Crystallization kinetics and thermal degradation behavior of
low-density polyethylene blended with poly(bispropoxyphosphazene)
Wen-Yen Chiu
a,*, Fa-Tai Wang
a, Leo-Wang Chen
a, Trong-Ming Don
a,1,
Ching-Yuan Lee
baDepartment of Chemical Engineering, National Taiwan University, Taipei, Taiwan, ROC
bChemical System Research Division, Chung Shan Institute of Science and Technology, Taoyuan, Taiwan, ROC
Received 2 April 1999; accepted 9 June 1999
Abstract
Low-density polyethylene (LDPE) was melt blended in a Brabender with dierent amounts of poly(bispropoxyphosphazene) ¯ame retardant, MFR. The crystallization kinetics and thermal degradation behavior of the blends were investigated. It was found that MFR was incompatible with LDPE, where MFR particles dispersed in LDPE matrix with very weak interfacial bonding. In the blends, polyethylene crystals were still in an orthorhombic form and spherulitic structure was observed. Under isothermal crystallization, the crystallization rate and crystallinity were both increased with the addition of MFR up to 20 phr, probably due to the decrease of viscosity and the increase in the number of nucleation sites. The Avrami parameter n was in the range of 2.2 to 2.7 for the blends. In thermal degradation behavior, the addition of MFR increased the maximum-rate degradation temperature (Tmax), char yield at 650C and the limiting oxygen index (LOI) of LDPE, indicating that MFR additive was bene®cial in ¯ame
retardancy. Yet, it deteriorated the tensile mechanical properties of LDPE. # 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Low-density polyethylene; Poly(bispropoxyphosphazene); Blends; Flame retardant; Crystallization kinetics; Thermal degradation behavior
1. Introduction
The thermal degradation behavior of polyethylene (PE) has been studied for many years. As early as in 1949, Straus et al. [1] believed that the degradation of PE was due to the random chain scission, since the pyrolysis products were composed of mainly hydro-carbons from C-1 to C-50. Tsuchiya and Sumi [2] pro-posed a free-radical chain scission mechanism to explain the thermal degradation behavior of a high density-polyethylene (HDPE). Grassie [3] also reported that PE degraded randomly along the main chain. PE was very
thermally stable before 290C; after that, it started to
degrade yet with very small amount of volatile until
360C. As the temperature was raised above 360C, the
degradation rate increased quickly and a quantity of volatile, seldom monomer-type compounds, was pro-duced. Seeger and Barrall [4] studied the thermal degradation behavior of HDPE, LDPE and ethylene± propylene copolymer. They pointed out that the scission rate of C±C bond at a and b position next to the tertiary carbon was twice as fast as the rest of linear chain, i.e. the more the branched chains, the more the iso-alkane compounds produced during degradation.
Flame-retardants are generally added into the organic polymers, since most of them do not have sucient ¯ame retardancy and burn easily. Among them, phos-phorous compounds are increasingly popular over their halogen counterparts, since they generally give o nontoxic and non-corrosive volatile combustion pro-ducts. Therefore, in this study, a poly(bispropoxypho-sphazene) ¯ame retardant, MFR, with high phosphorus and nitrogen content was melt blended with a LDPE. The eect of MFR additive on the crystallization kinet-ics of LDPE was ®rst investigated. Then, the thermal degradation behavior of the blends with dierent amounts of MFR was studied. Finally, the ¯ammability 0141-3910/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0141-3910(99)00116-0
* Corresponding author. Fax: +886-2-2362-3040. E-mail address: ycchiu@ccms.ntu.edu.tw (W-Y. Chiu).
1 Present address: Department of Chemical Engineering, Mingchi
of these blends was evaluated by measuring their limit-ing oxygen index (LOI) values and smoke densities. 2. Experimental
2.1. Materials
Low-density polyethylene (LDPE) having a density of
0.914 g/cm3was supplied by USI Far East Corporation.
It has a melting index of 46 g/10 min. Poly(bisprop-oxyphosphazene), MFR, was kindly supplied by Chung-Shan Institute of Science and Technology. It has
a viscosity of 35 000 cps at 25C and chlorine content
less than 1%. The chemical structure of this MFR is indicated below.
Prior to blending, LDPE pellets were dried in an oven
at 105C for at least 24 h. Various parts of MFR was
blended with 100 parts of LDPE in a Brabender at
135C for 30 min. The rotation speed was set at 30 rpm.
After blending, the mixture was removed from the mixer and placed in a mold sitting on a hot-press machine. The dimension of the mold was 15 cm15 cm3 mm. The composition and the code of the blends are listed in Table 1
2.2. Structure analysis
A Fourier transform infrared (FTIR) spectro-photometer from JASCO FTIR 300-E was used to analyze the chemical and/or physical interactions between LDPE and MFR in the blends. The crystal structure and its morphology were studied using an X-ray diractometer from Phillips PW-1710 and an optical microscope equipped with a polarizer from Zeiss Inc.
2.3. Crystallization kinetics
A dierential scanning calorimeter (DSC) from TA Company, TA 2000, was used to study the crystallization behavior of MFR/LDPE blends. In the isothermal crystallization process, samples were ®rst
heated to 135C and hold for 10 min, followed by
cool-ing down to a speci®c temperature at a rate of 40C/
min. The crystallization exothermic heat was then observed with time. The relative crystallinity, Xr , wast calculated from the ratio of integrated area at some speci®c time to the total area under the exothermic curve. In the dynamic mode, samples were ®rst heated to
135C and hold for 5 min to erase the thermal history,
and then either were quenched to ÿ140C or cooled
down slowly at 5C/min to room temperature. The
samples were then heated again at 5C/min to 140C.
Crystallization temperature (Tc) and melting temperature
(Tm) were calculated from the various thermograms.
2.4. Thermal degradation behavior
The thermal degradation behavior of MFR/LDPE blends was investigated using a Perkin±Elmer TGA 7 thermal gravimetric analysis at various heating rates under nitrogen and air atmosphere, respectively. 2.5. Mechanical properties
Tensile properties: ultimate tensile strength, initial modulus and elongation at break, were measured using Universal Tensile Testing Instrument, RTM-1, from Yashima Works Company, according to ASTM D638 with specimen type I!. The tensile speed was 20 mm/min. 2.6. Morphology observation
A scanning electron microscope, model JEOL TSM-6300, was used to observe the morphology of the blends. Samples were ®rst quenched in liquid nitrogen and immediately were fractured. The fractured surfaces were coated with gold using a sputter coater.
2.7. Flammability
A limiting oxygen index (LOI) tester from Polymer Lab. Ltd. was used to measure the LOI values of various blends following ASTM D2863 method. The specimen dimension was 140523 mm. The smoke density was determined by measuring the speci®c optical density
(Ds) of various blends following ASTM E-662 method.
The size of specimen was 75753 mm and the heat ¯ux
was 25 kW/m2. The maximum smoke density (D
m), the rate of increase of smoke density (R) and the time required to reach Ds=16 (tD16) were calculated. The rate R was calculated from the equation indicated below: Table 1
The composition of various blends and their phosporus content (wt%) Sample Composition (phr MFR)a Phosphoruscontent (%) (theoretical value) Phosphorus content (%) (experimental valueb) MFR ± 21 21 LDPE0F 0 0 <0.2 LDPE5F 5 1.05 0.93 LDPE10F 10 2.01 1.92 LDPE15F 15 2.21 2.67 LDPE20F 20 3.68 3.45 LDPE25F 25 4.42 4.03 a Based on 100 phr of LDPE.
R 1 4 0:9D mÿ 0:7Dm t0:9ÿ t0:7 0:7Dmÿ 0:5Dm t0:7ÿ t0:5 0:5Dtmÿ 0:3Dm 0:5ÿ t0:3 0:3Dmÿ 0:1Dm t0:3ÿ t0:1
where t0:9; t0:7; . . . :; t0:1 were the time needed when Ds
reached 0.9 Dm, 0.7 Dm,. . ..., 0.1 Dm, respectively.
Another parameter, index of the reduction in visibility or smoke obscuration index (SOI), was also calculated by the following equation:
SOI 100 tDm R D16
3. Results and discussion 3.1. Structure analysis
Fig. 1 shows FTIR spectra of MFR, a neat LDPE (LDPE0F) and a blend containing 20 phr MFR (LDPE20F). It can be seen that the spectrum of LDPE20F is only the addition of the two spectra from the corre-sponding MFR and LDPE components. This indicates that there is no chemical bonding between the MFR and LDPE. There is even no shift in the characteristic absorp-tion peaks indicating that LDPE did not have strong inter-action with MFR. Several characteristic absorption peaks are shown in Table 2.
The X-ray diraction patterns of various blends are shown in Fig. 2. There is no shift or apparently decrease in intensity of the peaks corresponding to the (110) and (200) re¯ections from the orthorhombic polyethylene crystals. Therefore, the presence of MFR and the blending method had no eect on the polyethylene crystalline structure. Fig. 3 shows the optical micro-graphs of various blends under isothermal
crystal-lization at 94C for 3 h. Spherulitic structure was
observed for all the blends. Yet with the more addition of MFR, the size of spherulites decreased and the dis-tribution of spherulites was less uniform. This is prob-ably because the addition of MFR increased the heterogeneous nucleation.
3.2. Morphology observation
Fig. 4 shows the SEM micrographs of the fracture surface of various blends. From the pictures, it can be seen that MFR and LDPE were not compatible, where MFR particles dispersed in LDPE matrix. During the fracture, some of MFR particles were ripped o, leaving the voids on the surface. This indicates that MFR did not have strong bonding with LDPE. As the amount of MFR additive increased, the size of particles or voids increased as well.
3.3. Crystallization kinetics
Various MFR/LDPE blends were heated to 135C to
the melt state, and then cooled down at 40C/min to the
pre-set temperature. The crystallization exothermic heat
then was observed with time, and the relative crystallinity
(Xr) was calculated from the area under the curve. Table
3 shows the values of several parameters calculated from these crystallization exothermic peaks of various samples at some speci®c crystallization temperatures
(Tc). In the Table, the Avrami parameters, n and k, were
determine by plotting ln(-ln(1 ÿ Xr)) versus ln(t)
trans-formed from the Avrami equation, 1 ÿ Xr exp ÿkt n.
A typical plot was shown in Fig. 5 for the blend LDPE20F. The n values were in the range of 2.7 to 2.2 for all blends and decreased with the addition of MFR. Table 3 also shows tmaxand t1=2, which were the times to reach the maximum of the exothermic peak and to
reach Xr=50% (half-life time), respectively. Both of
them were decreased with the addition of MFR up to 20 phr, indicating that the crystallization rate increased with the addition of MFR. This was probably caused by the decrease of viscosity and the increase in the number of nucleation sites, when the MFR was added into LDPE. Yet, with more addition of MFR, it started to decrease. This is because the quantity amount of MFR started to aggregate due to the incompatibility between the MFR and LDPE. The aggregation of MFR hinders the crystallization of LDPE. After isothermal
crystal-lization, samples were cooled down to 30C and scanned
again to 140C. The crystallinity was then calculated from
the ratio of endothermic melting heat, Hm, to the
Hm, the melting heat of 100% perfect LDPE
crystal-lites, 294 J/g. Table 4 shows the results of calculated crystallinity of LDPE in various blends after isothermal
crystallization at 94C. It can be seen that the
crystal-linity increased with the addition of MFR. This is also the result of the decrease of viscosity and the increase of the heterogeneous nucleation with the addition of MFR oligomer.
3.4. Thermal degradation behavior
Fig. 6 shows the thermal gravimetric curves of various
blends under nitrogen at a heating rate of 10C/min. It
can be seen that the thermal degradation of MFR
occurred 120C earlier than that of the neat LDPE.
With the addition of MFR increasing in amount, the blend started to degrade earlier and earlier. Table 5 lists all the calculated values from the curves including the
on-set degradation temperature (Ton), maximum-rate
Table 2
Characteristic FTIR absorption peaks of MFR and LDPE Functional
group Wavenumber(cmÿ1) Type
a No. in Fig. 1 PN 1440±1170 S (4) P±O±C 1050±970 S (2) 830±740 PO 1250±1150 S (3) P±O±CH2CH3 1850±1440 ± (6) >N±H 1640 B (7) CH2(hydrocarbon chain) 720 R (1) CH2 1465 S* (5) 1350±1150 T, W CH2 2853 (symmetry) S (8) 2926 (asymmetry) S End CH2 1416 S* ±
a S: stretch vibration; B: bending vibration; R: rocking vibration;
T: twist vibration; W: wagging vibration; S*: scissor vibration.
degradation temperature (Tmax) and the char yield at
650C. It can be seen from the table that the early
degradation of the blend most probably caused by the degradation of MFR component did not accelerate the degradation of LDPE matrix. On the contrary, a slight
increase in Tmax was observed. In addition, the char
yield at 650C increased steadily with the addition of
MFR. It has been pointed out that the degradation
products of MFR component, such as phosphoric acid or poly(phosphoric acid), could form some thermally stable compounds on the condensed phase [5,6]. As a result, these compounds could protect the LDPE matrix from further degradation. In other words, MFR did have an eect in ¯ame retardancy.
The relationship between weight loss fraction () and
apparent activation energy (Ea) can be calculated by
using Ozawa's method [7]. Fig. 7 shows several thermal degradation curves of LDPE10F at dierent heating rates (A); whereas Fig. 8 is their logarithm of heating rates (Log A) versus reciprocal of temperature (1=T) at some certain conversions (). The activation energies thus can be calculated from the slopes of these lines. Table 6 lists all the apparent activation energies at dif-ferent conversions for various blends. As MFR was blended with LDPE, the activation energy at the initial
stage of degradation (5% conversion) was decreased, after that, it was increased. These results support that the early degradation of MFR component would cause
a decrease in Ea at the beginning. Yet, it resulted in the
formation of thermally stable compounds on the con-densed phase, thus increasing the activation energy afterwards.
Fig. 9 shows the TGA curves of various blends at
10C/min heating rate under air atmosphere. Compared
with the TGA curves in N2 (Fig. 6), the degradation
temperature of LDPE main component and the char yield were both higher for the blends degraded in air.
For example, the Tmax and the char yield of LDPE20F
were 490.4C and 8.6%, respectively, which were 24C
and 3.5% higher than the corresponding values in N2.
All the calculated results are shown in the previous Table 5.
3.5. Flammability
The last column in Table 7 shows limiting oxygen index (LOI) values of various blends following ASTM D2863 method. It can be seen that the LOI value increased from 21.0 for a neat LDPE to 23.6 for a LDPE blended with 20 phr MFR. Thus, the addition of
Table 3
The calculated parameters of various samples under isothermal crys-tallizationa
Sample Tc(C) tmax(min) t1/2(min) n k
LDPE0F 97 6.2 6.5 2.65 4.9910ÿ3 96 5.2 5.4 2.66 7.7010ÿ3 95 4.1 4.3 2.67 1.4210ÿ2 94 3.2 3.3 2.68 2.9210ÿ2 93 2.3 2.4 2.70 6.6410ÿ2 LDPE10F 97 5.9 6.3 2.53 6.7010ÿ3 96 5.1 5.4 2.54 9.9010ÿ3 95 4.2 4.4 2.55 1.6010ÿ2 94 3.1 3.3 2.55 3.3910ÿ2 93 2.2 2.3 2.57 7.9610ÿ2 LDPE20F 97 4.7 5.0 2.44 1.3710ÿ2 96 3.6 3.8 2.42 2.5610ÿ2 95 2.8 2.9 2.47 4.9110ÿ2 94 2.1 2.2 2.48 9.7110ÿ2 93 1.7 1.8 2.50 1.6710ÿ1 LDPE25F 97 5.1 5.7 2.21 1.4810ÿ2 96 4.0 4.5 2.23 2.4810ÿ2 95 3.3 3.6 2.24 3.9010ÿ2 94 2.8 3.0 2.24 5.7310ÿ2 93 2.1 2.3 2.26 1.1010ÿ1 a T
c, crystallization temperature; tmax: the time needed to reach the
maximum of the exothermic peak; t1/2: the time needed to reach 50%
relative crystallinity (half-life time); n and k are the exponent and rate constant in Avrami equation.
Fig. 5. Several plots of ln(-ln(1 ÿ Xr)) versus ln(t) in order to
deter-mine the Avrami parameters, n and k.
Fig. 6. The TGA curves of various LDPE blends at a heating rate of 10C/min under N
2.
Table 4
The calculated crystallinity of various samples under isothermal crys-tallization at 94C
Sample Hm(J/g) Hm(J/g LDPE) Crystallinity (%)
LDPE0F 32.4 32.4 11.1 LDPE5F 49.7 52.2 17.8 LDPE10F 43.8 48.1 16.4 LDPE15F 41.8 50.1 17.1 LDPE20F 47.4 54.5 18.6 LDPE25F 39.7 49.6 16.9 Table 5
The calculated results from Tga curves of various samples (heating rate was 10C/min)
Sample In N2atmosphere In air atmosphere
Tona (C) Tmax b (C) Yc c (%) Ton a (C) Tmax b (C) Yc c (%) MFR 288.7 299.9 38.1 265.3 315.3 42.3 LDPE0F 409.0 452.5 0.9 365.9 460.7 3.5 LDPE5F 407.1 462.3 1.5 ± ± ± LDPE10F 370.8 464.8 2.0 358.3 471.8 7.8 LDPE15F 325.8 465.3 3.7 ± ± ± LDPE20F 295.0 466.0 5.1 315.9 490.4 8.6 LDPE25F 260.3 466.3 8.6 290.5 499.7 8.7 a T
onwas the on-set degredation temperature (temperature at 5%
weight loss).
b T
maxwas the maximum-rate degredation temperature. c Y
MFR improved the ¯ame retardancy of LDPE. Table 7 also lists all the calculated values from the smoke
den-sity measurement, including tD16, Dm, R and SOI. It
shows that tD16 was decreased as the amount of MFR
additive increased, indicating that the addition of MFR
did not improve the reduction of smoke density at the initial stage. This is because MFR degraded earlier than
LDPE. Yet, the maximum optical density Dm and the
rate of increase in smoke density R were decreased with the addition of MFR, especially for LDPE5F. There-fore, the early degradation of MFR, though causing an early production of smoke, had reduced the overall smoke density and the rate of increase in smoke density. This is bene®cial in ¯ame retardancy.
3.6. Tensile mechanical properties
The tensile mechanical properties of the various blends were measured at a speed of 20 mm/min. The modulus, ultimate tensile strength and the elongation at break were all decreased with the addition of MFR up to 20 phr (Fig. 10). This is because the soft inclusion of MFR dispersed in the LDPE matrix without any strong bonding between the particles and matrix. Therefore, these MFR particles, like voids, acted as stress con-centrators, which deteriorated the mechanical properties of LDPE.
Fig. 9. The TGA curves of various LDPE blends at a heating rate of 10C/min under air.
Fig. 7. The thermal degradation curves of a LDPE10F blend at dif-ferent heating rates.
Fig. 8. Several plots of logarithm of heating rate (A) versus reciprocal temperature (1=T) at dierent conversions for a LDPE10F blend.
Table 6
The apparant activation energies of various samples under thermal degredation in N2
Sample Activation energy Ea(kJ/mol)
Conversion, a 0.05 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 MFR 90.9 110.7 125.8 143.9 153.0 68.5 85.2 100.7 108.9 ± LDPE0F 141.7 142.4 135.2 136.1 137.8 135.7 135.6 134.0 133.6 128.9 LDPE10F 128.3 148.3 142.0 140.8 139.9 139.9 140.4 135.5 132.4 125.3 LDPE20F 112.4 176.8 145.9 142.9 146.4 147.5 147.4 148.6 149.3 116.8 LDPE25F 73.7 92.8 193.8 163.6 154.7 148.6 146.3 141.2 120.1 ± a =[100ÿresidual wt (%) in TGA]/100. Table 7
The limiting oxygen index and various parameters calculated from smoke density measurementa
Sample tD16(s) Dm R SOI LOI
LDPE0F 304 141 19.0 5.3 21.0 LDPE5F 208 77 6.8 1.5 21.4 LDPE10F 182 105 11.3 3.9 22.8 LDPE15F 141 107 12.5 5.7 23.0 LDPE20F 126 123 16.9 9.9 23.6 a t
D16, the time required to reach speci®c optical density=16; Dm:
the maximum smoke density; R: the rate of increase in smoke density; SOI: index of the reduction in visibility or smoke obscuration index; LOI: limiting oxygen index.
4. Conclusions
1. LDPE was melt blended with dierent amount of MFR ¯ame retardant in a Brabender. It was found
that MFR phase separated from LDPE matrix with very weak interfacial bonding.
2. The X-ray diraction patterns revealed two peaks corresponding to the (110) and (200) re¯ections from the orthorhombic polyethylene crystals for all the blends up to 25 phr MFR. In addition, polarized optical microscope shows spherulitic structures for all the blends, yet, the size of spher-ulites decreased with the addition of MFR. 3. The crystallization rate and the crystallinity of
LDPE were increased with the addition of MFR under isothermal crystallization at various
tem-peratures from 93 to 97C. This was attributed to
the decrease in viscosity and increase in the num-ber of nucleation sites due to the addition of MFR oligomer.
4. With the addition of MFR in LDPE, the max-imum-rate degradation temperature, char yield at
650C and the limiting oxygen index values were
all increased. These results indicated that the addition of MFR was advantageous in ¯ame retardancy.
5. Because of the soft inclusion of MFR and weak interfacial bonding, the MFR additive deterio-rated the tensile mechanical properties of LDPE. Acknowledgements
The authors would like to thank Chung Shan Insti-tute of Science and Technology (CSIST), Taiwan, ROC for their ®nancial support.
References
[1] Straus SSL, Madorsky D, William L. J Polym Sci 1949;4:639. [2] Tsuchiya Y, Sumi K. J Polym Sci, Part A-1 1968;6:415. [3] Grassie N. In: Mark HF, editor. Encyclopedia of polymer
sci-ence and technology, vol 14, 2nd ed. New York: Wiley, 1985. p. 65.
[4] Seeger M, Barrull EM. J Polym Sci, Polym Chem 1975;13:1515. [5] Grassie N, Scott G. Polymer degradation and stability.
Cam-bridge: Cambridge University Press, 1985 [Chapter 6].
[6] Wang P-S, Denq B-L, Chiu W-Y, Don T-M, Chiu Y-S. Polym Degrad Stab, in press.
[7] Ozawa T. Bull Chem Soc Jpn 1965;38:1881. Fig. 10. The tensile mechanical properties of various LDPE blends