1. Introduction
Polyoxymethylene (POM), structural formula HO(CH
2O)
nH, is the smallest polymerization product of formaldehyde, with a typical degree of polymerization of 8–100 units. Due to the fact that POM has the characteristics of high strength, wear resistance, fatigue resistance, good dimensional stability, self-lubrication, etc., it has been applied increasingly in mechanical engineering, the automotive industry, precision instruments, etc., either alone or as a composite material. For example, it is used to make parts for telephones, tape recorders, and computers, as well as shafts and gears [
1,
2,
3]. Its role has become more important over the years due to its excellent physical and mechanical properties [
4]. However, the unstable hemiacetal end groups on both sides of the molecular chain begin to break down when POM decomposes into formaldehyde gas once heated above 100 °C [
5]. Therefore, in practical applications POM must be treated with appropriate processing. The prerequisite for processing is the knowledge of its pyrolytic behavior, and the study of pyrolysis of combustible solids is the basis for its combustion. Therefore, the study of the pyrolysis process of POM can provide recommendations for its processing as well as for further studies of its combustion properties. Studies on pyrolysis of POM have appeared in a number of works. Most of these studies have been developed with the calculation of kinetic parameters. The thermal degradation profiles of POM/thermoplastic polyurethane (TPU) blends were investigated using TG and FTIR by Pielichowski et al. It was found that the incorporation of TPU into the POM matrix improved the thermal stability of the blends compared to the original material, and it was found that the thermal degradation of POM is basically a one-step process [
6]. Fayolle et al. investigated the thermal oxidation of unstable POM using gravimetric and infrared spectroscopic methods at 90, 110, and 130 °C and at various oxygen pressures from 0 to 2.0 Mpa [
7]. Studies on reaction models of POM had been relatively sparse until Lüftl S et al. confirmed the applicability of the Flynn-Wall method, and thus determined the kinetics of the pyrolysis reaction of POM [
8].
But identification of reaction models using the Coats–Redfern and Criado methods has not been performed. In addition, few researchers have studied the estimated lifetime and degradation mechanism of POM under different atmospheres. In this paper, the pyrolysis process and kinetic analysis of POM are investigated with thermogravimetric tests and Fourier transform infrared spectroscopy (FTIR) tests under nitrogen and air.
Combustion is the next step in the thermal decomposition of the material, and some research has been carried out on the combustion properties of POM. Shaklein et al. explored the effect of reactor geometry on polymer combustion [
9]. Korobeinichev et al. investigated the thermal decomposition and combustion of horizontally placed POM sheets, and determined the kinetic parameters of thermal degradation of POM assuming two parallel reactions [
10]. By using temperature profiles measured in the gas phase with the micro thermocouple technique, the combustion surface temperature and mass burning rate of POM were determined by Glaznev et al. [
11]. It is necessary to improve the combustion performance of POM in order to make it better for use; flame retardants are often used in the field of materials to improve the combustion properties of materials. Commonly used flame retardants are boron and molybdenum system, nitrogen system, and phosphorus–nitrogen mixed flame retardants [
12,
13,
14]. Phosphorus–nitrogen flame retardant is a compound of phosphorus and amine, and the nitrogen oxide released during combustion can play the role of flame suppression, as well as prevent contact between oxygen and carbon, which is more suitable for POM; therefore, phosphorus–nitrogen flame retardant was chosen in this study.
Research on the application of flame retardants in the combustion behavior of POM is still relatively sparse, so this paper selects phosphorus–nitrogen flame retardant to investigate the combustion performance of two kinds of polyacetal materials (with and without flame retardants added) to supplement some experimental data for research in the field of polyacetal. This paper can provide references and suggestions in terms of pyrolysis reaction model, lifetime estimation, combustion behavior of POM, and the effect of flame retardants.
2. Experimental Procedure
The POM ingredient (density: 1.43 g cm−3, purity: 99%) used was a commercial product of high purity manufactured by the DuPont Co. of Wilmington, DE, USA. TG tests were obtained by using a SDT-Q600 thermal analyzer (TA Co., New Castle, DE, USA) with a microbalance sensitivity of 0.1 μg and a temperature sensitivity of 0.001 °C. The sample mass of each experiment was 2 mg and the initial mass error of the sample did not exceed 0.1 mg. Conventional constant heating rate TG measurements were run at 5, 10, 15, and 20 °C min−1. The flowing gas was N2 and air at a flowing rate of 100 mL min−1. The FTIR technique (PerkinElmer TL8000, Waltham, MA, USA) with a purge flow of 35 mL min−1 nitrogen or air at 20 °C min−1 heating rate was employed to analyze the evolved gas. The combustion performance of two different types of POM was studied by using a conical calorimeter model 0007 from FTT (Fire Testing Technology), UK, with three settings of thermal radiation intensity (25 kW/m2, 35 kW/m2, and 50 kW/m2).
3. Theoretical Consideration
The kinetics of polymer degradation are described by Equation (1) [
15], where α is the level of conversion,
T means the environmental temperature,
β means the heating rate,
A the pre-exponential factor,
Ea the activation energy,
f (α) the different function of conversion, and
R the gas constant. The conversion,
α, is calculated in terms of weight loss by Equation (2) [
16], where
W0 is the initial weight of the sample,
Wt is the weight of the sample at time
t, and
Wf is the weight of the completely decomposed sample. The algebraic expression for kinetic models commonly used [
17] are exhibited in
Table 1.
Without any assumption on the decomposition model, the isoconversional methods (Friedman (FR), Flynn–Wall–Ozawa (FWO), and Kissinger–Akahira–Sunose (KAS) methods) can give activation energy
E as a function of conversion by using different heating rates. The Friedman method [
18] utilizes the TG data of different heating rates to calculate
Ea. As the mathematical plot of ln(d
α/d
t) against 1/
T shows in Equation (3), a linear relationship could be obtained with a slope equal to −
Ea/R. The FWO method [
19] is an integral method, which is independent of the degradation mechanism. The logarithmic form can be given as Equation (4). The activation energy can be obtained from plot of ln
β versus 1/
T at a fixed conversion with the slope being 0.4567
Ea/RT. Analogous to the FWO method, KAS [
20] is an integral calculation method and the equation of it can be given as Equation (5). Plotting ln(
β/
T2) against 1/
T allows
Ea to be calculated for each degree of conversion value.
The Criado model is applied to further confirm reaction mechanism through comparing the fitting degree of curves. By combining Equations (1), (5) and (6) is obtained [
21], where 0.5 refers to the conversion of 0.5. The left side of Equation (6), Z(
α)/Z(0.5), is a reduced theoretical curve which is characteristic of each reaction mechanism, whereas the right side of Equation (6) is associated with reduced rate. Through comparing the coincidence degree of plots, we can conclude that the kinetic model describes an experimental reactive process. Once a reaction mechanism is determined, the reaction order (
n) could be quantitative, which is essential to predicting thermal lifetime.
Then, based on a single heating rate measurement, the Chang method is used to evaluate the activation energy
Ea and the frequency factor
A without making any assumptions. The Chang method [
22] is a one-heating-rate treatment method. As shown in Equation (7), a plot of ln[(d
α/d
t)/(1 −
α)
n] versus 1/
T yields a straight line if the decomposition order is right. The slope and intercept of this line can provide the
Ea and ln
A values, respectively.
After the three necessary parameters (
n,
Ea, and
A) are obtained, the half-life time
t1/2 and estimated lifetime
tf can be predicted eventually. The
tf and estimated
t1/2 are defined to be the time when weight losses reach 0.5 and 0.05 as shown in Equations (8) and (9), respectively [
23].
Data such as ignition time, heat release rate, total heat release, smoke release rate, and total smoke release can be obtained using a cone calorimeter, and repeatability of operation can also be achieved [
24]. Time to ignition (TTI) is one of the main parameters used to characterize the fire hazard of a material. TTI reflects the ease with which a material can be ignited. A longer ignition time indicates that it is harder for the material to be ignited, and that it is less of a fire hazard and has better resistance to fire [
25]. The heat release rate (HRR) represents the rate of heat release per unit area after ignition of a sample under a certain heat flow intensity. HRR can be further subdivided into peak rate (PkHRR) and mean rate (MHRR) [
26]. The total heat release refers to the heat released per unit area of the material combustion process [
27], and the total heat release is almost independent of the value of other external factors when ventilation is adequate, so the total heat release is usually used as one of the parameters to evaluate the fire hazard of a material. Smoke release rate and total smoke release can reflect the degree of pollution of the material to the environment, and can also reflect the degree of combustion of the material [
28].