Non-isothermal crystallization and thermal degradation kinetics of MXene/linear low-density polyethylene nanocomposites

2.1 Materials and preparation of Ti₃C₂

Commercial grade linear low density polyethylene (LLDPE) (with a melt flow index of 2.0 g/10 min at 120°C) was supplied by Sinopec Corp (China). MXene (Ti₃C₂ with OH or F termination) was produced by immersing Ti₃AlC₂ in 49% hydrofluoric acid (HF) at 60°C for 24 h. Ti₃AlC₂ powders were prepared from the blending of TiH₂, Al and C reacted in a tube furnace (13).

2.2 Preparation of nanocomposites

2D nanosheet/linear low-density polyethylene (MXene/LLDPE) blends were prepared by mixing the polymer and MXene at different ratios by an extrusion process. The blends were done in an open refined rubber (plastic) machine [X(S) K-100, Jiangsu Tianyuan Test Equipment Co., Ltd., China] at a barrel temperature of 110°C–125°C and at a screw speed of 20 rpm. After preparation, the samples were melted at 120°C and were cooled under a pressure subsequently to remove the internal stress.

2.3 Analysis methods

Crystallization studies were conducted on a differential scanning calorimeter (DSC) under argon atmosphere with cooling rates of 2.5, 5, 7.5, 10°C/min, respectively, in the range of 20°C–220°C, and then held for 3 min at 220°C to eliminate all the thermal history effects. After that, about 10 mg of the nanocomposites sample was sealed in an Al₂O₃ pan, while another empty Al₂O₃ pan was used as a reference.

The thermal gravity (TG) analysis was done in a SETARAM Evolution 24 under flowing argon at the rate of 20 mL/min. All measurements were performed under nitrogen. Samples of 10–20 mg were heated from room temperature to 800°C with the heating rates of 5, 10, 15 and 20°C/min.

X-ray diffraction (XRD) patterns were obtained using D8 X-ray diffractometer equipment with Cu K_α radiation. The sample was scanned over the range (2θ) 5°–65° to evaluate the structure. Morphological analysis was performed by scanning electron microscopy (FESEM, S4800, Hitachi, Japan) with an accelerating voltage of 3 kV, which equipped with an energy dispersive spectrometer (EDS, EMAX ENERGY EX-250, Horiba, Japan).

3.1 The non-isothermal crystallization behavior

Figure 2 shows the crystallization thermograms for pure LLDPE and MXene/LLDPE nanocomposites at four different cooling rates (2.5, 5, 7.5, 10°C/min). From these curves, some useful kinetic parameters are listed in Table 1, such as onset temperature of crystallization (T_o), temperature of exothermic peak (T_p) and crystallization enthalpy (ΔH_c). As shown in the DSC thermograms, it is not surprising that the crystallization temperature shifted to lower temperatures and ΔH_cdecreased from 103.72 to 71.46 J/g for pure LLDPE, from 99.45 to 81.34 J/g for 2 wt% MXene composite and from 91.89 to 77.23 J/g for 4 wt% MXene composite with an increase on the cooling rates. For the samples crystallizing at a higher cooling rate, there was not adequate time to allow the nuclei to be produced or activated, and the chain segments could not be arrayed into a lattice without delay (16). On the contrary, at a lower cooling rate, the nuclei activation would occur at a higher temperature. In other words, the higher cooling rate would delay crystallization, including nucleation and crystal growth. Moreover, when cooling rates were 2.5, 5, 10°C/min, the T_o and T_pof MXene/LLDPE nanocomposites were higher than those of pure LLDPE, which demonstrated MXene could be effective in accelerating the nucleation rate. But composites exhibited lower ΔH_c values and wider temperatures from Table 1 indicating there might were more lattice defects, poor mobility of molecular chains and low crystallinity.

Figure 2:

Crystallization thermograms and relative degree of crystallinity with time for pure LLDPE and MXene/LLDPE nanocomposites at different cooling rates.

The relative degree of crystallinity (X_t) as a function of temperature (T) can be defined as:

where T_o and T_∞ are the onset and final crystallization temperature, respectively, and dH_c/dT is the heat flow rate. In non-isothermal crystallization, the temperature can be transformed to time scale by using the following equation:

where T is the crystallization temperature at time t, and D is the cooling rate. Thus, the value of T on the X-axis can be transformed into t as shown in Figure 2. All these curves have the similar sigmoidal shape, implying the crystallization process is divided into two stages of different speed, a fast primary process in the initial stage and a slower secondary process in the later stage. This could be explained by the growth site impingement and spherulites crowding in the final stage (17). Additionally, the transition between the two crystallization processes become less pronounced with the higher cooling rate. These curves also show the higher the cooling rate is, the shorter is the time is demanded for accomplishing crystallization.

3.2 Non-isothermal crystallization kinetics

The Avrami equation (18) is used to understand the evolution of crystallization and expressed as:

where n is the Avrami exponent, which depends on the nucleation mechanism and crystal growth dimension; Z_tis the crystallization rate constant including both nucleation and growth rate parameters. The double logarithmic form of Eq. [3] is:

Figure 3 shows the plots of log[–ln(1–X_t)] versus log t for each cooling rate, and the plots can be poorly fitted into a straight line. The value of Z_t and n can be calculated as the intercept and slope of the line, and each curve shifts to a shorter time with the increasing cooling rate. Considering the rate of non-isothermal crystallization depends on the cooling rate, Jeziorny (19) suggested the parameter Z_t should be corrected as:

Figure 3:

Jeziorny analysis plots and Mo analysis plots for all samples at different cooling rates.

The kinetics parameters obtained from Jeziorny theory are listed in Table 1.

It is well known that the Avrami equation can fit the data in primary stage of crystallization appropriately. But it cannot explain the situation in late stage for nanocomposites due to the secondary crystallization process (16). As shown in Figure 3, curves of each samples exhibit a poor linear relationship that signifies that the modified Avrami equation cannot describe the non-isothermal crystallization process of MXene/LLDPE nanocomposites accurately. The Avrami exponent n is in the range of 1.68–2.85 for LLDPE and 1.41–2.80 for the nanocomposites. So MXene serves a function as a nucleating agent and acts as a heterogeneous nucleation center in the polymer matrix (20). Additionally, polymer chains are liable to be absorbed on the active surface of nanomaterials. The interactions between LLDPE and MXene are also the dominant factors in affecting nucleation ability of composites. As the interactions could break the isotropic property of original polymer and transform interfacial molecular chains to energetic forms (21).

Z_c increases with the increasing cooling rate for each sample. It is worth noting the Z_c values of nanocomposites are lower than that of pure LLDPE at 7.5 and 10°C/min cooling rate, showing that the incorporation of MXene could retard the crystallization of LLDPE. Combined with the conclusion mentioned above, MXene could accelerate the nucleation process due to the fact that it acted as a nucleating agent in composites and provided more nucleation sites or surface area, but on the other hand, decreased the chain activity of LLDPE and limited the chain segments rearrangement so that the crystal growth process was decelerated. Comparable behaviors can be seen in the research work carried out by Lonkar et al. (20), Cheng et al. (22) and Ferreira et al. (23).

Mo et al. (24) proposed a novel kinetic equation by integrating the Ozawa and modified Avrami equations. Mo model was widely applied in the research of non-isothermal crystallization processes of polymers, and expressed as follows:

where F(T)=[K(T)​/​Zt]1/m refers to the value of the cooling rate at unit crystallization time when the measured system amounts to a certain degree of crystallinity, K(T) is the Ozawa crystallization rate constant, and a=n/m is the ratio of the Avrami exponent n to the Ozawa exponent m. At a given degree of relative crystallinity, the relationship between cooling rate D and crystallization time t can be obtained.

Plots of log D versus log t for pure LLDPE and MXene/LLDPE nanocomposites are shown in Figure 3. The plots show a better linear fit for all samples, illustrating that the Mo method could describe the non-isothermal crystallization kinetics of the system accurately. Values of F(T) and a could be obtained from the intercept and slope of the line at a certain relative crystallinity, and are listed in Table 2.

The a values for neat LLDPE are lower than those of 4 wt% MXene nanocomposites, while they are larger than those of 2 wt% MXene composites, the same variation trend occurs in F(T) values which increase in the order of: 2 wt% MXene<pure LLDPE<4 wt% MXene, at the same relative crystallinity. Higher a and F(T) values stand for a more difficult crystallization process. In other words, the crystallization rate of pristine LLDPE was slower than that of 2 wt% MXene nanocomposites at a given cooling rate, but larger than that of 4 wt% MXene. These results, mainly because the superfluous incorporation of MXene, would be harmful to forming a crystalline structure. With further increasing concentration of MXene, the molecular chain on the surface of MXene might be entangled with LLDPE which could restrict the mobility and diffusion of LLDPE chains (14). As a result, the expected promoting effect in increasing the overall crystallization rate is offset by the restraining effect. These consequences are consistent with the conclusion of the Jeziorny analysis to some degree. Yet 2 wt% MXene filler content accelerated the crystallization rate of composites that is not reflected in Jeziorny analysis, which indicate Mo analysis is more suitable to evaluate the non-isothermal crystallization kinetics in this work.

The activation energy can be calculated by several mathematical approaches that have been proposed in former publications. The Kissinger method is one of the most popular methods among them. By considering the variation of peak temperature with cooling rate, the activation energy can be obtained from the following equation:

where R is the universal gas constant and ΔE is the activation energy for non-isothermal crystallization. ΔE the activation energy for non-isothermal crystallization can be evaluated from the slope of the fitting straight lines of ln(D​/​Tp2) versus 1/T_p plots for all samples.

As can be seen in Table 3, 2 wt% MXene/LLDPE nanocomposites show a lower ΔE value than that of LLDPE in a pure form and it increases when the loading of MXene increases. The ΔE value of 4 wt% MXene/LLDPE is even higher than that of primitive LLDPE. ΔE represents the energy demanded for transporting the crystalline molecular chains through phases in polymers (25). Generally, the lower ΔE value, the easier crystallization process of polymer system will be. At this high content, the layered structures of MXene have the same function with physical branched-cross linked molecules so that the ΔE values increase and crystallization velocity is slowed down (26), the interaction between components in composites also leads the same conclusion. But 2 wt% of MXene would play a role as a nucleating agent and initiates the heterogeneous nucleation mechanism which can promote the crystallization process of LLDPE. These effects are smoothly consistent with the Jeziorny and Mo methods analyzed above and have been reported in many previous reports (14, 20, 21, 22, 23, 26).

3.3 Thermal stability

The thermal stability of the nanocomposites can be evaluated from thermogravimetric data measured at four different heating rates (5, 10, 15 and 20°C/min). Figure 4 presents TG and differential thermal gravity (DTG) curves of all the samples. Curves before 400°C were not presented in the figure as there was no obvious weight loss for all the samples under this temperature. As expected, TG curves shift to higher temperatures with increasing heating rates because of the lag effects on heat transfer. This method forges connections between time and temperature and facilitates the simulation of thermal degradation kinetics by supplying useful data.

Figure 4:

TG and DTG thermograms of all samples at various heating rates.

Generally, thermal degradation started after ~5% weight loss, so the temperature at which weight loss is 20% is determined as the initial decomposition temperature. Parameters T_20%, T_40% and T_max (DTG peak values) are summarized in Table 1. The three characteristic average temperature values of composite samples are higher than those of pure LLDPE. More precisely, the increase of temperature in the late stage is more considerable than that in 20% and 40% weight loss. That is, T_max increased from 477.94°C to 487.45°C, while T_20% and T_40% varies between 469.52°C and 472.26°C and between 475.51°C and 480.92°C, respectively. The results indicate that the nanocomposites show a better thermal stability and the ability to retard thermal decomposition is more notable in the maximum degradation rate (27).

3.4 Thermal degradation kinetics

Based on the non-isothermal kinetics theory and the Arrhenius experience equation, the kinetics equation for degradation of polymers is:

where α is the degree of conversion, dα/dt is the rate of the degradation reaction, A is the pre-exponential/frequency factor, R is the molar gas constant, n_r is the reaction order (for distinguishing the Avrami exponent n mentioned above, here we denote reaction order as n_r) and E_a as the apparent activation energy.

For non-isothermal experiments, the heating rate is constant, β=dT/dt, so the Eq. [8] becomes:

The thermal degradation kinetics of polymers can be investigated by integral and differential methods that have been proposed in recent years. The apparent activation energy and other degradation kinetics parameters of MXene/LLDPE nanocomposites which were established through the Kissinger method, are expressed as follows:

where T_max, and α_max are the temperature and conversion at the maximum of DTG curves where the maximum conversion rate appears, respectively. Kissinger supposed the factor nr(1−αmax)nr−1=1, and it is irrelevant to heating rates. In this case, Eq. 10 can be written as:

E_a can be evaluated from the slope (–E_a/R) of the fitting straight lines of ln(β​/​ Tmax2) versus 1/T_maxat various heating rates. The frequency factor A can be estimated from the intercept.

Values of E_a and A are established in Table 3. It is apparent that A showed a reduction when MXene filler was incorporated in the composites. Furthermore, it seems the mobility of polymer chains were restrained by MXene, and the thermal degradation reaction was suppressed as consequence. As is known, higher E_a values indicate a more difficult decomposition reaction due to the fact that E_a is related to the least energy demand in the decomposition process (8). MXene led an upward trend in E_a, which increased to 248.301 kJ/mol from 247.280 kJ/mol with increasing amounts of MXene. The slight variation suggests nanocomposites need higher activation energy than LLDPE in its pure form for thermal degradation and MXene has a positive effect on improving the stabilization effect in composites. These results are in agreement with the analysis of the degradation characteristic temperatures presented above.

It is known that the thermal stability of composites can be improved due to the increased thermal conductivity which causes better heat transfer in the system. Besides, the thermal conductivity of bulk samples can always be enhanced by the existence of thermal conductive fillers. There are several published reports that some kinds of MXene have favorable thermal conductivity (28, 29). On the other hand, the interactions between nanofillers and LLDPE lead to an increase in activation energy resulting in the enhancement of thermal stability of nanocomposites (22). Layered Ti₃C₂, which is a carbon-based filler, might restrain the volatile products produced in the degradation process to being released from bulk polymers by hindering the effects, which give rise to a decrease in the decomposition rate (30). Based on these, we can deem that the representative MXene, Ti₃C₂, can improve the thermal stability in MXene/LLDPE nanocomposites.