evaluation of poly(4-methyl-1-pentene) as a dielectric

Evaluation of Poly(4-Methyl-1-Pentene) as a Dielectric Capacitor Film for

High-Temperature Energy Storage Applications

Sahil Gupta,1* Ido Offenbach,1 JoAnne Ronzello,2 Yang Cao,2 Steven Boggs,2

Robert A. Weiss,1 Mukerrem Cakmak1

1Department of Polymer Engineering, University of Akron, Ohio 443252Institute of Materials Science, University of Connecticut, Storrs, Connecticut 06269

Correspondence to: M. Cakmak, Purdue University, Materials and Mechanical Engineering Departments, Birck Nanotechnology

Center, 1205 W State St, West Lafayette, Indiana 47906. E-mail: [email protected]

Received 20 March 2017; accepted 12 June 2017; published online 00 Month 2017

DOI: 10.1002/polb.24399

ABSTRACT: Poly(4-methyl-1-pentene) (P4MP) was characterized

to evaluate its viability as a high-temperature dielectric film for

capacitors. Detailed investigation of thermal, mechanical, rheo-

logical, and dielectric properties was carried out to assess its

high-temperature performance and processability. P4MP was

melt-processable below 270 8C without degradation and appli-

cation temperatures as high as 160–190 8C can be achieved.

The dielectric constant and loss of melt-processed P4MP films

was comparable to biaxially oriented polypropylene (BOPP)

capacitor films, although the dielectric strength was lower.

Enhancements in dielectric strength up to 250–300% were

achieved via solution-processing P4MP films, which could be

easily scaled up on a roll-to-roll platform to yield isotropic,

free-standing films as thin as 3–5 lm. The influence of crystal

structure, crystallinity, and surface morphology of these films

on the dielectric properties was examined. The dielectric

strength was further increased by 450% through biaxial stretch-

ing of solution-cast films, and a Weibull breakdown field of

514 V/lm was obtained. The dielectric constant was very stable

as a function of frequency and temperature and the dielectric

loss was restricted to <1–2%. Overall, these results suggest

that BOP4MP is a promising candidate to obtain similar energy

density as a BOPP capacitor film but at much higher operating

temperatures. VC 2017 Wiley Periodicals, Inc. J. Polym. Sci., Part

B: Polym. Phys. 2017, 00, 000–000

KEYWORDS: capacitor; dielectric; energy storage; morphology;

poly(4-methyl-1-pentene); roll-to-roll (R2R)

INTRODUCTION With the world’s ever-growing energydemands, effective energy storage, utilization, and manage-ment have become one of the most important and challeng-ing areas of the twenty-first century. Research in the area ofenergy storage devices such as batteries, fuel cells, superca-pacitors, and dielectric capacitors has increased significantlyin the last two decades.1–4 The requirements for new genera-tion telecommunication devices, electric/hybrid vehicles,electromagnetic weaponry, and other power electronicsapplications for commercial, industrial, and military use areprojected to drive the growth of the existing technologies forenergy storage devices. Compared to batteries and fuel cells,dielectric capacitors provide a high power density but rela-tively low energy density. Increased power density and oper-ating frequency of power electronics required developmentof high energy density capacitors. Therefore, recent researchin the area of dielectric capacitors is intended to increaseenergy density while minimizing the energy loss.5–11 Themaximum energy that can be stored in a dielectric mediumis given by

W5

ðDmax

0

EdD51

2E0ErE

2ds (1)

if the dielectric constant is independent of field, as is usuallythe case. Here E is the externally applied electric field (input),D is the electric displacement field (output) within themedium, E0 is the permittivity of vacuum (8.85 3 10212 F/m),Er is the relative permittivity or dielectric constant of themedium, and Eds is the dielectric strength or the breakdownfield strength of the medium.

The present state-of-the-art polymer for high-voltage, high-energy-density capacitor is dominated by metallized biaxiallyoriented polypropylene (BOPP).10–12 BOPP provides a highbreakdown field of 600–750 V/lm for �10-lm-thickfilms,5,13 a very low dielectric loss (<0.01%), and good self-healing characteristics.11,12 BOPP is especially suitable forapplications that require high voltage, short pulse length,and high repetition rates. BOPP, however, has two major limi-tations. First, it has a low dielectric constant (�2.2) that

*Present address: Parflex Division, Parker Hannifin Corporation, 1300 N. Freedom St., Ravenna, Ohio, 44266VC 2017 Wiley Periodicals, Inc.

WWW.MATERIALSVIEWS.COM JOURNAL OF POLYMER SCIENCE, PART B: POLYMER PHYSICS 2017, 00, 000–000 1

JOURNAL OFPOLYMER SCIENCE WWW.POLYMERPHYSICS.ORG FULL PAPERS

limits the stored energy density to 3–5 J/cm3 at break-down.5,13 Second, the usage of BOPP capacitors is limited totemperatures below 105 8C because PP has a peak meltingpoint of �162 8C with a relatively low melting onsetbetween 85–100 8C. The thermomechanical stability, andtherefore, the dielectric strength decreases dramaticallyabove 85 8C.14 In spite of these drawbacks, PP consumptionfor capacitor films has almost doubled in the last 5 years.

Although polymers such as poly(ethylene terephthalate)(PET) and poly(vinylidene fluoride) (PVDF) have shownpromise, these polymers suffer from critical property defectsthat limit their use in industrial capacitors. While PET pro-vides a higher dielectric constant than BOPP, and has ahigher operating temperature of 125 8C compared to 105 8Cfor BOPP, it suffers from poor thermal stability and a highdielectric loss that increases with temperature and pres-sure.11,15 PVDF possesses one of the highest dielectric con-stants (�10) among polymers due to the polar –CF2 groups,with a dielectric breakdown strength of 200 V/mm. How-ever, PVDF suffers from very high dielectric and hysteresislosses that severely limits the maximum energy density thatcan be obtained from this polymer. In addition, PVDF exhib-its an unstable dielectric constant which varies with voltage,temperature and frequency. Some of these problems havebeen mitigated with the use of PVDF based copolymers, suchas poly(vinylidene fluoride-co-trifluoroethylene-co-chlorotri-fluoroethylene)), and energy density as high as 17 J/cm3 hasbeen reported6–8,10 but the energy losses are still high. Useof such copolymers is therefore limited to low repetitionrate (low-voltage, low-frequency) applications.

Apart from PET and PVDF, alternate polymer chemistries arealso being explored to find a suitable match to the capacitorindustry requirements such as those of high-energy high-temperature dielectrics (>200 8C).3,14,16–19 Poly(ether imide)(PEI), poly(phenylene sulfide) (PPS),14 polyacrylamides,16

poly(ethylene naphthalate) (PEN),17 poly(carbonate) (PC),poly(tetrafluoroethylene) (PTFE), and poly(sulfone)18 arefew of those polymers that have been studied. However,compared to BOPP, most of these polymers suffer from limi-tations as described above for PET or PVDF, or provide com-parable energy densities but with poorer mechanical andthermal properties so that BOPP is still considered to be asuperior dielectric.

One of the polymers that seems interesting as a high temper-ature dielectric is poly(4-methyl-1-pentene) (P4MP). P4MP isa high temperature semicrystalline polyolefin that has notyet been evaluated properly as a high-temperature polymerdielectric for capacitor applications. It has been cited as acapacitor film in several patents,20,21 but it does not appearfrom those patents that the dielectric properties of P4MPfilm have been evaluated or any capacitor has actually beenbuilt with P4MP. In addition to high temperature, P4MP pro-vides the benefit of being one of the lowest density (0.83 g/cm3) polyolefins available commercially and is therefore use-ful for critical applications such as aerospace and militarywhere weight reduction is attractive.

In this work, the thermal, mechanical, and rheological prop-erties of two commercial grades of P4MP were investigatedto evaluate its processability for melt extrusion and furtherbiaxial stretching. Thin films were prepared through meltprocessing and solution processing and their dielectric prop-erties were compared. The real-time dynamics during dryingof solution-processed P4MP films were studied in a custom-ized instrumented platform which yielded insights into thesurface morphology formation in these films and providedoptimum conditions for scaling thin film manufacturing on aR2R platform. Finally, the effect of crystal structure, percentcrystallinity, surface morphology and biaxial stretching onthe dielectric properties of P4MP was evaluated.

EXPERIMENTAL

MaterialsTwo grades of P4MP were provided by Mitsui ChemicalsLtd.—TPXTM MX002O and TPXTM DX845 (hereafter referredto as MX and DX, respectively). The two grades varied intheir comonomer content, which was �8 mol % for MX and<5 mol % for DX. However, the comonomer chemistry wasproprietary and was not provided. Typical comonomers usedwith 4-methyl-1-pentene during polymerization include pen-tene, hexene, octene, decene, and octadecene.22 The poly-mers were provided in pellet form and were used as-is.Cyclohexane (�99.8%, C100307) and toluene (�99.5%,179418) were purchased from Sigma-Aldrich Co., and cyclo-pentane was purchased from Fisher Scientific Co. (99% pure,AC111480025). All solvents were used as-is.

CharacterizationThe viscosity-average molecular weight (Mv) was calculatedfrom the intrinsic viscosity measurements performed with aCannon-Ubbelohde viscometer using decahydronaphthalene(Decalin) solution at 135 8C with a standard of pure P4MP.Mv was estimated by the Mark–Houwink equation using theconstants for P4MP, [g]5K 3 Ma

v , where K 5 1.94 3 1025

(L/g) and a 5 0.81.22

Thermal characterization was performed using a TA Instru-ments Q-200 differential scanning calorimetry (DSC), a TAInstruments Q-50 thermogravimetric analysis (TGA), and TAInstruments Q-800 dynamic mechanical analysis (DMA).

DSC was used to measure the glass transition temperature(Tg), the crystallization temperature (Tc), the melting point(Tm), and the heat of fusion (DHf) of the polymers. Beforethe measurements, temperature calibration was performedusing an Indium standard (Tm 5 156.6 8C). All heating andcooling scans were made from 220 8C to 260 8C at a rate of10 8C/min. Tg was defined as the inflection point in thechange of the heat capacity associated with the glass transi-tion. The melting point (Tm) and the crystallization point(Tc) were defined as the peak of the melting endotherm andthe cooling exotherm, respectively. Mass fraction crystallinityof the polymer (/m) was calculated from the ratio of DHf/DHP4MP, where DHf was the area under the DSC meltingendotherm and DHP4MP is the enthalpy of fusion for a 100%

FULL PAPERS WWW.POLYMERPHYSICS.ORGJOURNAL OF

POLYMER SCIENCE

2 JOURNAL OF POLYMER SCIENCE, PART B: POLYMER PHYSICS 2017, 00, 000–000

crystalline P4MP, which is reported to be 5.20 kJ/mol (or61.85 J/g).23

The thermal stability of the polymers was checked by TGAusing a heating rate of 20 8C/min and a nitrogen atmo-sphere. Isothermal hold-up measurements were also carriedout from 180–280 8C under nitrogen atmosphere. The tem-perature was increased in increments of 20 8C, and isother-mal conditions were maintained for at least 10 hours tocheck long-duration thermal stability. DMA was used to mea-sure the temperature and frequency-dependent dynamic,E0(T,f), and loss, E00(T,f), moduli and to identify thesecondary-transition temperature associated with the slip-page of crystal planes, (Ta*).

24 Ta* was defined as the tem-perature of the peak in the loss modulus, E00(T,f), associatedwith that transition. The DMA measurements were per-formed in tension using a multiple-frequency strain modeand frequencies of f 5 1, 100 Hz from 2150 8C to 230 8C, aheating rate of 2 8C/min and a strain of 1%.

In addition to the thermal properties, rheological properties ofa polymer such as elasticity and viscosity and theirtemperature-dependence are also critical for thin filmmanufacturing. Hence rheological characterization was per-formed to check the melt processability of P4MP and to findthe correct temperature range for the same. Linear viscoelastic(LVE) properties were measured with a TA Instruments ARESG2 rheometer using small amplitude oscillatory shear (SAOS)measurements from 230 to 300 8C at frequencies (x)5 0.1–100 rad/s and strain amplitudes of 5–10% on a 25 mmparallel-plate fixture. Strain-sweeps were performed at 250 8Cand 100 rad/s to determine the LVE limit, and master-curveswere constructed using time-temperature superposition (TTS)with a reference temperature, Tr 5 250 8C. Steady shearexperiments were carried out from 230 to 280 8C with theARES G2 rheometer using a 25 mm cone-and-plate fixturewith a cone angle of 0.04 rad. Viscosity at higher shear ratesfrom 102 to 104 s21 was measured using a Bohlin (Malvern)RH7 capillary rheometer from 250 to 270 8C. A capillary diewith L:D 5 32:1 was chosen so that the Bagley correction wasnegligible and could be ignored.25 The Rabinowitch–Mooneycorrection was applied to the apparent shear rate to calculatethe true shear rate and the true viscosity.

The crystal structure in P4MP thin films (10–20 lm) wascharacterized with wide-angle X-ray diffraction (WAXD) usinga Bruker AXS D8 Goniometer with a CuKa radiation(k 5 0.1542 nm). The surface morphology of those filmswas characterized using a JEOL-7401 (Japan Electron OpticsLaboratory) scanning electron microscope (SEM).

The dielectric constant and dielectric loss were measuredisothermally using time-domain dielectric spectroscopy(TDDS). Approximately 10–20-lm-thin film with 2 cm2

exposed area was used, and measurements were conductedat 10 V from a frequency of 1023–104 Hz at room tempera-ture (RT), 50, 100, and 150 8C. The dielectric breakdownmeasurements were performed at room temperature underthe ramp DC high voltage with a constant ramping rate of

300 V/s. The high voltage was generated by HV power sup-ply (Stanford Model PS375) which was controlled by a rampsignal generator. The P4MP films were sandwiched betweentwo metallized BOPP films which were used as electrodeswith the metallization side in contact with the P4MP filmsurface. An active area of 2 cm2 was used and it was con-trolled through a 100 lm polyimide mask to protect againstdischarge between the two electrodes. The dielectricstrength was characterized using the two-parameter Weibullcumulative distribution function (CDF),26

F xð Þ512exp 2x

g

� �b" #

(2)

where g is the Weibull characteristic breakdown fielddefined as the breakdown field at 63.2% probability ofbreakdown (x5 g, F(x)5 0.632), and b is the Weibull slopeparameter which is a measure of dispersion in the data. TheF(x) value for each x (breakdown strength in the currentcase) was assigned based on its position (i) among the Nordered x-values, and was calculated using the Benard’s(empirical) approximation,

Fi5i20:3

N10:4(3)

Sample PreparationMelt Processing Thin FilmsMelt processing of P4MP films was carried out by compressionmolding the polymer pellets at 260 8C under vacuum using theTMP vacuum compression press. The mold was placedbetween preheated platens and allowed to thermally equili-brate for 10 min. followed by compression at 20,000 psi for 5min. The sample was then cooled back to room temperature at40 8C/min. Approximately 50-lm-thin films were obtainedusing this procedure. The films were prepared either withouta substrate (i.e., the polymer film was in contact withstainless-steel surface of the mold on both sides) or using a100-lm-thick polyimide (KaptonVR ) substrate.

Solution Processing and Real-Time DynamicsApart from melt processing, P4MP thin films can also beprocessed from a solution under appropriate conditions.While most of the common polyolefins such as polypropyl-ene or polyethylene are not solution-processed commercially,P4MP offers a unique opportunity due to its relatively highsolubility in common nonpolar solvents at temperatures aslow as 35–40 8C. Thin films (10–20 lm) of P4MP were pre-pared with a combination of different nonpolar solvents andcasting temperatures. The real-time drying dynamics of thesefilms were studied using a custom-built solvent casting plat-form whose schematic is shown in Figure 1. The instrumen-tation and working details are provided elsewhere.27

With this setup, film thickness, weight, surface temperature,birefringence (in-plane and out-of-plane), and % transmis-sion through the film were measured while the film was dry-ing. These measurements provided information about thereal-time development of local chain orientation, suitable



conditions for solvent-casting (e.g., casting temperature orthe initial polymer concentration in solution), and the timerequired to reach a steady-state behavior. This informationwas then used to replicate favorable conditions for preparinglarger films on a roll-to-roll platform as described below.

Roll-to-Roll (R2R) FabricationLarge-scale R2R fabrication of P4MP films was successfullyaccomplished by blade-coating polymer solution under favor-able conditions obtained from the batch processing studiesdescribed above. A nominal thickness of 10–20 lm was used,although films as thin as 3–5 lm could also be obtained. Filmswere cast on a PET (MylarVR ) substrate with a carrier speed of50 cm/min using a 1200 wide doctor blade. The films weredried overnight under atmospheric pressure, detached fromthe substrate, and collected on a separate roll. TGA measure-ments were performed to confirm the absence of solvent inthese films. The films were then used for dielectric measure-ments and further processing by biaxial stretching.

Biaxial Stretching of Solution Cast FilmsBiaxial stretching of P4MP thin films was performed using amodified Iwamoto biaxial stretcher. The instrumentation andworking details are provided elsewhere.28 Samples withdimensions 14 3 14 cm were cut from the cast P4MP films.The biaxial stretching was performed at 50 8C with a con-stant stretch rate of 15 mm/min to different stretch ratios.The sample was held between the pneumatic clamps withsandpaper sandwiched between the film and clamp to avoidslip and tear, heated up to the stretching temperature, ther-mally equilibrated for at least 10 min, equibiaxially stretchedand cooled back slowly to room temperature.

RESULTS AND DISCUSSION

Thermal PropertiesThe physical and thermal properties of the MX and DXgrades are summarized in Table 1, and their DSC endo-therms are shown in Figure 2. MX had a lower crystallizationtemperature (Tc), melting point (Tm), and crystallinity (/m)

FIGURE 1 (a) A customized solvent-casting platform to track real-time changes in the physical parameters during drying of poly-

mer films. (b) Blade coating of a polymer film on a glass substrate using a motorized drawdown coater. (c) Polymer solution (yel-

low) cast on a glass substrate (black). P1, P2, P3, and P4 are the locations of surface temperature measurements, and L1, L2, and

L3 are the locations of thickness measurements. Circular region at the center was used for birefringence measurements. (d) A

schematic for spectral birefringence measurement. (Adapted with permission from ref. 27, CopyrightVC 2012, AIP Publishing LLC.).

[Color figure can be viewed at wileyonlinelibrary.com]

TABLE 1 Physical and Thermal Properties of the Two Commercial Grades of P4MP

Grade

Comonomer

(mol %)

Mva

(kDa)

q(g/cm3)

Tdegb

(8C)

Tgc

(8C)

Tcc

(8C)

Tm,onsetc

(8C)

Tmc

(8C)

/mc

(%)

Tbd

(8C)

Ta*d

(8C)

MX �8 115 0.834 341 21.1 204 160 225 36 2139 200

DX <5 125 0.833 361 27.6 207 190 234 56 2141 217

a Calculated from [g] 5 K 3 Mav, where K 5 1.94 3 1025 and a 5 0.81.31 [g]

was measured in decalin at 135 8C.b From TGA at 20 8C/min.

c From DSC, heat and cool at 10 8C/min.d From DMA, heating at 2 8C/min, f 5 100 s21.


POLYMER SCIENCE


http://wileyonlinelibrary.com

compared to DX, which was attributed to a higher comono-mer concentration in MX and similar to our previous obser-vations from polypropylene copolymers.29,30 A highercomonomer concentration also resulted in a higher plastici-zation of MX, thereby, lowering its glass transition (Tg) aswell.

While the Tm was higher than 220 8C for both the grades,the onset of melting occurred at a much lower temperature(Tm,onset): 160 8C for MX and 190 8C for DX. A difference of10 8C between Tm, but 30 8C between Tm,onset of the twogrades indicated that the MX grade had a smaller averagesize of the crystal but a much broader crystal size distribu-tion. A low Tm,onset might limit the use of a polymer film totemperatures much lower than Tm due to its thermo-mechanical stability. For example, a stretched polymer filmheated above its Tm,onset might shrink, or experience consid-erable shrinkage stresses if spatially constrained. That occursbecause the oriented polymer chains experience an entropicforce to revert back to their coiled/unoriented configuration(commonly referred to as the “memory” effect). Simulta-neously, melting of the smaller crystals and their

recrystallization into bigger crystals abets the above process.Shrinkage of a dielectric polymer film or development ofmechanical stresses during a capacitor operation is highlyundesirable and should be avoided. It is because of this rea-son that the maximum operational temperature of BOPPdielectric film is restricted to 105 8C, the Tm,onset and Tm ofPP being �85–100 8C and ~162 8C, respectively. The capaci-tor then needs to be suitably derated at higher temperatures.One of the methods to enhance the thermomechanical integ-rity of a polymer film beyond Tm,onset is to stretch it at atemperature between Tm,onset and Tm, and subsequentlyheat-set the film where the crystallinity is enhanced and ori-ented amorphous chains are relaxed. Such a procedure mayensure thermomechanical stability till the stretch tempera-ture. An optimum stretch temperature can be measuredusing dynamic mechanical analysis and is discussed later.Heat-setting, however, was not pursued in this study andwill be a subject of future studies.

The thermal degradation temperature (Tdeg), which wasdefined as the temperature corresponding to a 1% mass lossin TGA, was >340 8C for both the grades (plots not shown).Although such a high degradation temperature seems prom-ising for high-temperature dielectric applications, isothermalhold-up measurements suggested otherwise. Figure 3 showsthe weight loss as a function of time for MX and DX gradesfrom 180 to 280 8C. Both the grades exhibited reasonablethermal stability up to 240 8C but degradation acceleratedbeyond that temperature.

The dynamic mechanical storage modulus (E0) and loss tan-gent (tan d) in three flow regimes—glassy, rubbery andmelt—are shown in Figure 4. In the glassy regime, E0 forboth MX and DX grades was similar even though the crystal-linity was higher for DX by 20%. This indicated an almostequivalent contribution from the crystalline and the amor-phous domains in the glassy state. In the rubbery region, theamorphous regions were much softer and mobile comparedto the crystals, which resulted in an increase in E0 for the DXgrade. In the melt state, where the crystals were no longerpresent, E0 was again higher for DX due to a higher molecu-lar weight. As most polymer films are required to be

FIGURE 2 DSC endotherms for the MX and DX grades of

P4MP. The curve for DX is shifted vertically for clarity. Inset

shows the chemical structure of P4MP. [Color figure can be

viewed at wileyonlinelibrary.com]

FIGURE 3 Isothermal TGA measurements for P4MP grades MX and DX. Both grades lose thermal stability above 240 8C. Thermal

stability for DX is relatively higher compared to MX. [Color figure can be viewed at wileyonlinelibrary.com]





mechanically wound under high tension in capacitors toensure uniform contact between the wrapped layers withoutany air entrapment, they should possess good mechanicalproperties such as a high elastic modulus at room tempera-ture. From Figure 4, the elastic modulus at room tempera-ture (25 8C) is measured to be �1.7 GPa for P4MP which isclose to the value reported previously by us for PP.29

The loss tangent showed three distinct thermal transitions forP4MP. These values are tabulated in Table 1. The b-relaxationpeak at the lowest temperature (Tb) is associated with thelocalized thermal motion of the side-chains. Tb was similar forMX and DX grades indicating that the presence of copolymerdid not affect the localized motion of the side-chains. The sec-ond peak corresponds to the glass-transition temperature, anda higher Tg for DX was consistent with the DSC results. Thethird peak corresponds to the a*–transition that is generallyobserved in semi-crystalline polymers between the Tg and Tm,and is associated with the slippage of crystal planes over oneanother or the motion of the crystallites.24 Practically, Ta* pro-vides an estimate of the processing temperature that is opti-mum for orienting the polymer chains through uniaxial orbiaxial stretching. Ta* was higher for the DX grade probablydue to its higher crystallinity and bigger crystal size whichrequired a higher thermal energy for producing crystal–crystalslip at a given frequency of deformation.

The results from Figures 2–4 show promise for the use ofP4MP thin films at high temperatures up to at least 160 8C.The high temperature limit is controlled by the melting onsetof the polymer which can be substantially increased bydecreasing the comonomer content. In addition, the polymerfilm can be biaxially stretched above 200 8C in the semi-

molten state and thermally annealed in the stretched stateto further raise the usage temperature.

Rheological Properties and Processing TemperatureThe storage (G0) and loss (G00) moduli obtained using small-amplitude oscillatory shear (SAOS) measurements are shownin Figure 5 for the two P4MP grades at a reference temperature(Tr) of 250 8C. Each curve is a master-curve constructed usingtime-temperature superposition (TTS) by shifting the isother-mal data horizontally and vertically using the shift factors aTand bT, respectively. The validity of TTS implied that the pres-ence of 5–8 mol % comonomer did not affect the temperaturescaling of relaxation times for the polymer chains, even thoughit had a plasticizing effect on the polymer flow behavior. Bothelasticity and viscosity, that is, G0 and G00, were higher for theDX grade. The relaxation time (sx), calculated from the inverseof the cross-over frequency of G0 and G00, was also higher for DX(0.10 s) than MX (0.05 s). Higher G0 , G00, and sx for DX wereattributed to its higher molecular weight and lower comono-mer content than MX. The terminal regime (corresponding to aNewtonian-like behavior of the polymer chains, where G0 / x2and G00 / x)32 was not observed for either of the polymers.

The flow activation energy, DE, of the polymers was deter-mined by using Arrhenius analyses,33 eqs 4 and 5, of thetemperature dependence of the frequency shift factors (aT)used to construct the LVE master-curves,

st5Aexp ðDE=RTÞ (4)

lnaT5lnðst=st;rÞ5lnðgo=go;rÞ5DER

1

T2

1

Tr

� �(5)

where st is the terminal relaxation time, go is the zero-shearviscosity, T is the temperature of interest, Tr is the reference

FIGURE 4 Storage modulus (E0) and loss tangent (tan d) as a

function of temperature for the MX and DX grades at f 5 100

Hz. Thermal transitions are observed at Tb, localized motion of

side-chains; Tg, segmental motion; Ta*, crystal-crystal slip; and

Tm, melting. Data above the melting point were calculated

from small amplitude oscillatory shear (SAOS) measurements

(E05 3G0, E0 05 3G0 0). [Color figure can be viewed at wileyonline-

library.com]

FIGURE 5 Master curves at 250 8C for the storage (G0) and loss

(G00) moduli of MX and DX grades of P4MP. The curves were pre-

pared by TTS of isothermal data taken at different temperatures

as mentioned in the legend. aT is the horizontal shift factor and

bT is the vertical shift factor used for TTS. Slopes of 1 and 2 in G00

and G0, respectively, correspond to the terminal Newtonian

behavior. [Color figure can be viewed at wileyonlinelibrary.com]


POLYMER SCIENCE





temperature, and R is the universal gas constant. DE is a mea-sure of the potential energy barrier to translation of the mono-mer segments and should be independent of the molecularweight for a specific homopolymer above a critical molecularweight of 2–3 times the entanglement molecular weight,Me,

34–36 which is �8.8 kDa for P4MP based on the value forpoly(pentene)).37 The Arrhenius plots from the SAOS measure-ments and calculated values of DE are shown in Figure 6(a) forMX and DX. The linearity of the plots indicates an activated pro-cess for flow. Although an Arrhenius dependency was followed,two different values of DE were obtained depending upon thetemperature range. DE was 77 kJ/mol for DX from 240–270 8Cand 85 kJ/mol for MX from 230–280 8C. Above those respectivetemperatures, DE increased to 211 kJ/mol for MX and 157 kJ/mol for DX, which was attributed to the polymer degradation athigher temperatures. Although the SAOS measurements weremade within a short time span of 10–20 min, polymer degrada-tion was still possible above 270–280 8C in that time span, asindicated by the TGA data in Figure 3. It is interesting to notethat the activation energy of other polyolefins (PE, PP, LDPE,etc.) is almost half that of P4MP,38,39 that is, the melt viscosityof P4MP is highly sensitive to a temperature change.

The shear viscosity for the two grades at a reference temper-ature of 250 8C is compared in Figure 7. The steady-shearviscosity (g _cð ÞZ) was obtained from the cone-and-plate andcapillary flow measurements. The dynamic viscosity (g*(x))was calculated from SAOS measurements using the equation,

g � xð Þ5G�x

5

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiG0

x

� �2

1G00

x

� �2s

(6)

Each curve in Figure 7 is a master curve that was con-structed using time-temperature superposition (TTS) byshifting the isothermal data horizontally and vertically usingthe shift factors aT and bT, respectively. Both the polymers

displayed a plateau in the shear viscosity at low shear rates,called the zero-shear viscosity (g0), beyond which a shear-thinning behavior was observed. g0 was higher for DX (9621Pa s) compared to MX (6356 Pa s) which was consistentwith a higher molecular weight of DX. An overlap of the vis-cosity data at 250 8C in Figure 7 indicated consistencybetween the three independent rheological measurements,and validity of the Cox–Merz rule40 for P4MP. An Arrheniustemperature-dependency described by eq 4 was alsoobserved from the steady-shear measurements, and theArrhenius plots are shown in Figure 6(b). The calculatedactivation energies and the degradation temperatures for MXand DX grades were consistent with those obtained from theSAOS measurements within experimental error.

FIGURE 7 Shear viscosity master curves for MX and DX grades

at 250 8C, prepared by TTS of the isothermal data taken at dif-

ferent temperatures, as mentioned in the legend. Data are col-

lected from steady-shear, small-amplitude oscillatory shear

(SAOS), and capillary measurements. [Color figure can be

viewed at wileyonlinelibrary.com]




The rheology data for P4MP indicates that a precise temper-ature control in the extruder is required while processingP4MP in the melt state, and that the melt extrusion tempera-ture should be restricted to <270 8C. Given the viscous andelastic behavior of P4MP, its processing into thin films isclearly possible but care has to be taken to prevent any poly-mer degradation to maintain reproducibility and consistencyin the final product.

Real-Time Dynamics during Drying of Thin FilmsWhile melt-processing is a viable and easy route for prepar-ing thin P4MP films, another route is solution-processing.Solution processing can yield free-standing P4MP films asthin as 3–5 lm besides providing a better control on thethickness uniformity. Figure 8 shows the drying behavior ofa P4MP (MX) film cast from a 5 wt % cyclohexane solutionat 22 8C. The mass and thickness changes were nearly linearfor the first �300 s after casting the film [Fig. 8(a)]. The sur-face temperature dropped by 1–2 8C due to convective cool-ing by solvent diffusion, but reached back to the initial casttemperature. The film was optically clear until that time ascan be observed from the transparent circular region in Fig-ure 8(b). A pseudo “steady-state” was reached after �500 sof casting the film, where the rate of reduction in the weightand thickness dropped drastically. This was attributed to theformation of a polymer-rich skin layer at the film–air inter-face as shown by the pictographs in Figure 8(b) (note thechange in color of the circular region from transparent toopaque white). That observation was also supported by adecrease in the light transmission through the polymer filmfrom an initial 100% to 43% at 500 s. The diffusion mecha-nism at t � 400 s changed from Case I diffusion, which isgoverned by linear Fick’s first and second law, to a Case IIdiffusion, which is described by a nonlinear diffusion equa-tion also referred to as the Thomas–Windle model.41,42

The in-plane (Dn12) and the out-of-plane (Dn23) birefringencewere zero before the skin formation indicating an isotropicstate of polymer chain orientation in the cast film [Fig. 8(c)].However, during and after the skin formation, Dn12 remainednegligibly small, whereas Dn23 increased to 0.003 (intrinsicbirefringence for isotactic P4MP � 0.0143). A nonzero Dn23might suggest that the polymer chains aligned preferentially inthe thickness direction while drying. However, several factorssuch as the skin layer morphology, skin layer thickness, andthe tendency of the polymer to crystallize or phase-separatewhile drying complicate this observation.

Similar measurements as above were made for different sys-tems involving a combination of polymer concentration, cast-ing temperature, and type of solvent. The results for bothMX and DX grades were qualitatively similar. The birefrin-gence values for three different systems—P4MP (MX) filmscast with cyclohexane at 22 8C and 65 8C, and with cyclopen-tane at 22 8C—are compared in Figure 9(a). The in-planebirefringence (Dn12) for all the systems was low, but theout-of-plane birefringence (Dn23) increased upon the forma-tion of the skin. However, unlike in the case of cyclohexane

at 22 8C, Dn23 decreased after the skin formation for theother systems. The surface morphology at the skin–air inter-face observed using scanning electron microscope (SEM) isshown in Figure 9(b) for all the systems. The surface rough-ness decreased with faster drying conditions that may delaythe polymer phase separation for kinetic reasons. A lowersurface roughness also explains the higher light transmissionthrough those polymer films, as shown in Figure 9(b), due tolower scattering of light at the skin–air interface. In addition,periodic asperities at the polymer–air interface with an aver-age length-scale comparable to the wavelength of light cangive rise to form-birefringence, which may be the case inFigure 9(a). Hence, a nonzero Dn23 for P4MP film cast withcyclohexane at 22 8C was an artefact caused by the surface(skin) morphology and not due to any preferential chainalignment in the thickness direction. Overall, P4MP thin filmshad negligible anisotropy in the polymer chain orientationand could be considered isotropic. The isotropy was alsoconfirmed from wide-angle X-ray diffraction (WAXD) meas-urements discussed in the next section.

Surface Morphology and Crystal StructureBased on the results discussed in previous section, thin films(10–20 lm) of P4MP were prepared on a R2R platform byvarying casting conditions such as the temperature of casting(substrate and/or the solution temperature), type of solvent,and the initial polymer concentration in the solution. The sur-face morphology of these films was characterized using scan-ning electron microscopy (SEM). Figure 10 shows the effect ofvarying the casting conditions on the surface morphology ofP4MP films at the polymer–air interface. A smoother surfacemorphology was obtained under faster drying conditions,which was a result of the delay in phase separation of P4MPfrom the polymer solution due to kinetic reasons.

In addition to the surface morphology, P4MP films exhibiteda number of crystal structures that were studied using wide-angle X-ray diffraction (WAXD) and the diffractograms areshown in Figure 11. These diffractograms confirm the overallisotropy of solution-cast films. P4MP exists in 5 crystal modi-fications with differing crystal structure and/or chain confor-mation,38,44–49 which are summarized in Table 2. All 5modifications were obtained by varying the casting condi-tions. However, good-quality films were obtained only forcrystal modifications I, II, and III. Crystal structure and mor-phology is known to affect the dielectric properties of ferro-electric polymers such as poly(vinylidene fluoride) (PVDF)and PVDF-based copolymers.8,50 An interesting question,therefore, is the effect of surface morphology and the crystalstructure on the dielectric properties of P4MP, which is dis-cussed below.

Dielectric PropertiesA representative data for the dielectric constant (E0), dielec-tric loss (E00) and the loss tangent (tan d 5 E00/E0) of solution-cast P4MP films is shown in Figure 12. The sample used inFigure 12 is a MX film cast with cyclopentane at 25 8C; thedielectric data for other films were qualitatively similar. The


POLYMER SCIENCE


thermal transitions for this film, obtained from DSC (exo-therm not shown), were a glass-transition (Tg) of 27 8C, peakmelting point (Tm) of 225 8C, and a crystallinity (/m) of18%. The dielectric constant [Fig. 12(a)] varied from 2.8 to3.0 upon increasing the temperature from RT to 150 8C, butwas very stable as a function of frequency and temperature,

which is desirable for the capacitor application. A higher E0

than the reported value of 2.1251 could be due to the pres-ence of copolymer or some residual solvent in the polymerfilm. The absence of an upturn in E0 at lower frequencies(x< 1 Hz) indicates that the interfacial polarization (i.e.,electrode polarization (EP) or Maxwell–Wagner–Sillars

FIGURE 8 Real-time changes in physical properties during drying of the P4MP film cast from 5 wt % cyclohexane solution at 22

8C. [Color figure can be viewed at wileyonlinelibrary.com]




FIGURE 9 A comparison of the (a) in-plane (Dn12) and out-of-plane (Dn23) birefringence and (b) 08 transmission for P4MP thin films

under different casting conditions. The pictures on the right plot are SEM micrographs of the surface morphology (scale bar 5 5

lm). [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 10 The effect of varying solution-casting conditions on the surface morphology of P4MP (MX) thin films at the polymer–

air interface. A smoother surface morphology was obtained under faster drying conditions. Scale bar corresponds to 5 lm. [Color

figure can be viewed at wileyonlinelibrary.com]



(MWS) polarization) was absent in this sample and, there-fore, would not contribute to the dielectric loss.

The loss tangent [Fig. 12(c)] was restricted to <0.4% forx> 1 Hz, which is again a desirable characteristic from acapacitor film. A dispersion peak at x 5 5 s21 at RT wasattributed to the relaxation by segmental motion of polymerchains as Tg was close to RT. Upon increasing the

temperature above Tg, this dispersion peak moved to higherfrequencies, or lower relaxation times, as indicated in Figure12(c), but the losses were still limited to <0.4%. At lowerfrequencies, x< 1 Hz, the loss tangent was much higher andincreased upon decreasing the frequency and increasing thetemperature. In the same frequency range, E00 approached aslope of 21 as shown in Figure 12(b), which suggested thatthe losses occurred due to DC conduction.

FIGURE 11 Wide-angle X-ray diffractograms, and 2-D diffraction profiles, for various crystal structures of P4MP obtained under dif-

ferent solution-casting conditions. The 2-D diffraction profiles indicate an in-plane isotropy in all polymer films. [Color figure can

be viewed at wileyonlinelibrary.com]

TABLE 2 Different Crystal Modifications for P4MP and Corresponding Unit Cell Structure, Unit Cell Dimensions, and Chain

Conformation38,44–49

Crystal Modification Unit Cell Structure Unit Cell Dimensions Chain Conformation

I Tetragonal a 5 18.66 A

c 5 13.8 A

7/2 helix

II Tetragonal a 5 19.16 A

c 5 7.12 A

4/1 helix

III Tetragonal a 5 19.36 A

c 5 6.94 A

4/1 helix

IV Hexagonal a 5 22.17 A

c 5 6.69 A

–

V Hexagonal c 5 6.5 A –




Figure 13 compares the dielectric constant and the loss tan-gent as a function of temperature at x 5 100 Hz for differentP4MP films. These films were prepared under varying cast-ing conditions, and the associated crystal structures as wellas the surface morphology are indicated in Figure 13. E0 var-ied from 2.6 to 2.8 depending upon the film but the overall

variation in E0 was not large. A slight increase in E0 uponincreasing the temperature could be related to the increasedpolymer chain mobility at higher temperatures. The dielec-tric losses were restricted below 0.4%, and decreased uponincreasing the temperature due to a shift of the dispersionpeak to higher frequencies at a higher temperature (Fig.

FIGURE 12 (a) Dielectric constant, (b) dielectric loss, and (c) loss tangent as a function of frequency and temperature for P4MP

(MX) thin film cast from cyclopentane at 25 8C. [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 13 Dielectric constant and loss tangent as a function of temperature at x 5 100 Hz for various P4MP thin films. The crystal

modification and the surface morphology of the film for each data are indicated with arrows. Crystal modifications II and III

undergo an irreversible transition to a more stable crystal modification I upon annealing at 80 8C and (55–73) 8C, respectively.52,53

[Color figure can be viewed at wileyonlinelibrary.com]


POLYMER SCIENCE




12(c)]. For P4MP, an irreversible crystal–crystal transitionoccurs from modifications II and III into a more stable modi-fication I upon thermal annealing at 80 8C and (55–73) 8C,respectively52,53 which essentially changes the polymer chainconformation from a 4/1 to a 7/2 helix without changingthe tetragonal crystal structure (Table 2). A change in thecrystal modification did not influence the dielectric constantor the loss. This result was not very surprising consideringthat P4MP is a nonpolar polymer with an overall isotropicdipole orientation in contrast to PVDF, for example, where achange in the polymer chain conformation between differentcrystal modifications (a, d, c, b) causes a change from a non-polar (a) to a ferroelectric behavior (d, c, b).8 Unfortunately,good-quality P4MP films containing crystal modifications IVand V could not be obtained to compare the effect of tetrago-nal versus hexagonal unit crystal structure on the bulkdielectric properties in P4MP.

The breakdown strength variation in P4MP films wasdescribed by the Weibull distribution shown in Figure 14with the solid lines being a nonlinear least squares fit of eq2 to the experimental data. The values of Weibull breakdownstrength (g) and the Weibull slope parameter (b) along withthe film characteristics are summarized in Table 3. Film sur-face morphology was the most important factor that affectedthe breakdown field strength. In general, g was higher forP4MP films with a more uniform surface morphology whichis clearly evident from the comparison in Figure 15 for threefilms. P4MP films prepared by compression-molding withoutany smooth substrate possessed the least uniform surfacemorphology and, therefore, had the lowest breakdownstrength. Compression-molding with a smooth KaptonVR (pol-yimide) substrate reduced some of the surface defects, butfew defects such as pits and cavities (encircled in the micro-graph) still remained. Therefore the breakdown strengthslightly improved. Solution-casting produced films with the

most uniform surface morphology and g for those films wasalmost three times that of compression-molded films withouta substrate. This shows that uniformity, and not flatness(height variations), of the surface morphology controlled thebreakdown strength of P4MP thin films. While g increasedfor the solution-cast films, b decreased compared to thecompression-molded films indicating a wider dispersion inthe breakdown data. However, simultaneous increases in theminimum and maximum breakdown strength of the solution-cast films suggested that there is vast potential for increas-ing b by further improving the film quality. One method toachieve this, for example, would be to eliminate any externalcontamination by preparing films in a controlled environ-ment such as in a Clean Room. Overall, even though thesolution-casting improved the breakdown strength of P4MPfilms compared to compression-molding, their characteristicbreakdown strength was still one-half compared to that ofBOPP (gBOPP � 700 V/lm) and the slope parameter waseven lower (bBOPP � 15).54

It should be noted that thickness of the solvent-cast filmswas smaller than the compression-molded films, which itselfcould result in a higher breakdown strength of solvent-castfilms due to less pronounced volume effects.55–57 However,in this study, it was not possible to decouple the surface andvolume effects in P4MP films.

Similar to our observations from Figure 13, the type of crys-tal structure had no effect on the breakdown field strength.Interestingly, percent crystallinity also did not influence thebreakdown field strength as observed for compression-molded versus solution-cast films and MX versus DX grades.For example, MX film cast from a 6 wt % cyclopentane solu-tion at 25 8C had a similar Weibull breakdown strength as aDX film cast from 4 wt % cyclopentane solution at 25 8Ceven though their crystallinities were 18% and 33%, respec-tively. This result is contrary to the observations made forother polyolefins such as polypropylene58 or polyethylene,59

where the breakdown field strength is observed to increasewith increasing crystallinity and that is also consistent withthe predictions from the free-volume theory60,61 of dielectricbreakdown in polymers. The solution-cast films with lowercrystallinity also had a smoother surface morphology indicat-ing that the two were coupled, that is, faster drying condi-tions that kinetically inhibited phase-separation of P4MPfrom its solution producing a smother surface morphologyalso inhibited regular chain folding into a crystal structureresulting in a lower crystallinity.

Breakdown-Induced Surface Morphology and Mode ofBreakdownExamination of the surface morphology after the occurrenceof dielectric breakdown revealed an interesting phenomenonthat occurred during and after the breakdown. Figure 16compares the surface morphology of P4MP thin films thatwere prepared by compression molding at 260 8C without asubstrate (a–c) and solution casting at 25 8C with cyclopen-tane (d–f). The breakdown area appeared to be closer to a

FIGURE 14 Weibull plot for the dielectric breakdown strength

of P4MP (MX) thin films prepared by compression-molding

without a Kapton substrate at 260 8C and solution-casting with

cyclohexane at 25 8C. The solid lines are nonlinear least

squares fit of eq 2 to the experimental data. The dotted lines

represent 90% confidence limits. [Color figure can be viewed at

wileyonlinelibrary.com]




TABLE 3 Summary of the Weibull Breakdown Strength and the Slope Parameter for P4MP Films Along With Their Crystallinity,

Crystal Structural, and Surface Morphology

Film Preparation Sample

Crystal

Structure

Surface

Morphology

Crystallinity

/m (%)

Breakdown Strength (V/mm)

Weibell Slope

Parameter b

Minimum

Eds,min

Maximum

Eds,max Weibell g

Compression-

molding

MX

260 8C

No substrate

I 23 64 105 93 10.3

Compression-

molding

MX

260 8C

Kapton

I 23 95 258 180 11.3

Solution-casting MX

Cyclohexane

50 8C

III 21 84 365 309 5.5

Solution-casting MX

Cyclohexane

35 8C

III 20 196 405 355 3.2

Solution-casting MX

Cyclohexane

25 8C

III 21 134 390 312 4.0

Solution-casting DX

Cyclohexane

25 8C

III 38 83 330 239 2.0

Solution-casting MX

Cyclopentane

35 8C

II 17 48 407 193 1.2

Solution-casting MX

Cyclopentane

30 8C

II 19 72 535 311 1.9

Solution-casting MX

Cyclopentane

25 8C

II 18 206 445 362 3.5

Solution-casting DX

Cyclopentane

25 8C

II 33 146 408 336 4.3

circle rather than depicting a lateral tree growth as has beenobserved for several polymers. In general, it was observedthat the overall breakdown coverage area was higher for thefilm with a lower Weibull breakdown strength. Magnificationof the surface features in Figure 16(b,e) revealed fractals witha fourfold (tree fractals) or a sixfold symmetry (snow fractals)which corresponds well with the tetragonal or the hexagonalshape of the crystalline unit cell of P4MP, refer Table 2. Thebranches [shown in yellow color in Fig. 16(c,f)] emanatingfrom the trunk [shown in green color in Fig. 16(c,f)] also fol-lowed a similar symmetry as the trunk, that is, the growthsymmetry was hierarchal so that the surface features resem-bled a dendritic pattern. Such self-similar fractal growth hasbeen attributed to a diffusion-limited process.62

An explanation for the occurrence of such features is themelting of polymer at the surface in contact with the elec-trode due to extreme heat and high temperatures producedduring the electrical breakdown, which is followed by slowand homogeneously nucleated recrystallization of the poly-mer. A slow and homogeneous growth of the crystals at thesurface is also supported by the observation from Figure 16that these crystals are nonpervading over the breakdownarea and consistent with the diffusion-limited growth of suchfractals.62

While the growth of crystal modification I upon cooling fromthe melt state was expected as it is the most stable among

all five modifications, the growth of crystal modifications IVor V is not obvious. It is interesting to note that Hasegawaet al.49 have reported a transition from crystal modification I(tetragonal crystal structure) ! IV (hexagonal crystal struc-ture) for P4MP at temperatures above 200 8C and a pressureof 456 MPa. It is also known that the local temperature riseat the breakdown site can achieve values as high as1000 8C56 which is much higher than the melting point ofthe polymer (224–235 8C). Considering the above data, itappears that the electrical breakdown in P4MP films is notonly accompanied by high temperatures but also high pres-sures, which suggests that the breakdown is electromechani-cal in nature. Additionally, data from Figure 16 indicate thatthe mode of electrical breakdown is same for all the filmseven though the breakdown strength differs based on theprocessing method.

Effect of Biaxial Stretching on Dielectric PropertiesPolymer chain orientation through uniaxial or biaxial stretch-ing affects several polymer properties such as mechanical,thermal, optical, and dielectric and induces anisotropy inthese properties.63–68 Polymer film orientation is known toimprove the dielectric breakdown strength of polymers dueto several factors such as an enhancement in crystallinity,increase in the amorphous phase density, improvement inmechanical properties such as the tensile modulus, andreduction in the surface defects and thickness varia-tions.15,58,59,69 Therefore, the P4MP thin films prepared by

FIGURE 15 SEM surface micrographs of P4MP (MX) films prepared under different casting condition along with their crystallinity,

Weibull breakdown strength (g) and Weibull slope parameter (b). These micrographs show that a uniform surface morphology

with minimum defects (such as pits, microcrevices, cavities, bumps, etc.) increases the breakdown field strength. [Color figure can





solution-casting were also biaxially stretched to evaluate itseffect on the dielectric properties. Figure 17 shows thedielectric constant (E0) and loss tangent (tan d) for P4MP(MX) films that were solution-cast with cyclopentane at25 8C and then biaxially stretched by an elongation ratio(k15 k2) of 1.3 and 1.6 at 50 8C. An ideal temperature forbiaxial stretching of the MX film would be T>Ta* (5200 8Cfrom Fig. 4), but such a high temperature was not possiblewith our current instrumental set-up. The dielectric constantreduced from a value of �3 for the solution-cast film (Fig.12) to a value of �2 for the stretched films. All films werecompletely dried before TDDS measurements, and the TGAmeasurements did not detect the presence of any moistureor solvent in these films, hence the influence of residual sol-vent is less likely. The dielectric constant was very stable asa function of frequency and temperature as high as 150 8C

and the dielectric losses were restricted to <1–2% above1 Hz for the biaxially stretched films, which are the desiredproperties for the capacitor application.

Although the low-field dielectric properties remainedunchanged upon biaxially stretching the P4MP thin films, sig-nificant effect was observed on the breakdown field strength.Figure 18 compares the Weibull breakdown characteristics,surface morphology and crystallinity of P4MP (MX) filmsolution-cast with cyclopentane at 25 8C, and the same filmafter biaxial stretching by an elongation ratio (k1 5 k2) of 1.3and 1.6 at 50 8C. The Weibull breakdown strength (g)increased for the biaxially stretched films compared to theunstretched film, although not much effect was observed onthe Weibull slope parameter. Furthermore, g increased withan increase in the elongation ratio and an increase of 65%

FIGURE 16 Surface morphology after dielectric breakdown of P4MP (MX) thin films prepared by (a–c) compression-molding at 260

8C without substrate [corresponding to Fig. 15(a)] and (d–) solution-casting at 25 8C with cyclopentane [corresponding to Fig.

15(c)]. The green lines in (b, f) shows the growth of the trunk (primary process) and yellow lines in (c, f) shows the growth of the

branches (secondary process). [Color figure can be viewed at wileyonlinelibrary.com]


POLYMER SCIENCE



was obtained at an elongation ratio of only 1.6. The break-down strength is expected to increase even more at higherelongation ratios. From Figure 18, the surface morphology ofthe biaxially stretched film appeared coarser but more uni-form than the solution-cast film. Considering that the electri-cal breakdown is a localized phenomenon, one hypothesisfor an increase in the dielectric breakdown strength of biaxi-ally stretched films over solution-cast films is the improve-ment in their surface morphology by elimination oftopological defects such as pits or cavities, which otherwisemay cause a nonuniform localized concentration of electric

field around the defects and, hence, precipitate breakdownat lower voltages.15,69 The same factor is responsible for ahigher breakdown strength of solution-cast films comparedto compression-molded films as shown in Table 3.

As the energy density of a film capacitor is directly propor-tional to square of the breakdown strength of the dielectricfilm, our results suggest that solution-processing a P4MPfilm followed by biaxial stretching can increase the energydensity by a factor of 20 over a conventional melt-processedfilm.

FIGURE 17 Dielectric constant and loss tangent as a function of frequency and temperature for P4MP (MX) film cast from cyclo-

pentane at 25 8C and biaxially stretched at 50 8C by an elongation ratio of (a) 1.33 3 1.33 and (b) 1.63 3 1.63. [Color figure can


FIGURE 18 SEM surface micrographs of solution-cast and biaxially stretched P4MP films along with their crystallinity, Weibull

breakdown strength, and Weibull slope parameter. [Color figure can be viewed at wileyonlinelibrary.com]





Quality of Base ResinThe effect of cleanliness or purity of the base resin on thebreakdown strength of polypropylene is well known. Resid-ual catalyst, residual monomer, and contaminants present inthe base resin or introduced during the manufacturing pro-cess result in a significant decrease in the dielectricstrength.15,70 On the other hand, producing a base resin thatmeets the stringent purity requirement of a capacitor film isdifficult and expensive. Current state-of-the-art thin filmmanufacturing for dielectric capacitor involves using a highpurity “capacitor-grade” base resin along with processingand metallization of films in a clean environment. Acapacitor-grade BOPP film exhibits a higher Weibull break-down strength (g) and slope parameter (b) than apackaging-grade BOPP film. A Weibull strength of 514 V/lmfor a biaxially oriented P4MP (BOP4MP) film in Figure 18 iscomparable to that of a BOPP film made from a packaging-grade base resin.15,70 Therefore, a further increase in thedielectric strength of BOP4MP is expected upon switching toa capacitor-grade resin, and BOP4MP could be a viable high-temperature alternative to state-of-the-art BOPP capacitorfilms.

CONCLUSIONS

Studies were conducted to assess the viability of poly(4-methyl-1-pentene) (P4MP) as a suitable candidate for high-temperature capacitor applications. Thermal stability to deg-radation at high temperatures (>200 8C), melting onset at190 8C and a high thermal transition for crystallite motion(200–216 8C) suggested that P4MP thin films should with-stand operating temperatures up to at least 190 8C in dielec-tric capacitors. Investigation of the rheological behavior ofP4MP indicated that it could be melt-processed below270 8C in the extruder without degradation. In addition tomelt processing, P4MP offers a much higher flexibilitytoward thin film manufacturing by solution processing dueto favorable solubility characteristics, which provides anadvantage over other polyolefins such as PP. Isotropic, free-standing films as thin as 10 lm were successfully obtainedusing different solvents and casting temperatures in thisstudy, although the thickness could be easily reduced to 3–5lm. The solution processing route also enabled a control ofthe film surface morphology, percent crystallinity and differ-ent crystal structures by controlling the casting conditions.The surface morphology had the biggest influence on thedielectric strength of P4MP films, while the unit crystalstructure or the percent crystallinity had a negligible influ-ence. The mode of electrical breakdown in both melt-processed and solvent-cast films appeared to be electrome-chanical. Biaxial stretching of the solution-cast films furtherincreased the breakdown strength due to improvement inthe surface morphology. Compared to the melt-processedfilms, the Weibull breakdown strength increased by 290%for solution-processed films, and 450% for solution-processed films that were biaxially stretched to an elonga-tion ratio of 1.63. This implies that an increase in theenergy density by up to 20 times over conventional melt-

processed films could be theoretically obtained. In addition,P4MP has a dielectric constant and loss similar to PP andvery stable as a function of frequency and temperature. Ithas been recently reported that ionic functionalization ofP4MP can further increase the dielectric constant and break-down strength while maintaining low dielectric losses.71

Therefore, a similar or an even higher energy density thanBOPP can be obtained from biaxially oriented P4MP filmswith an added advantage of operating at much higher tem-peratures than PP.

ACKNOWLEDGMENTS

Financial support for this work through a MultidisciplinaryUniversity Research Initiative (MURI) grant from the Office ofNaval Research (Contract No. N00014–10-1–0944) is greatlyappreciated. The authors are thankful to Mr Yasuhiro Higuchiat Mitsui Chemicals for providing the polymer resins, Dr XuepeiYuan at the Pennsylvania State University for performing themolecular weight measurements, and Franck Jacobs at BorealisPolymer NV for useful discussions.

REFERENCES AND NOTES

1 W. J. Sarjeant, R. E., Dollinger, J. Zirnheld, F. W.

MacDougall, H. Goldberg, In 22nd International Power Modula-

tor Symposium 1996; 209–212.

2 M. Winter, R. J. Brodd, Chem. Rev. 2004, 104, 4245–4269.

3 W. J. Sarjeant, J. Zirnheld, F. W. Macdougall, IEEE Trans.

Plasma Sci. 1998, 26, 1368–1392.

4 H. D. Abru~na, Y. Kiya, J. C. Henderson, Phys. Today 2008, 61,

43–47.

5 X. Yuan, Y. Matsuyama, T. C. M. Chung, Macromolecules

2010, 43, 4011–4015.

6 B. Chu, X. Zhou, K. Ren, B. Neese, M. Lin, Q. Wang, F. Bauer,

Q. M. Zhang, Science (80-.). 2006, 313, 334–336.

7 S. Zhang, B. Zellers, D. Anderson, P. Rehrig, X. Zhou, D.

Kushner, R. Russo, AIR FORCE Res. Lab. Rep. AFRL-RZ-WP-TP-

2010–2126. 2010.

8 L. Zhu, Q. Wang, Macromolecules 2012, 45, 2937–2954.

9 A. Boufelfel, Storming Media: Pentagon Reports TR-04–15

2006.

10 S. Zhang, C. Zou, X. Zhou, D. Anderson, B. Zellers, Q.

Zhang, In IEEE International Power Modulator and High Volt-

age Conference 2010; 221–224.

11 G. Picci, M. Rabuffi, IEEE Trans. Plasma Sci. 2002, 30, 1939–

1942.

12 C. W. Reed, S. W. Cichanowski, IEEE Trans. Dielectr. Electr.

Insul. 1994, 1, 904–922.

13 C. C. Wang, R. Ramprasad, J. Mater. Sci. 2011, 46, 90–93.

14 J. Ho, R. Jow, Army Res. Lab. Rep. ARL-TR-4880. 2009,

15 J. L. Nash, Polym. Eng. Sci. 1988, 28, 862–870.

16 Y. Budianto, A. Aoki, T. Miyashita, Macromolecules 2003,

36, 8761–8765.

17 J. J. Martinez-vega, N. Zouzou, J. Guastavino, L. Boudou, In

IEEE 7th International Conference on Solid Dielectrics 2001;

171–174.


POLYMER SCIENCE


18 W. J. Sarjeant, In 19th Electrical Electronics Insulation Con-

ference 1989; 1–51.

19 C. K. Landrock, B. Kaminska, In Design, Automation & Test

in Europe Conference & Exhibition 2010; 1349–1352.

20 G. Crass, L. Bothe, G. Schloegl, US 4692837 1987.

21 T. Yoshida, US 2010/0062239 A1 2010.

22 J. E. Mark, Physical Properties of Polymers Handbook, 2nd

ed.; Springer: New York, 2007.

23 P. Zoller, H. W. Starkweather, G. A. Jones, J. Polym. Sci. B

Polym. Phys. 1986, 24, 1451–1458.

24 N. G. McCrum, B. E. Read, G. Williams, Anelastic and

Dielectric Effects in Polymeric Solids; John Wiley and Sons:

New York, 1967.

25 C. D. Han, Rheology and Processing of Polymer Materials.

Vol. I Polymer Rheology; Oxford University Press: New York,

2007.

26 W. Hauschild, W. Moss, Statistic Techniques for High Volt-

age Engineering; Peter Pergrinus Ltd.: London, 1992.

27 E. Unsal, J. Drum, O. Yucel, I. I. Nugay, B. Yalcin, M.

Cakmak, Rev. Sci. Instrum. 2012, 83, 025114. 1–10).

28 M. Cakmak, M. Hassan, E. Unsal, C. Martins, Rev. Sci. Ins-

trum. 2012, 83,

29 S. Gupta, X. Yuan, T. C. M. Chung, S. Kumar, M. Cakmak, R.

A. Weiss, Macromolecules 2013, 46, 5455–5463.

30 S. Gupta, X. Yuan, T. C. M. Chung, M. Cakmak, R. A. Weiss,

Polymer (Guildf) 2014, 55, 924–935.

31 J. Brandrup, E. H. Immergut, E. A. Grulke, Polymer Hand-

book, 4th ed.; John Wiley and Sons: New York, 1999.

32 J. D. Ferry, Viscoelastic Properties of Polymers, 3rd ed.;

John Wiley and Sons: New York, 1980.

33 S. Glasstone, K.J. Laidlev, H. Eyring, The Theory of Rate

Processes: The Kinetics of Chemical Reactions, Viscosity, Diffu-

sion and Electrochemical Phenomena, 1st ed.; McGraw-Hill:

New York, 1941.

34 J. Wang, R. S. Porter, Rheol. Acta 1995, 34, 496–503.

35 K. Jordens, G. L. Wilkes, J. Janzen, D. C. Rohlfing, M. B.

Welch, Polymer (Guildf) 2000, 41, 7175–7192.

36 R. G. Mayrick, C. L. Sieglaff, J. Appl. Polym. Sci. 1965, 9,

3917–3927.

37 L. J. Fetters, D. J. Lohse, W. W. Graessley, J. Polym. Sci. B

Polym. Phys. 1999, 37, 1023–1033.

38 L. C. Lopez, G. L. Wilkes, P. M. Stricklen, S. A. White, J.

Macromol. Sci. C 1992, 32, 301–406.

39 S. Gupta, X. Yuan, T. C. M. Chung, M. Cakmak, R. A. Weiss,

Polymer (Guildf) 2016, 107, 223–232.

40 W. P. Cox, E. H. Merz, J. Polym. Sci. 1958, 28, 619–622.

41 N. L. Thomas, A. H. Windle, Polymer (Guildf) 1982, 23, 529–

542.

42 M. Souche, D. Long, Europhys. Lett. 2007, 77, 48002.

43 C. H. Choi, J. L. White, J. Appl. Polym. Sci. 2005, 98,

130–137.

44 G. Charlet, H. Phuong-nguyen, G. Delmas, Macromolecules

1984, 17, 1200–1208.

45 F. C. Frank, A. Keller, A. O’Connor, Philos. Mag. 1959, 4,

200–214.

46 M. Litt, J. Polym. Sci. A Polym. Chem. 1963, 1, 2219–2224.

47 P. Zoller, J. Appl. Polym. Sci. 1977, 21, 3129–3137.

48 L. A. Novokshonova, G. P. Berseneva, V. I. Tsvetkov, N. M.

Chirkov, Vysok. Soedin. 1967, 9, 562.

49 R. Hasegawa, Y. Tanabe, M. Kobayashi, H. Tadokoro, A.

Sawaoka, N. Kawai, J. Polym. Sci. B Polym. Phys. 1970, 8,

1073–1087.

50 F. Guan, J. Pan, J. Wang, Q. Wang, L. Zhu, Macromolecules

2010, 43, 384–392.

51 J. E. Mark, Polymer Data Handbook; Oxford University

Press: New York, 1999.

52 A. Nakajima, S. Hayashi, T. Taka, Kolloid-Zeitschrift Zeits-

chrift Polym. 1969, 233, 869–878.

53 A. Nakajima, S. Hayashi, T. Taka, N. Utsumi, Kolloid-Zeits-

chrift Zeitschrift Polym. 1969, 234, 1097–1108.

54 C. Xu, J. Ho, S. A. Boggs, IEEE Electr. Insul. Mag. 2008, 24,

30–34.

55 J. H. Mason, IEEE Trans. Electr. Insul. 1992, 27, 1061–1062.

56 S. Diaham, S. Zelmat, M. Locatelli, S. Dinculescu, M. Decup,

T. Lebey, IEEE Trans. Dielectr. Electr. Insul. 2010, 17, 18–27.

57 G. Chen, J. Zhao, S. Li, L. Zhong, G. Chen, J. Zhao, S. Li, L.

Zhong, Appl. Phys. Lett. 2012, 222904,

58 L. Y. Gao, D. M. Tu, S. C. Zhou, Z. L. Zhang, IEEE Trans.

Electr. Insul. 1990, 25, 535–540.

59 K. Yahagi, IEEE Trans. Electr. Insul. 1980, EI-15, 241–250.

60 J. Artbauer, J. Polym. Sci. Polym. Symp. 1967, 16, 477–484.

61 J. Artbauer, J. Phys. D Appl. Phys. 1999, 29, 446–456.

62 T. Vicsek, Fractal Growth Phenomena; World Scientific Pub-

lishing Co.: Singapore, 1992.

63 I. M. Ward, Proc. Phys. Soc. 1962, 80, 1176.

64 S. Gupta, J. D. Schieber, D. C. Venerus, J. Rheol. (N. Y. N.

Y). 2013, 57, 427.

65 J. D. Schieber, D. C. Venerus, S. Gupta, Soft Matter 2012, 8,

11781.

66 H. Janeschitz-Kriegl, Polymer Melt Rheology and Flow Bire-

fringence; Springer Verlag: New York, 1983.

67 D. Boese, H. Lee, D. Y. Yoon, J. D. Swalen, J. F. Rabolt, J.

Polym. Sci. B Polym. Phys. 1992, 30, 1321–1327.

68 I. Offenbach, S. Gupta, T. C. M. Chung, R. A. Weiss, M.

Cakmak, Macromolecules 2015, 48, 6294–6305.

69 I. Galperin, W. White, IEEE Trans. Electr. Insul. 1985, EI-20,

55–59.

70 F. Jacobs, In Carts Europe (Passive Component Sympo-

sium) 2006; 27–34.

71 M. Zhang, L. Zhang, M. Zhu, Y. Wang, N. Li, Z. Zhang, Q.

Chen, L. An, Y. Lin, C. Nan, J. Mater. Chem. A. 2016, 4, 4797–

4807.



本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具

http://www.xuebalib.com/cloud/

http://www.xuebalib.com/

http://www.xuebalib.com/cloud/


http://www.xuebalib.com/vip.html

http://www.xuebalib.com/db.php

http://www.xuebalib.com/zixun/2014-08-15/44.html


evaluation of poly(4-methyl-1-pentene) as a dielectric

Documents