J. Cent. South Univ. Technol. (2008) 15(s1): 067-071

DOI: 10.1007/s11771-008-316-0

Effect of substrates on crystallization of high density polyethylene

FAN Yu-run(范毓润), LIN Yuan(林 渊), RUAN Mian-zhao(阮绵照)

(State Key Laboratory of Fluid Power Control and Transmission, Zhejiang University, Hangzhou 310027, China)

Abstract: The experimental observations about remarkable influence of the substrates on the isothermal crystallization rate of a high density polyethylene(HDPE) were presented. Two methods were used to characterize the crystallization rate: the change of turbidity of the HDPE specimen and the changes of the complex viscosity and storage modulus measured by a rotational rheometer, which gave consistent results showing that the isothermal crystallization rate decreased in sequence as the specimen contacted with aluminum, brass and stainless steel plates, respectively. As to the dominant influence factor, the chemical composition of the substrates can be excluded via insulating the plate by an aluminum foil. Instead, we propose the plate’s ability of removing the latent heat of crystallization from the specimen. Rheological measurement is sensitive to the crystallization process. The colloid like model proposed by BOUTAHAR et al for the crystallization of HDPE gives reasonable predictions of the crystallized fraction from the measured storage modulus.

Key words: high-density polyethylene; isothermal crystallization; heat conduction

1 Introduction

It is well known that macroscopic deformation has great impact on both the morphology and rate of crystallization of polymers. Nowadays, shear-induced crystallization is an active research field in polymer science and polymer processing^[1-9]. It is a common practice that in the experiments on shear-induced crystallization the melt-crystallized sample is in contact with metal walls with considerably large surface-to-volume ratio. For instance, characteristic time scales such as the physical gelation time in terms of the dynamic moduli^[1-2] or tensile normal force^[3], the induction times in terms of the level-upturn transition of the shear stress^[4-5] or compressive normal force^[6] can be measured on rotational rheometers by using the parallel plates or cone-and-plate pair with typical diameter of 15-40 mm and the sample thickness below 2 mm. However, little attention has been paid to the effect of substrate on the crystallization of semicrystalline polymers and the underlying mechanism. In this study, we report remarkable effect of several metal substrates on the isothermal crystallization rate of a high-density polyethylene(HDPE). We characterize the crystallization rate by two methods: the change of turbidity of the melt-crystallized specimen and the change of dynamic viscosity or storage modulus of the specimen in small-amplitude oscillatory shear.

2 Experimental

2.1 Materials

A commercial isotactic high density polyethylene (HDPE5306J, Yangtze Petroleum) was used in this study. The HDPE specimen has the following characteristics: relative molecular mass M_n=1.5×10⁴ and M_w=9.4×10⁵, thus the polydispersity is 6.1. An equilibrium melting temperature T_m of about 160 ℃ was obtained by using the conventional DSC measurement. Pellets of the polyethylene were molded into disks using a laboratory press at 160 ℃ and under 10 MPa for 10 min.

Three metal substrates in the form of round plates were used in the experiments of isothermal crystallization of HDPE: aluminum, brass and stainless steel. In the turbidity measurements, the plates of 40 mm in diameter and 5 mm in thickness were used. In the rheological measurements, the plates of 25 mm in diameter and 4.5 mm in thickness were used. In order to insulate the effect of the surface chemical composition, a soft aluminum foil of 40 μm in thickness was pasted onto these plates and the same experiments of the isothermal crystallization of HDPE as those using the bare plates were carried out.

2.2 Characterizing crystallization rate by turbidity

A temperature-control cell was established in our lab, which is composed of an optically transparent box, a

heating pipe with manually adjustable flow-rate of nitrogen and a TCG-808 temperature control unit based on the heating-voltage control and the feedback of a thermal couple mounted in the substrate plates. In the isothermal crystallization experiments, via appropriately adjusting the flow rate, the maximum temperature fluctuation of ±1 ℃ can be achieved.

The plate chosen was first carefully cleaned using acetone and dried, the HDPE specimen was then put on the plate and set in a vacuum oven at the temperature well above the equilibrium melting point (200 ℃) for at least 24 h. Before the experiment of isothermal crystallization, the temperature cell was set to a target temperature (110 ℃ or 123 ℃), then the plate-specimen pair was moved swiftly from the oven into the temperature cell. Within 2-3 min, the temperature of plate could drop to the target value with an overshoot of 1 or 2 ℃, and then kept constant. At the beginning, the specimen was perfectly transparent; after a certain induction time, white mist appeared, usually spreading rather quickly from the edge region to the center; eventually, the specimen became totally turbid. The change of turbidity of the specimen during the crystallization process was recorded by a CCD video camera and processed by the MATLAB software. Three typical images taken respectively at early, middle and final stage are shown in Fig.1.

2.3 Characterizing crystallization rate by rheological measurements

Rheological experiments were performed on a Gemini-200 rotational rheometer (Bohlin Instruments). Parallel-plates made of the foregoing mentioned brass, stainless steel and aluminum were used. In the small-amplitude oscillatory shear experiment, the gap of 0.75 mm, shear stress of 50 Pa, and angular frequency of 1 rad/s were chosen, and the change of storage modulus and dynamic complex viscosity was used to characterize the crystallization rate.

The HDPE specimen was first heated at 160 ℃ for 20 min between the parallel plates enclosed in the chamber of temperature control of the rheometer. Then the chamber was cooled to the target temperature of 123℃ as soon as possible. The transient period lasted about 3 min, and the temperature fluctuations were less than ±0.2°C in the later isothermal crystallization process as measured by a thermal couple imbedded in the lower plate.

2.4 Measurement of thermo-conductivity of sub- strates

The heat diffusion coefficient α and the heat capacity c_p of the metal plates were measured by the Flash method (LFA 447 NanoFlash, NETZSCH) and by

Fig.1 Three typical images at early (a), middle (b) and final (c) stages during crystallization process of HDPE specimen on Al plate recorded by CCD video camera

a differential scanning calorimeter (DSC 200 F1 Maja, NETZSCH), respectively, at the temperature of 123 ℃. The coefficients of heat conduction λ of the plates were calculated as α=λ/(ρc). The NETZSCH LFA 447 NanoFlash is based on the well-known flash method, in which the front side of a plane sample is heated by a short light pulse and the resulting temperature rise on the rear side is measured using an infrared detector. By analyzing the resulting temperature—time curve, the thermal diffusivity can be determined.

3 Results and discussion

Fig.2 shows the change of grayness of the HDPE specimen during the isothermal crystallization processes at 110 ℃ and 123 ℃, respectively. The curves have a common sigmoid shape, indicating an induction time followed by a rapid increase and eventually a level off of the grayness. Obviously, the characteristic crystallization rate decreased dramatically as the melt contacted with the aluminum, brass and stainless steel plates, respectively. Similar results were obtained by rheological tests of small-amplitude oscillatory shear as the HDPE melt crystallized at 123 ℃ between the aluminum, brass and stainless steel parallel plates with the gap of 0.75 mm and diameter of 25 mm^[10]. The dynamic viscosity η* exhibits the typical level-upturn behavior. At 110 ℃ the crystallization was so fast that a large tensile normal force due to the volume contraction was quickly built up and it overloaded the instrument’s sensor before sufficient data of the viscosity could be collected. Fig.3 shows the curves  of the crystallization processes contacting with the aluminum plates at temperature increment of 0.2 ℃ and using the stress control mode of 50 Pa. The repeatability tests also confirmed 0.2 ℃ resolution of the rheological measurement.

Fig.2 Changes of relative grayness of HDPE specimen during isothermal crystallization on plates made of aluminum, brass and stainless steel: (a) At 110 ℃; (b) At 123 ℃

Thermodynamically, the driving force of polymer crystallization is the free energy difference between the amorphous and crystalline phases^[11-12]. The interfacial morphology and energy of the polymer melt and solid substrate may affect the crystallization rate. In order to insulate the surface effect, a soft aluminum foil of 40 μm

Fig.3 Changes of complex viscosity η* of HDPE melt crystallizing in small amplitude oscillation shearing between parallel plates made of aluminum

in thickness was pasted onto these metal plates and the strain control mode was used in the small-amplitude oscillatory shear with the shear strain of 0.5% and angular frequency of 0.1 rad/s. The strain control mode gave smoother results at the latter stage of crystallization than the stress control mode did. Fig.4 shows the full crystallization curves of  and  on the aluminum, brass and stainless steel plates pasted with the same aluminum foil, respectively. For these pasted plates the big differences of the crystallization rates of HDPE remain and are similar to those with the bare plates^[5]. Furthermore, Fig.4 indicates that the final morphology of crystallized HDPE may be different on these plates. This situation motivates us to consider the substrate’s ability to remove the latent heat of phase transition as a key factor that influences the crystallization rate in apparent ‘isothermal’ conditions. As listed in Table 1, the values of thermal conductivity are in a decreasing sequence for the aluminum, brass and stainless-steel plates.

BOUTAHAR et al^[1] studied the relation between the transformed fraction and the dynamic moduli of the crystallization of a polypropylene and a HDPE. The rheological behavior of HDPE is governed by the immediate existence of a yield stress and appears like a colloid of small particles in a liquid matrix. They proposed the following relationship between the crystallization fraction α(t) and the storage modulus:

                          (1)

where  are the storage moduli at time 0, t and infinity, respectively. This formula is proved to fit the DSC data of their HDPE reasonably well. We transformed the storage modulus in Fig.4 according to this formula. Fig.5 shows the results and the regression by the Avrami equation:

Fig.4 Changes of complex viscosity η* and storage modulus G′ of HDPE melt crystallized at 123 ℃ in small amplitude oscillation shearing between parallel plates made of aluminum, brass and stainless steel pasted with aluminum foil and with gap of 0.75 mm, shear strain of 0.5% and angular frequency of 0.1 rad/s

Table 1 Thermal parameters of aluminum, brass and stainless steel used in experiments at 123 ℃

α(t)=1-exp(-Ktⁿ)                             (2)

where  K is the rate constant, and n the Avrami exponent. The kinetic parameters determined are listed in Table 2. The fitting is rough, especially not good at the early stage of crystallization. However, the values of Avrami exponent are reasonable (2-3). We can estimate the characteristic time τ of crystallization according to α(t)=1-exp[-(t/τ)ⁿ], and the values are listed in Table 2.

Fig.5 Evolutions of crystallization fraction α(t) of HDPE melt crystallized at 123 ℃

Table 2 Avrami exponent, rate constant and characteristic crystallization time

If it is assumed that the final degree of crystallization is the same for the HDPE samples (about 0.84 as estimated by BOUTAHAR et al^[1] contacting with the pasted aluminum, brass and stainless steel plates, respectively, then the reciprocal of the characteristic time represents the characteristic rates of latent heat of crystallization that must be removed, thus the ratio of the corresponding heat fluxes is estimated to be q(Al)?q(Brass)?q(SS)=7.7?4.7?1.0. On the other hand, from Table 1, λ(Al)? λ(Brass)?λ(SS)=9.7?7.6?1.0. The proximity of the two sets of ratio implies that a rough one-dimensional steady heat conduction model may explain the large difference of the crystallization rates on these different plates. We admit that the actual latent heat transfer is unsteady and accurate analysis of the unsteady heat transfer for the crystallization process is very difficult at present stage, but the dominant factor that causes large difference of the crystallization rates on these different metal substrates is identified to be the ability to remove the latent heat from the specimen under the apparent ‘isothermal’ condition.

4 Conclusions

Both the optical and rheological experiments gave consistent results for the isothermal crystallization process of a HDPE, indicating that the crystallization rate decreased considerably in sequence as the melt contacts with the aluminum, brass and stainless-steel plates, respectively. As to the dominant influence factor, the chemical composition of the substrates can be excluded via insulating the plate by an aluminum foil. Instead, we propose the plate’s ability of removing the latent heat of crystallization from the specimen. Rheological measurement is sensitive to the crystallization process. The colloid like model proposed by BOUTAHAR et al for the crystallization of HDPE gives reasonable predictions of the transformed fraction from the measured storage modulus.

[1] BOUTAHAR K, CARROT C, GUILLET J. Crystallization of polyolefins from rheological measurements relation between the transformed fraction and the dynamic moduli [J]. Macromolecules, 1998, 31: 1921-1929.

[2] POGODINA N V, WINTER H H. Polypropylene crystallization as a physical gelation process [J]. Macromolecules, 1998, 31: 8164-8172.

[3] CHEN Q, FAN YR, ZHENG Q. A novel approach to rheological characterization for the gelation in polymer crystallization [J]. Chin J Polym Sci, 2005, 23(4): 423-434.

[4] KOSCHER E, FULCHIRON R. Influence of shear on polypropylene crystallization: Morphology development and kinetics [J]. Polymer, 2002, 43: 6931-6942.

[5] HADINATA C, GABRIEL C, RUELLMAN M, LAUN H M. Comparison of shear-induced crystallization behavior of PB-1 samples with different molecular weight distribution [J]. J Rheol, 2005, 49: 327-349.

[6] CHEN Q, FAN Y R, ZHENG Q. Rheological scaling and modeling of shear-enhanced crystallization rate of polypropylene [J]. Rheol Acta, 2006, 46: 305-316.

[7] VLEESHOUWERS S, MEIJER H E H. A rheological study of shear induced crystallization [J]. Rheol Acta, 1996, 35: 391-399.

[8] COPPOLA S, GRIZZUTI H. Microrheological modelling of flow-induced crystallization [J]. Macromolecules, 2001, 34: 5030-5036.

[9] JANESCHITZ-KRIEGL H, RATAJSKI E. Kinetics of polymer crystallization under processing conditions: Transformation of dormant nuclei by the action of flow [J]. Polymer, 2005, 46: 3856-3870.

[10] LIAO H Y, LIN Y, FAN Y R. Rheological properties of high-density polyethylene in isothermal crystallization [J]. J Chemical Industry and Engineering, 2007, 58: 2500-2505. (in Chinese)

[11] SCHULTZ J. Polymer materials science [M]. Prentice-Hall Inc, 1974.

[12] STROBL G. The physics of polymers (2nd edition) [M]. Berlin: Springer-Verlag, 1997.

(Edited by YANG Bing)

Foundation item: Project(20050335050) supported by the Special Foundation of Education Ministry of China; Project(10472105) supported by the National Natural Science Foundation of China

Received date: 2008-06-25; Accepted date: 2008-08-05

Corresponding author: FAN Yu-run, Professor; Tel: +86-571-87952226; E-mail: yurunfan@zju.edu.cn