - {\rm{drying}}}}}}

Where, IC_i shows the molecular m/z signals for molecular ion fragments “i” (arbitrary unit), Argon (Ar) flow rate of 500 ml min⁻¹ during the isothermal period IC_Ar is the m/z signals for Ar, and Wt._{pre-drying MSW} is the weight of the PMSW sample (mg).

The kinetics gives valuable knowledge for the design and optimization of thermal apparatuses. Meanwhile, it calculates the weight loss performance, chemistry of the solid products, distribution of the products, and heat transfer information of the equipment [44, 45]. For kinetic assessment, two methods can be applied as isothermal and non-isothermal. Comparatively, the non-isothermal is considered the easier method due to preventing the variations of physical and chemical characteristics of the sample and giving useful information through a minimum number of experiments [46, 47]. Hence the conversion degree (α) can be calculated from experimental data by using the equation as follows: (2) α=w0−wtw0−wf \alpha = {{w0 - wt} \over {w0 - wf}}

Where w0 is sample mass at the starting point, wt. is the mass of PMSW at a time and temperature T, and wf is mass leftover after the reaction.

In the case of non-isothermal, the mass of any feed-stock is metered as a function of temperature deviates with the time T=T(t). Therefore, heating rate (β) may be stated as; (3) β=dT/dt \beta = {\rm{dT}}/{\rm{dt}}

At a linear temperature, the β is the independent function such as fractional conversion function f(α) and temperature function k(T), so the kinetic can be described as; (4) dαdt=βdαdT=k(T)f(α) {{d\alpha} \over {dt}} = \beta {{d\alpha} \over {dT}} = k\left(T \right)f\left(\alpha \right)

The rate constant (k) on the temperature can be calculated via the Arrhenius expression. (5) k(T)=Aexp(−EaRT) k\left(T \right) = A{exp}\left({- {{Ea} \over {RT}}} \right)

Where A is the pre-exponential factor (s⁻¹) Ea is the activation energy Jmol⁻¹, and R the gas constant 8.314 kJmol⁻¹.

By merging Eqns. (4) and (5), the reaction rate may be given as follows: (6) dαdT=(Aβ)exp(−EaRT)f(α) {{d\alpha} \over {dT}} = \left({{A \over \beta}} \right) exp \left({- {{Ea} \over {RT}}} \right)f\left(\alpha \right)

Integrating to Eq. (6) can be solved as. (7) G(α)=∫0αdαf(α)=∫0T(Aβ)exp(−EaRT)dT=(AEaβR)P(x) G\left(\alpha \right) = \int_0^\alpha {{{d\alpha} \over {f\left(\alpha \right)}} =} \int_0^T {\left({{A \over \beta}} \right) exp \left({- {{Ea} \over {RT}}} \right)dT = \left({{{AEa} \over {\beta R}}} \right)P\left(x \right)}

G(α) and P(x) are the integrated form of fractional conversions and temperature integrals. So, here P(x) may be written as;

Here x=EaRT x = {{Ea} \over {RT}} and P(x) denotes the temperature integral as: (8) P(x)=∫x∞−(exp(−x)x2)dx P\left(x \right) = \int_x^\infty - \left({{{{exp}\left({- x} \right)} \over {{x^2}}}} \right)dx

Based on the integral and differential form of Eqn. (6), the kinetic parameter activation energy can easily be calculated through Flynn-Wall-Ozawa (FWO) method. FWO is an approximate integration technique ensuing from the Doyle equation. It has been effectively applied in the inquiry of non-isothermal calculations. Using FWO, the Ea can be enthusiastically determined to agree to a specified α and the lack of any information regarding the reaction [48, 49, 50]. As a result, this technique may be employed to investigate the value of Ea with the assumption of the reaction mechanism.

Through the FWO technique, Eqs. (6–7) can be transfigured into: (9) lnβ=ln[AEaRG(α)]−2ln(EaRT)−(EaRT) ln \beta = \ln \left[ {{{AEa} \over {RG\left(\alpha \right)}}} \right] - 2\ln \left({{{Ea} \over {RT}}} \right) - \left({{{Ea} \over {RT}}} \right)

Where G(α) is a constant at a specified α; however; 2ln(EaRT) 2\,\ln \left({{{Ea} \over {RT}}} \right) changes with some degree of range. Though P(x) does not have the precise solution, therefore, it is approached by Doyle's expression through the FWO model as: (10) P(x)=0.00484*exp(−1.0516x) P\left(x \right) = 0.00484*{exp}\left({- 1.0516x} \right)

Consequently, Eq. (9) is transformed in the FWO method as below. (11) lnβ=ln[AEaRT(α)]−5.331−1.052EaRT {ln}\beta = {ln}\left[ {{{AEa} \over {RT\left(\alpha \right)}}} \right] - 5.331 - 1.052{{Ea} \over {RT}}

Noticeably, making log β versus 1T {1 \over T} should provide the linear plots, and the Ea may be directly computed commencing to the gradient at conversion rate points α = 5 to 35.

The quantitative information of the decomposed solid sample is commonly presented by thermogravimetric (TG) and differential thermo-gravimetric (DTG) thermographs. TG shows the total conversion (weight loss) in percentage (%) as an inclined curve, whereas; the DTG shows the rate of conversion (d α/dt min⁻¹) as peaks at specific temperature ranges. TG and DTG profiles for the non-isothermal pyrolysis of the biomass at heating rates; 5 and 15°C/min are presented in Fig. 2. Pyrolysis of disintegrating solid mass is a diverse process that usually involves three specific temperature sub-processes: evaporation, primary decomposition, and secondary decomposition [47, 59]. In the 1^st (evaporation) stage, the free moisture and some unstable hydrocarbons are released in the temperature range 100–200°C, resulting in minor weight loss mainly depending on nature and the conditions of the process. In the 2^nd stage, the devolatilization (primary decomposition), the release of unstable polymers-plastics and biopolymers-bio-mass has occurred in the temperature range ~172–532°C. This stage is accounted for the major loss in weight therefore sometimes also known as the active stage of the pyrolysis. In the 3^rd (secondary decomposition) stage, which begins from the end of the active stage, a secondary breakdown of the sample results in char formation and the char formation reactions sustained up to upper temperatures of the process. In the secondary phase, minor weight losses are achieved over the primary phase hence this stage is known as a passive stage. The 2^nd and 3^rd phases were found to depend on each other, primarily on the general chemical composition of the reacting solid, specifically the structural bonding. The substantial decomposition of ~43% and 39% occurred in the devolatilization phase in the temperature ranges of 200–700°C and 200–510°C at 5 and 15°C/min respectively obviously due to the presence of higher volatiles in the mixture. Below 200°C, the TG curves (Fig. 2a) under both the heating rates showed a similar weight loss ~ 3% i.e. witnessed the presence of a smaller quantity of water in the samples. Under low heating rate, the sample reached its maximum decomposition of 47% at temperature ~716°C whereas at a higher heating rate, the decomposition of 43% at temperature 737°C. Henceforth, the lower heating rate exhibited higher decomposition ~4% than the higher heating rate. Very similar observations of weight loss at the primary and final phases were also reported by [51, 60] during the co-pyrolysis of a binary mixture of two plastics; hence our sample was also the mixture of several types of plastics. TG pyrolysis profiles of the PMSW samples for both the heating rate are illustrated in Fig. 2a.

Figure 2

Differential thermogravimetric (DTG) thermographs of the PMSW samples at both the heating rates are shown in Fig. 2b. For low heating rate, the four conversion phases were observed in the temperature ranges from 50–180°C, 230–360°C, 360–500°C, and 605–690°C. However, only two conversion phases were observed at the lower heating rate in the temperature ranges from 50–182°C and 230–515°C. In the first conversion phase, both samples followed almost the same. i.e., small decomposition profile for water removal ~3% in the temperature ranges from 50–180°C and 50–182°C for each heating rate. The devolatilization, commonly represented by second and third conversion phases, behaved differently; however, it showed a major weight loss at both heating rates. For lower heating rates, the second and third conversions conjointly represented the devolatilization (primary decomposition) between 230–530°C, henceforth cause to a major loss in weight ~43%. On the other hand, the pyrolysis at a higher heating rate, the thermal conversion stratagem of the PMSW sample was completely different; henceforth could reach only up to the 2nd conversion phase with the total weight loss ~39% and left 56% as unreacted part. The reason behind the two conversion phases for higher heating rate instead of four in case of the lower heating rate attributed to the sample minimum residence time even under the same process condition. In addition, at a higher heating rate, several blurring (dull) peaks were demonstrated in the temperature ranges from 220–380°C, and 550–690°C evidenced that the sample endeavoured to decay under fast heating just penetrated it partially, hence degraded slightly less. Whereas, under the lower heating rate condition, the decomposition process continued and entered the fourth conversion phase. The fourth (secondary decomposition) conversion phase is known for some char formation and its decomposition. In this phase, the decomposition of ~1.5–2% in the temperature range from 605–690°C occurred, hence strongly attributed to the self-gasification of char which was not achieved at higher heating rate. The final residue from the fourth conversion phase was approximately 53% as char. The pyrolysis results of PMSW at two different heating rates reported that the maximum residence time permits more heat entry into the innermost core of the solid particles in the case of moderate and/or slow heating. On the contrary, there was insufficient heat entry; hence, less decomposition occurred at fast heating. Overall, the lower heating rate DTG curves exposed the PMSW decomposition in four conversion phases. The devolatilization region ascertained itself the most influential weight loss phase. The details of the conversion phases at the applied heating rates are given in Table 3.

The gas generation trends present the qualitative information of any decomposed solid sample during thermal degradation, which can easily be determined by a mass spectrometer (MS). MS identifies the ion intensity of each gas specie based on their mass to charge (m/z) ratio. A semi-quantitative approach that effectively works based on signals corresponding to each ion fragment was used to assess the gas arrangement during both processes [30, 31]. The release of detected common fragments included; H₂, CH₄, C₂H₂, CO, and CO₂. The aim was to investigate only combustible gases; hence CO₂ was excluded. In addition, the concentration of C₂H₂ was too little to be appropriately displayed; therefore, it was also excluded. Fig. 3 depicts the generation rates of H₂, CO, and CH₄ and their yields obtained from pyrolysis of PMSW at the heating rates of 5 and 15°C/min. As the sample contained a considerable amount of plastics (29% share) and lignocellulosic components, the pyrolytic decomposition followed the decay profile via a radical mechanism for plastics, whereas a chain of exothermic endothermic nature of reactions for the lignocellulosic fraction.

Figure 3

The H₂ generation trends of the pyrolysis of the PMSW are illustrated in Fig. 3a. For both heating rates, the generation of H₂ was elevated from 0.5 to 4 ml min⁻¹ g⁻¹ in the temperature ranges from 400–430°C but suddenly decreased to 2 ml min⁻¹ g⁻¹ at 500°C. Afterwards, a rapid increment from 2 ml min⁻¹ g⁻¹ to 12 ml min⁻¹ g⁻¹ and 2 ml min⁻¹ g⁻¹ to 9 ml min⁻¹ g⁻¹ was observed in the temperature ranges 500–760°C and 500–710°C for lower heating rate and higher heating rate, respectively.

It is strongly believed that the H₂ production is mainly occurred due to the primary cracking of hydrocarbons and their subsequent decomposition reactions [34, 47]. According to [52, 53], the upper-level temperature provides favorable circumstances to the thermal disintegration of carbon-carbon bonds hence; high probability of water gas shift and reforming of methane reactions has occurred, resulting in the boosted H₂ production. Apart from that, the secondary cracking of char/ash catalytic assistance by the paper and biomass fractions in the mix might be contributed to augmenting the H₂ generation. This fluctuating phenomenon for the H₂ generation is likely to be processed via water-gas shift, water gas, and steam methane reforming reactions. From low to higher heating rate, the H₂ generation rate was decreased after 760°C and 710°C which continued till the end of the process. The results acquired for H₂ are in good agreement with the study conducted on individual components such as; paper, plastics, yard trimmings in the pyrolysis atmosphere [29]. In conclusion, the H₂ generation from the pyrolysis of the PMSW was slightly favoured by the lower heating rate over the higher heating rate.

The CO presented different generation trends at both heating rates, as indicated in Fig. 3b. At a lower heating rate, the CO generation increased from 10 to 62 ml min⁻¹ g⁻¹ in the temperature range from 230–500°C and reached its maximum level. There was a clear shoulder displayed at 300°C in the temperature range from 290–310°C, which then ascended and merged with the ongoing trend line. After reaching its maximum peak level, a combination of decreasing and increasing trends persisted up to 800°C. Notably, after touching 800°C to the final process temperature i.e. 900°C, the generation trend remained in a constantly increasing position. On the other hand, at the higher heating rate the CO generation trend increased from 5 to 78 ml min⁻¹ g⁻¹ in the temperature range from 250–420°C and reached at its maximum level. A little peak was observed at 330°C in the ranges 290–310°C. At that moment, the trend line suddenly dropped from 78 to 12 ml min⁻¹ g⁻¹ and that declination continued till the end of the process. The similar resembling CO fluctuating tend also reported for the pyrolysis of polystyrene plastic component which first increased with the elevating temperature and suddenly decreased at higher temperatures [37]. A similar higher CO generation was also reported by the other researchers for the lingo-cellulosic (biomass) and/or cellulosic originated (paper) materials in the PMSW mixture [34, 37]. Hence, the thermal disintegration of ether bonds, carboxyl, and carbonyl groups exist in hemicellulose and cellulose, resulting in higher CO generation hence the results well agreed with the previous studies [54, 55]. On the other hand, the phenomenon of the lower CO generation at the lower heating rate could have also changed due to the primary reaction, essentially the gradual improvement in H₂ and CO₂ associated with the marginal catalytic characteristics of ashes [56]. The evolution of CH₄ generation followed relatively parallel generation trends i.e. higher generation rate at the lower heating rate and lower at the high heating rate as illustrated in Fig. 3c. At a lower heating rate, the CH₄ generation increased from 4 ml min⁻¹ g⁻¹ to 18 ml min⁻¹ g⁻¹ in temperatures 300–400°C. But, a quick leaping was observed in the temperature range 400–470°C, which reached its lowest generation level of 2 ml min⁻¹ g⁻¹ to the end of the process. For the high heating rate, the trend was the same as the lower heating rate, but it could only reach half of the maximum level of lower heating rate, i.e., around 9 ml min⁻¹ g⁻¹ in 290–410 °C ranges. In addition, several peaks were seen which showed the inconsistency of pyrolytic decomposition under the higher heating rate. Overall, the H₂ generation was merely slow for a lower heating rate, which decreased from 750°C to the end. On the other hand, the higher heating rate pyrolysis produced the maximum CO generation after 300°C, increasing continuously. The low heating rate took a wide span of CH₄ increment between 320–490°C. Regarding the higher heating rate, the CH₄ generation increased between 300–500°C, and the decreasing trend continued till the end of the process. The total yield of the released gases is shown in Fig. 3d. Overall, the lower heating rate favoured the high total yield for all the gases. i.e. H₂, CO, and CH₄. However, a minor difference in the total yield was observed for the H₂. The CO owned the highest total yield in the temperature ranges 230–500°C and 800°C to end of the processes temperature i.e. 900°C.

To check the energy required in the weight loss of PMSW during pyrolysis, the estimation of kinetic parameters is essential, calculated by the Arrhenius equation. The TGA results were evaluated by the model-free fitting method to calculate the kinetic parameters. Afterwards, the temperatures parallel to fixed values of conversion degree (α) from experiments at the applied heating rates were obtained. In this study, the activation energy (Ea) was estimated by the non-isothermal iso-conversion technique by Flynn-Wall-Ozawa (FWO) and employing Doyle's approximation of p(x) as follows: (12) logβ=log[AEaRT(α)]−2.315−0.4567EaRT {log}\beta = {log}\left[ {{{AEa} \over {RT\left(\alpha \right)}}} \right] - 2.315 - 0.4567{{Ea} \over {RT}}

Fig. 4a represents the fitting curve plots of ln (β) versus 10³/T corresponding to the different conversion degrees (α) at both heating rates. To calculate the Ea at a certain α that how the reaction rate varies with temperature at a particular α has to be identified; hence it can be established by changing the heating rate (β) [34, 47]. Therefore at the conversion α, the Ea can be calculated after simplifying and by taking the derivative of equation 5 concerning to (−1RT) \left({- {1 \over {RT}}} \right) . (13) Ea(α)=∂ln(dαdt)a,j∂(−RTa,j)−1 Ea\left(\alpha \right) = {{\partial \ln {{\left({{{d\alpha} \over {dt}}} \right)}_{a,j}}} \over {\partial \left({- R{T_{a,j}}} \right) - 1}}

Figure 4

Where the subscript j means the heating program selected, hence the obtained Ea are recorded in Table 4. The obtained Ea values and model validation are schemed in Fig. 4a and Fig. 3b, respectively.

Table 4 depicts the estimated Ea values at 5 and 15°C/min heating rates. The correlation coefficient (R²) ranged from 0.9701 to 0.9983 hence; validating the generated data well. Fig. 5b depicts the lower heating rate owned higher values of Ea over higher heating rate except from the conversion rate (α) of 10 for the pyrolysis. The higher Ea values of 266 and 263 kJmol⁻¹ at both heating rates at α of 10 were found at ~295°C can be ascribed may be due to the presence of plastic content in the mix, which usually starts to decompose from 290–350°C in the absence of gasifying agent required more energy. At a lower heating rate, the values of Ea lied between 26 kJmol⁻¹ to 142 kJmol⁻¹.

Figure 5

Comparatively, the Ea values estimated in our study for the PMSW are less than the individual MSW components i.e. hemicellulose, cellulose, lignin, polyethylene, and polystyrene ranged between 171–221 kJmol⁻¹ [30]. Hence this work has demonstrated that the pyrolysis of PMSW at 5 and 15°C/min pay considerable value in terms of Ea related to the thermal behaviour of the PMSW. From the validation of model curves as depicted in Fig. 5, it may be seen that the FWO model provides an agreeable reproducibility to the experimental results throughout all the conversion degrees. Nevertheless, higher deviations among the calculated and experimental information may be seen at 5°C/min between the conversion degree of 5 and 25. The experimental and calculated curves trend was smooth and overlapped, hence correspondingly following the same pathway along the trajectory. The experiment curves at 5°C/min (Fig. 4a) had certain discrepancies with calculated curves along the same trajectory in the ranges about 300–500°C, thus appearing as slow and continued fluctuating curves presenting slower and minimum decomposition. While the calculated curves at 15°C/min (Fig. 4b) exhibited smooth experimental and calculated curves with persistent flat parallel curves. In summary, the pyrolysis of PMSW was observed to be slower than the deviation among the PMSW with identical conditions was strongly linked to the heating rates.