Masks thermal degradation as an alternative of waste valorization on the COVID-19 pandemic: A kinetic study

Abstract

The COVID-19 pandemic generated a new dynamic around waste management. Personal protective equipment such as masks, gloves, and face shields were essential to prevent the spread of the disease. However, despite the increase in waste, no technical alternatives were foreseen for the recovery of these wastes, which are made up of materials that can be valued for energy recovery.

It is essential to design processes such as waste to energy to promote the circular economy. Therefore, techniques such as pyrolysis and thermal oxidative decomposition of waste materials need to be studied and scaled up, for which kinetic models and thermodynamic parameters are required to allow the design of this reaction equipment. This work develops kinetic models of the thermal degradation process by pyrolysis as an alternative for energy recovery of used masks generated by the COVID-19 pandemic.

The wasted masks were isolated for 72 h for virus inactivation and characterized by FTIR-ATR spectroscopy, elemental analysis, and determinate the higher calorific value (HCV). The composition of the wasted masks included polypropylene, polyethylene terephthalate, nylon, and spandex, with higher calorific values than traditional fuels. For this reason, they are susceptible to value as an energetic material.

Thermal degradation was performed by thermogravimetric analysis at different heating rates in N₂ atmosphere. The gases produced were characterized by gas chromatography and mass spectrometry. The kinetic model was based on the mass loss of the masks on the thermal degradation, then calculated activation energies, reaction orders, pre-exponential factors, and thermodynamic parameters. Kinetics models such as Coats and Redfern, Horowitz and Metzger, Kissinger-Akahira-Sunose were studied to find the best-fit models between the experimental and calculated data.

The kinetic and thermodynamic parameters of the thermal degradation processes demonstrated the feasibility and high potential of recovery of these residues with conversions higher than 89.26% and obtaining long-chain branched hydrocarbons, cyclic hydrocarbons, and CO₂ as products.

Highlights

• •The processes of mask thermal degradation by pyrolysis generated higher than 89 % conversion levels.
• •The mask thermal degradation is a viable way of the energetic valorization of the mask.
• •The kinetic models can be used to design pyrolysis reactors of the most used masks on the COVID-19 pandemic in Ecuador.

1. Introduction

During the COVID-19 pandemic, measures were implemented worldwide to prevent and avoid the SARS-CoV-2 contagion. One of these measures was the mandatory use of face masks regarded as personal protective equipment (PPE) and made of specific materials to prevent the absorption of spit drops. However, once these face masks are worn, they generate a large amount of plastic waste [1], like personal safety equipment such as gowns, gloves, and face shields [2].

During the pandemic, the production of sanitary materials such as masks has increased significantly. For instance, in 2020, China exported around 220 billion units of masks [3] compared with the 5 billion units reported in 2019 [4]. In addition, 3M corporation increased its production of N95-type masks by 2021 to 2 billion units per year in its various manufacturing facilities to meet product demand [5].

Masks materials include polyethylene, polypropylene, or cellulose, along with elastic bands of polyester, nylon, and spandex. In addition, the filtering facepieces such as N95 respirators can be composed of polypropylene, polyester, polyurethane, and cotton with thermoplastic elastomer fasteners.

The waste of these materials is considered biohazardous because it has a high biological risk and, for this reason, should be disposed of in appropriate containers [6]. The National Risk and Emergency Management Service of Ecuador [7] has guidelines for managing these types of wastes based on their origin, so when they come from medical centers, masks are treated as medical waste and undergo a specialized treatment such as incineration. Masks used at home, however, should be disposed of in closed plastic covers, sprayed with disinfectant solutions, and put in a bag to be thrown away with ordinary garbage.

The World Wide Fund for Nature [8] says that if only 1% of plastic masks globally are disposed of incorrectly, this would generate approximately 10 million units or about 40,000 kg of polluting plastic. In addition, images of mask waste in different places, such as beaches, streams, rivers, and even the ocean, are reported worldwide.

Another significant issue is its natural degradation time since surgical masks can take more than 400 years due to their polymeric components. Thus, for example, in our capital city Quito-Ecuador, ordinary waste has increased by 15% each month since the pandemic began [9].

Nonetheless, some processes and treatments can be applied to this kind of waste to become precursors of new chemical products due to the polymeric materials. One of these treatments is thermal degradation by pyrolysis of residual masks for potential valorization as an energy source. Pyrolysis is a process in which solid material is in contact with an inert gas such as nitrogen, and usually, temperatures between 300 and 800 °C are used [10]. Pyrolysis is a process that takes place in the total or limited absence of oxygen for a specific time. These conditions contribute to the breakdown of complex hydrocarbon compounds into smaller molecules. As a result, relatively essential products such as gases, liquids, and solid carbon are formed [11]. Heat and gas decomposed solids into fewer molecules like methane, carbon dioxide, and carbon monoxide. Several authors have studied pyrolysis of wastes as energy valorization of municipal solid waste [[12], [13], [14]], polymeric waste [15], bio sanitary waste [[10], [16]], plastic medical waste [17], and masks for protection of coronavirus [[10], [18], [19], [20], [21]].

Around COVID-19 waste, for example, thermo-catalytic pyrolysis using a new 3-layer pre-dissolved face mask catalyst has been studied to extract from different types of plastic waste, obtaining the pyrolytic oil of waste masks, gloves, and other PPE kits possess. A similar study showed that the pyrolysis liquid had a high amount of alkanes and alkenes and a higher concentration of mixed aromatics [22]. Some oxygen-containing compounds were also observed in the derivatized sample, such as cyclohexane, cyclooctane, 1-methyl-3-propyl, etc. However, in this study, the gases produced in this study were not analyzed [23].

Co-pyrolysis was investigated in a mixing scenario between the face mask and food waste. In this case, pyrolytic gas is mainly composed of hydrocarbons and H₂. The proportion of oxygenates increases as the food waste loading in the feedstock increases since the fatty acids and esters that make up most of the oxygenates are derived from food waste. Therefore, pyrolysis of the disposable mask with adjuvant food waste decreased the yield of hydrocarbons [24].

Therefore, based on this analysis, masks can be treated by thermal pyrolysis and valorized as an energetic alternative. Although experimental studies on the thermal degradation of masks have been carried out, there are few studies on the kinetics of this reaction.

The kinetic models are very important for designing reactors that allow us to carry out these processes at a real scale. In addition, the kinetic models provide us with information about the thermodynamic parameters of the reactions. In this sense, this work presents the results of the degradation process by pyrolysis at the laboratory scale of the most used masks during the covid pandemic in Ecuador.

2. Experimental section

This work involved performing pyrolysis at a laboratory scale of different types of masks used in our country - Surgical, hygienic, and MRF masks (KN95 and N95). We analyzed waste masks that, before experimentation, were isolated for 72 h, washed, and dried to protect the material and the user's biosecurity [6]. Samples of masks were cut to a particle size of 1 mm. Table 1 shows the usage time of each mask.

Table 1.

Masks used on the thermal degradation reactions by thermogravimetry.

Type of Masks	Total usage time (hours)
Surgical	8
Hygienic	90
KN95	24
N95	24

2.1. Characterization of masks

The analysis was performed in the Research area from Facultad de Ingenieria Quimica’ Universidad Central del Ecuador. The composition of the waste masks was analyzed by Fourier Transform Infrared Spectroscopy in the range of 4000 and 600 cm⁻¹ with an FTIR spectrometer (PerkinElmer Spectrum Two) coupled with an ATR accessory (PIKE Technologies).

The mask's higher calorific value (HCV) was quantified with a calorimetric pump (Parr-SERIE) applying the ASTM D-240 methodology [25] to study the potential for energy recovery of these wastes. This test was done in the DPEC laboratories of Universidad Central del Ecuador.

The elemental analysis was performed with Elementar Vario MACRO CUBE equipment, using He as the carrier gas, O₂ as the flue gas, sulfanilamide (C₆H₈N₂O₂S_(s)) as the reference substance, and tungsten oxide (WO_3(s)) as the process catalyst [26].

2.2. Thermal degradation tests by thermogravimetry (TGA)

Thermal degradation tests were performed with the thermogravimetric balance (TGA 1 STARe System). Mask samples (approx. 10 mg) were placed in alumina crucibles, heated from 25 to 800 °C in an N₂ atmosphere with a flow of 50 mL/min, and three heating rates were used: 5, 20, and 40 °C/min. Analyses were done in duplicate to ensure data representativeness. Mass loss data as a function of time and temperature was detected and used for the kinetic model development.

In addition, to identify the products of thermal degradation, tests were carried out by coupling thermogravimetry to a gas storage unit (ITS16 SRA Instruments Chromatographic Solutions) which is analyzed by the gas chromatograph (7820A GC System) and mass spectrometer (5977E MSD Agilent Technologies). For this measurement, thermogravimetric tests were carried out with a heating rate of 40 °C/min and heated from 340 to 500 °C in the atmosphere of N₂, with a gas flow of 15 mL/min. The components were identified with data from the NIST library.

3. Kinetic modeling

Kinetic modeling of the mask's thermal degradation was carried out considering the data from the mass loss of the masks as a function of the time obtained from thermogravimetry [[20], [21], [22]].

With this data, the conversion rate (α) is calculated (Eqs (1), (2), (3), (4), (5))):

$$\frac{d\alpha }{dt}=k\left(T\right)*f\left(\alpha \right)$$ (1)

$$k\left(T\right)=A*e^{\left(-\frac{Ea}{RT}\right)}$$ (2)

$$f\left(\alpha \right)={\left(1-\alpha \right)}^n$$ (3)

$$\alpha =\frac{w_0-w_t}{w_0-w_f}$$ (4)

$$\frac{d\alpha }{dt}=A*e^{\left(-\frac{Ea}{RT}\right)}*{\left(1-\alpha \right)}^n$$ (5)

Where:

k(T) is the kinetic decomposition constant or Arrhenius equation.

f(α) is the conversion function.

A is the pre-exponential factor (s⁻¹).

Ea is the activation energy (kJ/kmol).

R is the gas constant (8314 kJ/kmol. K).

T is the absolute temperature (K).

α is the fraction of mass loss or conversion.

w_o, w_f, and w_t correspond to the initial, final, and time-based mass (mg).

The heating rate (β) can be represented as Eq (6):

$$\beta =\frac{dT}{dt}$$ (6)

Combining Eq (5) and Eq (6), the expression that correlates the effect of heating with conversion is achieved:

$$\frac{d\alpha }{{\left(1-\alpha \right)}^n}=\frac{A}{\beta }*e^{\left(-\frac{Ea}{RT}\right)}dT$$ (7)

Integrating Eq (7), it is obtaining the mathematical expression Eq (8):

$$F\left(\alpha \right)={\int }_0^{\alpha }\frac{d\alpha }{{\left(1-\alpha \right)}^n}=\frac{A}{\beta }*{\int }_{T_0}^T{e}^{\left(-\frac{Ea}{RT}\right)}dT$$ (8)

The purpose of these mathematical models is to represent the thermal degradation process and obtain the following kinetic parameters: activation energy (Ea), reaction order (n), and pre-exponential factor (A) [27].

Some models were applied to find the better model that fits the experimental data.

Model 1. The kinetic model proposed by Coats and Redfern [28] (Eqs (9), (10), (11))) is of the non-iso-conversional integral type and can be applied to the kinetic study of the polymer's thermal degradation [[27], [29]].

$$\ln \left(\frac{F\left(\alpha \right)}{T^2}\right)=\ln \left[\frac{AR}{\beta Ea}\left(1-\frac{2RT}{Ea}\right)\right]-\frac{Ea}{RT}$$ (9)

$$F\left(\alpha \right)=-\ln \left(1-\alpha \right)$$ (10)

$$F\left(\alpha \right)=\frac{1-{\left(1-\alpha \right)}^{1-n}}{\left(1-n\right)}$$ (11)

Applies Eq (10) when n = 1 and Eq (11) when n ≠ 1.

This model is solved by correlating ln (F(α)/T²) vs. 1/T and through linear regressions so that the respective kinetic parameters are determined.

Model 2. The model proposed by Horowitz and Metzger [26] (Eq (12), (13))) is a non-iso-conversional integral model that applies an auxiliary temperature variable (θ) and with which the approximation can be applied [30]:

$$\frac{1}{T}=\frac{1}{T_m\left(1+\frac{\theta }{T_m}\right)}\cong \frac{1-\frac{\theta }{T_m}}{T_m}$$ (12)

$$\theta =T-T_m$$ (13)

T_m is the temperature of the maximum mass loss (at which the maximum conversion occurs).

Through Eq (12), an approximate solution to the integral expression of Eq (8) can be calculated, obtaining Eq (14):

$$F\left(\alpha \right)=-\frac{AR{T}_m^2}{\beta Ea}\mathit{EXP}\left[-\frac{Ea}{{RT}_m}\left(1-\frac{\theta }{T_m}\right)\right]$$ (14)

Equation (14) can vary with the assigned reaction order value, Eq (15) can be used in the case of a first-order reaction (n = 1) where T = Tm (or approaches) and F(α) = -1, while Eq (16) can be used when n ≠ 1.

$$\ln \left[-\ln \left(1-\alpha \right)\right]=\frac{Ea\theta }{R{T}_m^2}$$ (15)

$$\ln \left[\frac{1-{\left(1-\alpha \right)}^{1-n}}{\left(1-n\right)}\right]=\ln \left(\frac{AR{T}_m^2}{\beta Ea}\right)-\frac{Ea}{R{T}_m}+\frac{Ea\theta }{R{T}_m^2}$$ (16)

Ln (F(α)) vs. θ is correlated to determine the kinetic parameters.

Model 3. The method described by Kissinger-Akahira-Sunose (KAS) [31] Eq (17) consists of an integral iso-conversional model based on the approximations proposed by Coats and Redfern. Still, it uses data series at different heating rates. However, it does not allow for the direct achieve a kinetic model of reaction, so it is combined with model adjustment approaches to obtain the kinetic parameters and reaction mechanisms [32,33].

$$\ln \left(\frac{\beta }{T^2}\right)=\ln \left(\frac{AR}{EaF\left(\alpha \right)}\right)-\frac{Ea}{RT}$$ (17)

Ln (β/T²) vs. 1/T is analyzed for values of α between 0.1 and 0.9 to achieve linear regressions and estimate the activation energy (Ea).

The values of the pre-exponential factor (A) and reaction order (n) can be determined by applying Eq (18) and through F(α) vs. Ea.p(x)/(βR) to determine the respective linear regressions.

$$F\left(\alpha \right)=\frac{EaA}{\beta R}p\left(x\right)$$ (18)

$$x=\frac{Ea}{RT}$$ (19)

The value of p(x) or temperature integral can be determined using rational approximations. For example, Ding et al. [34] and Zhang et al. [35] point out that, after applying Doyle's approximation, p(x) achieves the following form of Eq (20).

$${p\left(x\right)}_{Ding,etal}=0.00484*\mathit{EXP}\left(-1.0516*x\right)$$ (20)

Using graphical methods, Tang, Liu, Zhang, and Wang [36] obtained another approximation [37,38] which is expressed in Eq (21).

$${p\left(x\right)}_{Cai,etal}=\frac{\mathit{EXP}\left(-x\right)}{x*\left(1.00198882*x+1.87391198\right)}$$ (21)

The p(x) values with which we will work in Eq (18) will be the averages between the two approximations as results in Eq (22).

$$p\left(x\right)=0.5*\left[{p\left(x\right)}_{Ding,etal}+{p\left(x\right)}_{Cai,etal}\right]$$ (22)

3.1. Fitting the model to experimental data

To determine the kinetic parameters of best fit to the experimental data, we applied the objective function value (OFV) that quantifies and minimizes the difference between the experimental data series and the calculated ones [39]. The best fit of the data series will be determined by minimizing the mean square error Eq (23).

$$OFV=\mathit{min}\left[\frac{1}{N}\sum _{i=1}^N{\left({\frac{d\alpha }{dt}}_{\exp i}-{\frac{d\alpha }{dt}}_{calci}\right)}^2\right]$$ (23)

N is the total number of data analyzed.

Based on the models with the best fit in each type of mask determined with ANOVA.

3.2. Calculation of thermodynamic parameters

Based on the kinetic parameters, the thermodynamic parameters of thermal degradation of all the waste masks were calculated: enthalpy variation (ΔH), Gibbs energy (ΔG), and entropy (ΔS) with equations (24), (25), (26)):

$$\Delta H=Ea-R*T$$ (24)

$$\Delta G=Ea-R*Tm*\ln \left(\frac{K_B*T}{h*A}\right)$$ (25)

$$\Delta S=\frac{1}{Tm}\left(\Delta H-\Delta G\right)$$ (26)

Where K_B and h are Boltzmann's constant (1.381*10⁻²³ J/K) and Plank's constant (6.626*10⁻³⁴ J s) [[34], [35], [40]].

4. Results and discussion

4.1. Characterization of masks

Fig. 1 shows the results of waste masked analyzed by FTIR spectroscopy. Fig. 1 shows the FTIR spectra of the surgical mask with several overlapping peaks in the band of 3000 to 2800 cm⁻¹ assigned to stretching movements of the bonds [C–H]. The peaks of 2952 and 2872 cm⁻¹ correspond to methyl groups, while in 2917 and 2839 cm⁻¹ correspond to methylene groups, observes peaks due to bending of bonds [C–H] of the methylene groups at 1455 and 1164 cm⁻¹, and by methyl groups at 1375 cm⁻¹. The peaks of 800–1300 cm⁻¹ correspond to the bond stretching [C–C]. The spectra and functional groups obtained in Fig. 1a are characteristic of polypropylene [41], and the values at 840, 1000, and 1164 cm⁻¹ are expected of isotactic polypropylene [42].

Fig. 1.

Fourier transform infrared spectroscopy (FTIR) spectra of masks (a) surgical (b) KN95, (c) N95 and (d) hygienic.

The spectra obtained with KN95 masks are presented in Fig. 1b having bands similar to those obtained with surgical masks. This type of mask would also be composed of isotactic polypropylene.

Fig. 1c shows the FTIR spectra of N95 masks where the presence of methyl groups in 2964 cm⁻¹ is demonstrated, and peaks are also observed in 2907, 2933, and 2854 cm⁻¹ corresponding to the stretching of bonds [C–H] of aliphatic groups of methylene. The peak at 1713 cm⁻¹ is due to bonds [C O] by possible aromatic esters, while for aliphatic esters, peaks are observed at 1240 and 1094 cm⁻¹ by stretching of bonds [C–O]. The presence of aromatic rings is notorious due to several peaks in 1610, 1579, and 1504 cm⁻¹ by bond stretching [C C] and also in 1016, 970, 871, and 794 cm⁻¹ by bond bending [C–H].

The spectra in Fig. 1d corresponding to hygienic mask have structures that coincide with polyethylene terephthalate (PET) [43] and with polypropylene due to peaks between 3000 and 2800 cm^−1, similar to Fig. 1a and b.

The results of higher caloric value (HCV) and waste mask elemental composition are shown in Table 2. Results showed that waste mask composition has high carbon content, indicating that these materials can be considered fuels with high energy potential. The average sulfur content detected in all samples was less than 0.17%. In the case of thermal valorization, the generation of SO₂ (polluting gas) can be minimal and would occur in surgical and hygienic wasted masks. The percentage of hydrogen is similar to that of polypropylene samples, which can be between 14 and 17% [15]. In the case of hygienic masks, they present a high percentage of nitrogen which is consistent with the presence of amides in the FTIR spectra.

Table 2.

Higher calorific value (HCV) and elemental composition of the waste masks.

Mask	HCV (MJ/kg)	%C	%H	%N	%S
Surgical	44.167	79.26	21.29	0.83	0.17
KN95	40.941	79.89	19.15	2.03	0.03
N95	31.781	75.01	14.23	1.99	0.01
Hygienic	30.925	62.34	16.44	10.57	0.16

The higher calorific value (HCV) for the waste masks is described in Table 2, presenting values like those of conventional fossil fuels. Surgical masks and KN95, composed of polypropylene, have higher HCV and resemble the values of fuels such as diesel (42.92 MJ/kg) [44] and gasoline (44.8–48.7 MJ/kg) [45]. The HCV of hygienic masks is higher than that of nylon 6 (28.03 MJ/kg), but it is also lower than that recorded for Spandex fibers (31.4 MJ/kg). The same tendency is for N95 masks, which have an intermediate HCV between those registered for PET (23.00 MJ/kg) and polypropylene (46.02 MJ/kg) [44]. The HCV of hygienic masks and N95 are higher than those of coal and other paper and cardboard waste (17.5 MJ/kg), rubber, and leather (22.5 MJ/kg). A study performed with single-use masks obtained values of >40 MJ/kg in co-pyrolysis with food residues [24].

4.2. Thermogravimetric analysis

The thermal degradation of mask samples was carried out with thermogravimetry at different heating rates under N₂ atmosphere, with which a pyrolysis process was simulated. There is a higher conversion rate when moving from a slow pyrolysis (5 °C/min) to an intermediate one (20 and 40 °C/min), reducing thermal degradation reaction time.

Fig. 2, Fig. 3 show the thermal degradation's mass loss (TGA) and mass loss derivative (DTG) curves for all waste masks studied. The most significant mass loss occurs in a temperature range of 325 and 500 °C. Depending on the heating rate, temperatures with the highest mass loss are displaced, with 5 °C/min tests demonstrating degradation peaks at lower temperatures than in tests of 20 and 40 °C/min.

Fig. 2.

Thermogravimetric analysis (TGA) of masks (a) surgical, (b) KN95, (c) N95 and (d) hygienic for temperature ramps rate: 5, 20 and 40 °C/min.

Fig. 3.

Derivative thermogravimetric curves (DTG) of masks (a) surgical, (b) KN95, (c) N95 and (d) hygienic for temperature ramps rate: 5, 20 and 40 °C/min heating.

The TGA diagrams, Fig. 2(a–d) show the mass loss curves as a function of temperature, which has more than one slope reflecting the degradation of more than one type of material, as is the case with the samples of the KN95 (Fig. 2a) and hygienic (Fig. 2d) masks. These results are confirmed with the DTG diagrams (Fig. 3a–d), where the mass loss velocities are presented with peaks in the temperatures where the most significant losses occur. As the heating rates increased, degradation temperatures of different compounds were approaching.

Overlapping conversion peaks are observed in the DTG diagrams of KN95 and hygienic masks (Fig. 3b and d). In KN95 masks, they would represent polypropylene and at least one secondary compound, while in hygienic masks, they would correspond to the decomposition of spandex and nylon polyamide 6, with nylon being the compound that would contribute most to sample pyrolysis since it makes up most of these masks. Conversion peaks in hygienic masks are more noticeable in trials with 5 and 20 °C/min, while at 40 °C/min, both peaks are closer overlapping in one. In tests with cotton masks and microfibers [20], there were overlapping peaks in decomposition with nitrogen at different temperatures. The experiments with hygienic masks demonstrated slight conversion peaks between 40 and 100 °C due to the humidity of the wasted masks.

In this type of test, products generated are mostly gases achieving conversions more significant than 89%. Tests with slow pyrolysis (5 °C/min) generated the highest conversions (greater than 96.7%) because they required more time to reach the limit temperature of 800 °C. Experiments with KF-94-type masks [19] presented peaks of maximum thermal degradation at 550 °C with a conversion of 81%, while surgical-type masks [18] have maximum degradation between 405 and 510 °C.

4.3. Kinetic models

For the kinetic model of the experimental data, three types of kinetic models proposed by Coats-Redfern (Model 1) [[26], [29]], Horowitz-Metzger (Model 2) [46], and Kissinger-Akahira-Sunose (Model 3) [40] were applied. Through the application of the objective function value (OFV), it was possible to determine the model that provides the best fit between the experimental data and the kinetic model.

Models 1 and 2 (Fig. 4a and b) were developed by assigning values to reaction orders (n), and by linear regressions, the activation and pre-exponential factor energies were determined. For the development of model 3, the activation energy for each conversion level was determined (Fig. 4c). After assigning values to the reaction orders (n), the pre-exponential factors were determined (Fig. 4d).

Fig. 4.

Kinetic model for mask pyrolisis (a) Model 1: ln (F(α)/T²) vs 1/T surgical mask; (b) Model 2: ln (F(α)) vs θ for KN95 mask; (c) Model 3: ln (β/T²) vs 1/T and (d) Model 3: F(α) vs Ea.p(x)/(R.β)*10²⁰ for N95 mask.

Table 3 presents the kinetic parameters of the best-fit models for each type of waste mask. With the results obtained and through the methodology of the objective function value (OFV) [39], it was evident that the calculated models represent the experimental behavior by which the models replicate the processes of mask thermal degradation.

Table 3.

Kinetic parameters for thermal degradation reaction of waste masks.

Type of Mask	β (°C/min)	Models	n	R2	Ea (kJ/mol)	A (s−1)	OFV
Surgical	40	1 – CR	1	0.993	182.23	3.7E+11	1.6E-06
		2 – HM	1	0.998	205.00	1.3E+13	1.4E-06
		3 – KAS	1	0.996	214.80	1.5E+16	1.0E-06
		Our model	1		236.03	9.3E+15	6.4E-07
KN95	20	1 – CR	1.1	0.983	174.25	6.8E+10	5.8E-07
		2 – HM	1.1	0.992	196.84	2.9E+12	3.8E-07
		3 – KAS	1.1	0.997	214.65	2.2E+15	3.1E-07
		Our model	1.1		243.92	7.0E+15	2.4E-07
N95	20	1 – CR	1	0.999	192.67	2.2E+12	2.4E-07
		2 – HM	1	0.999	212.01	5.7E+13	1.2E-07
		3 – KAS	1	0.991	260.29	8.7E+19	4.0E-07
		Our model	1		223.44	4.2E+14	8.8E-08
Hygienic	20	1 – CR	1.5	0.976	154.52	4.6E+09	8.2E-07
		2 – HM	1.5	0.988	184.92	7.9E+11	6.1E-07
		3 – KAS	1.5	0.980	191.91	6.7E+12	5.9E-07
		Our model	0.6		156.91	8.5E+09	1.0E-07
			1.5		329.55	1.4E+22

Studies carried out with similar methodologies for surgical masks [18,21,47] obtained activation energy values between 230 and 280 kJ/mol with kinetic models similar to those studied.

4.4. Proposed model

Based on the models with the best fit in each type of mask determined with objective function value (OFV) values and ANOVA results, it will be possible to estimate better kinetic parameters for conversion speeds with the application of the objective function value (OEV) methodology. Minimizing the mean quadratic difference between the calculated and experimental data series by varying the kinetic triplet: activation energy (Ea), pre-exponential factor (A), and reaction order (n), with the OFV value closest to zero reflecting the kinetic parameters of best settings.

The proposed models follow the conversion speed equation (18):

$$\frac{d\alpha }{dt}=A*\mathit{EXP}\left(-\frac{Ea}{RT}\right)*{\left(1-\alpha \right)}^n$$ (18a)

The kinetic models proposed as a function of the conversion speed for each type of waste mask are expressed in Table 4 and Fig. 5, evidencing the goodness of the kinetic model concerning the experimental thermogravimetry data.

Table 4.

Proposed kinetic parameters and best fit models by thermal degradation of waste mask.

n	Ea (kJ/mol)	A (s−1)	Kinetic equation
Surgical mask, β: 40 °C/min
1	236.03	9.28E+15	$\frac{d\alpha }{dt}=9.28x{10}^{15}*e^{\frac{-246.03}{RT}}*\left(1-\alpha \right)$
KN95, β: 20 °C/min
1.1	241.64	5.52E+15	$\frac{d\alpha }{dt}=5.52x{10}^{15}*e^{\frac{-241.64}{RT}}*{\left(1-\alpha \right)}^{1.1}$
N95, β: 20 °C/min
1	223.44	4.23E+14	$\frac{d\alpha }{dt}=4.23x{10}^{14}*e^{\frac{-223.44}{RT}}*\left(1-\alpha \right)$
Hygienic mask, β: 20 °C/min
0.7	149.54	1,566E+9	$\frac{d\alpha }{dt}=1.566x{10}^{09}*e^{\frac{-149.53}{RT}}*{\left(1-\alpha \right)}^{0.7}$
1.5	329.55	1.38E+22	$\frac{d\alpha }{dt}=1.38x{10}^{22}*e^{\frac{-329.55}{RT}}*{\left(1-\alpha \right)}^{1.5}$

Fig. 5.

Experimental and calculated from kinetic model data rate of pyrolysis for: a) surgical, b) KN95, c) N95 and d) hygienic masks.

For the hygienic mask models (Fig. 5d), due to the presence of two conversion peaks, results were improved by working separately on the peaks they present, determining the corresponding kinetic parameters.

According to the kinetic parameters obtained, the results of the activation energies of the surgical masks (Fig. 5a), KN95 (Fig. 5b), and N95 (Fig. 5c) were in the range of 220–250 kJ/mol, the N95 those that required less energy input to perform the pyrolysis processes and, therefore, presented a higher reaction rate.

In the models proposed for hygienic masks, the lowest conversion peaks (corresponding to lycra) required kinetic parameters of lower values, unlike nylon 6, which represented the highest conversion peaks requiring higher values in the kinetic parameters. The pre-exponential factors determined in the models were higher than 1009 s^−1, which means an increased complexity of samples pyrolysis reactions [48], with KN95 masks and second peaks in the hygienic masks representing the most complex reactions.

Fig. 5 graphically demonstrates the similarity of the proposed models at specific conversion rates for the experimental values.

4.5. Thermodynamic parameters

The thermodynamics parameters of thermal decomposition reaction as enthalpy (ΔH), Gibbs energy (ΔG), and entropy (ΔS) were calculated based on the results of the proposed kinetic models for a specific conversion rate in each type of mask (Table 5).

Table 5.

Thermodynamic parameters of thermal degradation waste masks reaction.

Type of Masks	β (°C/min)	ΔH (kJ/mol)	ΔG (kJ/mol)	ΔS (kJ/mol. K)
Surgical	40	230.06	204.68	0.03
KN95	20	235.73	206.86	0.04
N95	20	217.59	204.36	0.02
Hygienic	20 (Peak 1)	151.31	201.99	−0.07
	20 (Peak 2)	323.54	208.95	0.16

The enthalpy variation values (ΔH) reflect the reactivity in the heat exchange between the reactants and products [[34], [40]], which suggests the second conversion peaks in the hygienic masks (ΔH between 315.64 and 341.43 kJ/mol) required more significant amounts of energy to carry out the corresponding pyrolysis processes.

The results of the KN95 masks and the second peak of pyrolysis of the hygienic masks presented the highest values of Gibbs energy variation (ΔG between 206.2 and 209.4 kJ/mol). The second conversion peaks on the hygienic masks showed the most significant variation in entropy (ΔS between 0.1486 and 0.1686 kJ/mol. K) a similar study with medical masks obtained an average value of 0.107 kJ/mol.K [47]. This result reflects a more considerable complication to performing the samples pyrolysis and how complex the products generated tend to be since they require more outstanding energy contribution.

4.6. Analysis of exhaust gases of thermal degradation of masks

Through gas chromatography and mass spectrometry coupled to the thermogravimetry equipment, it was possible to identify products from the samples pyrolysis with a heating rate of 40 °C/min and in the range of 340–500 °C, conditions of more significant mass loss. The gases identified are presented in Table 6.

Table 6.

Compounds identified in masks thermal decomposition reaction at different temperature ranges.

Surgical mask			Hygienic mask		KN95 Mask		N95 Mask
Temperature range, °C	Relative percentage %	Compound	Relative percentage %	Compound	Relative percentage %	Compound	Relative percentage %	Compound
341–380	45.54	Carbon dioxide	77.27	2,4-dimethyl-1-heptene	62.44	2,4-dimethyl-1-heptene	17.58	Carbon dioxide
	34.42	Cyclopropane	17.23	6-amino-2-methyl-2-heptanol	34.67	2-butene	9.63	Cyclopropyl carbinol
	11.79	2-butene	2.10	Cyanic acid, 2-methylpropyl ether	21.98	Nitrous oxide	3.23	Acetaldehyde
	5.99	Cis-bicycle [4.2.0]-3,7-octadiene	2.30	5-methyl-1-heptanol	20.30	Cyclopropane	2.58	Cyclopropane
	2.27	Acetaldehyde	1.09	Propene	15.73	Carbon dioxide	1.41	3-butin-1-ol
					6.19	cis-Bicycle [4.2.0]-3,7-octadiene	1.19	3-methyl-2-butanamine
					1.13	Methylencyclopropane	1.09	2-butene
							0.85	Bicycle [2.2.1]-2-(2-propenyl)-heptane
381–420	46.56	Carbon dioxide	48.40	Carbon dioxide	46.59	2-amino-1-propanol	66.19	2,4-dimethyl-1-heptene
	40.55	Cyclopropane	43.45	2,4-dimethyl-1-heptene	21.88	Cyclopropane	24.44	Norpseudoephedrine
	8.60	2,3-dimethylhexane	6.50	Norpseudoephedrine	5.78	2,4-dimethyl 1,4-pentadiene	4.69	Cyclopropane
	4.29	3-methyl-2-hexene	0.66	E−1,5,9-decatriene	5.08	2,4-dimethyl-1-pentene	3.05	2-methyl-1-pentene
			0.48	Tetrahydrofuran	4.38	2,4-dimethyl-1,4-pentadiene	0.87	4Z-hexenyl tiglate
			0.27	(Propoxymethyl) benzene	3.87	(E,E)-2,4-hexadiene	0.76	Bicycle [2.1.0]-1,4-dimethyl-pentane
			0.24	(3-chloropropyl) methylene-cyclopropane	3.26	2,4-dimethyl-1-pentene
					2.78	3-Buten-1-ol
					2.68	Tetrahydrofuran
					2.28	2-methyl-(E)-1,3-pentadiene
					1.43	Benzene
421–460	30.70	Cyclopropane	46.36	2,4-dimethyl-1-heptene	53.38	Cyclopropane	70.94	Carbon dioxide
	26.49	Carbon dioxide	42.64	Carbon dioxide	45.09	Nitrous oxide	13.46	Cyclopropane
	24.16	2-methyl-1-pentene	3.09	Tetrahydrofuran	1.53	1-heptin-6-one	12.45	1-ethyl-2-methyl cis-cyclopropane
	13.15	Tetrahydrofuran	2.91	Propane			1.41	(E)-2-pentenal
	1.79	2,4-dimethyl-1,4-pentadiene	1.91	Oxalic acid, cyclobutyl hexyl ester			1.23	(Z)-3-hexen-1-ol
	1.60	2,4-dimethyl-1-pentene	1.86	Cyclopentene			0.52	2,4-dimethyl-1-pentene
	1.11	3-hexen-1-ol	0.91	2,3-dihydrofuran
	1.00	Bhutanal	0.33	(S)-3,4-dimethylpentanol
461–500	33.03	2-methyl-1-pentene	78.60	Carbon dioxide	34.26	2,4-dimethyl-1-heptene	75.38	Carbon dioxide
	14.57	Carbon dioxide	7.96	Tetrahydrofuran	29.60	2-methyl-1-pentene	8.28	2-propanamine
	13.49	2,4-dimethyl-1-heptene	3.49	2-cyan-acetamide	19.41	2,4-dimethyl-1-heptene	5.67	2,4-dimethyl-1,4-pentadiene
	12.28	3,7-dimethyl-1-octene	2.94	1-hepten-6-ino	9.63	Cyclopropane	3.26	2,4,6-trimethyl-1-noneno
	7.77	Cyclopropane	2.92	Cyclopentene	4.19	2-methylpentane	3.05	2,4-dimethyl-1-pentene
	5.16	Tetrahydrofuran	2.08	Propane	1.33	(Z), (Z)-2,4-hexadiene	2.17	1,3-hexadiene
	4.84	2,4-dimethyl-1,4-pentadiene	1.39	Ethyl-cyclobutane	0.88	2,3-dimethyl-2-butene	1.47	4,5-dimethyl-1-hexene
	3.21	2,4-dimethyl-1-pentene	0.63	3-hydroxy-butanal	0.72	Bhutanal	0.48	2,2-dimethyl-4-pentenal
	2.74	2-methyl-1-pentanol					0.23	4-acetoxy-2-azetidinone
	2.24	2-methyl-(E)-1,3-pentadiene
	0.69	butyraldehyde

The main products identified were unsaturated, branched hydrocarbons such as 2,4-dimethyl-1-heptene or 2,4-dimethyl-1,4-pentadiene. The compounds also presented groups of amines, methylene, alcohols, esters, and the recurrent presence of cyclic hydrocarbons was also determined, with carbon dioxide being one of the majority components.

Unsaturated hydrocarbons were significant due to the possible transfer of hydrogen radicals and free radical exchange with varying degrees of polymerization [34]. Some of the products observed with more recurrence in the decomposition gases of surgical masks and KN95 samples, composed of polypropylene, were CO₂, cyclopropane, and a variety of branched hydrocarbon products. N95 masks, composed of polyester (PET) and polypropylene, demonstrated the presence of CO₂ and 2,4-dimethyl-1-heptene as the most recurrent products. The results of hygienic masks, consisting of nylon 6 and lycra, show the presence of propene as a recurrent compound in the process. CO₂ and 2,4-dimethyl-1-heptene are also presented as compounds with a more significant generation.

Some of the products observed with more recurrence in the decomposition gases of surgical mask samples and KN95, composed mainly of polypropylene, were CO₂, cyclopropane, and a variety of branched hydrocarbon products. Similar results were obtained for medium and slow pyrolysis processes [10]. N95 masks, composed of polyester (PET) and polypropylene, demonstrated the presence of CO₂ and 2,4-dimethyl-1-heptene as the most recurrent products.

The results of hygienic masks, consisting of nylon 6 and lycra, show the presence of propene as a recurrent compound in the process. In addition, CO₂ and 2,4-dimethyl-1-heptene are also presented as compounds with a more significant generation.

The presence of nitrogen products in the pyrolysis of KN95 and N95 masks could be due to the interaction with the nitrogen content presented by both samples, belonging to compounds not present in the infrared spectra analyzed. The KN95 mask could be the precursor of 2-amino-1-propanol, and that in the N95 mask generated other compounds with a lower proportion.

The recurrent generation of CO₂ was due decomposition of the organic molecules that make up the samples, such as esters and carbonyls. In the cases of surgical masks and KN95, they would come from the secondary components of the waste mask or residual compounds due to their usage.

5. Conclusions

By analyzing the infrared spectra of the mask samples, it was determined that the main components of the masks currently selected in Ecuador were polypropylene polymers, polyethylene terephthalate (PET), nylon polyamide 6, and spandex.

In all mask samples, the carbon contents were predominant. In contrast, their sulfur content was minimal, indicating that these samples have a high potential to be considered combustible materials with minimal generation of SO₂. In addition, due to the mask sample composition, values of the higher calorific value (HCV) resemble values of recurrent fuels such as diesel or gasoline. They are even higher than those obtained with other types of waste, such as paper, cardboard, and rubbers.

The processes of mask thermal degradation by pyrolysis generated high conversion levels, higher than 89% and therefore low carbonaceous waste and ash, so that, if these processes were applied on a larger scale, spaces would occupy this type of waste in landfills would be significantly reduced. Kinetic parameters that generated models with better adjustments to the experimental data were proposed. Together with the thermodynamic parameters, energy contributions necessary for thermal degradation were characterized and quantified, demonstrating the viability for the valorization of the mask and determining activation energy values between 220 and 330 kJ/mol. The GC-MS essays confirmed the variety of products generated by mask samples pyrolysis, which could be energetically enhanced to create heat or transform into chemicals through more complex processes. The kinetic models obtained can be used to design reactors for pyrolysis processes of the most commonly used masks during the COVID-19 pandemic in Ecuador.

Author contribution statement

Carolina Montero-Calderón: Conceived and designed the experiments; Formulation of overarching research goals and aims; Wrote the paper.

Roger Tacuri: Performed the experiments; Wrote the initial draft paper.

Hugo Solís: Performed the experiments; Analyzed and interpreted the data.

Andrés De-La-Rosa: Analyzed and interpreted the data; Wrote the initial draft paper.

Gilda Gordillo: Conceived and designed the experiments; Wrote the paper.

Pablo Araujo-Granda: Conceived and designed the experiments; Analyzed and interpreted the data; Wrote the paper.

Funding statement

This work was supported by Universidad Central del Ecuador, COVID Projects Program [DI-COVID19-11].

Data availability statement

Data will be made available on request.

Declaration of interest's statement

The authors declare no competing interests.