Colloid & P o l y m e r Science

Rheo-optical

Colloid & Polymer Sci. 262, 223-229 (1984)

Fourier-transform

infrared (FTIR) spectroscopy

of p o l y m e r s

5. Strain-induced crystallization of crosslinked natural rubber H. W. Siesler Bayer AG, Werk Dormagen, Research & Development, Dormagen, West Germany

Abstract: Rheo-optical FTIR spectroscopy has been applied to monitor the onset, progress and decay of strain-induced crystallization in loading-unloading cycles of sulfurcrosslinked natural rubber at 300 K and 343 K. From the short-time spectroscopic data conclusions were also drawn with respect to the orientation of the average polymer and the polymer chains in the strain-induced crystal phase.

Key words: rheo-optics, FTIR spectroscopy, natural rubber, strain-induced crystallization.

Introduction Apart from the degree of crosslinking the mechanical properties of a polymer having network structure are strongly influenced by strain-induced crystallization [1-7]. This phenomenon is of great practical importance both during processing [8] and with regard to the technological properties of the product such as tear strength and maximum extensibility [9]. Elongation of a crosslinked elastomer decreases the entropy of the network chains and the additional decrease in entropy required for crystallization to occur is therefore relatively small. A schematic representation of strain-induced crystallization within a polymer network which has been elongated by a force in the specified direction is shown in figure 1. Crystallites thus formed act as crosslinks of high functionality and since they are nondeformable at the stress levels involved, diminish the amount of elastomeric material able to respond to the imposed stress [2, 9]. Additionally, the crystallites act as filler particles which generally increase the modulus of a rubber-like material [9, 10]. In the stress-strain diagram strain-induced crystallization effects a drastic increase of elastic force with strain as a consequence of a significant selfreinforcement of the elastomer during elongation [1, 8, 9, 11] (see also fig. 2, curve a). The elastic force fis given by equation 1 in terms of the elongation ratio 3. [2]: f = vkTA(3.

-

3.-2)

If*]

=

2C~ + 2C23.-'

(2)

in which C1 and C2 are constants independent of 3. and the reduced stress or modulus [f*] is given by [14-16]:

If*I= A(3._f 3.-2) 9

(3)

A linear relationship between [f*] and 3.--1, however, holds only at small elongations and an upturn in the reduced stress occurs at small reciprocal elongations

_

(

strain

(1)

where v and A are the density of network chains, i.e. their number per unit volume and the undeformed cross-sectional area, respectively. Contrary to numerK 705

ous experimental results equation 1 implies that f should be proportional to 3. at very large elongations. Such stress-strain data are customarily represented using the semi-empirical equation of Mooney and Rivlin [12, 13] :

Fig. 1. Schematicof strain-inducedcrystallizationin a crosslinked elastomer

Colloid and Polymer Sdence, 17ol. 262. No. 3 (1984)

224

less than 0.4. The deviation from the linearity is controversely interpreted on the one hand by the limited chain extensibility [4, 17-19] and on the other hand by strain-induced crystallization [5, 6, 9]. Recently, the technique of rheo-optical FTIR spectroscopy [20-22] has proved a valuable alternative to study the phenomenon of strain-induced crystallization in crosslinked natural rubber. Thus, with the aid of a specially designed stretching apparatus short-time spectroscopic and mechanical data have been acquired simultaneously to the loading and unloading of film samples at ambient and elevated temperature and were evaluated in terms of the onset, progress and decay of strain-induced crystallinity and the corresponding changes in polymer chain alignment during elongation and recovery.

Experimental The rheo-optical spectra were obtained on a Nicolet 7199 FTIR spectrometer equipped with a Nicolet 1280 64 K computer. The electromechanical apparatus constructed for the simultaneous measurement of FTIR spectra and stress-strain diagrams during elongation, recovery and stress-relaxation of polymer films at variable temperature has been described in detail elsewhere [22]. By exploiting the automated information processing capability of the dedicated computer in the FTIR system much of the routine analysis of spectra series has been alleviated by specificaUy developed BASIC software [23]. Generally, in rheo-optical polarization measurements a series of spectra are recorded alternately with light polarized parallel and perpendicular to the direction of elongation. For the evaluation of the individual spectra programs were applied which - based on the peak maximum intensity automatically calculate the structural absorbance .Ao [20]:

Ao=

AII + 2A• 3

In the mechanical treatment the film samples were subjected to a loading-unloading cycle up to 530% strain with an elongation and recovery rate of 85% strain per minute and between 12- to 15-scan spectra were taken in 9- to 11-second intervals with a resolution of 4 cm-L

Results and discussion In figure 2 the stress-strain diagrams of the sulfurcrosslinked natural rubber measured at 300 K (a) and 343 K (b) are shown alongside the corresponding data taken at 300 K with the radiation-crosslinked synthetic polyisoprene (c) which did not crystallize during elongation primarily due to the configurational irregularity. The increasing tendency for strain-induced crystallization from (c) to (a) is accompanied by a significant enhancement of stress hysteresis [7% (c), 21% (b), 36% (a)] in the loading-unloading cycles of the different experiments. The hysteresis arises from the tendency of the crystals formed on extension to persist as the tensile force is reduced [see also fig. 5 (a)]. Ultimately, however, the amorphous state is recovered and the stress approaches the corresponding values in the loading half-cycle. In figure 2 (a) the wide-angle X-ray diagrams taken at 200% and 500% strain, respectively, of the stress-relaxed samples have been included to demonstrate the change in the state of order during elongation. The FTIR spectra monitored during such an elongation-recovery cycle of sulfur-crosslinked natural rubber at 300 K with light polarized alternately parallel and perpendicular to the direction of stretch are shown separately for the two polarization directions in figure 3. Although the actual absorption intensities

(4) 6-

or the dichroic function DF [24]:

300K (a)

DF=

(Aff A• (Aff A•

- 1 + 2

(5)

of specified absorption bands for each spectrum by appropriately correlating the successively measured absorbance values All and A• The structural absorbance was selected as intensity parameter because it eliminates the influence of changing orientation on the actual intensity of an absorption band. Changes in sample thickness during elongation were eliminated by ratioing against a suitable reference band. Upon data processing the individual spectroscopic values can be subsequently plotted as a function of strain with a separate software routine. The results were obtained with film samples of sulfur-crosslinked (1.8% S) natural rubber (100% 1,4-cis-polyisoprene) and a radiation-crosslinked synthetic polyisoprene (93% 1,4-cis-isomer). The specimens were prepared by microtoming from sheets of the polymers under investigation film sections which were 4 mm in width, 15 mm in length and about 100/an in thickness.

343K (b)

8 300K (c)

-"

1(30

2(~0

300 strain (%)

400

500

600

Fig. 2. Stress-strain diagrams of various crosslinked rubbers: (a) sulfur-crosslinked natural rubber at 300 K; (b) sulfur-crosslinked natural rubber at 343 K; (c) radiation-crosslinked synthetic 1,4-cispolyisoprene at 300 K

Siesler, Rheo-optical FTIR spectroscopy of polymers, 5.

225

2.501

rain (%)

8 .Q

/

1800

1600

1200

1000

800

wavenumbers

2.50'

2.00.

1.50'

%)

1.00-

0.50.

1800

1600

12~)

1000

800

wavenum~3ers

Fig. 3. FTIR polarization spectra taken alternately with light polarized parallel (a) and perpendicular (b) to the stretching direction during an elongation-recovery cycle of a rubber-crosslinked natural rubber at 300 K

are superimposed by dichroic effects in both spectra series absorption bands can be detected which increase in intensity despite reduction of sample thickness during elongation. For the accentuation of these crystallinity-sensitive bands the spectrum of the original sample has been subtracted from the spectrum of the 500% elongated sample with an appropriate scaling factor (fig. 4). ~In order to eliminate the influence of orientation effects the absorbance subtraction was performed on the structural absorbance spectra synthesized from the individual polarization spectra according to equation (4). In the difference spectrum [fig. 4 (c)] at least four absorption bands at 1378, 1362, 1126 and 844 cm -1 haven been isolated which can be associated with the crystalline phase formed during stretching. The discontinuity in the 1450cm -1

wavenumber region is caused by overabsorption of the 6(CH2) and 6a,(CH3) absorption bands in the spectrum of the unstrained sample. The vibrational spectrum of polyisoprene has been discussed by several authors [25-30] but owing to the lack of data manipulation capabilities before the advent of FTIR spectroscopy discrepancies existed as far as the accentuation and the unambiguous assignment of the crystallinity bands was concerned. To monitor the onset and extent of strain-induced crystallization during the loading procedure the 1126 cm -1 absorption band which has been assigned to a C-CH 3 in-plane deformation vibration [25, 27] and which shows the largest relative increase as a consequence of crystallization was utilized in the present study. The v(C=C) absorption band at 16

226

Colloid and Polymer Science, VoL 262. No. 3 (1984) 1.5o

1.00c

0.50-

1800 1.50

lr~O

1400

1200

10120

800

600

16~)

lab0

rob0

ld00

8~o

6oo

+6~

+4bo

lioo

~obo

s6o

800

1.00

0.5O-

1800 0.45

0.30-

c)

J~

o.15-

moo

wavenumbers

Fig. 4. Accentuation of the crystallinity-sensitive absorption bands in the spectrum of natural rubber by absorbance subtraction: (a) structural absorbance spectrum of 500% elongated sample; (b) structural absorbance spectrum of unstretched material; (c) difference spectrum (a)-(b)

1662 c m - 1 is almost perfectly compensated by the subtraction procedure [fig. 4 (c)] and has been applied as reference band to compensate the reduction of sample thickness during elongation. Thus, the ratio of the structural absorbances of the 1126 cm -~ and 1662 cm -1 bands was evaluated w i t h the aforementioned software routine for the spectra taken during the loading-unloading cycles of the individual polymers (fig. 2) and plotted as a function of strain for the elongation and recovery procedure in figure 5. The phenomenon of strain-induced crystallization is most clearly reflected by the data of curve (a) which correspond to the loading-unloading cycle of sulfurcrosslinked natural rubber at 300 K [fig. 2 (a)]. Here, the onset of crystallization can be assigned to a strain value of about 230% where the structural absorbance ratio first deviates from the initial linear curve with a subsequent drastic increase at larger strains. Further-

more, a significant retention of the strain-induced crystallinity relative to the loading half-cycle is observable during recovery down to the threshold value of crystallization. Thus, basically, the process of crystallization and melting is reversible and can take place in the time scale of the experiment [31, 32]. In order to test the effect of temperature increase on strain-induced crystallization a sulfur-crosslinked natural rubber film specimen has been subjected to an analogous mechanical treatment at a temperature of 343 K, approximately 70 K above the melting point of the crystal phase in the unstressed state [7]. In agreement with the results of thermoelastic investigations of natural rubber [33] the applied stress is initially higher in the elevated-temperature experiment [fig. 2 (b)] and crosses over to lower values at elongations were the strain-induced crystallization commences at room temperature. The reason for this behaviour and the smaller hysteresis can readily be derived from the corresponding structural absorbance/strain-plot in figure 5 (b). Thus, despite a comparatively larger scatter of the structural absorbance ratios due to a lower signal-to-noise ratio at elevated temperature the spectroscopic data clearly reflect the much lower degree of strain-induced crystallization at 343 K and the shift of its onset towards higher strains. In actual fact, strain-induced crystallization has been detected by rheo-optical FTIR measurements up to 373 K [34]. When the chains in the amorphous network are stretched out because of the applied deformation the entropy of fusion is significantly diminished. Due to the inverse relationship to the entropy the melting temperature of the crystal phase is thereby increased and crystallization is induced in some of the network chains even at temperatures far above the melting point of the unstressed sample [35]. As expected from the stress-strain diagram no crystallization can be detected in the radiation-crosslinked synthetic 1,4-cis-polyisoprene [fig. 5 (c)]. The relative shift of the structural absorbance ratios to higher values is merely due to a different baseline evaluation of the reference band for this polymer. An interesting effect becomes evident when the structural absorbance of the ~(C = C) thickness reference band is plotted as a function of strain (fig. 6) for the loading-unloading cycle of the strongly crystallizing system [fig. 2 (a)]. Thus, a distinct asymmetry in the recovery of the original sample geometry could be detected whenever the polymer under examination exhibited extensive strain-induced crystallization. The phenomenon may be correlated with the composite nature of such a strain-crystallized system and must be interpreted in terms of a preferential recovery of the

Siesler, Rheo-optical FTIR spectroscopy of polymers,

5.

227

0.8-

300K (c) zx 9

0.7 ~'

z~

"

300K (a)

z~ z~

"~ 0.6.

zx A

9

A

9

oO o~" ooo

0.5. .~i

,t~|

A9 t9

9 9

O~i

i

343 K (b)

%o,i, ~
s

4oo

50o

strain (%)

Plot of the structural absorbance ratios A~1261A~~ versus strain corresponding to the stress-strain diagrams of figure 2 (closed symbols: elongation, open symbols: recovery) Fig. 5.

the crystal phase and not associated with perfectly elongated individual chains only, is supported by orientation measurements at elevated temperature [34]. Thus, despite a drastic reduction of straininduced crystallization the degree of polymer chain orientation evaluated in terms of the dichroic ratio of the ~,(C=C) absorption band at 1662 cm -1 did not significantly deviate from the measurements at ambient temperature. To monitor the crystalline and average orientation of the strongly crystallizing system during a loadingunloading cycle [fig. 2 (a)], the dichroic functions [24] of the 1126 cm -1 and 1662 c m - 1 absorption bands, respectively, have been plotted versus strain in figure 7. The representation of the spectroscopic data in this form has been chosen because no values for the transition moment direction of these bands were available from the literature. Nevertheless, the onset of strain-induced crystallization and its retention during unloading can be readily derived from the dichroic function plot of the 1126 cm -1 absorption band. Samuels [24] has shown in detail, that if the dichroic functions for two different bands (a, b) from the same sample are plotted against each other, the slope of the resultant line will depend on the relation

Q8- 'A

k 07

0.5.

i'

AA 9

k k 0.6-

A a

Az~ Az~ i

a

OJ--

A

Z~

AA a

AA

o05"

ZX

9

A A

A~A a AA Z~aX~

A

0.3. a

AAAAA~Z~ZaAZXi x AAAAA A

0.4-

A

5

9 A AAAZ~Z~ ~A

~ O2,

&A ~A

iA

.o.9. o

160

~

360 strain

460

9 9

500

660

(%1

Fig. 6. Structural absorbance/strain-plot ofthe 1662 cm -1 v ( C = C ) reference band (closed symbols: elongation, open symbols: recovery) of polymer (a) in figure 2

thickness relative to the transverse dimension in the stress-strain plateau region of the unloading half-cycle between about 400% and 200% strain [fig. 2 (a)]. The assumption that the 1126 cm -1 absorption band is characteristic of the three-dimensional order in

.e

,.~ 0.1.

0.0

160

260

360 (Y,,)

460

560

660

strain

Fig. 7. Dichroic function/strain-plot of the 1662 cm -1 (O) and 1126 cm -1 ( A ) absorption bands corresponding to the mechanical

treatment of polymer (a) in figure 2 (closed symbols: elongation, open symbols: recovery) 16"

Colloid and Polymer Science, Vol. 262. No. 3 (1984)

228

between their respective orienting phases and transition moment angles, since [ ( g - 1)/(g + _ s [(R - 1)/(R + 2)]b 5

[(R o + 2)/(Ro -

1)] b

[(Ro + 2)/(Ro -

1)]# (6)

Here, R is the dichroic ratio A f f A z , f t h e orientation function and R o = 2 cot 2 a, where a represents the transition moment angle of the absorption band under consideration. Figure 8 shows the computer-plots of equation (6) in terms of the 1126 cm -1 and 1662 cm -1 absorption bands for the elongation half-cycles of two independent experiments at 300 K. This curve can be roughly decomposed into two straight lines intersecting at the point which is characteristic of the onset of straininduced crystallization. While the straight line left of the intersection corresponds to the strain interval where the contributions of both absorption bands can be assigned to the amorphous phase (0-230% strain), the steeper slope to the right of the intersection is a consequence of the newly developing strain-induced crystal phase. Thus, from the smaller slope of the straight line corresponding to the amorphous phase and under the assumption that the transition moment angle a of the 1126 cm -1 band is zero (the dichroic function is then equivalent to the orientation function

O

LO 1

D

A

(0 o C~ ~D

CC

, D

\

D ~

I Q ff-t'd

if,

,

%

o f F ,00

~ '.20 '.~0 [R-I~/(R+2)

~.60 I~62

CM

~.80 -1

'1.00

'1.20

~I0~-I

Fig. 8. Computer-plot of the dichroic functions of the 1126 cm -1 and 1662 c m - 1 absorption bands monitored during two independent elongation half-cycles (~/[] ) in analogy to polymer (a) of figure 2

of this absorption band) an angle of approximately 40 ~ was derived for the transition moment direction of the v(C= C) absorption band at 1662 cm -1 and the polymer chain axis. The drastic improvement of chain alignment in the strain-crystallizing phase relative to the average polymer is reflected by the increase of the slope to the right of the inflection point by a factor of about 1.6. From the orientation and dichroic functions of the 1126 cm -1 and 1662 cm -1 bands, respectively, an average inclination angle of approximately 35 ~ was calculated for the polymer chains of the crystal phase and the direction of stretch for the 530% drawn sample whereas a value of about 44 ~ was determined for the chain alignment of the average polymer. Rheo-optical FTIR spectroscopy, therefore, not only provides a means to monitor strain-induced crystallization on-line to the mechanical treatment but also yields detailed information in terms of the orientation of the polymer chains in the average polymer relative to the strain-crystallizing phase. Acknowledgments The author gratefully acknowledges valuable discussions and the supply of the polymer samples by Dr. U. Eisele (Bayer AG, Leverkusen) and thanks Bayer AG for the permission to publish the experimental data. References

1. Flory PJ (1947) J Chem Phys 15:397 2. Flory PJ (1953) Principles of Polymer Chemistry, Cornell University Press, Ithaca, NY 3. Andrews EH, Gent AN (1963) Crystallization in Natural Rubber in "The Chemistry and Physics of Rubberlike Substances" (Bateman L ed), Wiley, New York 4. Treloar LRG (1975) The Physics of Rubber Elasticity, 3rd ed, Clarendon Press, Oxford 5. Davies CKL, Wolfe SV, Gelling IR, Thomas AG Polymer 24:107 6. Chiu DS, Su TK, Mark JE (1977) Macromolecules 10:1110 7. Shimomura Y, White JL, Spruiell JE (1982) J Appl. Polym Sci 27:3553 8. Eisele U (1979) Progr Collo & Polym Sci 66:59 9. Mark JE (1979) Polym Eng Sci 19:254, 409 10. Smith TL (1977) Polym Eng Sci 17:129 11. Gee G (1947) J Polym Sci 2:451 12. Mooney M (1948) J Appl Phys 19:434 13. Rivlin RS (1948) Phil Trans Royal Soc (London) A-241:379 14. Ciferri A, Flory PJ (1959) J Appl Phys 30:1498 15. Mark JE, Flory PJ (1966) J Appl Phys 37:4635 16. Mark JE (1975) Rubber Chem Technol 48:495 17. Doherty WOS, Lee KL, Treloar LRG (1980) British Polym J 3:19 18. Furukawa J, Onouchi Y, Inagaki S, Okamoto H (1981) Polym Bull 6:381 19. Treloar LRG, Riding G (1979) Proc Roy Soc A369:281 20. Siesler HW, Holland-Moritz K (1980) Infrared and Raman Spectroscopy of Polymers, Marcel Dekker, New York

Siesler, Rheo-optical FTIR spectroscopy of polymers, 5. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

Holland-Moritz K, Siesler HW (1981) Polym Bull 4:165 Siesler HW (1983) Polym Bull 9:382 Siesler HW, Schlemmer HP, unpublished results Samuels RJ (1981) Makromol Chem Suppl 4:241 Sutherland GBBM, Jones HV (1950) Trans Farad Soc 9:281 Saunders RA, Smith DC (1949) J Appl Phys 20:953 Binder JL (1963) J Polym Sci A1:37 Binder JL (1969) Appl Spectroscopy 23:17 Golub MA (1959) J Polym Sci 36:523 Gotoh R, Takenaka T, Hayama N (1965) Kolloid-Z Z Polym 205:18 31. De Candia F, Romano G, Russo R, Vittoria V (1982) J Polym Sci Polym Phys Ed 20:1525

229 32. 33. 34. 35.

Mitchell JC, Meier DJ (1968) J Polym Sci A2:6, 1689 Mark JE (1976) J Polym Sci Macromol Revs 11:135 Siesler HW, unpublished resuks Mark JE (1981) J Chem Ed 58:898 Received July 7, 1983; accepted September 30, 1983

Author's address : H. W. Siesler Bayer AG, Werk Dormagen Research & Development I)-4047 Dormagen