Printer Friendly

Effect of molecular weight and molecular weight distribution on weld-line interface in injection-molded polypropylene.

INTRODUCTION

Polypropylene (PP) is a commercially important semi-crystalline polymer and widely used in injection-molded automotive parts such as bumpers and instrumental panels owing to its highly balanced mechanical properties. The injection molding of such large and complex parts requires the use of a multi-gated molding system. A weld-line, formed when two or more polymer melt flows meet in a cavity, is often the weakest part of an injection-molded or extruded product and thus the initiation of crack propagation. As a result, the presence of weld-lines reduces the mechanical strength and affects the surface appearance of the products. In practical cases, the problems are increased by the need for efficiency in terms of short cooling cycles. With growing demand for high-quality plastic parts, improvement of the weld-line behavior has been strongly required. Although many potential reasons for the weld-line problem have been proposed such as poor intermolecular entanglement across the weld-line, molecular orientation induced by fountain flow, and the stress concentration effect of surface V-notch (1-7), a complete understanding has not yet been achieved. This is mainly because in actual cases they are inter-related in such a way that it is usually hard to correlate experimental findings with one of the factors only.

The surface V-notch is considered to be an obvious source of stress concentration that initiates fracture. Tomari et al. (8-10) have studied the effect of V-notch on the weld strength in polystyrene. The surface of the weld-line was partially eliminated by cutting with a milling machine to several levels of cutting depth. It was found that the weld strength increased with increasing the cutting depth. They concluded that the V-notch exhibited little influence on weld strength at least for this system and postulated a layer of poor bonding under the V-notch owing to the frozen-in molecular orientation near the mold.

The material at the flow front is known to experience fountain flow. Molecular chains close to the melt front are stretched parallel to the weld-line, and once the flow fronts meet, this molecular orientation makes the interfacial structure different from the bulk phase. Using polarized optical microscopy, Hobbs (1) found a continuous band of row-nucleated spherulites running completely through an injection-molded specimen at the weld. It was postulated that this interfacial morphology had to be attributed to the molecular orientation owing to the fountain flow. Many optical microscopic studies have been made to observe the micrometer scale structure in those injection moldings, but it is still difficult to determine the molecular orientation quantitatively. There are a few investigations to determine the molecular orientation by using X-ray diffraction and birefringence, but the spatial resolution of these experiments was limited to about 100 [micro]m. Akashige and Usami (11) found that the intensity ratio of Raman bands at 841/809 [cm.sup.-1] was sensitive to the molecular orientation of PP. These Raman bands have been widely used to estimate the micrometer scale distribution of molecular orientation of PP (12), (13).

To make quantitative understanding of the relationship between interfacial structures and weld strength, it is necessary to evaluate the interfacial strength at microscale. Although there have been many studies about the measurement of interfacial strength, it is still difficult to measure it at microscale. In recent years, SAICAS (Surface and Interfacial Cutting Analysis System) has been developed and applied to various researches with its ability of microscale cutting near the surface and interface (14). This method is considered to be effective as a quantitative measurement of the characteristics on microscale. In this study, SAICAS measurements have been carried out to characterize the property of weld-line interface.

In general, controlling a molecular weight and a molecular weight distribution is one of the most important techniques to obtain desirable properties and processability. Then, if the effects of molecular weight and molecular weight distribution on the weld-line formation are clarified, these findings would have a significant effect on the polymer processing technology. The objective of this study is to generate a procedure to characterize the microstructure or weld-lines in injection-molded PP and to identify more accurately the cause of weld-line strength. In this study, the effect of molecular weight and molecular weight distribution on the microstructure of weld-line interface was investigated by using a Laser Raman Spectroscopy and microscale mechanical property measurement. Molecular dynamics simulation was also conducted to understand the structure--property relationship.

EXPERIMENTAL

Sample Preparation

PP materials used in this study were produced by Japan Polypropylene. The molecular characteristics of these materials are summarized in Table 1. These samples have different molecular weights and molecular weight distributions. In Table 1, L and H describe lower and higher molecular weight, respectively. N, B, and M describe narrower, broader, and middle molecular weight distribution, respectively. Dumbbell-shaped test specimens containing a weld-line perpendicular to the tensile direction were injection molded by an injection-molding machine manufactured by Japan Steel Works Molding conditions are listed in Table 2. As a reference, a series of nonwelded specimens was molded under the same processing conditions as the welded specimen. The specimens with a weld-line at the center were obtained by double gate molding and nonwelded specimens were obtained by single gate molding.

TABLE 1. Characteristics of sample polypropylene.

Sample    [M.sup.w] (x [10.          [M.sup.w]     [M.sup.w]/
                   sup.4])    (x [10.sup.4])        [M.sup.n]

LN                    19.0                 6.4            3.0
LM                    19.0                 3.5            5.4
HM                    40.0                 7.4            5.4
HB                    41.0                 4.2            9.8

TABLE 2. Molding conditions.

Cylinder temperature    240[degrees]C
Mold temperature         40[degrees]C
Injection speed              100 mm/s
Holding pressure              2.0 Mpa


Molecular Orientation in the Weld-line

The microscale distribution of the molecular orientation parallel to the flow direction was observed in the cross-section by polarized micro-Raman spectroscopy. The cross-section normal to the transverse direction in the central part of the test specimen was exposed by using a diamond knife. A polarized Raman spectrum obtained from a laser Raman spectrometer, model NRS-3100 manufactured by JASCO, is shown in Fig. 1. The 532-nm line of a green laser was used as the excitation source and the laser beam was focused at a specimen position to approximately 2 [micro]m. The intensity ratio I844/I813 of 844 [cm.sub.-1] band to 813 [cm.sub.-1] band in polarized Raman spectrum was used to evaluate molecular orientation parallel to the flow direction (11). The observation was conducted at the position of weld-line and corresponding position of nonwelded specimen. Although the weld-line interface perpendicular to the flow direction is hardly defined, the molecular orientation in the weld-line interface was indirectly estimated by comparing the intensity ratio for the welded specimen with that for nonwelded one.

[K.sub.Imax] of Weld-line

Stress intensity factor ([K.sub.Imax]) of nonwelded and welded PP was measured with double-edge-notched tensile (DENT) test. Prenotches with a depth of about 1 mm were introduced manually into the weld-line using razor blade. Figure 2 shows the geometry of DENT specimens. The actual depth of the notch a in Eq. 1 was measured. Tensile test was conducted at a cross-head speed of 10 mm/min at room temperature. [K.sub.Imax] was evaluated by:

[K.sub.Imax] = [sigma][square root of ([pi]a)]f(2a/w) (1)

where [sigma] is the peak stress, a the crack length, and w the specimen width. f(2a/w) is given by (15):

f(2a/w) = (1.22 - 0.561x - 0.205[x.sup.2] + 0.477[x.sup.3] - 0.19[x.sup.4]/[square root of (1 - x)]

x = 2a/w (2)

Shear Strength of Weld-line

We have also evaluated the shear strength of weld-lines by the SAICAS produced by Daipla Wintes. Figure 3 shows a schematic diagram of micro-cutting measurement. The cutting blade was initially located on the weld-line and was moved parallel to the weld-line. The location of the cutting blade was carefully examined by a CCD microscope equipped on this system. Vertical force ([F.sub.v]) and horizontal force ([F.sub.H]) at the blade were measured during cutting of the test specimen until the horizontal force decreases abruptly and became constant. The shear strength of weld-line ([[tau].sub.s]) was evaluated by:

[[tau].sub.s] = [F.sub.H]/(2wd cot [phi]) (3)

where w is the blade width, d is a cutting depth from the surface of test specimen, and [phi] is a shear angle defined as [tan.sup.-1]([F.sub.H]/[F.sub.v])/2 (14). The cutting blade with 0.5 mm width was used with a horizontal moving speed of 1.0 [micro]m/s and a vertical moving speed of 0.1 [micro]m/s to keep a stable cutting condition.

Simulation

The microstructure and deformation behavior of weld-line was investigated based on the coarse-grained molecular dynamics simulation of the bead-spring polymer model of Grest and Kremer (16), (17). All of the simulations were conducted by using the molecular dynamics program COGNAC in the OCTA system (http://OCTA.jp) (18). In the coarse-grained models, several atoms and bonds along the backbone are combined into one effective segment. The polymer is treated as a chain that has several beads connected by a non-Hookian spring potential, and all beads interact with each other. The time evolution of the bead position [r.sub.n] was calculated by the Langevin equation (16), (17),

m [d.sup.2][R.sub.n]/d[t.sup.2] = -[partial derivative]U({[r.sub.n]})/[partial derivative][r.sub.n] - [GAMMA]d[r.sub.n]/dt + [W.sub.n](t) (4)

where m is the mass of bead, U is the total potential energy of the system, [GAMMA] is the friction constant, and [W.sub.n](t) is a Gaussian white noise generated by

< [W.sub.n](t)[W.sub.m](t') >= 2[k.sub.B][T.sub.m],[GAMMA][[sigma].sub.nm]I[sigma](t - t') (5)

The potential energy U({[r.sub.n]}) consists of three terms, the bead--bead potential [U.sup.bead--bead], the bond potential [U.sup.BOND], and the bead--wall potential. The bead--bead potential is given by

[U.sup.bead--bead] ([r.sub.n]) = [summation] [U.sup.LJ]([r.sub.mm])(6)

where [r.sub.mm] is the distance between the bead m and n, and [U.sup.LJ](r) is the Lennard--Jones potential given by

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (7)

where [sigma] is the diameter of LJ sphere and [epsilon] is the interaction energy. The bond potential is given by

[U.sup.BOND]({[r.sub.n]}) = [SIGMA] [U].sup.B] ([r.sup.B]i ) (8)

where [r.sup.B]i is the length of the ith bond that is connected to the bead n, and [U.sup.B](r) is given by the FENE--Lennard--Jones potential:

[U].sup.B](r) = [U].sup.FENE](r) + [U.sup.LJ](r) (9)

with

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (10)

where k is the spring constant, and [R.sub.0] is the maximum extension of the spring. These parameters are taken as k = 30.0[epsilon]/[[sigma].sup.2], [R.sub.0] = 1.5[sigma]. The process of polymer welding was treated as the diffusion of chains across the interface formed when two hard planer surfaces were brought together (19). The hard planer surfaces were produced using a simulation box which had periodic boundary condition applied to the x- and y-directions under the atomic wall boundary condition applied to the z-direction (20). The wall was represented by a set of bead placed on the square lattice on its surface, and the bead--wall potential was obtained as the summation of the interactions between the chain bead and the wall bead (http://OCTA.jp). The polymer melt was confined between the walls in the z-direction and was deformed by moving both walls with a fixed velocity in opposite directions. After merging a symmetric image of the deformed polymer on the original one in the z-direction, the weld interface parallel to the xy-plane was produced by a relaxation process. A schematic diagram of the modeled interface is shown in Fig. 4. The chains on the polymer surfaces were allowed to diffuse across the merged interface to form a weld. The interdiffusion dynamics was then simulated by solving an equation of motion using the velocity Verlet algorithm with time step At = 0.01T under the constant density [rho] = 0.85m/[[sigma].sip.3]condition (16), (17). The calculation was carried out in the reduced unit, in which [sigma], [epsilon], and m were taken as the units of length, energy, and mass, respectively. The unit of time was given by [tau] = [[sigma](m/[epsilon]).sup.1/2. The temperature [k.sub.B]T was set to 1.0[epsilon], and the friction constant [GAMMA] was set to 0.5 [[tau].sup.-1]for all cases. To make the initial state, chains were placed in unit cell using a simple random walk and the Metropolis Monte Carlo steps of several million times were carried out to get an energetically relaxed state, and then conducted the equilibrium molecular dynamics simulation for 30,000[tau] (18). To get highly stretched polymer chain as would be found in the weld line, the initial structure for the inter diffusion was prepared by shear deformation of the equilibrium polymer with very high shear strain rate 0.01 [[tau].sup.-1] under deformation time 5000[tau]. A binary blend with different chain length was used as the simplest polydisperse model to compare the monodisperse model. The width of the molecular weight distribution [M.sub.w]/[M.sub.n] of this binary blend system was set to 8.0 as follows.

[M.sub.w]/[M.sub.n] = ([N.sub.L][phi] + [N.sub.s] (1 - [phi]))([psi]/[N.sub.L] + (1 - [psi]) / [N.sub.s]) (11)

where [N.sub.L], [N.sub.s] and [phi] are the length of longer chain, that of shorter chain, arid a volume fraction of longer chain, respectively. In this simulation, [N.sub.L], [N.sub.s], and [psi] were 300, 10, and 0.5, respectively. The chain length of monodisperse model was set to the average value of the polydisperse model as 155. The interdiffusion was calculated until the interfacial thickness of monodisperse model was equivalent to the gelation radius of the chain whose length was the number of segments between adjacent entanglement points in the present MD simulation (16), (17). A weld-line structure created with the above procedure was then subjected to an external deformation. The stress--strain behavior was simulated by relaxation process under the external deformation, where the simulation box and the chains were deformed affinely in the perpendicular direction to the interface at every fixed time steps while keeping the size of the simulation box in the parallel direction to the interface constant.

RESULTS AND DISCUSSION

Molecular Orientation in Weld-line

The molecular orientation along the flow direction was characterized by the Raman intensity ratio. Figure 5 shows the depth profile of the Raman intensity ratio near the surface of both nonwelded and welded specimen. The intensity ratio was higher near the surface and decreased toward the interior for all nonwelded specimens. The intensity ratio near the surface increased with increasing molecular weight and reduced more significantly owing to the weld-line. Molecular chains close to the melt front are stretched normal to the flow direction owing to the fountain flow, and these stretched molecules result in an anisotropic molecular orientation in the weld-line. The relaxation of the molecular orientation is less near the surface because the molten polymer solidifies faster (21), (22). Therefore, these observations suggest that higher molecular weight PP result in anisotropy of the molecular orientation in the weld-line owing to its longer relaxation time.

[K.sub.Imax] of Weld-line

Figure 6 shows the relationship between [K.sub.Imax] for the test specimen with and without weld-line. For the non-welded specimens, [K.sub.Imax] increased with increasing molecular weight with the same width of the molecular weight distribution. [K.sub.Imax] increased with decreasing width of the molecular weight distribution for both molecular weights. The same effects of molecular weight and the width of the molecular weight distribution on [K.sub.Imax] were revealed for the welded specimens. However, the reduction of [K.sub.Imax] caused by the weld-line was different within these samples. Figure 5 shows the effect of molecular weight and molecular weight distribution on the reduction of [K.sub.Imax] owing to the weld-line. It was observed that [K.sub.Imax] ratio was significantly reduced with higher molecular weight and broader molecular weight distribution. Several researchers (8-10) pointed out that the weld-line strengths were significantly affected by molecular orientation in the weld-line interface. Comparing with the molecular orientation observed from the laser Raman spectroscopy, it was considered that the effect of molecular weight on the [K.sub.Imax] ratio was related to the molecular orientation in the weld-line interface. However, the molecular orientation was not a dominant cause of the effect of the width of molecular weight distribution on the [K.sub.Imax] ratio. Figure 7 shows SEM photographs of the fracture surface from DENT specimen. The upper figures show the surface of fracture and lower figures show their magnified area around the surface V-notch. In the upper figures, the pre-notch is shown at both ends and the surface V-notch of welded specimens is located at both sides. It was found that the fracture area of all samples was rough on the face inside. In the lower figures, it was found that the area just below the V-notch was smooth for higher molecular weight and broader molecular weight distribution PP. This observation suggests that the reduction of [K.sub.Imax] ratio of the injection-molded PP with higher molecular weight and broader molecular weight distribution was closely related to the fracture properties in the microscale area just below the surface V-notch in the weld-line interface.

Shear Strength of Weld-line

We evaluated the shear strength by micro-cutting measurement for both welded and nonwelded specimens. The shear strength profiles near the surface of the specimen are shown in Fig. 8. The steep increase close to the surface observed for all specimens shows an initial elastic resistant when the cutting blade is pressed into the material. This shear strength soon decreased and reached to steady state at almost 10 [micro]m depth. It was found that the shear strength kept increasing in the case of welded specimen of HM and HW samples, whereas the others reached constant value after the depth of 10 [micro]m. As molecular weight of HM and HW samples is larger than the others, the interdiffusion in this region would proceed more slowly for both HM and HW samples. Figure 9 shows the relationship between the average shear strength in the region from 10 to 15 [micro]m depth for both the specimen with and without weld-line. For the nonwelded specimen, the average shear strength increased with increasing molecular weight, and decreased with broadening of the molecular weight distribution. It was observed that the welded specimens have same effects of molecular weight and molecular weight distribution on the average shear strength as observed for the nonwelded specimens. However, the reduction of the average shear strength caused by the weld-line was different within these samples. Figure 9 shows the effect of molecular weight and molecular weight distribution on the reduction of the shear strength owing to the weld-line as the shear strength ratio. It was observed that the shear strength ratio decreased significantly with higher molecular weight and broader molecular weight distribution. According to these results, it can be said that the reduction of the [K.sub.Imax] ratio is closely related to the reduction of the shear strength ratio measured in the area below the surface V-notch.

Polydispersity Effect on Weld-line Structure

To investigate the effect of polydispersity on the interfacial structure, the spatial distribution of the simulated segment density was evaluated. The density profile was calculated by counting the number of segments in a sliced region of unit cell. Figure 10 shows the segment density profile of each side of chain around the weld-line interface as observed in the molecular dynamics simulation for the polydisperse polymer and monodisperse one. It was found that the density profile of the polydisperse polymer was almost the same as that of monodisperse polymer. Figure 10 also shows the total density profile of the polydisperse polymer. It was found that the depletion of longer chains took place at the interface, whereas the shorter chains are preferentially located at the interface. The origin of this result can be understood from the entropy point of view. As loss of conformational entropy owing to the chain orientation at the interface would be larger for the longer chains than the shorter chains, the longer chains are expected to diffuse away from the interfacial region.

Polydispersity Effect on the Deformation Behavior of Weld-line

The influence of polydispersity on the stress--strain curves of the interfacial region and the snapshot pictures of the interfacial region after elongation are shown in Fig. 11. The stress--strain curves for both systems appeared the yielding behavior. Corresponding to the stress--strain behavior, it was clearly observed from the snapshot picture that the yielding point was dominated by interfacial failure. It was found that both the yield stress and the yield strain for the polydisperse system become lower than those for the monodisperse system. As the number of segments between adjacent entanglement points along the chain is about 35 in the present MD simulation (16), (17), the shorter chains are expected to contribute virtually nothing to the interfacial strength. It is well known that chain length should be sufficiently long to show entanglement after a diffusion process takes place at an interface (3), (4). Therefore, the reduction of the yielding stress in the weld interface owing to the polydispersity can be explained by the segregation of the shorter chain at the interface. Based on the above discussion, it was suggested that the reduction of the mechanical property caused by weld-line in broader molecular weight distribution PP was considered to be responsible for the lack of entanglement owing to the chain segregation of low-molecular-weight component as would be formed in the weld-line interface. Therefore, it was postulated that decreasing the width of the molecular weight distribution was one of the desirable molecular designs to minimize the reduction of mechanical properties caused by weld-lines in case that the average molecular weight was kept constant.

CONCLUSIONS

In this study, microstructure of weld-lines in injection-molded PP with different molecular weights and molecular weight distributions was characterized by using polarized laser Raman spectroscopy, micro-cutting analysis, and DENT method. The mechanical property reduction induced by weld-line was evaluated as a [K.sub.Imax] ratio of the weld-line infested specimen to one without weld-line. It was revealed that the [K.sub.Imax] ratio decreased with increasing molecular weight or broadening molecular weight distribution. Using polarized laser Raman spectroscopy, it was observed that the effect of molecular weight on the ratio was related to the molecular orientation in the weld-line. The molecular orientation was not a dominant cause of the effect of the width of molecular weight distribution on the [K.sub.Imax] ratio. The micro-cutting analysis method revealed that the reduction of the [K.sub.Imax] ratio was closely related to the reduction of the shear strength ratio measured in the area below the surface V-notch. Based on the coarse-grained molecular dynamics simulation, it was suggested that the reduction of the mechanical property caused by weld-line in broader molecular weight distribution PP was considered to be responsible for the lack of entanglement owing to the chain segregation of low-molecular-weight component as would be formed in the weld-line interface. It was postulated that decreasing the width of the molecular weight distribution was one of the desirable molecular designs to minimize the reduction of mechanical properties caused by weld-lines in case that the average molecular weight was kept constant.

Correspondence to: Katsuyuki Yokomizo; e-mail: Yokomizo.Katsuyuki@mp.pochem.co.jp

DOI 10.1002/pen.23487

Published online in Wiley Online Library (wileyonlinelibrary.com). [c] 2013 Society of Plastics Engineers

REFERENCES

(1.) S.Y. Hobbs, Polym. Eng. Sc., 14, 621 (1974).

(2.) S.C. Malruarnera, A.I. Manisali, and D.C. Riggs, Polym. Eng. Sci., 21, 1149 (1981).

(3.) R.P. Wool and K.M. O'Connor, J. Appl. Phys., 52, 5953 (1981).

(4.) Y.H. Kim and R.P. Wool, Macromolecules, 16, 1115 (1983).

(5.) S. Gook and N.P. Sub, Polym. Eng, Sci., 26, 1200 (1986).

(6.) N. Mekhilif, A. Ait-Kadi, and A. Ajji, Polymer, 36, 2033 (1989).

(7.) B. Fisa and M. Rahmani, Polym. Eng. Sci., 31, 1330 (1991).

(8.) K. Tomari, S. Tonogai, and H. Hamada, Polym. Eng. Sci., 30, 931 (1990).

(9.) K. Yamada, K. Tomari, U.S. Ishiaku, and H. Hamada, Polym. Eng. Sci., 45, 1059 (2005).

(10.) K. Yamada, U.S. Ishiaku, and H. Hamada, Polym. Eng. Sci., 45, 1067 (2005).

(11.) E. Akashige and T. Usami, Anal. Sci., 7, 1655 (1991).

(12.) R.P. Paradkar, S.S. Sakhalkar, X. He, and M.S. Ellison, Appl. Spectrosc., 55, 534 (2005).

(13.) C. Minogianni, K.G. Gatos, and C. Galiotis, Appl. Spectrosc., 59, 1141 (2005).

(14.) Y. Kijima and I. Nishiyama, J. Adhes. Soc. Jpn., 41, 6 (2005).

(15.) T.W. Orange, ASTM STP-560, 122 (1974).

(16.) G.S. Grest and K. Kremer, Phys. Rev. A, 33, 3628 (1986).

(17.) K.K. Kremer and G.S. Grest, J. Chem. Phys., 92, 5057 (1990).

(18.) T. Aoyagi and M. Doi, Computational and Theoretical Polymer Science, 10, 317 (2000).

(19.) K. R. Haire and A. H. Windle, Computational and Theoretical Polymer Science, 10, 227 (2001).

(20.) K. Yokomizo, Y. Banno, and M. Kotaki, Polymer, 53, 4280 (2012).

(21.) M. Fujiyama, T. Wakino, and Y. Kawasaki, J. Appl. Polym. Sci., 35, 29 (1988).

(22.) M. Fujiyama and T. Wakino, J. Appl. Polym. Sci., 43, 57 (1991).

Katsuyuki Yokomizo, (1) Yoshihiro Banno, (2) Taketo Yoshikawa, (3) Masaya Kotaki (3)

(1) Japan Polychem Corp., Research and Development Division, Yokkaichi, Mie 510-8530, Japan

(2) Japan Polypropylene Corp., Products Technical Center One, Yokkaichi, Mie 510-8530, Japan

(3) Kyoto Institute of Technology, Matsugasaki, Sakyo-ku, Kyoto 606-8585, Japan
COPYRIGHT 2013 Society of Plastics Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2013 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Yokomizo, Katsuyuki; Banno, Yoshihiro; Yoshikawa, Taketo; Kotaki, Masaya
Publication:Polymer Engineering and Science
Article Type:Report
Geographic Code:9JAPA
Date:Nov 1, 2013
Words:4357
Previous Article:Oxygen-enriching properties of a new hybrid membrane cross-linked from PDMS and fluorine-containing silicone resin.
Next Article:Semi-interpenetrated network with improved sensitivity based on poly(N-isopropylacrylamide) and poly(aspartic acid).
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters