EAA Properties

8/3/2019 EAA Properties

1/6

244 E. P. OTOCKAN D T. K. KWEI Macromolecules

TABLEConstants in eq 5 I've S

Present work 6 10Ackers work& 5 13-

a See ref 33 .rv = r,,, + s erfcc1(Kd) ( 5 )

The cons tan t rqo s the posi t ion of the maximum ordi-nate of the distr ibution, s the measure of s tan dard dev-iation, and erfc-1 the inverse error function comple-ment.Ackers showed for several gels that a plot of rq GS .erfc-(Kd) was linear and that r,,, a n d s could be takento be cal ibrat ion constants for a given gel. I n presentwork, the available values of Kd = (V, - Vo)jViweregraphed in terms of the inverse error function comple-m ent, e r f ~ - l ( K ~ ) , ~ ~s. the corresponding estimatedmolecular radii, r7 , to yield the approximately l inear(34) Tables of the Erro r Funct ion and I t s Derivat ives ,Nat ional Bureau of Standards, Appl ied Mathemat ics Series 41 ,

U. S. Government Prin t ing Office, Washington , D. C., 1954.

relationship shown in Figure 8. Cons tan ts f ound f orthis line, together with those estimated from Figure 2of Ackers pap er33 re given in Table V. The proximityof rV aa n d s values found in the present work withthose repo rted by Ackers-although his plot is based ondat a from fractionation of dextrans-gives sup por t tothe above discussed co ncepts of the molecular size andthe swelling relationships of NaL S molecules. Ou rexperimental results seem to provide further evidencefavoring G orings interpretations as well as the presentideas concerning the mechanism of gel chromatograph yseparations.

Acknowledgments. The authors appreciate the help-ful counsel of Dr. Bjorn F. Hrutfiord of the College ofFores try and the De par tmen t of Chemical Engineer ing,and the substantia l ass is tance given by Mr. R oger W adeof the Department of Biochemistry, of the Universityof Washington. They are also grateful to D r . D. A. I.Goring of the Pulp and Paper Research Ins t i tute ofCanada , and to Dr . Ka j For s s and P r ofes sor W . Jensenof the Finnish Pulp and Paper Research Ins t i tute ofHelsinki for helpful discussions of their i mp ortan t wo rk.

Properties of Ethylene-Acrylic Acid CopolymersE. P. Otocka and T. K. KweiBell Telephone Laboratories, Inc., Mur ray Hill, ew JerseyReceived January 25, 1968 07974.

ABSTR ACT: The effects of carboxylic acid comonomers on polyethylene are studied using calorimetry, dynamicmechanical testing, and infrared spectroscopy. Infrared studies indicate that the fraction of acid groups hydrogenbonded is determined by a temperature-dependent equilibrium with thermodynamic p arameters similar to smallmolecule acids in ordinary solvents. During elongation, the planes of the acid dimer become perpendicular tothe direction of stretch. The resultsshow that the acid groups are not crystallizable, and the dimerization takes place in the amorphous phase. Theincreases in T, with increasing acid content as determined by E are more correctly described by cross-linkingequations than by the simple copolymer relationships.

The melting points of the copolymers are tested using Florys equation.

variety of effects arise from the incorporation ofA comonomers with functional pendent groups inotherwise nonpo lar vinyl polymers. In the cases ofcarboxylic acids, alcohols, amines, or amides, changesin physical properties are often identif ied with thehydroge n bon ded structures formed. Systematic in-vestigation of how relatively isolated functional gro upsalter polymer behavior is necessary in understandingthe behavior of systems where they are major con-stituents.Early investigations have shown that the presence ofcarboxylic acid or carboxylate comonomers generallyincreased matrix cohesion but did not lead to mechan-ically stabl e networks.1-3 Th e pictu re of acid dimeri-zation and acid-base interactions in amo rpho us poly-

( I ) H. P. Brown (to B. F. Goodrich Co.), U . S. Pa t en t( 2 ) W. C o o p e r , J . Polj,m. Sci., 28 , 195 (1958).( 3 ) W . E. Fitzgerald and L . Nielscn, Proc. RO J . SOC.London),

2,662,874 (1953).

mers has mo re recently been refined. Below T, thefunctional groups are highly associated and immobile.Above T , chain mobility allows relatively free diffusionof the gro ups and t he fraction associated is determinedby a temperature-dependent equilibrium. --6 Super-ficially this lat ter beha vior is identical with that of smallmolecule analogs in non pol ar solvents. Th e thermo-dynamic parameters of the association, however, maydiffer from th e low molecular weight systems.In this work the effects of acrylic acid comon omer onthe properties of branched polyethylene are discussedwhere th e additional variable of crystallinity is present.Several other authors have discussed such changes forthe case of methacrylic acid and its metal salts. Theseexperiments have shown that ionization of the aciddepresses crystal formation m ore severely than does the

(4 ) R. Longworth and 13. Morawetz , J . P O ~ J , V I .c i . , 29 , 307( 5 ) E. P. O t o ck a an d F. R . Eirich, ibid. , in press.( 6 ) E. P. Otocka a n d F. R. Eirich, ibid., i n prcss.

(1958).12282, 137 (1962).


2/6

V O ~ ., NO.3, MLiJV-June IYh8 ErHYLENt-AC'KYLIC ACID C O P O L Y M E R S 245

init ia l incorporat ion of the comonomer , yet s t i f fensthe sample as a whole.7-9 O n the othe r han d, it isindicated that the p t rans i t ion of the copolymers israised mo re by the incorpo rat ion of the acid alone thanby its univalent metal salt. 0Four ethylene-acrylic acid copolymers of varyingcompos i t ion wer e ob ta ined f r om the Dow Chemica lCo. and Union C ar b ide Cor p . and the ac id content waschecked by t i t ra t ion. The methyl grou p content wasdetermined by infrared analysis. These findings aresummarized in Table I .Experimental Section

Infrared Spectra. Three separate series of infrared ex peri-ments were carried o ut on the acid copolymers to determinethe effects of composition, temperature, and elongation ontheir spectra. All data were taken on a Perkin-Elmer Model21 prism instrument. For the elongation studies, a pair ofPerkin-Elmer AgCl polarizers were employed.Two special mounting techniques were used fo r thestretching and heating experiments. A steel frame wasconstructed to fi t the spectropho tometer mounting slide. Onthis frame was a single screw, threaded in opposite directionson its upper and lower portions. As the screw is turned,the sample grips travelling on each threaded section moveapart. allowing the midpoint of the film to remain i n thecenter of the ir beam. Spectra were taken a t a number ofelongations to 200 with polarizations parallel and per-pendicular to the direction of stretch.A heated cell1' was modified to allow study of the copoly-mer films above T,,,without changes in thickness. The gasspace of the existing design was eliminated and circularaluminum foil shims equal to the sample thickness wereused as spacers. Circular samples were cut from thecopolymer films to fit snugly within the area defined by thespacer. The sample and spacer were held by moderatemechanical pressure between salt plates during the experi-ments. Isothermal (+1') scans of the 2-15-p region weretaken at about 10' intervals, and again during cooling. Theintensities of the major bands of the spectra here virtuallyidentical before and after heating.Finally, series of three films of differing thickness fromeach copolymer were scanned to determine the extinctioncoefficient of one oft he acid absorptions.Calorimetry. The melting points and heats of fusion forthe branched polyethylene reference and the four acid co-polymers were determined on the Perkin-Elmer differentialcalorimeter, DSC-1B. The instrument was calibrated usingthe melting points of water and benzoic acid, 0 an d 122".Samples were cut from copolymer films, melted, then crystal-lized by cooling at 5.0c/min. After 24 h r at 25 ' the samplewas remelted at 10'/min. The areas under the latter curveswere determined by planimeter for crystallinity calculations.The maxim um excursion from the base line was taken as themelting point. In a like manner, the maximum i n the crystal-lization exotherm was taken as the crystallization tempera-ture. Solutions of the copolymers inxylene (0.01 solids) were used i n an effort to grow crystalsfrom solution. While no single crystals could be obtained.al l the samples yielded at least some very small, open,spherulitic crystal sheaths. Nucleation would not occur

Solution Crystallization.

(7 ) R. W. Rees ,and D. J . V au g h n , P o / r m . Prepr in t s , 6 , 28 7(1965).( 8 ) R . W . Rees an d D . J. Vaughn, ib id . , 6, 296 (1965).(9 ) W . J . M acK n i y h t , e t a/ . , . A p p l . Phj , s . , 38 , 4208 (1967).( I O ) W. J . M acI< n i g h t , e t a[ . , bid., in press.( 1 1 ) M . G. C h a n and W. L . Hawkins , P o / J ~ .a g . Sci . , 7, 264( I 967).

TABLEPHYSICALROPERTIESFETHYLENE-ACRYLICCIDCOPOLYMERS~ _ _Melt CH 3: COO H: Density,Polym er index lOOCH2 lOOCH2 glee

A 2 0 3 5 0 0 918B 4 8 - 2 0 0 66 0 926C 5 0 -2 0 I 57 0 935D 4 8 > I 5 2 26 0 94 6E 6 9 > 1 . 5 2 78 0 953


3/6

24 6 E. P. OT OC KA AND T. K. KWEI Mucroniolecules

TABLE1EXTI NCTI ONOEFFICIENTSF ACID D I M E R ST 940 CM-1

Acid, e,Polymer mol/cc cm2/m ol- ~BCDE

1008060

40

20- 0.vY 4

2

1.00.a0.6OX

4 24 x 10-4 3 . 5 8 x 1049.71 X 3 . 4 5 x 101.66 x 10-3 3 52 X 1042.00 x 10-3 3.46 X loJ

O / A

kcal /mole0

4 2.6 2. 8 3.0 3.2 3.4 3IO I T ( K - )

Figure 1.i n ethylene copolymers: 0 ,heating; 0 , ooling.Association constants for the carboxylic acids

28.0I- 2 .6 5 2.4: .2m 2.0-212 1.81.6I 41.2

C - AD -E - v

2.2 1

X E L O N G A T I O NFigure 2 . Dichroic ratios for the 940- and 1705-cm-1absorptions as a function of elongation.

Samples of varying thickness of each of the copoly-mers were scanned in this region and extinction coeffi-cients calculated f rom the known acid content on theassumption that dimer izat ion was complete at roomtemperature . These results are shown in Table 11.If the assumption was faulty, there should be a differ-ence in the extinction coefficients from sample to

sample. Since this is no t the case, the data may beinterpreted as an indicat ion that abou t the sam e f ract ionof acid is associated in each samp le at room te mperature,an d that this fraction is very close to 1.0. This agreeswith the behavior of the 1750-cm-l peak.Th e 940-cm- absorption intensity was followed withtemperature , and the results were used to calculateassociation constants, K, . T h e K,s for the fou r copoly-mers are shown in Figure 1.T h e A H for dimer formatio n deduced f rom the s lopesof th e lines is -11.5 kcal/m ol. Thi s value is in excel-lent agreement with previously determined values foracid association in both polymeric and small moleculesolvent The A S per hydrogen bond for thefour copolymers at 50 range from -12.5 to - 1.3 euwith increasing acid content. The se A S values ingeneral agree with those observed in solutions of smallmolecule carboxylic acids.

An impor tan t f ea tur e of F igur e 1 is the continuityof the bonding equil ibr ium through the melt ing region.This, along with the facile reversibility of the reaction,indicates that the bonding occurs within the amorp housregion exclusively.Finally, the dichroic ratios of the 940- and 1705-cm-absorptions were determined on thin films stretchedstepwise to abou t 200%. Figure 2 shows that whilethe 1705-cm- absorption gives the expected weakperpendicular dichroism, the 940-cm-I absorption isstrongly polarized parallel to the direction of stretch.A space-filling mo del of th e copolym er clearly explainsthis. If the ethylene chain is in an all-rrans zigzag,which is highly favored during elongation, the acids ide gr oup r o ta t ion is blocked. Th e plane defined byth e -C(=O)O of the acid must be roughly perpendicularto the plane def ined by the backbone. The incorpora-tion of two guuche conformations in the backboneallow free rotation, but this is energetically unfavorable.Thus , when the polymer chains are al igned, the dimerplanes tend to be perpendicular to them a nd th e out-of -plane transition moment is parallel to the direction ofstretch.The dichroic ratios for the 1705-cm- absorption arelower because of absorptions in the 2 plane which donot regis ter in this experiment. The trend fo r alldichroic ratios to decrease with higher acid conten t is atpresent unexplained, but could be a consequence ofincreased translational molecular motion because ofthe decrease in crystallinity (to be discussed below).Calorimetry. The melt ing po ints an d crys tal l izat ionpoints are given in Table 111. Also shown a r e thecrystallinities of each sam ple, calculated from t he heatsof fusion and using 66 cal/g as the value for 100%crystalline polyethylene. While these transition tem-peratures are by no means ul t imate or equil ibr iumvalues, they serve as a valid yardstick for comparisonwithin t he series.The decreasing crystallinity with increasing acidcontent agrees with the trend reported by others7, andwith the observations of solution crystallization resultsmentioned ear l ier.The melting points of the acid copolymers decrease

( I 5) G. Piiiientcl and A. McClellan, Th e Hydrogen Bond,Appendix A , W . H. Fr eeman and Co., Sa i l Francisco, Calif. ,1960.


4/6

VoI. I , No . 3, Maj*-Jirne 1968 E T H Y L E N E - A C R Y L I CC IDC O P O L Y M E R S47

with increasing acid content. In the theory of Floryi6the dependence of the melt ing point o n the concentra-tion of crystallizable units in a random copolymer isgiven by

where T, is the copolymer melting point, T,," that of thecrystalline homopolymer, N is the mole f ract ion ofcrystallizable units in the cop olymer, an d AH , the hea tof fusion per mclle of repeating homopolymer unit.In these copolymers , both methyl and carboxyl groupsare present , and the proper choice of T," a n d N rests onthe fol lowing considerat ions . I t is known that themelting point depressions of polyethylenes with methyl,ethyl, and n-propyl substituents do not str ictly followFlory's equations. 79 l 8 The recent work of Holdsworthand Keller'g show's that in crystallization fro m solution,chain folded lamellae can be grown from polyethylenecontaining sizable amou nts of methyl and ethyl branches.Such branches art: acco mm oda ted thro ugh less perfectpacking and crystal habits. While crystallization frombulk m ay no t m,ske such generous allowances, it ishighly pr oba ble that m ethyl group s are crystallizable tosome extent , even in the bulk. Using, then, thebranched polyethylene "A" as a reference for zeroacid content, we should be able to test w hether the acidunits enter the crystal. A s a first approximation weassume the acid groups are not crys tal l izable , andcalculate N on the bas is that a l l othe r units are crys tal-l izable to the same extent as in polymer A. The r e-sults of this approach are shown in Figure 3. A straightline results, the slope of which gives a AH , of 1510cal/mol of ( -CH2CH2-) or 755 cal/mol of (-CH2-).Thi s value, str ictly speaking , is not th e same as that of amethylene unit in an ethylene homopolymer becausethe reference poin t has been changed. Nevertheless,the magnitude compares favorably with the AH , of950 cal/mol (-CH2-) (di luent method ) and 925 cal /mol(-CH2-) (calorimetric). T h e T," deduced f r om theintercept in Figure 3 is 108", vir tual ly the sam e as themelt ing point of sample A . The melt ing points andAH , give a ASu of about 2 cal/deg mol for these sys-tems.It is fully recognized that thjs applicatjon is not theway in which the equa t ion was to be used. However ,the ap parent success of this approx imation car r ies thes trong implicat ion that the acid units are not crys tal-l izable and reside in the amo rphous region of the copoly-mers.

Mechanical Testing. The loss tangent curves of thepolymers are shown in F igur e 4. The peak whichoccurs below 0" i n the low-density PE c an be identifiedas the /3 t ransi t ion. On the other hand, the CY transition(-80") which is very pronounced in high-densi ty PEcann ot be easily located in the less crystalline polymer A .T h e t a n 6 curve for copolymer B containing 3 wtacid shows a shoulder near 0" and then pas ses to abr oad maximum a t about 40". We shall tentatively(16) P . J. F lo ry , J . Chem. Ph y s . , 17, 22 3 (1949).(17) B. Ice, J . Pol.i,m. Sci . , 61, 7 (1962).(18) M . J . Richardson , P. J. Flory , and J. B. Jackson, Po/.v-mer . 4. 221 (1963) .( i 9 ) ' P . J . ' H o ld s \ + o r th qtid A . Kellcr, J . Polytn. Sci . , Part R, ,60.5 (1967).

TABLE11DIFFERENTIALCANNINGALORI METRYTUDIES

Crystallinity,Polymer T,,." "C "/, T,,b"CA 108* 1 40 90 rk 1B 105 2iz 1 40 87 iC l o o k 1 35 8 5 * 1D 97 * 1 21 so* 1E 9.5+ 1 16 77 * 1

a Heating rate, IO"/min. Cooling rate, S"/min.r"-

0 .01 .02 .03 .0 4 .05 .06- 1 0 9 N

Figure 3. Melting point depressions of the copolymers.

0.011 1 , 1 I , I-6 0 -40 -20 0 20 40 W 80TEMPERATURE ("c)

Figure 4. Loss tangents at 110 cps: 0, polymer A;0 ,B ; A,C; A, D; 0 ,E.identify the shoulder as related to the p transition. I tis not cer tain, however , whether the broad maximum a t40" has the sam e molecular mechanism as the 01 t ran-s i t ion in high-density P E s ince the copolymer has aboutthe same degree of crystallinity as the low-density PE.I t appears that the interplay of two relaxation mech-anisms prevents the clear separation of the loss peaksoccurring near 0 a n d 40". As the acid content in-creases, copolymer C shows only a broad maximumnear 40" and copolymers D an d E exhibit single sharp


5/6

248 E. P. OTOOCKAND T, K. KWEI

10'.

Mucroinolecules

AA-AD A d 0T P t t t + + A

I I I I I

TABLEVGLASS RANSI TI ONE M P E R A T U R E SF TH E COPOLYMERS

p x 103,Polymer T g , 'C AT,, " C moliccB - 5 1 6 0 . 3 8C 12 3 3 C . 8 0D 20 41 1 . 1 0E 24 45 1 . 2 4

, I I I I 1 I

L -I

TEMPERATURE ("C)Figure 5.Same symbols as in Figure 4.Dynamic elastic and loss moduli a t 110 cps.

t

i0m o l e

( P AM ORPHOUS P O L Y M E RFigure 6 . Change in glass temperature as a function ofhydrogen bond cross-link density. p is expressed as molesof acid dimer per gram of amorphous polymer or as molesof covalent cross-link per gram of PS or P M M A .loss peaks near 30". Apparen tly in the latter twocopolymers, one relaxation mechanism predominates.The s torage moduli E' of these materials are shownin Figure 5 . As acid content increases and crystallinitydecreases, the drop in modulus associated with thetransition becomes more pronounced. I t is also noted

that, above 25", the value of E' at a given temperaturedecreases as the crystallinity of the polymer decreases.I t is perhaps more appropr iate to examine the lossdispersions of these low-crystallinity materials on thebasis of loss modulus E".20 These curves are alsoshown in Figure 5 . Well-defined loss maxima occur at-21, - 5 , 13 , 20 , and 24" for polymers A-E, respec-tively. The broad maxima in the tan 6 curves of co-polymers B and C near 40" now appear only as small, ill-defined shou lders (if they exist at all) in the loss moduluscurves. In the ensuing discussion we shall direct ourattention only to the well-defined maxima. Furthe r-more, we shall identify these loss dispersions as arisingfrom segmental motions in the amorphous region,similar to th e 0 ransition in low-density polyethylene.These considerations are consistent with (1) the lowdegrees of crystallinity of these materials and ( 2 ) th eobserved rapid decrease in E' of copolymer D or Eduring the transition.Since all four copolymers have approximately thesame number of methyl branches as the low-density P E,the increase in T gas one progresses from polymer A tothe copolymer series may be attr ibuted to acid contentalone. The increase may arise from two possible con-tributions, the copolymerization effect and the cross-linking effect. In th e first, since polyacrylic acid has ahigh T,, namely, 106", a copolymer of ethylene andacrylic acid is expected to have a higher Tg.21 ow -ever, the copolymerization effect, as calculated for th eamor phous phase f r om the equa t ion of Gor don andTaylo r,22 gives estimated T , values of -17, -13,-9 , and - 5" for the copolymer series. Therefore,it is not likely that the observe d increase in T, canbe accounted for by the copolymer compositioneffect alone. In th e second, the acrylic acid unitsexisting mainly in the dimer form at room tempera-ture may act as effective cross-links which restrictthe motion of the backbone chain. In order tocompare the observed increase in T, with literaturedata on the effect of covalent cross-links on T,, wecalculate the cross-link density as follow s: (1) all th eacid units res ide in the amorphous phase (based onmelting point depression data), and ( 2 ) dimerizationis essentially complete (based o n ir results) . Th ecross-link densities, p , expressed in moles per gram ofcirnorphous material, and the increases in glass temper-atures , AT,, defined as [T,(copolymer) - T,(polymer A)]are listed in Table IV. A plot of AT, DS. p gives astraight line as shown in Figu re 6. The linear relation-ship between ATg and p was derived by Fox andLoshaek23 rom a consideration of the volume shr inkagedue to the formation of a covalent cross-link and wasexperimentally verified in cross-linked polystyrenez4o rpoly(methy1 methacrylate) * 5 when the contributionto T , by the copolym er compo sition effect of the cross-link was taken into consideration. (These data are alsoshown in Figure 6 for the purpose of comparison.) We

(20) K. H. I l lers and H. Breuer, J . Colloid S c i . , 18 , 1 (1963).(21) L . A. W o o d , J . Polym. Sci., 28, 319 (1958).(22) M . Gordon an d J . S. Taylor, J . A p p l . Chem. (London), 2 ,(23) T. G Fox a n d S. Loshaek , J . Polj'm. Sci. , 15 , 371 (1955 .(24) I


6/6

Vol. 1, No . 3, ML~~ I - J U I I C968 J ) , p -RIPHFNOI,- DIANILINOSIL,ANF CONDFNSATION COPOLYVERS 249

have no t a t t empted to ca r ry o u t a similar correction forthe composit ional con tr ibution of the acid units to AT,,because extensive dimerization of the carboxyl groupsalso exists in polyacrylic acid and it is impossible toassess indepen dently the compositional effect. Evenwith these reservations in mind, it is still believed thatthe increase in T, can be a t tr ibuted mainly t o the cross-linking effect, as implied by the agreement of our ex-per imental data with the theory of Fox and Loshaek.The smaller propor t ion ali ty constant between A T , an dp in our copolymers in compar ison with that foun d forpolystyrene or poly(methy1 methacrylate) reflects thefact that volume shr inkage accompanying the dimer i-zat ion of carboxyl groups is less than the volume de-crease produced by th e introdu ction of a covalentcross-link. Ou r results also fit the empirical equa t ionof Shibayama and Suzuki26 which are applicable tomany covalently cross-linked polymers

T, = K l In K2pWe therefore conclude that the dimer izat ion of th e acidunits in these copolym ers raises the T, of a branched PEcha in by a mechanism analog ous to the act ion of cova-lent cross-links.Once the Tp as been attained the segmental mob ilityal lows the dimer izat ion equil ibr ium to respond tothermodynamic control . Abo ve this tempe rature thedimer izat ion equil ibr ium is dynamic, Le., while thefraction of groups associated is specified by temper-ature , the identi ty of individual par tners is los t. Thismeans that interchange loss mechanisms are present a t

(26) I

Documents

EAA Properties