JOURNAL OF RESEARCH of the National Bureau of Standards Vol . 84, No.6, November- December 1979

Room Temperature Poling of Poly(Vinylidene Fluoride) with Deposited Metal Electrodes*

James M. Kenney and Steven C. Roth

Center for Materials Science, National Bureau of Standards, Washington, DC 20234

May 21, 1979

High va lues of pyroelectri c a nd pi ezoelect ric activit y in full y-e lectroded films of poly(vin ylide ne fluoride)

were obta ined by "conventi ona l" (non-corona) poling at room te mpe rature with suffic ie ntl y high field s . The

avoida nce of breakdow n whil e obta in ing hi gh activit y requires a n understa ndin g of the time de pe nde nce of both

breakdow n a nd acti vit y. Time-t o-breakdown as a fun cti on of fi e ld , a nd roo m-te mperature pyroelectri c acti vit y

(7- 14 days a ft er poling at 23.5 °C) as a fun cti on of poling time a nd fi eld were obtained for 25-,."m biaxi a ll y

stretched film s with evaporated a luminum electrodes. The hi ghest acti vities we re obta ined by poling a t the hi ghest

fi elds a nd poling to breakdow n. A pyroelec tri c ac ti vit y of 36 ,."CK- 1m- 2 was measured two weeks a fte r poling with

a nominal fi e ld of 550 MVm- 1 for a nominal po ling time of 10 s (reduced by multiple breakdow n). The hi ghest

ac ti vity obta ined with no a ppa rent breakdown (3 1 ,."CK- 1m- 2) was measured a week aft er poling with a fi eld of 400

MVm- 1 for 10 s . These values a re compa rable with the highest tha t have been report ed for thi s ma te ria l us ing a ny

poling te mperature or us ing coro na po ling.

Key words: Di electric stre ngt h; electrets; pi ezoelectri c it y; pola ri zati on; poling proced ures; polymers;

poly(vin ylide ne fluoride); pyroe lectri c it y; tra nsd ucers.

1. Introduction

Man y authors have sought , with varying degrees of suc­cess, to account for th e pyroelec tri c and piezoelectri c prop­eI1i es of poly(vinylide ne fluori de) (PVDF). In the recent mod el by Broad hurst, et al. [1] , t th ese properti es are directl y proporti onal to remanent polarizati on. The ac hi eve ment of a give n degree of polarization , however, depends in a complex way, not ye t understood , upon a number of variables assoc i­ated with the poling process. Optimization of thi s process to achieve pyroelectric and piezoelec tri c ac tiviti es tha t are both large and stable are therefore entirely empirical at this time.

One of the major obstac les in the development of poling techniques has been a strong tendency toward generalization based upon inadequate data. In particular, the following statements have in the past been commonly accepted:

1) An elevated temperature during poling is essential for high acti vity.

2) Poling is virtuall y complete after about 30 minutes. 3) Because form II (a-phase) c rys tallites are not pola r they

the refore cannot contribute to pyroelectri c or pi ezoelectri c ac tivity.

* This pape r was also a Final RerKU1 , to the Naval Ocean Systems Center, San Diego, Catifornia 92152, dated Sept e mber 30, 1978 entitled "Fabrication ofPiezoelec· tnc Polymer Film. "

1 Figures in brackets indicate lite rature refere nces at the end of this paper.

4) " ... breakdow n of elec troded samples occurs at room

temperature before fi elds hi gh enough to pole the mate rial can be reached [2]. "

The " traditional" requireme nt of an elevated po ling te m­perature has been la rgely de mol ished by the sll ccess of corona poling [2, 3]. Earl y s tudi es (4-6] of the e ffect of poling time seemed to indicate a saturati on of ac ti vity with time after 30 minutes or so, but Blev in [7] has shown how activity continues to increase (at 100 MVm- l ) over a long peri od , with faster poling rates a t hi gher tempera ture . X-ray studi es [3] of corona poled PVDF have indicated th at form II (a-phase) crystallites can be converted to a polar form , since the molecular cha in does have a dipole moment. That full y­electroded PVDF can be "conventi onall y" poled to high activity at room tempera ture was first reported , to our knowledge, by Sharp and Gam [8], who obta ined 25 fLCK- tm - 2 with a fi eld of 300 MVm- t (in 6-fLm biaxiall y

oriented PVDF with ni chrome elec trodes). A fuller explora­ti on of the room temperature poling of fully-electroded samples is reported here.

2. Procedures

Initial attempts to increase the pyroelectric and piezoelec­tric activity of PVDF made apparent the importance of electrical breakdown as the limiting factor. Flashover to the walls of a brass sample cell used for elevated temperature

447

poling, and large thermal inertia of the cell, were experimen­

tal difficulties which were eliminated with room temperature

poling, for which no cell was needed. Large unelectroded margins of PVDF around the electrodes, to prevent edge flashovers, and the use of grease or oil over and around the

electrodes, permitted very high voltages to be applied, resulting in higher pyroelectric and piezoelectric activity

than was generally obtained by elevated temperature poling. Breakdown, however, appeared erratic, making difficult the

achievement of high activity in undamaged samples. It became apparent that the breakdown at a given voltage

depended on the time of application of that voltage, so a

study was made of time-to-breakdown as a function of field. Circular electrodes, generally aluminum 100 nm (1000 A)

thick and 37 mm in diameter, were vacuum deposited by evaporation onto 25 J.Lm (nominal) films of biaxially stretched

(blow molded , as supplied) "capacitor grade" PVDF from Kureha Chemical Co. , Ltd. 2 A large "tab" extension of each

electrode allowed tinfoil leads to be attached with silver­

filled rubber cement. Orientation of the tabs 1800 apart kept them out of the field, so that self-healing of the electrodes by evaporation of the aluminum during breakdown would not be inhibited by the attached leads.

A grounded guard ring, 136 mm inner diameter, was

applied to the film around one electrode in order to intercept surface leakage. One electrode was connected to ground via

a combination current and charge meter (designed for this work to have very low drift) which drove a dual-channel chart

recorder. To avoid instantaneous sample breakdown initiated by

external flashover, it became necessary to immerse the

samples in a light paraffin oil (Saybolt viscosity 125/135) when poling at high fields. The samples were later cleaned with methanol.

Following the breakdown study, other samples were poled with a wide range of fields for specific lengths of time using

a microprocessor-type controller to program the poling volt­age source. As one second was the minimum ramp time for this controller, the shortest poling time used was ten seconds at the nominal (constant) field. A different sample was used for each combination of poling time and field. Pyroelectric specimens of 28 mm diameter were punched from the poled

samples and stored within folded tinfoil prior to their measurement between 7 and 14 days after poling. When

breakdowns occurred, they almost invariably took place at the edges of the electrodes, except after many breakdowns

(when damage progressed inward). The smaller diameter of the punc hed area, therefore, allowed the areas damaged by breakdown to be excluded from measurements. The pyroelec­

tric activity of the punched spec imens was determined by

measuring the (reversible) current resulting from ramping the

sample temperature over a range of about two kelvin at close

to room temperature, as described by Broadhurst , et al. [9]. A small irreversible component of sample current was

compensated for by measuring currents for both increasing and decreasing temperatures and taking the difference be­tween these at the same temperature . The assumption that

the small temperature cycle did not alter the irreversible

current was verified by comparing the pre-cyc ling and post­cycling currents of several samples at the same steady temperature .

Previous to these measurements a number of samples,

poled at room temperature at 100 MVm- 1 and 200 MVm- J ,

had been measured for pyroelectric activity using the same current amplifier and also using a low-drift charge amplifier. These samples were also measured [9] for hydrostatic piezoe­

lectric activity in the same sample cell, using the charge

amplifier, with the cell temperature monitored to permit a correction for pyroelec tric charge .

3. Results

As time did not permit the measurement of piezoelectric

activity for most samples, it is useful to have a conversion

factor between the pyroelectric coefficient, PY' and hydro­static piezoelectric coefficient, dp , for this particular mate­rial, poled and measured at room temperature. As shown by

figure 1, the ratio of dp to py is quite consistent to within the

precision of measurement. The lower group of points repre­sent samples poled at 100 MVm-l, and the upper group

~ 10 :> i= () <t ()

a: t> ~ 5 w o N w ii:

10 20 30

PYROELECTRIC ACTIVITY - ... CK-'m-2

F IGURE 1. Determination of the ratio between hydrostatic piezoelectric

2 Identificat ion of a commercial product is made only to facilitate experimental activity and pyroelectric activity for samples poled arui measured at room reproduc ibility and does not imply endorsement by NBS. temperature.

448

samples poled at 200 MVm- l, with poling times ranging from 50 s to 3 h. The average ratio of dp to py is 0.478 ,uKPa- l, with a sample standard deviation of 0.0157. Dropping the single most deviant datum (which also represented the lowest activity), the average ratio is 0.474 ,uKPa- l , with a sample standard deviation of 0.0081. Th ese ratios are in good agreement with the predicted value [1] of 0.5 ,uKPa- l. It must be noted, however, that subsequent poling was done at much higher fi elds. Virtually identical values of pyroelectric activity were obtained by measuring charge as by measuring current.

The breakdown data for aluminum-electroded samples is shown in figure 2; most of the data fit a straight line with a slope of 100 MVm- 1 per time decade. At the lowest fi eld strengths there is an upward deviation, as would be expected , but the testing was not continued to field s low enough to closely approach a vertical asymptote (limit of time depen­dence). A common line seems to fit the data for both air­poled and oil-poled samples, although the latter appear to have a slightly longer time to breakdown at the same fi eld . Above 420 MVm- 1 breakdown , even under oil , appeared to be instantaneous, probably occurring in most cases before the nominal voltage was reached. Subsequent poling of other samples for fixed times shorter than the breakdown times of figure 2 resulted in some rapid bre akdowns at fi elds below

en o z o u w en

z ~ o o ~ <l: w a:: IC

o I­w :2; i=

LIM ITS OF EXPANDED CHART

----1/ I I I I I I 1 1 I I

0- POLED IN AIR \ t,. - POLED IN OIL \ 25 I'm PVDF \ 23 V, DC \

I \

100~----1~00~--~2~00~--~3OO~--~400~~1~~=

POLING FIELD · MVm "

1 WEEK

1 DAY

1 HOUR

1 MINUTE

fiG URE 2. Time to breakdown as a function of field strength.

420 MVm- 1, perhaps due to contamination of the oil from previous breakdowns or from exposure of the oil to the air. Improved techniques, such as oil degassing, or the use of a more su itable type of oil, may extend the straight portion of figure 2 to higher fi elds or at least provide better consistancy.

A number of samples with electrodes of metals other than aluminum were poled to breakdown in air at 300 MV m - I.

The results are shown in figure 3, which is an enlargement of the area outlined near the center of figure 2. Four aluminum­electroded samples were also poled to breakdown at this field, all of which, including one poled in oil, had unusually short breakdown times. There appears to be too much scatter to generalize about these electrode differences with regard to breakdown time. With regard to current during poling, however, the results are more conclusive, as shown by figure 4. As the field was first applied there was a large c urrent which gradually dropped to an almost constant value, which is plotted in figure 4, before gradually ris ing monotonicall y until breakdown occurred. The currents at breakdown and at fixed intervals after applying the fi eld were also noted, but did not provide any more prec ise predicti on of time to breakdown than does figure 2. Comparing work -functions [10] with the ordering of metals in figure 4, the higher currents correspond to the larger work-functions, except for gold whi ch is badly out of sequence (bu t which does not bond

1~r----.-----r----'----'-----r----.-----'

en o z o u w en

z ~ o o ~ <l: w a:: IC

~ w :2; i=

o

0- POLED IN AIR

t. - POLED IN OIL

o

- GOLD - GOLD

COPPER

25 I'm PVDF CHROMIUM :::::t- t. 23V, DC TIN - ."'" CHROMIUM ALUMINUM ELECTRODES t. COP\ER EXCEPT AS INDICATED - COPPER

'2......GOLD

oTIN o-COPPER

1 HOUR

103L-__ ~ ____ ~ ____ -L ____ -L ____ ~ ____ ~ __ ~.

200 220 240 260 280 300 320 340 POLING FIELD · MVm "

FIGURE 3. Enlarged portion of figure 2 with added data for samples electroded with various metals , as shown, and poled in air.

449

E 200 <l: "-

>­I-Cii 100 Z w o I-Z w

~ 50 :::l U Cl Z :::; o "-

~ 20 :2 ~ :2

10

100

CHROMIUM~ COPPER-(Z~

TIN~

GOLD~

o POLED IN AIR

t:. POLED IN OIL

25 Jim PVDF 231', ·C

ALUMINUM ELECTRODES EXCEPT AS INDICATED

200 500 1000

POLING FIELD· MVm "

FIGURE 4. Minimum poling current density as a

function of poling field.

well to PVDF). Superimposed on the slowly varying current were small and very narrow spikes, usually positive but frequently negative. The frequency (average spacing) of these spikes was reduced when poling in oil, so they are presumed to be associated with corona from exposed leads; this has not been studied , however.

The pyroelectric activity as a function of field strength of samples poled (generally) for times shorter than those of the curve in figure 2 are shown in figures 5 and 6, the latter using decibel notation (20 10glO py) of arbitrary reference (0 dB: 1 p,CK - 1m-2) to provide a better indication of the significance of activity differences. The decibel increments

should be the same for piezoelectric and pyroelectric activi­ties. There is a discontinuity between data for air-poled and oil-poled samples, probably as a result of a smaller temper­ature rise in oil due to heating of the sample by the poling current. Considering separately the data above and below this discontinuity, there appears to be no abrupt saturation of ac tivity with field in figure 5, although, as shown by figure

6, the relative increase in activity due to poling above 250 MVm- 1 is only about 3 dB, and may not be large enough in

many applications to compensate for the problems of poling at higher fields. The highest-field sample (550 MVm- l )

experienced almost continuous breakdown, reducing the average poling field below the nominal value for most of the

450

35

30

~

.E :..: u 25

::&.

~ :> 20 i= u <l: u a: I- 15 u w .... w 0 II: >- 10 "-

10·

~10" 5 z

81()3 W

(/)102

0 10

0 100 200

r+ POLED IN OIL I B B I B I I I

B • BREAKDOWN

25 /1m PVDF 23'h ·c

300 400 500 600

POLING FIELD· MVm"

FIGURE 5. Pyroelectric activity as a function of poling field with

poling time as a parameter.

B 30

~ 25 J :.: u ::&.

~ 20

10· m .... w CD U 15 w 0

~ :> i= 10 m U 0 « z u 0 a: u

w I- m u 5 w ....

10' w B . BREAKDOWN 0

II: 25/1m PVDF >-

"- 23'h ·C 0

10

.5~O----1~00~--~2~00~--~300~--~400~--~5OO~--~6~00

POLING FIELD · MVm"

FIGURE 6. As in figure 5, but with the pyroelectric activity

expressed in decibels.

?

"E :.:: u :1-

~ :> i= u « u a: I-u w -' w 0 a: > 0.

35

30

25

20

15

10

5

~B 550 0 - POLED IN AIR

A - POLED IN OIL 7 B45O B B - BREAKDOWN

~B"'_ '" B 350 / B

B~: 150

100 MVm "

25 pm PVDF

23% 'c O~ __ -L ____ ~ ____ ~ __ ~ ____ -L __ ~

10 10' 103 1()4 10' 10'

POLING TIME - SECONDS

FIGURE 7. Pyroelectric activity as a fun.ction of poling time (on

a log scale) with poling field as a pammeter.

poling time; this may accou nt for the relatively small increase in ac ti vi ty relative to 500 MVm- ' . Breakdowns in other samples were generally isolated and probably caused little deviation from nominal conditions, especially fo r poling times of 100 seconds or more. The pyroelectric ac tivi ty of these same samples as a fun ction of poling time is shown in figures 7 and 8, the latter using decibel notation. As wi th fi eld , there appears to be no abrupt saturation with time although at 200 MVm- 1 and above there is less than I-dB change in activity for each decade of poling time. At the highest fields, a scarci ty of data and possible errors due to breakdowns preclude generalization.

By in te rpola ti on and extrapolation from the (unsmoothed) activity dependence plots, constan t activity values were derived and combined with the breakdown curve (with extrapolations) of figure 2 to produce figures 9 and 10. (Values derived from extrapolati ons are connected by broken lines). Of significance here are the in tersec tions between the breakdown curve and the constant-activity curves. It is readily apparent that the highest activ ity is obtained at the highest fi elds, even tho ugh poling time may be severely limited by breakdown. T he optimum poling conditions and the maximum obtainable ac tivity depend upon the highest fi eld strength that can be reached without an " instantaneous" breakdown and which, in turn, depends upon the precautions

~ 30 B~ B400 B~SIC

B- B350 250

~200 150

E 25

:.:: u :1-

~ 20 100 MVm .,

en -' w III U w 15 c

~ :> i= 10 u « u a: I-u w 5 -' 0 - POLED IN AIR w 0 A - POLED IN OIL a: > B - BREAKDOWN 0-

0 25 I' m PVDF

23% 'c

-5 ~ __ ---'--____ -'--____ -'--__ ---'-____ -'--__ ----'

10 10' 103 1()4 10' 10'

POLI NG TIME - SECONDS

F IGURE 8. As in figure 7, but with the pyroelectric activity

expressed in decibels.

taken to prevent external flashovers that seem to initiate breakdowns. For the particular breakdown curve shown, with the "instantaneous" breakdown limit a vert ical line, opt imum poling time would be about 100 s, s ince the small slope in the constant activity lines reduces the ac tivity by a very slight amount for shorter poling times along the (vert ical) limiting portion of the breakdown curve. For practical applications one may wish to avoid breakdown damage by poling outside the breakdown region by a reasonable safety margin. Pyroelectric ac ti vities approaching 30 p,CK- l m- 2

without breakdown appear to be possible. (In oil , py > 31 p,CK- 1 m- 2 was ac tually obta ined with 400 MVm- ' for 10 s, and in a ir, Pu > 29 p,CK- l m- 2 was obtained with 250 MVm- ' for 104 s.) If breakdown damage can be tolerated, as when damaged areas can be trimmed off, even higher activiti es are possible. (Pu > 36 p,CK- 1m- 2 was obtained with 550 MVm- ' for 10 s, in oil, although damage was extensive, and Pu > 35 p,CK-' m- 2 was obtained with 450 MVm- ' for 100 s, in oil, with only one small hole.)

Theory [1] predicts a linear dependence of pyroelectric (and piezoelectric) activi ty on remanent polarization. In figure 11, pyroelectric activi ty has been plotted against the total irreversible poling charge, i.e., the net charge that had passed through the external electrodes of the sample after the poling field had been reduced to zero and after the

451

o

~925 , I , I

10 \1 '? 'I

0- POLED IN AIR A- POLED IN OIL

1159\

20,

25 I'm PVDF 23% ·C

, 30

t, ? \\ \ \' \

\ 32

28'~ 9 '~ I , I

\\~\ I ~ ~?1: I I\~a ? b~\ ~~

32 \" 'to I ,,,32I'CK 'm-' I 30' T I ' I

q \ \ 1\.t36

? ~ I \ I \ I \ I \ I \

5101520 25 2830 30 34 \36

100 200 300 400 500

POLING FIELD - MVm -'

600

1 WEEK

1 DAY

1 HOUR

1 MINUTE

FIGURE 9. Constant pyroelectric activity curves, taken from figure 5

and 7, and breakdown, taken from figure 2, as functions of poling field and poling time.

current transient had disappeared. This represents the re­

manent polarization plus the integral of the leakage current and instrumentation (error) currents. For many samples, this

charge was not determined due to saturation of the first

(current-to-voltage) stage of the current and charge meter by transients during either turn-on or turn-off of the poling

supply, or to saturation of the integrator. When breakdowns

occurred, a correction was made for charge lost (the usual

case). Because of scaling problems and the need to avoid overload, the accuracy of the poling charge measurement was

frequently poor, even neglecting error currents. It may be seen from figure 11, however, that the activity of the samples poled at low-fields is reasonably proportional to charge . Notable exceptions include a 200 MVm-1 sample where a

poling time of 104 seconds may have allowed a sizable charge accumulation from error currents . The other badly deviant samples were all poled at 300 MVm- 1 and above, and all experienced breakdown. Total current, however, was not

visibly changed by breakdowns, following recovery, and a

correction was made for any charge disturbance, as has been mentioned. Large error currents should result in a clear separation of points by poling time (and by field, for large

leakage currents) but there is only slight evidence of such separation.

<J) c z o <..l w <J)

w :E i= (!) Z ::::l o Il.

10'

o

I I I I I , I

21 h7 189 ~9'\

I I ,

100

I ' 1\ 1\ I ,

28 '

200

0- POLED IN AIR A - POLED IN OIL 25 I'm PVDF 23% ·C

300 400 500

POLING FIELD - MVm "

1 WEEK

1 DAY

1 HOUR

1 MINUTE

600

FIGURE 10. As in figure 9, but using pyroelectric activity expressed in

decibels lakenfromfigures 6 and 8.

35

30

~ E

:><: 25 <..l :1.

> I-:> 20 i= <..l <I: <..l a: 15 I-<..l w ....I W 0 a: > 10 Il.

50 100

0450B V300 B

350 BV

010 SECOND

A 10' V 10' c 1()4

B- BREAKDOWN 25 I'm PVDF 23% ·C

150 200 250

POLING CHARGE DENSITY - mem-'

FIGURE 11. Pyroelectric activity as a function of poling charge density.

452

4. Conclusions

High pyroe lectric and piezoelec tric ac tivity can be ob­

tained with bi axially stre tc hed PVDF by poling full y-elec­

troded samples at room te mpera ture us ing a suffic iently high

field. Time-to-breakdown and constant-ac ti vity plots can be

used to optimize ac tivity while avoiding or minimizing

breakdown damage. Hi ghest ac tivity at other poling te mper­

atures and for other mate rials can be obta ined in s imilar

fashion- by determining the time and fi e ld dependence of

both activity and breakdown.

For the particular mate rial inves tigated, poling at room

tempe rature does not yield a hard satura ti on of ac ti vity with

e ither fi eld or time, although the re is little practi cal benefit

in u sing e ither the hi ghest fi e lds or the longest times. The

lack of hard satura ti on may be d ue to the grad ua l alte ration

in crystal packing of TGTG' polyme r cha in s from an tipolar

form II (a-phase) to polar ali gnment , as recently proposed

[3], and perhaps at the highest fi eld to a conversion to the

all-trans polar form I (,B-phase).

There is no evidence th a t breakdown pe r se alte rs sample

activity ou ts ide the damaged area, but it does limit both th e

field and the poling time.

Although room temperature is parti cularl y conven ie nt for

poling, and may have commerc ial importa nce in the la rge­

scale production of poled materi al or in the fabrica ti on of

large "monolithic" transducer a n 'ays, e tc . , highe r ac ti vity

may be attainable at some othe r te mperature , although it has

yet to be de monstrated that the improveme nt would be

s ign ifi cant. A d ete rminat ion of the long- te rm stab ility of

ac tivity as a fun c tion of poling conditions would probably be

more important than a small improve ment in initi al act ivity.

Higher ac ti vity might be obtained by the use of muc h shorte r

poling times, whic h may allow highe r fi elds to be applied.

TI~e large curre nt needed to rapidly c harge and discharge a

sample of reasonable area, however, presents experimental

difficulties . Finally, the depend e nce of ac tivity on electrode

material needs to be bette r unde rstood.

Partial s upport of thi s work by the Naval Ocean Syste ms

Center is gratefull y ac knowl edged.

5. References

[1] Broad hurst , M. C., Davi s, C . T. , a nd Mc Kinney, J. E., Pi ezoelectric­

it y and pyroelectri c ity in polyvin ylide ne fluoride - a mode/. J. ApI'/. Phys. 49,4992 (1978).

[2] Sout hgate, P. D. , Room temperature poling and morphology changes

in pyroe lectri c polyvinylidene fluoride. ApI' /. Phys. Lett. 28, 250 (1976) .

[3] Davi s, C. T. , Mc Kinney, J. E. , Broad hurst, M. C., and Roth , S. c., Electri c-fje ld-induced phase c harges in poly(vinyl idi ne fluoride). J. ApI' /. Phys. 49,4998 (1978).

[4J Muraya ma, N., Oikawa, T. , Katt o, T. , and Nakamura , K. , Pers istent

polarizati on in po ly(vinylidene fl uori de). II . Pi ezoe lec tri c it y of

poly(v in ylidene fl uoride) therm oe lectrets. J. Polymer Sc i. , 13,

1033 (1975).

[5J Shufo rd , R. J., Wi lde, A. F. , Ri cca, J. J. , and Thomas, C. R. ,

Characte ri zati on and piezoelectri c activit y of s tretched and po led

poly(v inylidene fluoride) . Part I: Effect of draw rat io and poling

conditions. Polymer Eng. & Sci. 16,25 (1976).

[6] Day, C . W. , Ha milton, C. A. , Peterson, R. L., Phelan, R. J., Jr. , a nd Mullen, L. 0., Effects of po ling conditions on respons ivit y and uniformit y of pola ri zati on of PVF, pyroelectri c detectors. ApI' /.

Phys. Lett er. 24,456 (1974).

[7] Blev in, W. R., Poling rates for films of polyv in ylidene fluoride . ApI' /.

Phys. Lett. 31 ,8, (1977) .

[8] Sharp, E. J., a nd Cam , L. E. , Effects of agingon thermall y stimula ted

current s in poly(vinylidene fl uoride). ApI' /. Phys . Let t. 29, 480

(1976).

[9] Broadhurst, M. C. , Malmberg, C. C. , Mopsik , F. /. , and Harris, W.

P., Piezo- and pyroelec tric it y in polymer e lectrets. Electrets,

Charge Storage wul Tmnsporl. in Dielectrics, M. M. Perlman, ed. , Electrochem. Soc. , Princeton, N. J. 492-504 (1973).

[10] Handbook of Chemistry and Physics, 47th ed., Che mi ca l Rubber Co.,

Cleveland , Ohio, E67.

453