32

NE

HS

a

ARRAA

KNCDM

1

amtipiaimccteaetdl

rcp[

0h

Thermochimica Acta 545 (2012) 148 156

Contents lists available at SciVerse ScienceDirect

Thermochimica Acta

journa l h o me page: www.elsev ier .com/ locate / tca

on-isothermal crystallization kinetics of peroxide-crosslinked polyethylene:ffect of solid state mechanochemical milling

ejun Wu, Mei Liang , Canhui Lu

tate Key Laboratory of Polymer Materials Engineering, Polymer Research Institute of Sichuan University, Chengdu 610065, China

r t i c l e i n f o

rticle history:eceived 21 April 2012eceived in revised form 13 June 2012

a b s t r a c t

The influence of solid state mechanochemical milling process on the non-isothermal crystallization ofperoxide-crosslinked-low-density polyethylene (peroxide-XLPE) was investigated via differential scan-ning calorimetry (DSC) at different cooling rates. Peroxide-XLPE was partially decrosslinked duringccepted 10 July 2012vailable online 16 July 2012

eywords:on-isothermal crystallization kineticsrosslinked polyethylene

mechanochemical milling process. The Avrami method modified by Jeziorny, the Ozawa method andthe Mo method were applied to describe the non-isothermal crystallization process of peroxide-XLPE,decrosslinked XLPE (de-XLPE) and pure LDPE successfully. The results revealed that the decrosslink-ing effect could accelerate the crystallization process of XLPE due to the higher chain mobility aftermechanochemical milling, which was confirmed by the values of half crystallization time (t1/2), crystal-

paraecrosslinkingechanochemical milling

lization rate exponent Zc,

. Introduction

Crosslinked polyethylene (XLPE) has been widely used asn insulation material in power cables because of its excellentechanical properties and dielectric properties [13]. However,

he three-dimensional network structure formed after crosslink-ng not only improves its desirable properties [4], but also changesolyethylene from a thermoplastic to a thermoset [5]. Therefore, its very difficult to recycle XLPE due to its low fluidity and poor mold-bility caused by crosslinking. Currently, as the country updatests electrical infrastructure, waste XLPE from scrap power trans-ission and distribution cables is becoming a significant disposaloncern. There are some ways to recycle the waste XLPE, includingonversion into fuel by thermal and catalytic cracking [6,7], andhe use as a powdery filler by blending with thermoplastics in anxtruder to make a thermoplastic recycled polymer [8]. Thermalnd catalytic cracking might be accompanied by a large amount ofnergy consumption since it is usually carried out at high tempera-ures [9], and the poor mechanical properties caused by randomecomposition of the polymer chain during extrusion may alsoimit the applications of powdered XLPE as a filler [10].

Considering the prevention of environmental pollution and

esource conservation, the development of more effective recy-ling technologies of XLPE is desired. Some researchers haveroposed the application of supercritical fluids such as water11], methanol [12], and alcohol [13] as a recycling method to

Corresponding author. Tel.: +86 28 85460607; fax: +86 28 85402465.E-mail addresses: [email protected] (M. Liang), [email protected] (C. Lu).

040-6031/$ see front matter 2012 Published by Elsevier B.V.ttp://dx.doi.org/10.1016/j.tca.2012.07.008meter K(T) and F(T). 2012 Published by Elsevier B.V.

achieve selective decrosslinking at crosslinking points of XLPEwhich is difficult in conventional processes. For example, Lee etal. [12] studied a decrosslinking reaction of XLPE in supercriti-cal methanol using a batch reactor and developed a first-orderkinetics on the basis of experimental results in which the reac-tion rate was linearly proportional to the gel concentration andrelated exponentially to the temperature model. However, prob-lems still persist because the critical temperature and pressureof these supercritical fluids are high [14], relating to the highcost of equipment and difficulty for continuous processing. Thus,it remains a technological challenge to develop a more effec-tive method for recycling XLPE without any additional materialsand chemicals at ambient temperature. Recently, an eco-friendlymechanochemical technology called solid state mechanochemicalmilling using a pan-mill equipment to recycle crosslinked poly-mers has been developed in our laboratory [1519]. This methodhas been proved to have many advantages over other conventionalrecycling methods, such as the fact that it can be conducted at ambi-ent temperature, is chemical free, has low energy consumption, andis more favorable for the environment. The pan-mill equipment wasdesigned in the authors laboratory for solid-state mechanochem-ical reactions of polymers [20], showing multifunction, such aspulverizing, dispersion, mixing, and activation [2125]. In our pre-vious work [18], it has been demonstrated that the solid statemechanochemical milling process can be utilized as an effec-

tive method to recycle waste XLPE by partial decrosslinking. Theselective decrosslinking was successfully achieved at crosslinkingelements by mechanochemical milling, and consequently led to thehigher chain mobility of XLPE which might affect its crystallizationprocess.

ica Acta 545 (2012) 148 156 149

tpctnnttdtppccpdsom

tpwmdatSc

2

2

DspP

2

fgtA1tl

ttuAododthTsl

H. Wu et al. / Thermochim

It is well-known that crystallization behaviors play an impor-ant role in the physical, chemical, and mechanical properties ofolymers. Moreover, from the practical view the non-isothermalrystallization is more useful than the isothermal crystalliza-ion because most processing techniques actually occur underon-isothermal conditions [26]. Therefore, it is of great sig-ificance to understand the effect of milling process on thehermal behavior and the crystallization kinetics of XLPE, dueo the fact that the resulting physical properties are stronglyependent on the morphology formed and the extent of crys-allization occurring during processing [27]. There are extensiveapers on studying the non-isothermal crystallization kinetics ofolyethylenes [2836], but few publications have been found con-erning to the studies of non-isothermal crystallization kinetics ofrosslinked polyethylenes [37,38], especially peroxide-crosslinkedolyethylene (peroxide-XLPE), not to mention the case with theecrosslinked peroxide-XLPE (de-XLPE). Because of this, more workhould be done to study the non-isothermal crystallization kineticsf peroxide-XLPE, as well as the influences of the mechanochemicalilling process on its non-isothermal crystallization behavior.In this article, to investigate the effect of the decrosslinking on

he non-isothermal crystallization kinetics of XLPE, low-densityolyethylene (LDPE) was crosslinked by dicumyl peroxide (DCP)ithout further additives, and then subjected to solid stateechanochemical milling using a pan-mill equipment to obtainecrosslinked XLPE (de-XLPE). Several theoretical models werepplied to describe the process of non-isothermal crystalliza-ion based on the differential scanning calorimeter (DSC) data.imultaneously, pure LDPE was analyzed in the same process foromparison.

. Experimental

.1. Materials

Low density polyethylene (LDPE 1254NT) was obtained from theow Chemical Company, with a melt index of 8 g/10 min and a den-ity of 0.918 g/cm3. The crosslinking agent employed was dicumyleroxide (DCP, >99% purity), provided by SinopecShanghaiGaoQiaoetrochemical Corporation.

.2. Preparation of samples

LDPE was mixed with 1.5 wt % DCP on a two roll mill at 110 Cor 58 min. At this state the DCP was mixed with LDPE withoutoing through chemical reactions because the mixing tempera-ure was not sufficient to activate the peroxide initiation reactions.fterwards, the samples were compressed molded at 170 C for5 min under 15 MPa, and then cooled to ambient temperature athe same pressure, obtaining the peroxide-crosslinked polyethy-ene (peroxide-XLPE) sheets of 2 mm thickness.

A pan-mill type mechanochemical reactor was used to performhe decrosslinking of XLPEs at ambient temperature, as described inhe previous article [18]. The sheets of XLPE were chipped into gran-les (510 mm in diameter) with a rotary blade chipping machine.fter that, the chips of XLPE were fed in hopper set at the middlef the moving pan at a rotating speed of 30 rpm from the inlet andriven by shear force, moving along a spiral route toward the edgef the pan until they came out from the outlet. The discharged pow-er was collected for the next cycle of milling. The average retention

ime of the powder during milling was 2540 s per cycle, and theeat generated during milling was removed by water circulation.he chain-transmission system and a screw-pressure system wereet to regulate the rotation speed of the moving pan and imposedoad, respectively, which could strictly control two major dynamicFig. 1. Effect of mechanochemical milling on the gel content of XLPE.

parameters during milling: the velocity and force. The repetitionoperation continued for 20 cycles to produce decrosslinked XLPE.

2.3. Characterization

2.3.1. Determination of gel contentThe gel content of crosslinked and decrosslinked samples was

determined gravimetrically, according to ASTM D 2765, using a 24-h soxhlet extraction cycle, with p-xylene as the solvent at 140 C.Approximately 0.3 g weighed samples were cut into small piecesand placed in a pre-weighted stainless steel, fine wire mesh. Afterthe extraction cycle, the samples were washed with acetone andvacuum-dried to a constant weight. The gel fraction was calculatedas the percentage ratio of the final weight of the polymer to itsinitial weight.

2.3.2. Differential scanning calorimetry (DSC) measurementThe non-isothermal crystallization was carried out by a differ-

ential scanning calorimeter (DSC-204, Netzsch, Germany). Samplesweighing about 5 mg were cut off for characterization. Nitrogenpurge gas with a flux of 40 ml/min was used to prevent thermaldegradation of samples during scanning. All samples were firstquickly heated at 40 C/min up to 180 C and held isothermally for5 min to eliminate the thermal history of the polymers, and thenthey were cooled to room temperature at four different coolingrates: 2.5, 5, 10, and 20 C/min, respectively.

3. Results and discussion

3.1. Effect of mechanochemical milling on the gel content of XLPE

To elucidate the effect of mechanochemical milling on the non-isothermal crystallization kinetics of XLPE, LDPE was crosslinked by1.5 wt% DCP, and then subjected to 20 cycles of mechanochemicalmilling using a pan-mill equipment to obtain decrosslinked XLPE.The effect of mechanochemical milling on the gel content of XLPEis shown in Fig. 1. It was evident that the gel content decreasedremarkably after mechanochemical milling, from its initial 62.3 to14.6% after 20 cycles of milling. This indicated that the crosslinkedstructure of XLPE was broken during the mechanochemical millingprocess, which made the soluble fractions of the samples increase.As the results of our previous article [18], it was confirmed by the

SEC test that the molecular weights of the sol fraction of XLPE beforeand after decrosslinking were close, which further demonstratedthat the XLPE was fairly selectively decomposed at the crosslinkpoints rather than at random sites [1012]. As illustrated in Fig. 2,the crosslinking points of XLPE additionally made the chains

150 H. Wu et al. / Thermochimica Acta 545 (2012) 148 156

ecross

pafitces

3

tsespbtblJ

Fa

Fig. 2. Scheme for the d

restressed to move physically close enough to become crosslinkednd also restricted the movement of the molecular chains. There-ore, when the molecular network and the entire structure wasn motion during mechanochemical milling, the crosslinked struc-ure of XLPE would most likely have had a higher opportunity forhain scissions and been broken preferably, because it had the high-st stress and moment under the action of the adequately strongqueezing and shearing forces exerted by the pan-mill equipment.

.2. Non-isothermal crystallization behavior

The non-isothermal crystallization of XLPE, de-XLPE, LDPE fromhe melt at four cooling rates of 2.5, 5, 10, and 20 C/min wastudied using DSC. Fig. 3 shows the nonisothermal crystallizationxotherms of all samples at a cooling rate of 2.5 C/min. It can beeen from Fig. 3 that both pure LDPE and de-XLPE has a singleeak, whereas XLPE seems to have double peaks. The appearance ofimodal peaks for XLPE at low cooling rate is caused by the forma-

ion of network after crosslinking, which may work as a physicalarrier for the crystal growth and consequently yield peak withower crystallization temperature. Similar results are reported byiao et al. for silane-crosslinked LLDPE [37] and Hutzler et al. for

ig. 3. DSC thermograms of non-isothermal crystallization for XLPE (a), de-XLPE (b),nd LDPE (c) at the cooling rate of 2.5 C/min.linking process of XLPE.

irradiated LDPE [39]. The onset temperature (To), which is the tem-perature at the crossing point of the tangents of the baseline andthe high temperature side of the exotherm, the peak temperature(Tp), the crystallization enthalpy (Hc) and the relative crystallinity(Xc), defined by the ratio of Hc to the heat of fusion of the purelycrystalline form of PE, H0f (H

0f = 289.9 J/g [40]) derived from

the DSC curves are listed in Table 1.It could be found that Tp shifted toward lower temperature and

the exothermic peak became broader with increasing cooling ratesfor all samples. As is known, the faster cooling rate means theshorter time stayed at every temperature for the polymer matrixto develop crystallization from the melt, and thus the crystalliza-tion process is restrained. Meanwhile a wider temperature rangewas needed for the samples to accomplish their crystallization asthe cooling rate increased. Furthermore the increase of coolingrate could also make the crystal structure imperfect and irregu-lar, which then induced the lower Hc and Xc. It should be noticedthat for a given cooling rate, values of these aforementioned param-eters listed in Table 1 followed the order: XLPE < de-XLPE < LDPE,which could be explained by the partial decrosslinking effect duringmechanochemical milling. The crystallization process of LDPE washindered after the crosslinking because XLPE formed a net struc-ture. In contrast, mechanochemical milling could primarily resultin partial decrosslinking at the crosslinking points of XLPE, whichdestroyed its crosslinked structure to make the macromoleculechains more flexible. So the crystallization became easier, and con-sequently the crystallization temperature and relative crystallinityof de-XLPE became higher than those of XLPE.

3.3. Non-isothermal crystallization kinetics

In order to further analyze the non-isothermal crystallizationprocess, the crystallization kinetics of XLPE, de-XLPE, and LDPEwere compared. The relative crystallinity as a function of crys-tallization temperature, XT, was calculated as the ratio of theexothermic peak areas [4143] according to Eq. (1):

T(dH /dT)dTXT = Toc T

To(dHc/dT)dT

(1)

where To and T are the onset and end crystallization tempera-ture, respectively, T is an arbitrary crystallization temperature at

H. Wu et al. / Thermochimica Acta 545 (2012) 148 156 151

Table 1Characteristic data of nonisothermal crystallization exotherms for XLPE, de-XLPE, and LDPE at different cooling rates.

Sample Cooling rate (K/min) To (C) Tp (C) Hc (J/g) Xc (%) t1/2a (min)

XLPE 2.5 112.8 109.1 100.0 34.5 1.605 112.0 106.8 97.3 33.6 1.13

10 110.8 103.0 91.7 31.6 0.9120 109.0 97.1 85.8 29.6 0.61

de-XLPE 2.5 116.2 111.2 135.2 46.6 1.945 115.2 109.5 132.7 45.8 1.09

10 114.1 107.8 128.2 44.2 0.7020 112.3 104.5 112.7 38.9 0.43

LDPE 2.5 118.7 116.6 194.3 67.0 1.045 118.0 114.7 192.9 66.5 0.78

n to t

tatf

t

aittfcrwpgftdcfufwtiTrcm

aetc

3

ttc

X

wtt

Jeziorny method are summarized in Table 2 as well.On the whole, the good linearity of the curves (correlation coeffi-

cient R2 > 0.99) in Fig. 5 demonstrates that the Avrami equation fitswell with the non-isothermal crystallization kinetics of all samples.For LDPE, the Avrami plots shown in Fig. 5c were also linear over the

Table 2Non-isothermal crystallization kinetic parameters based on the Avrami and Jeziornymethod.

Sample (K/min) n Zt R2 Zc t1/2a (min)

XLPE 2.5 2.63 0.18 0.998 0.50 1.675 2.48 0.49 0.998 0.87 1.15

10 2.35 1.06 0.997 1.00 0.8320 2.30 2.22 0.996 1.04 0.60

de-XLPE 2.5 2.40 0.17 0.997 0.49 1.805 2.38 0.67 0.997 0.92 1.01

10 2.32 2.09 0.996 1.08 0.6220 2.26 5.03 0.997 1.08 0.42

LDPE 2.5 2.09 0.70 0.996 0.87 1.0010 117.0 20 114.9

a The half crystallization time, defined as the time from the onset of crystallizatio

ime t, and dHc is the enthalpy of crystallization released duringn infinitesimal temperature interval dT. The horizontal tempera-ure scale can be transformed into the time domain based on theollowing equation in which the stands for the cooling rate:

=To T

(2)

As a representation, the plot of the relative degree of crystallinitys a function of time, Xt, for all samples at different cooling rates isllustrated in Fig. 4. It could be seen clearly from Fig. 4 that the higherhe cooling rate, the shorter the time for completing the crystalliza-ion process. The half crystallization time (t1/2), defined as the timerom the onset of crystallization to the time at which the relativerystallinity is 50%, can be directly estimated from Fig. 4 and theesults are listed in Table 1. As expected, the value of t1/2 decreasedith increasing cooling rates for all samples, indicating that theolymer crystallized faster when the cooling rate is increased. At aiven cooling rate, the value of t1/2 for de-XLPE was lower than thator XLPE except at 2.5 C/min. On the basis of the molecular motionheory, the crystallization of XLPE is affected by the crosslinkingensity of its molecules. The mechanochemical milling processould destroy the crosslinked structure of XLPE and increase the solraction (decreased the crosslinking density) with different molec-lar weights of de-XLPE. When the cooling rate was low enough,or example at 2.5 C/min, it might allow more molecular chainsith different molecular weights of de-XLPE to entrance into crys-

alline regions, resulting in a higher t1/2. Nevertheless, this differentn t1/2 between de-XLPE and XLPE is almost negligible, and both theo and Tp of de-XLPE are higher than those of XLPE at the coolingate of 2.5 C/min as listed in Table 1. In this case, it can still beoncluded that the decrosslinking effect due to mechanochemicalilling could accelerate the overall crystallization process.For the analysis of the experimental results of XLPE, de-XLPE,

nd LPDE under non-isothermal conditions, the modified Avramiquation, the Ozawa analysis, and the Mo method were appliedo describe their crystallization process to obtain the parametersharacterizing the kinetics of non-isothermal crystallization.

.3.1. Avrami methodThe Avrami method [44,45] is most widely applied in describing

he isothermal crystallization kinetics of polymers. According tohe Avrami theory the relative crystallinity at a time t, Xt, can bealculated from the following equation:t = 1 exp(Zttn) (3)

here n is the Avrami crystallization exponent dependent on theype of nucleation mechanism and growth parameters, t is theime taken during the crystallization process, Zt is a composite rate112.0 189.5 65.4 0.62108.2 168.9 58.3 0.41

he time at which the relative crystallinity is 50% from Fig. 4.

constant involving both nucleation and growth rate parameters. Eq.(3) can be liberated in its double logarithmic form as:

ln[ ln(1 Xt)] = ln Zt + n ln t (4)Plotting ln[ln(1 Xt)] versus ln t for each cooling rate, a straightline would be obtained and the values of n and Zt can be calculatedfrom the slopes and intercepts of the lines, respectively.

It should be noted that n and Zt in non-isothermal crystalliza-tion do not have the same physical significance as in the isothermalcrystallization because temperature is lowered constantly dur-ing non-isothermal crystallization. Considering the non-isothermalcharacter of the process investigated, Jeziorny [46] extended theisothermal Avrami equation to the nonisothermal situation byproposing that the rate parameter Zt should be corrected by coolingrate as follows:

ln Zc = ln Zt

(5)

where Zc is the corrected kinetic rate constant. According to theprevious reports [47], the half crystallization time, can also be cal-culated from the corrected constant using the equation

t1/2 =[ln 2Zt

]1/n(6)

The plots of ln[ln(1 Xt)] versus ln t at different cooling rates aregiven in Fig. 5 and the results obtained from the Avrami plots and5 2.12 1.31 0.997 1.06 0.7410 2.12 1.79 0.996 1.06 0.6420 1.97 4.52 0.994 1.08 0.39

a The half crystallization time calculated from the corrected kinetic rate constantZc by Eq. (6).

152 H. Wu et al. / Thermochimica Acta 545 (2012) 148 156

Fv

waodcic

bt

ig. 4. Relative crystallinity versus time of XLPE (a), de-XLPE (b), and LDPE (c) atarious cooling rates.

hole range implying that it had only one crystallization process,nd this was different from the result of Ref. [48] in which the sec-ndary crystallization of LDPE was observed. This difference may beue to the initial portion of short ethylene branches in LDPE whichould inhibit its overall crystallization behavior [35,36]. LDPE usedn this study might have fewer short ethylene branches since its

rystallinity as listed in Table 1 was almost more than 65%.

The values of Avrami exponents n for all the samples wereetween 2 and 3, suggesting that crystallites are two- orhree- dimensional and nucleation may be between instantaneousFig. 5. Plots of ln[ln(1 Xt)] versus ln t for the non-isothermal crystallization ofXLPE (a), de-XLPE (b), and LDPE (c) at various cooling rates.

and sporadic [49]. These values of n were consistent with corre-sponding literature data reported for polyethylene in range from 2to 4 (mostly for isothermal crystallization) [5052]. It could be seenthat for all samples the values of Zc increased and t1/2 decreasedas the cooling rate increased, which meant an increase of crystal-

lization rate. For a given cooling rate, Zc for de-XLPE was higherthan that for XLPE and t1/2 for de-XLPE was lower, signifying thatde-XLPE had a faster crystallization rate than XLPE. This could beexplained by the enhanced mobility of the polymer chains due to

H. Wu et al. / Thermochimica Acta 545 (2012) 148 156 153

Fig. 6. Ozawa plots of ln[ln(1 Xt)] versus ln for the non-isothermal crystalliza-t

dtific

Ozawa plot for polyethylene was also found in some literaturesion of XLPE (a), de-XLPE (b), and LDPE (c) at different temperatures.

ecrosslinking of XLPE after mechanochemical milling. In addition,he uniformity of measured and calculated values of t1/2 (listedn Tables 1 and 2, respectively) also indicated clearly that modi-

ed Avrami equation can be used to describe the non-isothermalrystallization kinetics of XLPE, de-XLPE and LDPE.Fig. 7. Plots of the ln versus ln t for the non-isothermal crystallization of XLPE (a),de-XLPE (b), and LDPE (c) at different degrees of crystallinity.

3.3.2. Ozawa methodThe Ozawa equation has been successfully used for describ-

ing the non-isothermal crystallization of many polymers [5357],including some polyethylene [37]. However, the nonlinearity of the[32,58,59]. Thereby, it is necessary to investigate the validity of theOzawa model for the samples used in the current paper.

154 H. Wu et al. / Thermochimica Acta 545 (2012) 148 156

Table 3Non-isothermal crystallization kinetic parameters based on the Ozawa method.

Temperature (C) XLPE de-XLPE LDPE

m K(T) R2 m K(T) R2 m K(T) R2

107 1.75 8.76 1.00 1.38 16.78 0.99 108 1.89 6.30 1.00 1.42 12.43 0.99

9.49 0.98 1.60 47.47 0.998.08 0.95 1.67 40.85 1.007.17 0.99 1.71 33.12 1.002.59 0.98 1.80 27.11 1.00

tir

1

wttiw

l

Slsor

FltoOinVfisiftvtiaa

Rtpipaw

3

mt[

l

Table 4Non-isothermal crystallization kinetic parameters based on the Mo method.

Sample Xt (%) F(T) R2

XLPE 20 2.05 2.66 0.9840 2.16 5.31 0.9960 2.23 9.12 0.9980 2.28 15.64 0.99

de-XLPE 20 1.33 3.13 1.0040 1.38 5.00 1.0060 1.40 7.10 1.0080 1.45 9.87 0.99

LDPE 20 1.93 1.05 0.97109 2.01 3.82 1.00 1.53 110 2.13 1.75 0.98 1.70 111 1.65 112 1.73

According to the Ozawa theory [60], the non-isothermal crys-allization process is the result of an infinite number of smallsothermal crystallization steps and the degree of conversion orelative crystallinity Xt can be calculated as follows:

Xt = exp[K(T)

m

](7)

here is cooling rate, m is Ozawa exponent that depends onhe dimension of crystal growth, and K(T) is the cooling crystalliza-ion function, which is related to the overall crystallization rate andndicates how fast crystallization occurs. The above equation can beritten as

n[ ln(1 Xt)] = ln K(T) m ln (8)

tudying the process at different cooling rates and plottingn[ln(1 Xt)] against ln at a given temperature, a straight linehould be obtained if the Ozawa method is valid, and parametersf m and K(T) can be determined from the slope and intercept,espectively.

The Ozawa plots of XLPE, de-XLPE and LDPE are presented inig. 6. In order to obtain a reliable and comparable result fromeast-square lines drawn through plots of at least three points, theemperature range varies from sample to sample, lying in the rangef 107112 C. It was shown obviously in Fig. 6 that the lines of thezawa plots for a certain sample were linear and nearly parallel,ndicating that that the Ozawa model is suitable for describing theon-isothermal crystallization kinetics of XLPE, de-XLPE and LDPE.alues of m and K(T) as well as the corresponding correlation coef-cient R2 for each sample are summarized in Table 3. For all theamples, the Ozawa exponent m was found to roughly increase withncreasing temperature, while the Ozawa rate constant K(T) wasound to decrease, suggesting that the higher the temperature was,he slower the polymers crystallized. For a given temperature, thealue of K(T) for de-XLPE was higher than that for XLPE, suggestinghat de-XLPE had a faster crystallization rate. This result was alson agreement with the previous observations of the parameters Zcnd t1/2 obtained through the modified Avrami method mentionedbove.Moreover, based on the corresponding correlation coefficient

2, it can be concluded fairly that the Ozawa method is a satisfac-ory description of the non-isothermal crystallization kinetics ofolyethylene studied in this article, regardless of whether or not its crosslinked [61]. The most possible reason for this may be thatrimary crystallization these polymers played a predominant rolend the second crystallization could be neglected [32,36,62], whichas in favor of the results of the Avrami analysis.

.3.3. Mo methodBy combining the Ozawa and Avrami equations, a methododified by Mo [63] has been successfully employed to describehe non-isothermal crystallization kinetics of various polymers6467]. The final form of modified equation is given as follows:

n = ln F(T) ln t (9)40 2.15 2.12 0.9760 2.21 4.10 0.9880 2.20 7.85 0.97

where the parameter F(T) = [K(T)/Zt]1/m refers to the value of cool-ing rate chosen at unit crystallization time when the system hasa certain degree of crystallinity, m is the Ozawa exponent, and refers to the ratio of the Avrami exponent n to the Ozawa exponentm ( = n/m).

According to Eq. (9), at a given degree of crystallinity the plotof ln versus ln t for a certain sample would yield a straight lineas shown in Fig. 7, and the value of a and F(T) could be obtainedby the slope and the intercept of the line as listed in Table 4. Itcould be seen clearly that the plots showed a good linear relation-ship between ln and ln t, verifying obviously that Mo methoddescribed appropriately the non-isothermal crystallization kineticsof all samples.

The values of F(T) and systematically increased with a rise inthe relative degree of crystallinity. F(T) has a definite physical andpractical meaning, for a given degree of crystallinity, the higher thevalue of F(T), the higher the cooling rate needed within unit crys-tallization time, indicating the difficulty of polymer crystallization.By comparing the values of F(T) of different samples, it was foundthat the values of de-XLPE were almost lower than those of XLPE,indicating that the crystallization rate of de-XLPE was faster thanthat of XLPE. This was in good accordance with the results obtainedfrom the Avrami and Ozawa approach.

4. Conclusions

It have been demonstrated that solid state mechanochemicalmilling using a pan-mill equipment is a good method to recycleXLPE by partial decrosslinking. In order to investigate the effect ofmilling process on the crystallization kinetics of XLPE, three dif-ferent kinetic models of the Avrami, the Ozawa and the Mo wereapplied to systematically study the non-isothermal crystallizationkinetics of XLPE, de-XLPE and LDPE based on the DSC technique.All the three models were found to describe the experimental data

very well.

The parameters Zc and t1/2 derived from the modified Avramimethod suggested that de-XLPE had a faster crystallization ratethan XLPE due to the higher chain mobility resulted from partial

ica A

dtttcMtmt

A

dRdlIf

R

[

[

[

[

[

[

[

[

[

[

[[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

H. Wu et al. / Thermochim

ecrosslinking. K(T) from the Ozawa method revealed a tendencyhat accorded well with results of the Avrami method. In addition,he Ozawa method was found to be a satisfactory description ofhe non-isothermal crystallization process of LDPE for its secondaryrystallization was not obvious and could be neglected. From theo method, the value of F(T) of de-XLPE was found to be lower than

hat of XLPE, which also indicated that the decrosslinking effect ofechanochemical milling could accelerate the overall crystalliza-

ion process.

cknowledgments

This study is supported by the National Natural Science Foun-ation of China (50903053 and 51073108) and the Applied Basicesearch Programs of Science and Technology Commission Foun-ation of Sichuan Province (2010JY0022), and the authors wouldike to thank Analytical and Testing Group of Polymer Researchnstitute of Sichuan University for providing DSC measurementsacility.

eferences

[1] F. Ciuprina, G. Teissedreand, J.C. Filippini, Polyethylene crosslinking and watertreeing, Polymer 42 (2001) 78417846.

[2] A.S. Vaughan, Y. Zhao, L.L. Barre, S.J. Sutton, S.G. Swingler, On additives, mor-phological evolution and dielectric breakdown in low density polyethylene,Eur. Polym. J. 39 (2003) 355365.

[3] A. Smedberg, T. Hjertberg, B. Gustafsson, Effect of molecular structure andtopology on network formation in peroxide crosslinked polyethylene, Polymer44 (2003) 33953405.

[4] P. Rizzo, F. Baione, G. Guerra, L. Martinotto, E. Albizzati, Polyethylene unit celland crystallinity variations as a consequence of different cross-linking pro-cesses, Macromolecules 34 (2001) 51755179.

[5] R.W. Waldron, H.F. McRae, J.D. Madison, The effects of various monomers oncrosslinking efficiency, Radiat. Phys. Chem. 25 (1985) 843848.

[6] G. Manos, I.Y. Yusof, N.H. Gangas, N. Papayannakos, Tertiary recycling ofpolyethylene to hydrocarbon fuel by catalytic cracking over aluminum pillaredclays, Energy Fuels 16 (2002) 485489.

[7] R.V. Grieken, D.P. Serrano, J. Aguado, R. Garcia, C. Rojo, Thermal and catalyticcracking of polyethylene under mild conditions, J. Anal. Appl. Pyrolysis 5859(2001) 127142.

[8] C.C. White, J. Wagenblast, M.T. Shaw, Separation, size reduction, and processingof XLPE from electrical transmission and distribution cable, Polym. Eng. Sci 40(2000) 863879.

[9] D.S. Achilias, C. Roupakias, P. Megalokonomos, A.A. Lappas, E.V. Antonakou,Chemical recycling of plastic wastes made from polyethylene (LDPE and HDPE)and polypropylene (PP), J. Hazard. Mater. 149 (2007) 536542.

10] T. Goto, T. Yamazaki, T. Sugeta, I. Okajima, T. Sako, Selective decomposition ofthe siloxane bond constituting the crosslinking element of silane-crosslinkedpolyethylene by supercritical alcohol, J. Appl. Polym. Sci. 109 (2008) 144151.

11] S. Watanabe, K. Komura, S. Nagaya, H. Morita, T. Nakamoto, S. Hirai, F. Aida,Development of cross-linked polymer material recycling technology by super-critical water, in: Proceedings of the 7th International Conference on Propertiesand Applications of Dielectric Material, Nagoya, Japan 2, 2003, pp. 595598.

12] H.S. Lee, J.H. Jeong, H.K. Cho, C.M. Koo, S.M. Hong, H. Kim, Y.W. Lee, A kineticstudy of the decross-linking of cross-linked polyethylene in supercriticalmethanol, Polym. Degrad. Stab. 93 (2008) 20842088.

13] T. Goto, Y. Yamazaki, Recycling of silane cross-linked polyethylene for insula-tion of cables using supercritical alcohol, Hitachi Cable Rev. 23 (2004) 2427.

14] M. Goto, Chemical recycling of plastics using sub- and supercritical fluids, J.Supercrit. Fluids 47 (2009) 500507.

15] X.X. Zhang, C.H. Lu, M. Liang, Devulcanisation of natural rubber vulcanisatethrough solid state mechanochemical milling at ambient temperature, Plast.Rubber Compos. 36 (2007) 370376.

16] X.X. Zhang, C.H. Lu, M. Liang, Preparation of rubber composites from groundtire rubber reinforced with waste-tire fiber through mechanical milling, J. Appl.Polym. Sci. 103 (2007) 40874094.

17] X.X. Zhang, C.H. Lu, M. Liang, Properties of natural rubber vulcanizates con-taining mechanochemically devulcanized ground tire rubber, J. Polym. Res. 16(2009) 411419.

18] H.J. Wu, M. Liang, C.H. Lu, Morphological and structural development of recy-cled crosslinked polyethylene during solid-state mechanochemical milling, J.Appl. Polym. Sci. 122 (2011) 257264.

19] X.X. Zhang, C.H. Lu, M. Liang, Preparation of thermoplastic vulcanizates basedon waste crosslinked polyethylene and ground tire rubber through dynamic

vulcanization, J. Appl. Polym. Sci. 122 (2011) 21102120.

20] X. Xu, Q. Wang, Mechanochemical reactor, Chin. Pat. ZL95111258.9 (1995).21] W.G. Shao, Q. Wang, F. Wang, Y.H. Chen, The cutting of multi-walled carbon

nanotubes and their strong interfacial interaction with polyamide 6 in the solidstate, Carbon 44 (2006) 27082714.

[

[

cta 545 (2012) 148 156 155

22] C.H. Lu, Q. Wang, Preparation of ultrafine polypropylene/iron composite pow-ders through pan-milling, J. Mater. Process. Technol. 145 (2004) 336344.

23] W. Zhang, M. Liang, C.H. Lu, Morphological and structural development of hard-wood cellulose during mechanochemical pretreatment in solid state throughpan-milling, Cellulose 14 (2007) 447456.

24] W. Zhang, X.X. Zhang, M. Liang, C.H. Lu, Mechanochemical preparation ofsurface-acetylated cellulose powder to enhance mechanical properties ofcellulose-filler-reinforced NR vulcanizates, Compos. Sci. Technol. 68 (2008)24792484.

25] M. Liang, C.H. Lu, Y.G. Huang, C.S. Zhang, Morphological and structural devel-opment of poly(ether ether ketone) during mechanical pulverization, J. Appl.Polym. Sci. 106 (2007) 38953902.

26] M.L. Di Lorenzo, C. Silvestre, Non-isothermal crystallization of polymers, Prog.Polym. Sci. 24 (1999) 917950.

27] M.T. Run, Y.J. Wang, C.G. Yao, J.G. Gao, Non-isothermal crystallization kinet-ics of poly(trimethylene terephthalate)/poly(ethylene 2,6-naphthalate) blends,Thermochim Acta 427 (2005) 129135.

28] K. Nakamura, T. Watanabe, K. Katayama, T. Amano, Some aspects of non-isothermal crystallization of polymers. I. Relationship between crystallizationtemperature, crystallinity, and cooling conditions, J. Appl. Polym. Sci. 16 (1972)10771091.

29] K. Nakamura, K. Katayama, T. Amano, Some aspects of nonisothermal crystal-lization of polymers. II. Consideration of the isokinetic condition, J. Appl. Polym.Sci. 17 (1973) 10311041.

30] L. Minkova, M. Mihailov, Kinetics of nonisothermal crystallization and meltingof normal high density and ultra-high molecular weight polyethylene blends,Colloid. Polym. Sci. 267 (1989) 577582.

31] L. Mandelkern, M. Glotin, R.A. Benson, Supermolecular structure and ther-modynamic properties of linear and branched polyethylenes under rapidcrystallization conditions, Macromolecules 14 (1981) 2234.

32] M. Eder, A. Wlochowicz, Kinetics of non-isothermal crystallization of polyethy-lene and polypropylene, Polymer 24 (1983) 15931595.

33] A.K. Gupta, S.K. Rana, B.L. Deopura, Crystallization kinetics of high-densitypolyethylene/linear low-density polyethylene blend, J. Appl. Polym. Sci. 51(1994) 231239.

34] P. Supaphol, J.E. Spruiell, Nonisothermal bulk crystallization studies of highdensity polyethylene using light depolarizing microscopy, J. Polym. Sci. Part B36 (1998) 681692.

35] M.A. Islam, I.A. Hussein, M. Atiqullah, Effects of branching characteristics andcopolymer composition distribution on non-isothermal crystallization kineticsof metallocene LLDPEs, Eur. Polym. J. 43 (2007) 599610.

36] I.A. Hussein, Nonisothermal crystallization kinetics of linear metallocenepolyethylenes, J. Appl. Polym. Sci. 107 (2008) 28022809.

37] C.M. Jiao, Z.Z. Wang, X.M. Liang, Y. Hu, Non-isothermal crystallization kineticsof silane crosslinked polyethylene, Polym. Test. 24 (2005) 7180.

38] L.X. Tang, M.Q. Yan, C.H. Liu, P.H. Wang, B.J. Qu, Effect of UV-photocrosslinkingon the nonisothermal crystallization kinetics of polyethylene, J. Appl. Polym.Sci. 108 (2008) 174180.

39] B.W. Hutzler, L.D.B. Machado, A.-L.C.H. Villavicencio, A.B. Lugao, Crystal-lization of irradiated polyethylene, Radiat. Phys. Chem. 57 (2000) 431434.

40] X.P. Xia, C.S. Xie, S.Z. Cai, Non-isothermal crystallization behavior of low-density polyethylene/copper nanocomposites, Thermochim. Acta 427 (2005)129135.

41] J.N. Hay, M. Sabir, Crystallization kinetics of high polymers. PolyethyleneoxidePart II, Polymer 10 (1969) 203211.

42] J.N. Hay, P.A. Fitzgerald, M. Wiles, Use of differential scanning calorimetry tostudy polymer crystallization kinetics, Polymer 17 (1976) 10151018.

43] P. Cebe, S.D. Hong, Crystallization behaviour of poly(ether-ether-ketone), Poly-mer 27 (1986) 11831192.

44] M. Avrami, Kinetics of phase change. I. General theory, J. Chem. Phys. 7 (1939)11031112.

45] M. Avrami, Kinetics of phase change. II. Transformationtime relations for ran-dom distribution of nuclei, J. Chem. Phys. 8 (1940) 212224.

46] A. Jeziorny, Parameters characterizing the kinetics of the non-isothermal crys-tallization of poly(ethylene terephthalate) determined by d.s.c, Polymer 19(1978) 11421144.

47] Z.Y. Zhang, C.F. Xiao, Z.Z. Dong, Comparison of the Ozawa and modified Avramimodels of polymer crystallization under nonisothermal conditions using acomputer simulation method, Thermochim. Acta 466 (2007) 2228.

48] J. Jin, S.J. Chen, J. Zhang, Non-isothermal crystallization kinetics of partiallymiscible ethylenevinyl acetate copolymer/low density polyethylene blends,Epress. Polym. Lett. 4 (2010) 141152.

49] E. Piorkowska, A. Galeski, J.M. Haudin, Critical assessment of overall crystal-lization kinetics theories and predictions, Prog. Polym. Sci. 31 (2006) 549575.

50] J.N. Hay, Z.J. Przekop, On a nonexponential transformation equation forspherulitic crystallization, J. Polym. Sci. Part B 16 (1978) 8189.

51] J.N. Hay, P.J. Mills, The use of differential scanning calorimetry to study polymercrystallization kinetics, Polymer 23 (1982) 13801384.

52] J.L. Chen, Q. Sun, Y.K. Zou, G. Xue, DSC studies on the melting crystallizationof polyethylenes prepared from alkanes of varying molecular size, Polymer 43

(2002) 68876891.

53] A. Hammami, J.E. Spruiell, A.K. Mehrotra, Quiescent nonisothermal crystalliza-tion kinetics of isotactic polypropylenes, Polym. Eng. Sci. 35 (1995) 797804.

54] L.C. Lopez, G.L. Wilkes, Non-isothermal crystallization kinetics of poly(p-phenylene sulphide), Polymer 30 (1989) 882887.

1 ica A

[

[

[

[

[

[

[

[

[

[

[

[66] J.W. Huang, Y.L. Wen, C.C. Kang, W.J. Tseng, M.Y. Yeh, Nonisothermal crystalliza-tion of high density polyethylene and nanoscale calcium carbonate composites,56 H. Wu et al. / Thermochim

55] P. Sajkiewicz, L. Carpaneto, A. Wasiak, Application of the Ozawa model to non-isothermal crystallization of poly (ethylene terephthalate), Polymer 42 (2001)53657530.

56] P. Supaphol, N. Apiwanthanakorn, Non-isothermal crystallization kinetics ofpoly(p-phenylene sulphide), J. Polym. Sci. Part B 42 (2004) 41514163.

57] G. Antoniadis, K.M. Paraskevopoulos, D. Bikiaris, K. Chrissafis, Non-isothermalcrystallization kinetic of poly(ethylene terephthalate)/fumed silica (PET/SiO2)prepared by in situ polymerization, Thermochim. Acta 510 (2010) 103112.

58] J.S. Qian, P.S. He, Non-isothermal crystallization of HDPE/nano-SiO2 composite,J. Mater. Sci. 38 (2003) 22992304.

59] P. Zou, S.W. Tang, Z.Z. Fu, H.G. Xiong, Isothermal and non-isothermal crys-tallization kinetics of modified rape straw flour/high-density polyethylenecomposites, Int. J. Therm. Sci. 48 (2009) 837846.60] T. Ozawa, Kinetics of non-isothermal crystallization, Polymer 12 (1971)150158.

61] P. Supaphol, Nonisothermal bulk crystallization and subsequent meltingbehavior of syndiotactic polypropylenes: crystallization from the melt state,J. Appl. Polym. Sci. 78 (2000) 338354.

[

cta 545 (2012) 148 156

62] X.L. Chen, L.M. Wang, Y.L. Liu, J.G. Shi, H. Shi, Nonisothermal crystallizationkinetics of high-density polyethylene/barium sulfate nanocomposites, Polym.Eng. Sci. 49 (2009) 23432349.

63] T. Liu, Z. Mo, S. Wang, Nonisothermal melt and cold crystallization kinetics ofpoly(aryl ether ether ketone ketone), Polym. Eng. Sci. 37 (1997) 568571.

64] M.Y. Liu, Q.X. Zhao, Y.D. Wang, C.G. Zhang, Z.S. Mo, S.K. Cao, Crystallizationkinetics and morphology of nylon 1212, Polymer 44 (2003) 25372545.

65] Y. Tao, Y. Pan, Z. Zhang, K. Mai, Non-isothermal crystallization, meltingbehavior and polymorphism of polypropylene in [beta]-nucleated polypropy-lene/recycled poly(ethylene terephthalate) blends, Eur. Polym. J. 44 (2008)11651174.Polym. Eng. Sci. 48 (2008) 12681278.67] Y. Liu, G.S. Yang, Non-isothermal crystallization kinetics of polyamide-

6/graphite oxide nanocomposites, Thermochim. Acta 500 (2010) 1320.