crystallinity and dielectric properties of peek, poly(ether ether ketone) - 1994

IEEE Transactions on Dielectrics and Electrical Insulation Vol. 1 No. 6, December 2994 QQ1

Crystallinity and Dielectric Properties of PEEK,

Poly(ether ether ketone)

T. W. Giants Polymers and Composites Section, Mechanics and

Materials Technology Center, The Aerospace Corporation, Los Angeles, CA

ABSTRACT The dependence of specific dielectric properties of polymers on their fundamental molecular parameters was investigated as part of a study involving HV insulation in spacecraft systems. As part of an effort to understand the causes of dielectric breakdown, this investigation focused on the effect of polymer morphology on the electrical characteristics of a poly (ether ether ketone) (PEEK), a high temperature insulating thermoplastic. Amorphous PEEK was thermally annealed to produce a range of crystallinities, and its dielectric properties were measured. As the crystalline/amorphous ratio of the polymer increased, the dielectric breakdown strength was reduced. The electrical conductivity of the material was also shown to decrease with increasing crystalhity. The effect on the permittivity was inconclusive.

1. INTRODUCTION S part of an ongoing effort in evaluating HV insu- A lation materials, attention was directed toward at-

tempting to understand the fundamental causes of dielectric breakdown in such materials [l]. This work was based on the assumption that charge trapping is a key parameter in dielectrics and that structural disorder is the basis for trap formation [2,3]. This disorder can in- clude crosslink sites, free-volume holes, impurities, and disorder at the crystalline/amorphous interface. Phase- separated polymer microstructures are hypothesized to force breakdown paths to assume more tortuous routes around traprich two-dimensional phase boundaries, compared to polymers in which there are only point defects (free-volume holes and crosslinks). Phase boundary net- works help prevent an uninterrupted electron path par- allel to an applied field, thereby reducing chances for impact ionization. Strong trapping is also expected to limit overall current, i.e., dc conductivity, as well as to impact the ultimate dielectric breakdown properties.

Our approach was to assess the effect of polymer mor-

phology on the dielectric properties of insulating materials, with a focus on elucidating the interrelationship between the crystalline/amorphous nature of the material and its effect on dielectric breakdown strength. To implement this approach, a polymer system was selected where thermal aging resulted only in a change in crystalline/amorphous morphology.

r

L J n

Figure 1. Repeat unit of a poly (ether ether ketone).

Significant electrical property data have been obtained from studies involving polyethylene, where controlled annealing conditions change only the crystalline/amorphous nature of the polymer and allow an unambiguous evaluation of the relationship between morphology and electrical characteristics [4,5]. For many applications, however,

1070-9878/94/ $3.00 @ 1994 IEEE

QQ2 Giants: Crystallinity and Dielectric Properties of PEEK

property requirements other than electrical characteristics often drive the choice of material, and conditions often exceed the thermal and mechanical properties where polyalkylenes are useful [6]. Many of the high temperature polymers desired for their insulating characteristics do not lend themselves well to an evaluation of this kind [7]. However, one high temperature thermoplastic, P E E P , a poly (ether ether ketone) shown in Figure 1, can be thermally annealed to produce varying degrees of crystallinity. Moreover, annealing can be accomplished without any attendant chemical changes that can com- plicate the interpretation of the results.

PEEK is a high molecular weight thermoplastic with good thermal stability whose chemical structure is essentially unchanged when exposed to thermal cycling or annealing. The only change exhibited by this material when the amorphous form is subjected to annealing is in its physical morphology [8]. Thus, the study of dielectric properties of this material is not complicated by other parameters, such as chemical degradation or incomplete cure. The crystalline/amorphous ratio in PEEK was varied by selective annealing of the film in an inert atmosphere. The resultant samples were characterized by thermal analysis, density gradient, and X-ray crystallographic techniques. The electrical conductivity, permittivity, and dielectric breakdown strength of the annealed samples were measured, and the relationship between the morphological nature of PEEK and its effect on the electrical properties was examined. The experimental details of this work will be described, and the property relation- ships will be compared.

2. EXPERIMENTAL The PEEK material used in this work was supplied

in an unoriented amorphous film form with a nominal thickness of - 25 pm. According to the manufacturer, the PEEK film is considered to be < 1% crystalline in the amorphous form. Crystallinity in this material can be induced either by thermal annealing above its glass transition temperature or by cooling from the melt. In order to avoid batch-to-batch property variations, espe- cially in concentration of ionic species due to processing, comparable measurements were made on separate single sheets of annealed material.

Differential scanning calorimetry (DSC) was used as a guide in assessing the annealing conditions that would produce the widest range of crystallinities. DSC curves were obtained with a Mettler TA 3000 differential scanning calorimeter.

The amorphous, as-received PEEK was subjected to thermal annealing under an atmosphere of argon in a

convection oven using a variety of time/temperature parameters. Sample strips of PEEK film taken from a single sheet were placed between clean glass plates to maintain planarity, and the temperature was monitored with ther- mocouples placed in close proximity to the film. Sam- ples were annealed in a manner designed to produce the widest variation in density without affecting the integrity of the material. The material undergoes rapid crystallization above its glass transition temperature, and conditions must be controlled carefully in order that suitable samples be obtained. On completion of the annealing process, the samples were cooled to room temperature at a constant rate and stored under ambient conditions.

The most precise determination of the degree of crystallinity was made by measuring the density of the samples and relating the values obtained to the measured density of amorphous PEEK and the calculated density for fully crystalline PEEK. A Techne density gradient column that meets the temperature and dimensional requirements of ASTM D1505-60T was used to determine the density of the annealed PEEK samples prepared for this study. An aqueous sodium bromide solution was selected to encompass the range of densities expected for the amorphous/crystalline range of the material.

X-ray diffraction (XRD) studies were also performed to gain additional information on polymer morphology and to obtain an independent estimate of the crystalline/amorphous ratio. Unfortunately, the XRD method requires the measurement of completely amorphous and crystalline samples. Since the 100% crystalline form of PEEK cannot be obtained because of slow crystallization kinetics above 30%, a peak intensity approximation technique was used. 28 - 8 scans were recorded with copper radiation using a computer controlled Philips Electron- ics Instruments APD 3720 vertical powder diffractome- ter equipped with a theta compensating slit and a scin- tillation detector. Samples were cut to 5 x 5 mma and mounted to a zero background plate that consisted of single crystal quartz cut 6' off the c-axis. This plate had no XRD reflections in 8 - 28 scans and produced very little background scatter. Scans were recorded from 2 to 60', with a step size of 0.02' 28 and a count time of 1 s. In order to measure the crystallinity of the PEEK samples, the XRD pattern of a completely amorphous PEEK specimen was superimposed on the pattern of a semicrystalline specimen. An intensity scaling factor was applied to the amorphous pattern so that it best approx- imated the intensity of the amorphous component in the semicrystalline sample. This scaling was performed so that the discrimination of the amorphous and crystalline contributions to the XRD patterns would be as repro- ducible as possible. On a hardcopy plot, a baseline was then drawn underneath the amorphous pattern, and the

IEEE Transactions on Dielectrics and Electrical Insulation Vol. 1 No. 6, December 1994 993

amorphous and crystalline contributions were then cut out and weighed on an analytical balance in order to measure the percent crystallinity. For the measurement of the full width at half maximum (FWHM) of the crystalline reflections, the pattern of an amorphous pattern was similarly fitted to the amorphous component of a mixture and digitally subtracted to produce the XRD pattern of the crystalline component. The results from the density gradient method were compared with those from the XRD method.

The electrical conductivity of a sheet material is calculated using the expression U = ( L / A ) ( I / V ) , where L is the thickness of the sheet, A is the cross-sectional area conducting current I, and V is the voltage applied across the sheet. Therefore, the conductivity of PEEK was measured by monitoring I as a function of V . These current vs. voltage studies were performed on 7.5x7.5 cm2 samples of PEEK having a nominal thickness of 25 pm. Each sample was prepared for test by sputter depositing a gold film on the bottom surface and a similar film in the form of a 5 cm diameter spot enclosed by a guard ring on the top surface. The sample was inserted in the test fixture, with the bottom surface in contact with the flat surface electrode at HV and the circular spot and guard ring in contact with the spherical electrodes a t ground potential. The entire assembly was placed under vacuum in a chamber with feedthroughs to accommodate the electrical inputs and outputs. For measurements of I a t various values of V between 0 and 500 V (or stress from 0 to 20 kV/cm), each sample was conditioned and tested as fol- lows. For each voltage measurement, starting with the lowest voltage VI, the voltage from the dc power supply (Electronic Measurement, Inc., TCR 600 SI) was set at 50 V above the value to be measured (Vz) and applied to the sample for 1 h. The voltage was then reduced to that value V,, producing a change in the current (monitored with a Keithley 610C electrometer) that decreased with time. Current I1 corresponding to voltage VI was recorded only after the current had stabilized. This se- quence was repeated for several values of V up to and including 500 V. Conductivity for each sample of known crystallinity was calculated at each value of V to indicate how cr varied with voltage (or electrical stress) and with percentage crystallinity. The permittivity for each sample is calculated using the expression E = (C/&,)(L/A). The value E , = 8.85 x 10-la F/m is the permittivity of space, and C is the capacitance measured at 1 kHz and 23'C using an Andeen-Hagerling model 2500A capacitance bridge.

The dielectric strength measurements were performed using a breakdown test fixture consisting of two 25 mm diameter brass cylindrical rod electrodes with rounded edges. Each sample sheet was placed between the two

electrodes and immersed in a dielectric fluid (Fluorinert FC 77). A 60 Hz ac voltage increasing a t a constant rate was applied to the top electrode using a Biddle 40 kV breakdown test supply with automated ramp capability. The bottom electrode was tied to ground potential. Dur- ing test, the Biddle test voltage was monitored as a function of time using an XY recorder, generating a straight line with constant slope up to the breakdown point. At breakdown of the sample, a drop in Biddle voltage occurred and was recorded on the plot, giving precisely the breakdown voltage. In these breakdown studies, a minimum of 10 breakdown measurements per sample was obtained.

I

'IEMPERKIVRE OC Figure 2.

DSC curve at 20C/min of unoriented amorphous PEEK.

3. RESULTS The DSC curve (Figure 2) for the as-received amor-

phous PEEK shows its behavior as the temperature is raised from ambient through its melting temperature. An endotherm is observed in the region of 153'C defin- ing the glass transition temperature Tg of PEEK. With increasing temperature, a crystallization exotherm is observed at - 185'C. As the temperature increases further, a point is reached where the crystalline material melts, as shown by the endotherm peaking at 340'C.

By thermal annealing the sample for different lengths of time a t selected temperatures, amorphous PEEK can be converted into a material with varying degrees of crystallinity. The behavior of the material resulting from four different annealing processing conditions is seen in the DSC scans in Figure 3. The DSC thermograms in Figure 3 are not normalized. They are representative of the changes that occur to PEEK film during the annealing process and indicate the extent of formation of the crystalline region.

In these examples, the material has been isothermally crystallized at different temperatures and then cooled to room temperature. All scans in Figure 3 exhibit a similar Tg at 153'C as that seen in Figure 2. In scans A through

004 Giants: Crystallinity and Dielectric Properties of PEEK

,----

1 I l ~ l l l I l l I I ' 1 I I I I , " I I , , " I l l , , I , , [ , , g : 8 s g s g a

TEh4F.E-r Figure 3.

DSC curve at 2O'C/min for annealed PEEK crystallized isothermally at (A) 160% for 1 h, (B) 210% for 1 h, (C) 260% for 1 h, (D) 310% for 1 h.

D, in addition to the melt endotherm at 340'C, a lower temperature endotherm is also observed. The lower temperature endotherms that occur in these DSC scans represent melting of crystals formed during the annealing process. In each instance, the lower melting endotherm occurs at a temperature slightly higher than the annealing temperature. As the isothermal annealing temperature is increased, it appears that the polymer experiences increasing crystalline perfection, culminating in a higher melting endotherm [9, lo].

Using the results from Figure 3, the annealing process was varied to produce as wide a range of crystallinities as possible. The lower values are limited by high crystallization rates, while the higher values are limited by film distortion. A separate set of film samples was pro- cessed uniformly by heating to predetermined temperatures. The hold times were varied, and the samples were cooled to room temperature a t the same rate, all in an argon atmosphere. The crystallinity of these samples was initially determined by the density gradient method [ll]. X-ray data for the as-received PEEK showed no indica- tion of a crystalline phase. This is in agreement with the amorphous nature of the material. Utilizing the calibration curve obtained from the density gradient method, the densities of the annealed PEEK samples were compared with the density of the as-received material. The results are shown in Table 1.

It was found from the annealing experiments that lit-

Table 1. Density of PEEK samples obtained from varying annealing conditions.

Anneal T'C

Amorphous Amorphous Amorphous

160 160 210 210 210 210 260 260 310 310 310

Time h - - - 1 1 1 1 2 2 1 2 1 1 2

Density

1.2649 1.2685 1.2732 1.2879 1.2898 1.2967 1.2994 1.2950 1.2982 1.3036 1.3012 1.3025 1.3084 1.3073

g/cm3 Cryst.

% 0.6 3.8 7.9 20.2 22.5 28.6 30.9 27.1 29.9 34.6 32.5 33.7 38.8 37.9

tle change occurred to the morphology of PEEK when annealed below its Tg for any length of time. Above the T,, crystallization was found to occur rather rapidly [12]. This result is indicated from the small variations in crystallinity observed for different times at the same temperature. Also, this work and that of previous authors, indicate that by annealing above Tg but below the melting temperature, it is extremely difficult to produce a low degree of crystallinity. It was pointed out that the results would be enhanced significantly with data points at the lower crystallinities. Although this would be most desir- able, the results of this work and the references therein have shown that the crystallization kinetics due to annealing are too rapid to control the crystallinity in this region. More recent literature reports on PEEK also con- firm this difficulty [13-151. However, three low density values were obtained from different batches of as-received PEEK. These values are included in Table 1. The measured density of the 100% amorphous as-received PEEK is taken to be 1.2642 g/cm3 and that for a 100% crystalline sample has been calculated to be 1.378 g/cm3 [16]. The crystallinity of the annealed PEEK samples was determined from a calibration curve of percent crystallinity vs. density for amorphous and crystalline PEEK.

X-ray crystallographic studies were also performed on the annealed PEEK samples in order to compare their crystallinity values with those obtained from the density data. Although DSC would be expected to provide useful data for percent crystallinity, the nature of the PEEK material is such that further annealing would occur during the measurement and the results could not be considered reliable. The X-ray technique is not inva-


3.80 1

2.70

2.49

x 1 ~ 3

Z.:B '

density g/cm3 1.2649 1.2685 1.2732 1.2827 1.2967 1.2982 1.2994 1.3025 1.3036 1.3084

s ~ 1.88 ' - '3 1.58'

1 2 B '

8.98'

0.68 '

Crystallinity % from dens. from X-ray

0.6 0.0 3.8 0.0 7.9 0.0

16.3 19.8 28.6 24.6 29.9 33.8 30.9 26.4 33.7 38.2 34.6 32.7 38.8 40.5

(b) 1J.s ' 15:. ' k r ' 2510 ' &e ' 35:s ' w h ' 4.~

Figure 4. (a) X-ray data for PEEK annealed at 210'C and (b) 310'C normalised to amorphous PEEK.

ze (dew=)

sive but was somewhat handicapped by the unavailabili- ty of a 100% crystalline sample for comparative analysis. However, it was felt that the use of this technique would be instructive as a comparative study with the density gradient method. Examples of X-ray diffraction curves for annealed and unannealed PEEK samples are shown in Figure 4. The computer program that acquires the XRD data expresses intensity as counts accumulated for a specific count time. For comparative purposes, all samples were analyzed with the same count time. Estimates of crystallinity were obtained by scaling an amorphous curve under the diffraction peaks [4,16]. The ratio of the areas of the crystalline peaks to the total area was taken as a weight fraction index for the X-ray crystallinity. The X-ray data were found to compare favorably with the crystallinity values derived from density measurements.

From the calibration curve of crystallinity vs. density and the data from Table 1, the crystallinity of samples annealed from the same film batch of controlled thickness is compared with the crystallinity from the X-ray data. These results are shown in Table 2. Unfortunately, a lim-

ited supply of material of controlled thickness prevented a broader range of crystallinities from being generated for this work. However, the results are in fairly good agreement, given the relative deficiencies in the analytical techniques used to establish the extent of crystallinity in polymeric thin film samples of this kind.

Percent cr

- - 1DOv -0- mv -*-4oov - - 0 . - mv

6.00

5.00 . . . - . . . . .

0 10 20 30 40 50 crystauw (9)

Figure 5. Conductivity vs. crystallinity for annealed PEEK, (from density data).

In the conductivity study, current/voltage measurements were obtained under ambient temperature conditions for nominally 25 p m thick PEEK samples. These samples were obtained from the same material batch pro- cessed for the density measurements. Each sample was placed under vacuum for 24 h prior to application of voltage for the conductivity experiments. The results of these experiments were in agreement with the hypothesis that increased charge trapping a t the crystalline/amorphous interface results in decreased conductivity. This behavior was indeed observed. Conductivity vs. crystallinity determined from density data is shown in Figure 5.

Increasing crystallinity a t the expense of the amorphous phase presents restrictions to pathways for the


transport of charge. The trend of the data in Figure 5 is consistent with the increased trapping hypothesis [17]. An alternate sample set using crystallinity values derived from X-ray data instead of density data also confirmed this trend. Unfortunately, the sensitivity of amorphous PEEK to annealing temperature made it difficult to obtain samples with crystallinity below 25%. The measurement errors associated with both crystallinity and conductivity in Figure 5 are insignificant.

3.90 3 I

The effect of crystallinity on the permittivity proved to be inconclusive. A number of measurements made on a variety of samples of varying crystallinity showed no particular trend (Figure 6). Thus, the permittivity obtained by this technique did not reflect the results of the conductivity experiments and was not a sufficient predic- tor of the effect of crystallinity on ultimate breakdown properties.

Dielectric breakdown studies were performed on a selected sheet of as-received PEEK that provided the lowest thickness variability. In general, the dielectric breakdown strength decreases with increasing sample thickness. Usually this has been ascribed to the increasing lev- el of defects in materials providing sites for the propaga- tion of a breakdown route. Recent studies have attempted to correlate the relationship between film thickness and dielectric breakdown and the causes of breakdown in polymers [18-201. It was deemed essential to minimize this variation for the sample under study. The sheet was cut into strips and annealed under similar conditions at different temperatures to provide a broad range of crystallinities. After annealing, the samples were subjected to breakdown conditions described previously. The results are shown in Figure 7 .

II

0 10 m 30 40 50 *-tY (%)

Figure 7. Dielectric breakdown strength vs. crystallinity (from density), for annealed PEEK.

The results in Figure 7 distinctly show a deterioration in dielectric breakdown strength when PEEK changes from an amorphous to a crystalline morphology. At- tempting to narrow the number of variables that con- tribute to breakdown was a major emphasis of this work and one of the reasons for choosing PEEK. The wide er- ror limits displayed are indicative of the difficulty in ob- taining a uniformly flat film, and a minimum of 10 measurements were randomly made throughout the surface of each sample to provide reasonable statistics. Values of percent crystallinity below 25% were limited by the annealing process conditions that did not allow precise control of crystallinity. Values exceeding 35% resulted in brittle films that begin to deviate from planarity, thereby introducing large errors in thickness measurements. These limitations prevented a more thorough analysis of the effect of crystallinity on electrical properties.

4. DISCUSSION The morphology of PEEK for crystallization tempera-

tures up to - 300'C has been described as being in the form of spherulites consisting of narrow lathlike lamellae that adopt an edge-on orientation in thin films. At higher crystallization temperatures, the lamellae have more thickness and are more perfectly ordered [21]. These samples were also described as exhibiting much higher electron densities in the center of the spherulites than in their periphery [22]. It has been mentioned previously that the samples annealed at 310'C tended to be more brittle when the crystallinity exceeded - 30%. If the de- scription of the morphology developed above and below 300'C is as described, there appears to be little effect on the trend of the conductivity observed in Figure 5. This result is consistent with a gradual nucleation and growth of spherulites and of increasing crystallite size, where the crystalline/amorphous interface dominates the conductivity. The change in morphology could not be corre-


lated with any change in permittivity. Studies of chain dynamics of PEEK indicate that most of the aromatic rings are essentially immobile in the amorphous and crystalline phases [23]. This evidence supports a frozen morphology after annealing, where locked-in polarizable groups might explain the lack of any trend in the values of the permittivities observed. The conductivity results are consistent with the original hypothesis, but no relationship with the permittivity was revealed.

The dielectric breakdown data (Figure 7) show that the value for the sample annealed at 310'C is noticeably lower on the curve than the samples annealed a t temperatures < 300'C. A study of isothermal crystallization of PEEK for annealing times > 20 min has shown that the relative fractional free volume remains constant [24]. Since free-volume holes have been mentioned as providing a breakdown pathway, these data argue for an alternate breakdown pathway. Assuming nominal average sample thickness, the lower breakdown strength observed at 310'C may be due to defects arising from the different morphologies generated above and below 300'C. Support for this view derives from the controversy regarding the origin of the multiple endothermic behavior of PEEK and related materials (Figure 3). Two models have been pro- posed to explain this endothermic behavior. One model describes the initial formation of low melting crystals, which are then transformed to a higher melting form during annealing [8]. The other model describes the double melting endotherm as arising from separate populations of crystals [IO]. Additional reports have added to the controversy by giving supporting evidence for each of the models presented [25,26]. Each model, however, provides an opportunity for the formation of an increasing pop- ulation of defect sites during annealing. The dielectric breakdown strength appears to decrease with increasing crystallinity, indicating that changing morphology introduces breakdown pathways not present in the amorphous material. The relationship between decreasing breakdown strength and decreasing conductivity as morphology changes does not appear to be consistent with the original hypothesis. An additional point was made con- cerning the effect of water on the dielectric properties. The sorption of water in PEEK films of varying crystallinity has been reported to be negligible [27,28], and the low uptake is further supported by results for thick specimens [29]. Thus moisture is not expected to have an impact on the results.

Recent reports on the electrical properties of polyethylene provide some insight for understanding breakdown mechanisms and the relationship of polyethylene behavior to the study of PEEK. In a report on polyethylene cable insulation, the breakdown strength was found to vary either directly or inversely with the degree of

crystallinity, depending on the aging process used [30]. It is important to note that the various aging processes also introduced additional structural changes in the polymer. As described previously, different structural changes in the material affect the breakdown process in different ways, leading to ambiguous results. Similarly, a report on partial discharge experiments involving polyethylene blends concluded that thermally induced mechanical stresses can introduce cracking in materials with increased crystallinity and thereby shorten the life of the polymer [31]. Although the binary blends were designed to introduce varying degrees of crystallinity, a change in the molecular weight distribution of the various blends was an inherent part of the process. The use of blends introduces an additional molecular parameter, which has the potential for compromising the degradation mechanism.

In a report that minimizes structural variations in polyethylene and is most closely related to the experimental approach taken in this study, it was found that the breakdown strength decreases with increasing crystallinity [32]. The report describes annealing of vapor-deposited polyethylene films to varying degrees of crystallinity in an inert atmosphere, thereby avoiding the introduction of chemical changes that can affect the breakdown mechanism. Further study of these films also found that the mean-free-path of injected electrons was longer in the samples with increased crystallinity and that these ener- getic electrons were responsible for the reduced dielectric strength of the polymer. This contrasts with an earli- er report in which similar results were obtained but the mechanism of breakdown was given a different interpretation [33]. Here the reduced dielectric strength with increasing crystallinity was attributed to a lower density pathway developed at the boundary layer between the crystalline and amorphous phases. Both reports im- ply that electron transport is more favorable with increasing crystalline content. However, the mechanism through which increasing crystallinity leads to reduced breakdown strength continues to be a subject requiring further investigation.

5. CONCLUSIONS T is difficult to assess which of the subtle effects of I changes in morphology are responsible for changes in

electrical properties. Both free-volume holes and imper- fections a t the crystalline/amorphous interface are potential contributors to breakdown. The relationship of crystallite size on electrical breakdown has been discussed by other researchers and has been considered here. A gradual nucleation and growth of a large number of crystalline sites are in keeping with the annealing conditions used. The consistency of the annealing process is not


believed to have introduced variation that would have influenced the results from the standpoint of crystallite size. However, experiments that can introduce variation in crystallite size while simultaneously maintaining the same degree of crystallinity would clarify the effect of crystallite size on breakdown strength. The resolution of the controversy regarding the origin of the melting endotherms also may be of considerable importance to the understanding of defect structure and its relationship to dielectric properties.

This work has attempted to limit impurity variations and the introduction of decomposition products by choosing a thermoplastic material of high thermal and chemical stability. This choice was to ensure that any effect on breakdown strength of the material could be limited to the aforementioned parameters and a more fundamental relationship could be examined. In this study, it was found that, in the absence of other processing variables, there is a reduction in the dielectric breakdown strength as the morphology of the material changes from a purely amorphous nature to one of increasing crystallinity. This result is in agreement with that obtained for polyethylene arrived at through different approaches and with con- trasting explanations as to the breakdown mechanism. Further complications will arise when one attempts to determine the influence of polar groups on the breakdown mechanism, as is the case with PEEK. The precise source of the reduction in dielectric strength with increasing crystallinity is still open to question, as the results of the detailed morphological characterization of PEEK and the more recent studies with polyethylene have indicated.

ACKNOWLEDGMENT The author wishes to thank Robert Castaneda for ma-

terials preparation, density gradient, and thermal and dielectric property measurements; Paul Adams for the X-ray data and interpretation; and Dr. Frank Hai for in- sightful discussions on the dielectric measurement techniques. Funding for this work was provided by the Miss- ion-Oriented Investigation and Experimentation program of The Aerospace Corporation.

REFERENCES M. Ieda, Dielectric Breakdown Process of Poly- mers, IEEE Trans. Electr. Insul., Vol. 15, No. 3,

P. Fischer and P. Rohl, Transient Currents in Lin- ear Low-density Polyethylene , Progress in Colloid and Polymer Science, Vol. 62, pp. 149-153, 1977.

pp. 206-224, 1980.

[3] R. M. Keyser, K. Tsuji and F. Williams, ESR and Optical Studies of Trapped Electrons in Glasses and Polymers, in Radiation Chemistry of Macro- molecules, Vol. 1, Chapter 9, pp. 145-191, M. Dole, ed., Academic Press, New York 1972.

[4] R. Nath and M. M. Perlman, Effect of Crystallini- ty on Charge Storage in Polypropylene and Poly- ethylene, IEEE Trans. Electr. Insul., Vol. 24, No.

[5] G. Krause, K. Moller, and D. Meurer, Dielectric Phenomena in Semicrystalline Polymers, IEEE Trans. Electr. Insul., Vol. 24, No. 2, pp. 215-222, 1989.

[6] M. F. Rose, Electrical Insulation and Dielectrics in the Space Environment, IEEE Trans. Electr. Insul., Vol. 22, No. 5, pp. 555-571, 1987.

[7] P. J . Phillips, Morphology - Electrical Property Relations in Polymers , IEEE Trans. Electr. Insul.,

[8] D. J. Blundell and B. N. Osborn, The Morphology of Poly (Aryl-ether-ether-ketone) , Polymer, Vol.

[9] P. Cebe and S. Hong, Crystallization Behavior of Poly (Ether-ether-ketone), Polymer, Vol. 27, pp.

[lo] D. J. Bassett, R. H. Olley, and I. A. M. A1 Raheil, On the Crystallization Phenomena in PEEK, Polymer, Vol. 29, pp. 1745-1754, 1988.

[ll] G. M. K. Ostberg and J. C. Seferis, Annealing Ef- fects on the Crystallinity of Polyetheretherketone (PEEK) and its Carbon Fiber Composite, Jour- nal of Applied Polymer Science, Vol. 33, pp. 29-39, 1987.

[12] H. Marand and A. Prasad, On the Observation of a New Morphology in Poly (Arylene Ether Ether Ketone). A Further Examination of the Double En- dothermic Behavior of Poly (Arylene Ether Ether Ketone) , Macromolecules, Vol. 25, pp. 1731-1736, 1992.

[13] M. J. Folkes, G. Kalay and A. Ankara, The Effect of Heat Treatment on the Properties of PEEK and APC2 , Composites Science and Technology, Vol.

[14] R. A. Chivers and D. R. Moore, The Effect of Molecular Weight and Crystallinity on the Me- chanical Properties of Injected Molded Poly (Aryl- ether-ether-ketone) Resin , Polymer, Vol. 35, No.

[15] P. Damman, Ch. Fougnies, J.-F. Moulin and M. Dosikre, Crystallinity of Poly (Aryl Ether Ether

3, pp. 409-412, 1989.

Vol. 13, NO. 2, pp. 69-81, 1978.

24, pp. 953-958, 1983.

1183-1192, 1986.

46, pp. 77-83, 1993.

1, pp. 110-116, 1994.

IEEE Transactions on Dielectrics and Electrical Insulation Vol. I No. 6 , December 1994 999

Ketone) by Vibrational Spectroscopy, Macro- molecules, Vol. 27, pp. 1582-1587, 1994.

[16] J. N. Hay, D. J. Kemmish, J . I. Langford, and A. I. M. Rae, The Structure of Crystalline PEEK, Polymer Commun., Vol. 25, pp. 175-178, 1984.

[17] D. K. Das-Gupta and K. Doughty, Dielectric and Conduction Processes in Polyetherether Ketone (PEEK), IEEE Trans. Electr. Insul., Vol. 22, No.

[18] J . H. Mason, Effects of Thickness and Area on the Electric Strength of Polymers, IEEE Trans. Electr. Insul., Vol. 26, No. 2, pp. 318-322, 1991.

[19] B. Helgee and P. Bjellheim, Electric Breakdown Strength of Aromatic Polymers, IEEE Trans. Electr. Insul., Vol. 26, No. 6, pp. 1147-1152, 1991.

[20] J. H. Mason, Electric Breakdown Strength of Aro- matic Polymers, IEEE Trans. Electr. Insul., Vol.

[21] S. D. Hudson, D. D. Davis, and A. J. Lovinger, Semicrystalline Morphology of Poly (Aryl Ether Ether Ketone)/poly (Ether Imide) Blends, Macromolecules, Vol. 25, pp. 1759-1765, 1992.

[22] A. J . Lovinger, S. D. Hudson, and D. D. Davis, High-temperature Crystallization and Morphol- ogy of Poly(Ary1 Ether Ether Ketone), Macro- molecules, Vol. 25, pp. 1752-1758, 1992.

[23] M. D. Poliks and J. Schaefer, Characterization of the Chain Dynamics of PEEK by CPMAS 13C NMR., Macromolecules, Vol. 23, pp. 3426-3431, 1990.

[24] H. Nakanishi and Y. C. Jean, Dynamics of Excess Free Volume in Semicrystalline PEEK Studied by Positron Annihilation, Macromolecules, Vol. 24,

[25] M. P. Lattimer, J. K. Hobbs, M. J . Hill, and P. J. Barham, On the Origin of the Multiple En- dotherms in PEEK, Polymer, Vol. 33, pp. 3971- 3973, 1992.

1, pp. 1-7, 1987.

27, NO. 5, pp. 1061-1062, 1992.

pp. 6618-6621, 1990.

[26] M. E. Nichols and R. E. Robertson, The Origin of Multiple Melting Endotherms in the Thermal Analysis of Polymers , Journal of Polymer Science: Part B: Polymer Physics, Vol. 30, pp. 305-307, 1992.

[27] A. Arzak, J . I. Eguiazabal and J . Nasabal, The Effects of Solvent Sorption on the Properties of Poly(Ether Ether Ketone), Journal of Materials Science, Vol. 28, pp. 3272-3276, 1993.

[28] E. J . Stober, J. C. Seferis and J. D. Keenan, Char- acterization and Exposure of Polyetheretherketone (PEEK) to Fluid Environments, Polymer, Vol. 25,

[29] M. A. Grayson and C. J . Wolf, The Solubility and Diffusion of Water in Poly(Ary1-etherether-ketone) (PEEK), Journal of Polymer Science: Part B: Polymer Physics, Vol. 25, pp. 31-41, 1987.

[30] S. Grzybowski, A. Rakowska, and J . E. Thomp- son, Aging of Polyethylene for Cable Insulation , IEEE Trans. Electr. Insul., Vol. 22, No. 6, pp. 729- 734, 1987.

[31] A. Gustafsson, M. T. Conde Braiia, and U. W. Gedde, Electrical Degradation of Binary Poly- ethylene Blends, Polymer, Vol. 32, pp. 426-431, 1991.

[32] Y. Tanaka, N. Ohnuma, K. Katsunami, and Y. Ohki, Effects of Crystallinity and Electron Mean- free-path on Dielectric Strength of Low-density Polyethylene, IEEE Trans. Electr. Insul., Vol. 26,

[33] S. N. Kolesov, The Influence of Morphology on the Electric Strength of Polymer Insulation, IEEE Trans. Electr. Insul., Vol. 15, No. 5, pp. 382-388, 1980.

pp. 1845-1852, 1984.

NO. 2, pp. 258-265, 1991.

Manuscript w u received on 1 December 1993, in revised form 23 June 1994.

crystallinity and dielectric properties of peek, poly(ether ether ketone) - 1994

Documents