tensile and dynamic mechanical properties of improved ultrathin polymeric films

Tensile and Dynamic Mechanical Properties of ImprovedUltrathin Polymeric Films

BHARAT BHUSHAN, TIEJUN MA, TAKUJI HIGASHIOJI*

Nanotribology Laboratory for Information Storage and MEMS/NEMS, Department of Mechanical Engineering, The OhioState University, 206 West 18th Avenue, Columbus, Ohio 43210

Received 18 March 2001; accepted 7 July 2001Published online 21 December 2001

ABSTRACT: Tensile and dynamic mechanical properties of improved ultrathin poly-meric films for magnetic tapes are presented. These films include poly(ethylene tereph-thalate) or PET, poly(ethylene naphthalate) or PEN, and aromatic polyamide (ARAMID).PET film is currently the standard substrate used for magnetic tapes; thinner tensil-ized-type PET, PEN, and ARAMID were recently used as alternate substrates withimproved material properties. The thickness of the films ranges from 6.2 to 4.8 �m.Young’s modulus of elasticity, F5 value, strain-at-yield, breaking strength, and strain-at-break were obtained at low strain rates by using a tensile machine. Storage (orelastic) modulus, E�, and the loss tangent, tan �, which is a measurement of viscousenergy dissipation, are measured by using a dynamic mechanical analyzer at temper-ature ranges of �50 to 150°C (for PET), and �50 to 210°C (for PEN and ARAMID), andat a frequency range of 0.016 to 29 Hz. Frequency–temperature superposition was usedto predict the dynamic mechanical behavior of the films over a 28 decade frequencyrange. Results show that ARAMID and tensilized films tend to have higher strengthand moduli than standard PET and PEN. The rates of decrease of storage modulus asa function of temperature are lower for PET films than those for PEN and ARAMIDfilms. Storage modulus for PEN films are higher than that for PET films at highfrequencies, but this relationship reverses at low frequencies. ARAMID has the highestmodulus and strength among the films in this study. The relationship between poly-meric structure and mechanical properties are also discussed. © 2002 John Wiley & Sons,Inc. J Appl Polym Sci 83: 2225–2244, 2002

Key words: dynamic mechanical analysis; storage modulus; loss tangent; polymerfilm; magnetic tape

INTRODUCTION

Poly(ethylene terephthalate), or PET, is currentlythe most widely used polymeric substrate mate-

rial for magnetic recording tapes.1 Thinner sub-strates and higher areal densities (track density� linear density) are required to meet the de-mand for advanced magnetic storage devices withhigh volumetric densities, especially for computerdata storage tapes. For higher areal densities, asubstrate with high dimensional stability undervarious environmental conditions is required. Forhigh track densities, lateral contraction of thesubstrates due to viscoelastic, thermal, hygro-scopic, and shrinkage effects must be minimalduring storage on a reel and use in a drive.1 To

* On leave from Toray Industries, Otsu, Shiga, Japan.Correspondence to: B. Bhushan ([email protected]).Contract grant sponsor: Nanotribology Laboratory for In-

formation Storage and MEMS/NEWS at The Ohio StateUniversity.Journal of Applied Polymer Science, Vol. 83, 2225–2244 (2002)© 2002 John Wiley & Sons, Inc.DOI 10.1002/app.10194

2225

minimize stretching and damage during use, thethinner substrates should be a high-modulus andhigh-strength material with low creep andshrinkage characteristics. Furthermore, becausehigh-coercivity magnetic films on metal evapo-rated tapes are deposited and/or heat treated atelevated temperatures, a substrate with stablemechanical properties up to 100–150°C or evenhigher is desirable. Elastic, viscoelastic, andshrinkage measurements were performed on PETand alternative substrates by various investiga-tors.1–5

Viscoelasticity refers to the combined elasticand viscous deformation of a substrate when ex-ternal forces are applied, and shrinkage occurswhen residual stresses present in the substrateare relieved at elevated temperatures. If a sub-strate of a magnetic tape shrinks or deforms vis-coelastically, then the head cannot read informa-tion stored on the tape. Various long-term reli-ability problems including uneven tape-stackprofiles (or hard bands), mechanical print-through, instantaneous speed variations, andtape stagger problems can all be related to theviscoelastic properties of substrates.1 To mini-mize these reliability problems, it is not only im-portant to minimize creep strain, but the rate ofincrease of total strain needs to be kept to a min-imum to prevent stress relaxation in a woundreel. The elastic and viscoelastic behavior of thesubstrate is also important to determine how thetape responds as it is unwound from the reel andtravels over the head. This elastic/viscoelastic re-covery and subsequent conformity of the tapewith the head occurs in just a few millisecondsand requires optimization of the substrate’s dy-namic properties. The properties are measured byusing dynamic mechanical analysis (or DMA),and the information acquired from this analysisincludes storage or elastic modulus, E�, and losstangent, tan � (a measure of the amount of vis-cous or nonrecoverable deformation with respectto the elastic deformation). Both storage modulusand loss tangent are measured as a function oftemperature and deformation frequency, and re-sults can be used to predict the dynamic responseof substrates over several orders of magnitude.Substrates of magnetic tapes were recently sub-stantially improved.1–5 Weick and Bhushan6 alsoused DMA data to predict tape-to-head confor-mity. A lack of tape-to-head conformity can leadto an increase in wear of the head, as demon-strated by Hahn.7

The main objective of this study was to mea-sure tensile and DMA properties of improved ul-trathin polymeric films that are either used orbeing developed for magnetic tape substrates forultrahigh recording media applications. DMA ex-periments were carried out at temperatures rang-ing from �50 to 150°C (for PET) and �50 to210°C [for poly(ethylene naphthalate) (PEN) andaromatic polyamide (ARAMID)], which cover theglass transition temperatures (Tg) of PET andPEN films. On the basis of these data, mastercurves were generated by using frequency-tem-perature superposition to predict the mechanicalproperties of the films. Tendencies of storagemodulus, loss tangent, and effects from advancedprocessing were related to the structural charac-teristics of each polymeric film.

EXPERIMENTAL

Test Apparatus and Procedure

Tensile measurements were made by using aMTS (MTS Systems Corp., MN) model 804 tensilemachine with a 50-lb. load cell manufactured bySensotech Ltd. (Ohio). The tests were conductedat ambient temperature (19–22°C) and uncon-trolled humidity [25–33% relative humidity(RH)]. In a previous work,8 the ASTM D 1708standard was used to measure the tensile proper-ties of dog-bone-shaped microtensile specimens.The dog-bone-shaped test specimens were used soas to ensure failure in the thinner region. In thisASTM method, the gauge length was assumed tobe equal to the distance between the grips. Forthin plastic sheets thinner than 1 mm, anotherstandard test method according to ASTM D882–97 is recommended for measurement of ten-sile properties and is used here. According to thisstandard, a width–thickness ratio of at least 8shall be used, because narrow specimens magnifyeffects of edge strains or flaws, or both. Rectan-gular 10-mm-wide � 150-mm-long samples withthe grip distance of 100 mm were selected tomeasure the Young’s modulus (modulus of elas-ticity), F5 value (the stress at 5% elongation),strain-at-yield, breaking strength, and strain-at-break. The sample length was limited by theavailable sample sizes. Young’s modulus of allsamples was measured at the strain rate of 0.1/min. For the properties other than Young’s mod-ulus, according to this standard, a strain rate of0.1/min should be used if the strain-at-break of

2226 BHUSHAN, MA, AND HIGASHIOJI

the specimen is lower than about 20%, and a rateof 0.5/min should be used if the strain-at-break is20–100%. Based on this recommendation, the se-lected strain rate was 0.1/min for ARAMID and0.5/min for other samples. Incidentally, as thebreaking strength and yield strength may varyaccording to the sample geometry and test condi-tions, the F5 value proved to be stable and inde-pendent and is commonly used in industry as therepresentative of mechanical strength for poly-mers. A minimum of five tests were performed foreach sample. The reproducibility of the data waswithin about 5% for Young’s modulus and F5value and about 10% for strain-at-yield, breakingstrength, and strain-at-break.

A Rheometrics (Piscataway, NJ) RSA II dy-namic mechanical analyzer was used to measurethe dynamic mechanical properties of the poly-meric films. Figure 1 shows the functional blockdiagram of this test apparatus. The analyzer wasused in a tension/compression mode, and rectan-gular 6.35-mm-wide � 22.5-mm-long sampleswere used. In this mode, rectangular samples arefastened vertically between the grips and a sinu-soidal strain is applied to the specimen. Frequency/temperature sweep experiments were performedfor a 0.1–182 rad s�1 (0.016–29 Hz) range, and 14data points were taken. For each frequency, testssweep at 11 different temperature levels rangingfrom �50 to 150°C for the PET films, and 14temperature levels ranging from �50 to 210°C forPEN films and ARAMID film. The temperatureincrement was 20°C, and the soak time for eachtemperature level was 10 s. For PET and PENfilms, the upper limits for the test temperatureswere set to cover the peak temperatures for losstangent related to glass transition, which are justabove the glass transition temperatures. Theglass transition temperature for PET, PEN, andARAMID are 80, 120, and 280°C, respectively.

The analyzer was operated in autotensionmode with a static force on the samples. Thisprevented buckling of the thin films by applyingthe peak dynamic forces (corresponding to astrain of 0.0025) while using the static force as amean, as shown in Figure 2(a).9 It was found thata static strain of more than 0.25% was needed toprevent buckling [see Fig. 2(b) for an example forstandard PET]. Therefore, the initial static forceof 110 g was used for standard PET and ARAMIDin the machine direction (MD), and 65 g was usedfor other samples, which is equivalent to strainoffset of about 0.25%.

Equations used to calculate the storage (orelastic) modulus, E�, and loss tangent, tan �, areas follows:

Figure 1 Functional block diagram of DMA test ap-paratus–Rheometrics Inc. RSA II.

Figure 2 (a) Schematic showing the effect of preten-sion (static force) on the sample buckling and (b) anexample of storage modulus (E�) measurement in astrain–sweep test of a Standard PET sample. E� re-mains steady after a strain of about 0.25%; samplebuckles below this strain value.

IMPROVED ULTRATHIN POLYMERIC FILMS PROPERTIES 2227

E� � cos ��

�� (1a)

E� � sin ��

�� (1b)

�E*� � ��E��2 � �E��2 (1c)

tan � �E�

E�(1d)

� �DL � � FgK� (1e)

where E� is the storage modulus, E� is the loss (orviscous) modulus, �E*� is the magnitude of thecomplex modulus, � is the applied strain, � is themeasured stress, � is the phase angle shift be-tween stress and strain, D is the displacementfrom the strain transducer, K� is the a stressconstant equal to 100/wt,9 w is the width of thesample, t is the thickness of the sample, g is thegravitational constant (9.81 m s�2), and F is themeasured force on the sample from the load cell.A minimum of two tests was performed on eachsample and reproducibility was within a few per-centages.

In simple terms, at each temperature level theanalyzer operates by applying a strain on thesample in a sinusoidal fashion for each of the0.1–182 rad s�1 frequencies. The strain is mea-sured by a displacement transducer, and the cor-responding sinusoidal load on the sample is mea-sured by a load cell. Because the polymeric filmsare viscoelastic, there will be a phase lag betweenthe applied strain and the measured load (orstress) on the specimen. The storage modulus, E�,is therefore a measure of the component of thecomplex modulus which is in-phase with the ap-plied strain, and the loss modulus, E�, is a mea-sure of the component which is out-of-phase withthe applied strain. The in-phase stress and strainresults in elastically stored energy which is com-pletely recoverable, whereas out-of-phase stressand strain results in the dissipation of energywhich is nonrecoverable and is lost to the system.The loss tangent, tan �, is simply the ratio of theloss modulus to the storage modulus.1

Test Samples

Table I provides a list of the polymeric films ex-amined in this study, along with thicknesses and

symbols used throughout the article. PET filmsinclude three kinds of films: standard PET, ten-silized PET, and supertensilized PET. A 14-�m-thick standard PET film is the typical substrateused for videocassette recorder (VCR) tapes.Standard PET is drawn biaxially about 300%(stretch ratio � 4) in both MD and transversedirection (TD) during processing. Thinner tensil-ized-type PET films, which include tensilized PET(T-PET) and supertensilized PET (ST-PET), arefurther drawn in MD and have a 6.1 �m thick-ness. The tensilized-type PET films are used foradvanced magnetic tapes, especially computerdata storage tapes. Similarly, PEN films have 6.2�m thickness and include three kinds of films:standard PEN, tensilized PEN (T-PEN), and su-pertensilized PEN (ST-PEN). PEN films have be-gun to be used as substrates for advanced mag-netic tapes, especially long-play videotapes andcomputer data storage tapes. Both PET and PENfilms are manufactured by a biaxial drawing pro-cess. On the other hand, ARAMID film is manu-factured by using a solution-casting process fol-lowed by a slight drawing process. ARAMID filmused for advanced magnetic tapes has a thicknessof 4.8 �m.

The unit structures of the polymer films areillustrated in Figure 3. PET and PEN have iden-tical hydrocarbon backbones indicative of polyes-ter materials. PET contains a single benzene ringin each repeating unit, whereas PEN contains anaphthalene ring that is slightly more rigid. How-ever, naphthalene groups have high mobility fromambient temperature to 60°C.10 Typical crystal-linities of PET and PEN films are 40–50 and30–40%, respectively. ARAMID contains amidegroups with intermolecular hydrogen bonds thatare stronger than the intermolecular interactions

Table I List of Substrates Used in This Study

Sample SymbolThickness

(�m)

Standard PET Standard PET 14Tensilized PET T-PET 6.1Supertensilized PET ST-PET 6.1

Standard PENStandardPEN 6.2

Tensilized PEN T-PEN 6.2Supertensilized PEN ST-PEN 6.2ARAMID ARAMID 4.8

The glass transition temperatures for PET, PEN, andARAMID are 80, 120, and 280°C, respectively.


for PET and PEN. As a result, ARAMID enablesthe formulation of high-strength, high-modulusfilms.

Because of the different stretch ratios along theMD and the TD during manufacture, all the ma-terials were anisotropic and were tested along theMD and TD. It was confirmed that the MD corre-sponds to the direction of orientation of crystals inbiaxial films by using a cross-polarizer in whichoptical axes show the extinction of light. Two lin-ear polarizing films of 50.8 mm (2 in.) square(Tech SpecTM Quality J43-781, Edmund Scien-tific) were used. As examples, Figure 4 shows theorientation of the major axis with respect to theMD of PET films. The difference between the ma-jor axis and the MD is below �5 degrees.

RESULTS AND DISCUSSION

Tensile Properties

The engineering stress–strain curves of the filmsare shown in Figure 5 along the MD and TD. Allthe samples extended uniformly during the test,and no neck was formed. The yield point is thenapproached by determining the intersection of thetwo tangents to the initial and final parts of thestress–strain curve.11 Table II and Figure 6(a)give summaries of tensile properties of variousfilms. Figure 6(b) shows the effect of strain rate

on tensile properties on T-PET and T-PEN. Anincrease in strain rate results in increases inYoung’s modulus, F5 value, breaking strength,and decreases in strain-at-yield and strain-at-break. T-PEN appears to be more strain-rate de-pendent than T-PET: when strain rate increasesfrom 0.1 to 0.5/min, the Young’s modulus for T-PEN (MD) increases by 1.2 GPa, whereas that forT-PET increases only by 0.3 GPa [Fig. 6(b)]. Thestrain rate could be related to the frequency inDMA tests, where the storage modulus of PEN inDMA tests appears to be more frequency depen-dent.

Tensilization significantly increases the Young’smodulus, F5 value, and breaking strength of PETfilms in the MD. For example, the Young’s mod-ulus of Standard PET (3.5 GPa) is doubled aftertensilization (T-PET: 7 GPa) and supertensiliza-tion (ST-PET: 7.35 GPa). However, the deforma-tion properties, strain-at-break and strain-at-

Figure 3 Chemical unit structures of various poly-meric films.

Figure 4 Orientation of the optical major axis withrespect to the MD of PET films (a) along the TD and (b)along the MD.


yield, are decreased after tensilization. These ten-dencies reverse in the TD, as the Young’smodulus changes from 4.8 GPa for Standard PETto 4.5 GPa for T-PET, and strain-at-breakchanges from 79% for Standard PET to 130% forT-PET.

Standard PET shows a fairly low work-harden-ing ratio (d�/d�) after the elastic limit is reachedfor the MD samples. This increases as a result ofthe stretching process in the case of T-PET andST-PET. The increase in the hardening ratio as aresult of stretching arises because the chain seg-ments in the amorphous regions are already ex-tended in a variety of directions, thereby limitingthe freedom that can occur during additionalstretching (e.g., during tension test). The stress–strain curve for T-PET in the TD is similar to thatfor Standard PET along the MD. Furthermore, itis supposed to have been stretched at a low ratioin this direction during the manufacturing pro-cess. Comparatively, ST-PET may have a higherstretch ratio in both the MD and the TD thanT-PET.

Tensilization has similar effects on PEN filmsas on PET films, but is not that significant. Forexample, the modulus for Standard PEN in-creases from 7 to 7.7 GPa (T-PEN) and 8.5 GPa(ST-PEN) in the MD and decreases from 7.25 to6.1 GPa (T-PEN) and 5.5 GPa (ST-PEN) in the TDafter tensilization.

ARAMID has a high Young’s modulus,strength, and low elongation, and it is more aniso-

tropic than PET and PEN films. The Young’smodulus along the MD and TD are 20.4 and 11.3GPa, respectively, as compared to 9.5 GPa (MD)and 10.5 GPa (TD) in a previous article.8 This isprobably due to an improvement in tensilizationalong the MD during the manufacturing process.

DMA Properties

PET Data

Results from the DMA for PET films are shown inFigure 7; storage moduli are plotted in Figure7(a), and the loss tangents are plotted in Figure7(b). In addition to the two-dimensional represen-tations of storage modulus and loss tangent as afunction of frequency and temperature, a three-dimensional representation of these parametersis also shown for each material. From the three-dimensional surfaces, it can be seen that higherelastic moduli correspond with higher deforma-tion frequencies and lower temperatures,whereas lower elastic moduli correspond withlower deformation frequencies and higher tem-peratures. This can also be seen from the two-dimensional representations of the data.

For all the PET films, the rate of decrease ofstorage moduli as a function of temperature is lowbefore a certain temperature is reached; then therate suddenly increases and the storage modulidrop to a very low level. This change correspondsto the character of the loss tangents in Figure7(b), where the loss tangents remain low below

Figure 5 Engineering stress–strain curves of various polymeric films. Strain rate forPETs and PENs was 0.5/min, strain for ARAMID was 0.1/min.


Table II Summary of Mechanical Properties of Various Substrates

Data Resource

Manufacturers’ Data

Data From This Study

DMATensile

Modulus ofElasticity

(GPa)

F5Value(MPa)

BreakingStrength

(MPa)

Strainat Break

(%)

Modulus ofElasticitya

(GPa)

F5Valueb

(MPa)

Strain atYieldb

(%)

BreakingStrengthb

(MPa)

Strain atBreakb

(%)

Modules ofElasticityc

(GPa)

Standard PETMD 4.4 113 295 130 3.30 95 3.00 200 115 4.12TD 6.2 143 340 100 4.54 117 4.10 266 79 5.05

T-PETMD 7.2 180 440 70 6.30 172 3.10 350 44 6.40TD 4.5 108 270 130 4.10 106 2.80 230 108 4.20

ST-PETMD 8.5 229 533 50 7.15 195 3.30 461 45 7.38TD 5.2 121 327 106 4.47 107 3.20 257 75 4.96

Standard PENMD 7.0 173 438 69 6.25 175 3.20 334 42 6.42TD 7.3 202 450 62 6.90 200 3.20 384 42 6.70

T-PENMD 8.0 184 465 61 6.50 190 2.60 340 30 7.30TD 6.5 161 410 81 5.60 158 2.90 300 40 6.50

ST-PENMD 9.0 217 623 52 7.80 220 2.70 452 36 7.79TD 6.0 139 340 103 5.42 144 2.65 293 75 5.37

ARAMIDMD 19.3 590 655 10.6 20.4 628 2.80 638 6.4 17.2TD 11.5 390 414 6.4 11.3 338 3.80 433 11 12.1

a Data for all samples taken at 0.1/min strain rate.b Data for all PETs and PENs taken at 0.5/min strain rate, and for ARAMID at 0.1/min strain rate.c Data taken at 0.016 Hz, 20°C.

IMP

RO

VE

DU

LT

RA

TH

INP

OL

YM

ER

ICF

ILM

SP

RO

PE

RT

IES

2231

Figure 6 (a) A summary of tensile properties of various polymeric films. For Young’smodulus, data for all samples were taken at 0.1/min strain rate. For other properties,data for PETs and PENs were taken at 0.5/min strain rate, and for ARAMID were takenat 0.1/min strain rate. (b) Effects of strain rate on tensile properties for T-PET andT-PEN.


50°C, indicating that a small amount of energy isdissipated. The amounts of loss tangent generallyreflect the mobility of molecules, and larger losstangent shows high mobility of molecules. Theloss tangents begin to rise at 70–90°C (corre-sponding to different frequencies) and peak at theelevated temperatures of 110 to 130°C, whichmeans the molecules begin to change their shape/orientation on a large scale. For example, rotationcould occur along the junctions in the backbone ofthe chain, such as the OOO bond in PET. Thetemperature at which the loss tangent starts toincrease corresponds to the glass transition tem-perature,12 Tg. Above this temperature, the poly-meric material behaves as a viscoelastic rubberdue to the rotation of molecules. The glass tran-sition temperature of PET measured in this studyis consistent with that reported by Ward,11 whichwas 80°C, for amorphous PET. Below Tg, PETfilms have lower loss tangent. This result indi-cates that PET films have less mobility at ambi-ent temperature below Tg.

The temperature at which the peak of the losstangent occurs is a function of frequency13; itshifts to a lower value and higher temperature asthe frequency increases [Fig. 7 (b)], because of thefrequency dependence of relaxation processes.For example, the loss tangent peak for StandardPET shifts from 0.22 at 110°C to 0.12 at 130°C asthe deformation frequency increases from 0.016to 29 Hz. Essentially, the faster the applied stim-ulus, the less time the molecules have to respondto it. At lower frequencies, the molecules have alonger time to respond to the applied strain,whereas at high frequencies, the time is too shortand the response is considered to be glassy (i.e.,the molecules cannot move rapidly). Therefore, ingeneral terms, a high temperature is needed toenergize molecular movement at high frequen-cies. On the other hand, at a constant tempera-ture, as the stimuli are applied over a range offrequencies, the glass transition is seen first forthe lower frequencies. Minimal movement of themolecular chain at high frequencies also explainsthe continuous decrease in loss tangent as thefrequency increases. From the frequency depen-dence of the loss tangent peak, it is possible toevaluate the activation energy for the process.13

PEN Data

Figure 8(a,b) shows the storage modulus and losstangent for Standard PEN, T-PEN, and ST-PEN.There are some common points among PET and

PEN films: the storage modulus increases as thefrequency increases and temperature decreases,the magnitude of the loss tangent peak shiftstoward a lower value, and it occurs at highertemperature as the frequency increases. Unlikethe PET films of which the storage modulus keepsalmost constant at the glassy state, the storagemodulus of PEN films decrease in a certain slopein the whole temperature range. A high rate ofdecrease of storage modulus of PEN films corre-sponds to a high value of loss tangent, as shown inFigure 8(b). For example, the loss tangents forPEN films at 30°C are 0.05 to 0.09, whereas thatfor PET films are only 0.01–0.03. Also for PENfilms, there is a minor loss tangent peak at 50–70°C in addition to the peak that is related to theglass transition (140 to 170°C), and this minorpeak is absent for PET films. These peaks areresponsible for a drop in the storage modulusnear these temperatures.

ARAMID

The storage modulus and loss tangent for ARAMIDas a function of frequency and temperature arepresented in Figure 9. ARAMID has a signifi-cantly higher storage modulus than PET andPEN films. It also has good temperature resis-tance; for example, the storage moduli remainsabove 6 GPa even at 200°C in both the MD andthe TD.

The glass transition temperature for ARAMIDis typically reported as 280°C, so there is no glasstransition peak in the loss tangent-temperaturediagram in this study; instead, only a steady in-crease is present in the 170–210°C range. How-ever, similar to the curves of PEN films, a minorloss tangent peak at 50°C (at 0.016 Hz) to 130°C(at 29 Hz) is present, which is responsible for adrop in the storage modulus near this tempera-ture. The loss tangent for ARAMID at elevatedtemperatures (above ambient) is 0.5–0.8, similarto that for PEN films and higher than that forPET films. As a result, the storage modulus forARAMID keeps on decreasing at elevated temper-atures, and there is no sudden drop as in thestorage modulus-time-temperature 3D diagramfor PET.

Comparison of DMA Data at 30°C and 0.016 Hz

Figure 10 shows the effects of frequency (andtemperature) on the DMA properties of variouspolymeric films at 30°C (and 0.16 Hz) along the


Figure 7 (a) Storage modulus (E�) and (b) loss tangent (tan �) of PET films asfunctions of frequency and temperature.


Figure 7 (Continued from the previous page)


Figure 8 (a) Storage modulus (E�) and (b) loss tangent (tan �) of PEN films asfunctions of frequency and temperature.

Figure 8 (Continued from the previous page)


Figure 9 Storage modulus (E�) and loss tangent (tan �) of ARAMID as functions offrequency and temperature.


Figure 10 Effects of frequency (at 30°C) and temperature (at 0.016 Hz) on storagemodulus (E�) and loss tangent (tan �) of various polymeric films.


MD and TD, so that we can have a general com-parison with different films.

ARAMID has significantly higher storage mod-ulus than that for PET and PEN films in both theMD and the TD. The storage moduli for PEN filmsare slightly higher than that for PET films, butthe difference is reduced as the deformation fre-quency decreases during the test. In general, thestorage modulus at higher frequency is mainlydetermined by the elastic elements, and the vis-coelastic deformation, creep deformation, pre-dominates at lower frequency. Tensile test data inFigure 6(b) also show that frequency dependenceon modulus for PEN is higher than for PET. Thestorage moduli of the same material (such asstandard PET, T-PET, and ST-PET) have similarslope against frequency. In the previous re-search,4,5 PEN films have slightly more creep de-formation compared with tensilized-type PET.Thus, it appears that PEN is superior to PET inelastic properties, whereas it is inferior in viscousproperties. In this study, the loss tangents of PENfilms are the highest among the three kinds ofmaterials at ambient temperature.

Tensilization increases the storage modulus inthe MD and decreases the storage modulus in theTD, which makes the material more anisotropic.It also generally increases the loss tangent forboth PET and PEN. Temperature has a reverseeffect on the storage modulus of the polymeric

films. The same material (such as PET films)show a similar rate of decrease of storage modu-lus versus temperature.

Prediction of Mechanical Behavior UsingTime–Temperature Superposition

A technique known as frequency-temperature su-perposition was used to predict the storage mod-uli over a wider frequency range at a specificreference temperature.14,15 The superposition iscarried out by using DMA data taken over a rel-atively narrow frequency range at different tem-perature levels. For instance, the storage modu-lus versus frequency curves in Figure 7(a) areused as the starting point, and a reference tem-perature of 30°C is selected. Curves at tempera-tures higher than 30°C are shifted to the left untilthey fit together smoothly, and the curves corre-sponding to the temperatures lower than 30°Care shifted to the right. The shift direction corre-sponds with the viscoelastic nature of polymericfilms. Storage moduli measured for a polymer athigh frequencies under ambient conditions will beequivalent to those measured at lower frequen-cies and colder temperatures.

Based on the time–temperature superposition,master curves are generated. Two plots areshown in the MD and TD in Figure 11. The re-sults are shown on a logarithm–logarithm scale,

Figure 11 Master curves of storage modulus (E�) of various polymeric films at 30°C.


which indicates the storage modulus along theMD and TD over a frequency range from 10�18 to1010 Hz. The storage modulus curves for PETfilms can be identified as two regions according tothe slopes separated at about 10�10 Hz. As amatter of fact, the storage modulus data for PETfilms at frequencies lower than 10�10 Hz comefrom the storage moduli at temperatures higherthan Tg. Thus, the curves can be regarded ascomposed of two regions: a glassy region and arubbery region. This has more academic meaningrather than practical meaning, because 10�10 Hzmeans the load cycle lasts 317 years. The data forPEN films can also be identified as glassy andrubbery regions, separated at about 10�7 Hz, al-though the change of slope is not as significant asthat observed for PET films. This is more realis-tic, or, more critical, because the PEN films wouldbegin to act rubbery when the load cycle lasts forabout 4 months. In the MD, the storage modulusfor PEN films is higher than that for PET films athigh frequencies, and this relationship reversesat low frequencies. The storage modulus curvesfor Standard PET and Standard PEN cross at thefrequency of 10�5 Hz (cross-frequency). Tensiliza-tion shifts the storage modulus curves for PETand PEN films to a higher value. Tensilizationalso shifts the cross frequency toward higher fre-quencies. However, tensilization does not obviouslyaffect the threshold frequencies for transition fromglassy to rubbery. There is no glass transition forARAMID in the frequency range corresponding tothe temperature range used in this study; the wholecurve is within the glassy region.

ARAMID, over the available frequency range(10�10 to 105 Hz), has a superior storage modulusthan other polymeric films, although the decreas-ing rate of storage modulus as a function of fre-quency is relatively high. PEN films appear to bestiffer than PET films at the frequencies higherthan 10�2 Hz (for MD) and 10�6 Hz (for TD), butPET has a lower storage modulus slope in theglassy range. It is interesting to find that at lowfrequencies (10�2 Hz for MD and 10�6 Hz for TD),the storage moduli for PET films are higher thanthose for PEN films. The modulus at the cross-frequency is about 3–5 GPa.

Relationship between Mechanical Properties andPolymeric Structure

PET Films

The mechanical properties of polymeric films aredependent on their chemical composition and

physical structure, such as crystallization andmolecular orientation, which in turn are affectedby the manufacturing process.

The crystallinity (the portion of crystal regionin the whole material) of PET is higher than thatof PEN and ARAMID, as schematically shown inFigure 12. However, the molecules of StandardPET have a higher mobility because of less ori-ented chains, and distortion of the intermolecularbonds such as van der Waals attractions couldcontribute to the lower modulus of elasticity. Ten-silized-type PET films, including T-PET and ST-PET, contain more oriented chains than StandardPET, which effectively restrain the molecular de-formation along the drawing direction. Cakmakand Wang16 suggested that the increase in crys-tallinity and orientation of chains in the amor-phous region results in a reduction in creepstrains for PET. At the same time, tensilizationstores mechanical energy in the material as itincreases the entropy and leaves residual stressesin the system. As a result, the loss tangent has ahigher peak magnitude for T-PET and ST-PETthan Standard PET [Fig. 7(b)]. The long-rangemolecular motion at a temperature higher thanTg is less affected by the molecular orientationthan by the deformation frequency. As presentedin Figure 10, the loss tangent peak positions arenot affected by tensilization; that is, Tg is inde-pendent of manufacturing processes such asstretching.

In Figure 10, PET films show less temperaturedependence than PEN films and ARAMID exceptat the range from 90 to 130°C where PET trans-fers from a glassy to a rubbery state. The reasonfor this is that crystallization can effectively sta-bilize the polymer. Mascia and Fekkai’s work17

indicated that when the degree of crystalline of

Figure 12 Schematic of structures of various poly-meric films.


PET reached 43%, the shrinkage would be com-pletely suppressed.

Tensilization also increases the stress neededto start the molecular rotation; because the anglebetween molecular orientation and applied load issmall, a higher load is required to reach the crit-ical partial shear stress to rotate the segment. Asa result, the yield strength is increased.

PEN Films

PEN has a similar molecular structure to PET (asshown in Fig. 3). PEN contains a naphthalenering that is slightly more rigid than the benzenering for PET. This is why the PEN films have ahigher Young’s moduli and storage moduli thanPET films. The glass transition temperature Tgfor PEN is about 110–130°C (120°C as typicallyreported), higher than that for PET. Also, the zerostrength temperature for PEN films (170°C) ishigher than that for PET films (130–150°C). Allthese ensure that PEN films are superior to PETfilms at high temperatures. However, the largenaphthalene segment inhabits the movement ofthe molecules during the crystallization and re-sults in a low crystallinity. The typical crystallini-ties of PET films and PEN films are 40–50 and30–40%, respectively. Crystalline and orientedchain segments play important roles in determin-ing mechanical properties of polymers. The net-work structure formed by crystal region signifi-cantly improves the stiffness and mechanicalstability by anchoring polymeric chains in thehigh-density and less mobile crystalline region.The lower density amorphous region is more tem-perature and frequency dependent than the crys-tal region and behaves viscoelastically. Higashiojiand Bhushan5 calculated the apparent activationenergy of creep of some tensilized PETs andPENs, assuming that the viscoelastic deformationprocess has an Arrhenius temperature depen-dence. The result showed that tensilization ofPET increased the activation energy from 290 to400 kJ/mol, but this increase for PEN was onlyfrom 300 to 310 kJ/mol. It is possible that theminor loss tangent peak at 50 to 70°C for PENfilms eliminates the improvement from advancedprocessing such as tensilization.

The investigations of Chen and Zachmann12

and Ezquerra et al.18 on dynamic mechanicalproperties of amorphous PEN revealed that thereexist three peaks (or relaxations) in the loss tan-gent versus temperature diagram: � (at �50°C),�* (at 60°C), and � (at 80–130°C), corresponding

to short- and long-range molecular motion. The �relaxation is attributed to motion of theOCOOOgroups; the �* relaxation is attributed to motionsof naphthalene rings and the � relaxation is at-tributed to glass transition. The secondary mo-tions of � and �* relaxation involve localized mo-tion such as the so-called crankshaft mechanismand local-mode process, and distortion of the poly-meric chains occurs through intermolecular dis-tances.19,20 Because the relaxations, especiallythe �* relaxation, represent the deformation abil-ity and determine the mechanical properties ofthe polymer close to ambient temperature, theywere extensively investigated both experimentallyand theoretically.12,18–20 Gillmor and Greener10

suggested that a possible localized motion in PENis interlayer slippage of naphthalene ring stackedin order of a liquid crystal parallel to the plane ofthe film because of the rigid and planar confor-mation of the naphthalene ring. Cakmak andLee21 suggested that, during the deformation ofPEN, the naphthalene rings are rapidly alignedparallel to the surface of the films, and they arealso highly localized. Ezquerra et al.18 indicatedthat the �* relaxation could not be detected bynuclear magnetic resonance (NMR) because thefrequency was too high (� 1 MHz). Alhaj-Moham-med et al.22 also found that at high frequency themotion of the naphthalene ring (�*) only becomesdetectable around Tg. Localized molecular motiondoes not have significant effects on the Young’smodulus, which is mainly determined by the poly-mer backbone. However, the localized motion re-sults in high-loss tangent value (associated withthe large energy dispersion) and high tough-ness.23 This relaxation also has reversal effects onthe long-term stability, as concluded in Higashiojiand Bhushan’s work,5 that PEN films have highercreep compliance than that of tensilized PET atambient and elevated temperatures. This can alsobe confirmed from the storage modulus–tempera-ture 2D diagram in Figures 7(a) and 8(a) thatwhen the frequency changes from 29 to 0.016 Hz,the extension of storage moduli distributions (ver-tically) are much wider in PEN films than in PETfilms. In Figure 10, at ambient temperature, thestorage modulus of T-PEN drops about 3 GPafrom 29 to 0.016 Hz, whereas the modulus forT-PET drops only 1 GPa. Thus, the mechanicalproperties of PEN are more frequency dependent.The modulus for tensilized-type PET films couldbe comparable with that of PEN films at lowerfrequency and lower strain rate.


ARAMID Film

ARAMID consists of rigid molecular chains with alinear structure. Molecular chains normally ori-ent themselves in a single direction during theformation of liquid crystals in a doped solution.Therefore, the molecules in solid ARAMID arehighly oriented and compose the so-called rigidrod structure. There is no oxygen atom in thebackbone chain; instead, it consists of an aromatichydrocarbon group (benzene ring) combined byamide bonds, para-linked by intermolecular hy-drogen bonds that are stronger than the intermo-lecular interactions for PET and PEN. As a result,ARAMID enables the formation of high-strength,high-modulus, and low strain-at-yield, strain-at-break polymer film. The factors that determinethe superior mechanical properties of ARAMIDalso include the following:

Rigid aromatic (NHOCO) bondingIntermolecular hydrogen bond that comes from

amide groupHighly symmetric configurationHigh interlinked density from substituent Xm

and Yn atomsPresence of additional polar group

Electron delocalization also contributes to highchain stiffness and various threshold tempera-tures. This is easily identified from the colorsbecause color increases both in occurrence and indarkening as delocalization increases.23 The PETfilm is transparent, whereas PEN is pearl andsemi-opaque, and ARAMID appears to be yellowand opaque.

The ARAMID film used in this study (TorayInc., MictronTM) is manufactured by using a solu-tion-casting process followed by a slight drawingprocess. The solution casting process virtuallyeliminates the shrinkage that is common in thedrawing processes. However, the cost of the filmproduced by casting is higher than by drawing.The study on the secondary relaxation of ARAMIDfilm MictronTM has not been accumulated in de-tail. Different aramids have different substituentXm and Yn atoms; thus, they have different inter-molecular reacting mechanisms and differenttemperatures at which the secondary relaxationsoccur. Studies on other aramids such as aliphaticpolyamide and poly(p-phenylene terephathal-amide) (PPTA) fiber Kevlar® and poly(m-phe-nylene isophthalamide) fiber (Nomex®) suggested

that the �* relaxation was due to the hinderedlocal motion of phenylene group.13,24 Kunugi andhis colleagues13 studied the loss tangent peaks forPPTA fiber over �100 to 500°C and indicated thatthe peak near 60°C was a � relaxation. Frosiniand Butta24 suggested that the relevant relax-ation process was due to the motion of the freeamide group which did not form inter- or intramo-lecular hydrogen bonds and that the presence ofwater molecules or other low-molecular-weightpolar substances played an important role in therelaxation process.

CONCLUSION

Tensile and dynamic mechanical properties of ul-trathin polymeric films, including PET, PEN, andARAMID, as well as tensilized-type PET andPEN films, were measured and analyzed. Tensiletests at room temperature show that the AR-AMID has a higher Young’s modulus, F5 value,breaking strength, and lower strain-at-breakthan other films. Additional tensilization of PETand PEN films along the MD increases theYoung’s modulus, breaking strength, F5 value,but reduces the strain-at-yield and strain-at-break along this direction, because more chainsegments were oriented and aligned in the stretchdirection. It has reversed effects on the propertiesin the TD.

Storage modulus of polymeric films increasesas the deformation frequency increases and testtemperature decreases. PEN films have higherfrequency dependence than PET films. PET filmshave a low rate of storage modulus decrease astemperature increases before its glass transitiontemperature is reached. This is also confirmed bythe low value of loss tangents for PET films at lowtemperatures. A sudden drop in storage modulioccurs at 80–130°C when PET films transformfrom a glassy to a rubbery state, corresponding toa loss tangent peak at 110°C (0.016 Hz) to 130°C(29 Hz). No secondary relaxation peak for PETfilms was detected in the temperature range usedin this study (i.e., �50–150°C). Although forPENs, besides the loss tangent peak related toglass transition at 150°C (0.016 Hz) to 170°C (29Hz), a secondary relaxation, �* relaxation, wasidentified at about 50–70°C, which is responsiblefor drop in the storage modulus near this temper-ature. The loss tangents for PEN films are higherthan that for PET films at the temperatures lowerthan 70°C. All the loss tangent peaks shift toward


low value and high temperature as test frequencyincreases. Tensilization does not change the posi-tion of these loss tangent peaks, but slightly in-creases their absolute value. ARAMID also has asecondary relaxation (� relaxation) at 50°C (0.016Hz) to 130°C (29 Hz), which results in the quickdrop of storage modulus at these temperatures.This is consistent with previous work5 thatshowed ARAMID has an apparent activation en-ergy which is lower than that for tensilized PET,which does not have a relaxation in this temper-ature range.

A frequency–temperature master curve wasgenerated from the DMA data that covers thestorage modulus of polymeric films over a fre-quency range from 10�18 to 1010 Hz. ARAMID hasa superior storage modulus when compared toother materials through the available frequencyrange. The storage moduli for PET films are thelowest at high frequencies, but its rate of decreaseas frequency decreases is also low until 10�10 Hz.The storage moduli for PET films become higherthan that of PEN films when frequency is lowerthan 10�2 Hz (MD) and 10�5 Hz (TD). It suggeststhat elastic properties of PEN films are superior,whereas viscous properties are inferior as com-pared to PET films. Tensilization raises the stor-age modulus–frequency curves to higher posi-tions, but generally does not affect the tendenciesand slope.

Financial support for this research was provided by theindustry membership of Nanotribology Laboratory forInformation Storage and MEMS/NEMS (NLIM) at TheOhio State University. The samples and relevant infor-mation were kindly supplied by Dr. K. Abe of Toray andDr. T. Osawa and Dr. H. Murooka of Teijin. We alsothank Dr. Liming Zhang for his work in DMA tests.

REFERENCES

1. Bhushan, B. Mechanics and Reliability of FlexibleMagnetic Media, 2nd ed.; Springer-Verlag: NewYork, 2000.

2. Weick, B. L.; Bhushan, B. J Inf Storage Proc Syst2000, 2, 207–220.

3. Weick, B. L.; Bhushan, B. J Appl Polym Sci toappear.

4. Higashioji, T.; Bhushan, B. IEEE Trans Magn toappear.

5. Higashioji, T.; Bhushan, B. J Appl Polym Sci toappear.

6. Weick, B. L.; Bhushan, B. Wear 1996, 202, 17–29.7. Hahn, F. W. Wear of recording heads by magnetic

tape, in Tribology and Mechanics of Magnetic Stor-age Systems, Vol. 1; B. Bhushan et al., Eds., ASLE:Park Ridge, IL, 1994; SP16, pp 41–48.

8. Weick, B. L.; Bhushan, B. IEEE Trans Magn 1996,32, 3319–3322.

9. Owners Manual for the Rheometrics RSA-II Dy-namic Mechanical Analyzer, Rheometrics, Inc., 1-TPossumtown Rd., Piscataway, NJ, 08854-9990,(201) 560-8550.

10. Gillmor, J. R.; Greener, J. Proc Soc Plastic EngANTEC97, 1997; pp 1582–1585.

11. Ward, I. M. Mechanical Properties of Solid Poly-mers, 2nd ed.; Wiley: Chichester, NY, 1983.

12. Chen, D.; Zachmann, H. G. Polymer 1991, 32,1612–1621.

13. Kunugi, T.; Watanabe, H.; Hashimoto, M. J ApplPolym Sci 1979, 24, 1039–1051.

14. Williams, M. L.; Landel, R. F.; Ferry, J. D. J AmChem Soc 1955, 77, 3701–3707.

15. Ferry, J. D. Viscoelastic Properties of Polymers,3rd ed.; Wiley: New York, 1980.

16. Cakmak, M.; Wang, Y. D. J Appl Polym Sci 1990,41, 1867–1890.

17. Mascia, L.; Fekkai, Z. Polymer 1993, 34, 1418–1422.

18. Ezquerra, T. A.; Balta-Calleja, F. J.; Zachmann,H. G. Acta Polym 1993, 44, 18–24.

19. Ahumada, O.; Ezquerra, T. A.; Nogales, A.; Blta-Calleja, F. J.; Zachmann, H. G. Macromolecules1996, 29, 5002–5009.

20. Blundell, D. J.; Buckingham, K. A. Polymer 1985,26, 1623–1627.

21. Cakmak, M.; Lee, S. W. Polymer 1995, 36, 4039–4054.

22. Alhaj-Mohammed, M. H.; Davies, G. R.; AbdulJawad, A. J Polym Sci, Part B: Polym Phys 1988,26, 1751–1760.

23. Segmour, R. B.; Carraher, C. E. Structure–Proper-ties Relationships in Polymers; Plenum Press: NewYork/London, 1984.

24. Frosini, V.; Butta, E. J Polym Sci, Part B: PolymPhys 1971, 9, 253–260.


tensile and dynamic mechanical properties of improved ultrathin polymeric films

Documents