EpoxyRubber Interactions - American Chemical Society


EpoxyRubber Interactions - American Chemical Societypubs.acs.org/doi/pdfplus/10.1021/ba-1993-0233.ch012additional amount...

1 downloads 98 Views 1MB Size

12

Downloaded by UNIV OF CALIFORNIA SANTA BARBARA on March 20, 2016 | http://pubs.acs.org Publication Date: May 5, 1993 | doi: 10.1021/ba-1993-0233.ch012

Epoxy-Rubber Interactions F. J. McGarry and R. B. Rosner Department of Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139

Films containing amine-terminated butadiene-acrylonitrile rubber and diglycidal ether of bisphenol A (DGEBA)

(ATBN)

epoxy, cross-

-linked with amine curing agent, exhibit tensile extensibility over the composition range of 50-600 parts by weight rubber to 100 parts by weight epoxy. This tensile extensibility suggests the presence of ductile behavior in the second-phase particles of ATBN DGEBA

rubber-toughened

epoxy systems, even if the particles contain substantial

amounts of epoxy. Such cured films also are capable of absorbing large additional amounts of liquid epoxy that contains the cure agent. When the epoxy is cured in situ, the film tensile behavior is consistent with the overall proportions of rubber and epoxy present. The solubility behavior also suggests that the glassy epoxy matrix immediately surrounding a precipated particle contains rubber in solid solution and thereby can plastically yield under shear-stress

action. As observa-

tions confirm, such flow would be heat recoverable.

CROSS-LINKED EPOXY RESINS OFFER EXCELLENT PROPERTIES AT MODERATE cost, and because of this they are widely used in coatings, adhesives, and fiber-reinforced composites. The same cross-linking that produces strength, stiffness, and chemical resistance often results in brittleness. However, several decades ago it was found that the addition of certain low-molecular-weight elastomers could confer additional cracking resistance and fracture toughness (1). These elastomers are carboxy- or amine-terminated copolymers of butadiene and acrylonitrile ( C T B N and A T B N , respectively) that are liquid at room temperature and soluble in most liquid epoxies. The reactive end groups form chemical bonds with the epoxy, and as the system cures, the rubber separates out to form small particles in the epoxy matrix. The 0065-2393/93/0233-0305$06.00/0 © 1993 American Chemical Society

Riew and Kinloch; Toughened Plastics I Advances in Chemistry; American Chemical Society: Washington, DC, 1993.

Downloaded by UNIV OF CALIFORNIA SANTA BARBARA on March 20, 2016 | http://pubs.acs.org Publication Date: May 5, 1993 | doi: 10.1021/ba-1993-0233.ch012

306

RUBBER-TOUGHENED PLASTICS

two-phase morphology results in improved fracture toughness without reducing the other attractive properties of the glassy matrix. (If the rubber remains in solid solution, reductions in modulus, strength, hardness, and softening temperature take place.) Another route to the same end product involves prereacting the rubber and epoxy with certain catalysts such that better control over the (preformed) particle size can be realized (2). In both instances, it is necessary to achieve the discrete second phase. Furthermore, it is believed that chemical bonding between the rubber particles and the surrounding epoxy matrix is essential to the toughening action. Progress has been made in understanding the toughening mechanism and how it relates to the morphology of rubber-modified epoxies. The stress-whitening observed in failed samples suggested the occurrence of plastic flow, possibly in the form of crazing (3), but further studies showed that the areas around the rubber particles undergo shear-yielding and cavitation at the fracture surface (4, 5). Other work demonstrated that crazing is unlikely because of the high strand density present in the epoxy (6). A theory that explains the shearing and cavitation as the result of a two-step mechanism has been presented (7, 8). Initially, the dilatational stresses that develop near the crack tip cause debonding at the particle-matrix interface and microcavitation around each rubber particle. The failure of the particles induces further shear-yielding in the adjacent matrix. The theory emphasizes that the particles also keep the shear forces localized, thereby delaying catastrophic failure, but it is not clear why the adjacent matrix epoxy seems especially prone to such large-scale plastic flow. Further, the flow is recoverable if the material is heated above its glass-transition temperature. Most research in the area has concentrated on the relationships between the morphology and the physical properties of rubber-toughened thermosets. Comparatively little has been done to investigate the properties of the rubber particles themselves, although the rubber particles are the critical part of the toughening mechanism. For example, estimates of the composition of the particles range from a high rubber content of 80% (9) to less than 30% by weight (JO). A curious observation is also found in studies of the rubbermatrix interface: nuclear magnetic resonance relaxation data indicate that the interface is very sharp (I J), yet under moderately different curing conditions the rubber may never precipitate at all (12). To better understand the nature of the rubbery occlusions, a number of elastomer films were cast from mixtures of A T B N and catalyzed epoxy resin. The homogenous single-phase films were used as model materials to determine the relationship between the composition and the mechanical properties of the rubber particles. The solubility of the rubber in the epoxy was also investigated by observing the swelling behavior of the films in liquid epoxy. These results may be useful in better understanding the interfacial region in such rubber-modified thermosets.

Riew and Kinloch; Toughened Plastics I Advances in Chemistry; American Chemical Society: Washington, DC, 1993.

12.

MCGARRY AND ROSNER

307

Epoxy-Rubber Interactions

Downloaded by UNIV OF CALIFORNIA SANTA BARBARA on March 20, 2016 | http://pubs.acs.org Publication Date: May 5, 1993 | doi: 10.1021/ba-1993-0233.ch012

Experimental Details The epoxy resin used in all the experiments was a diglycidal ether of bisphenol A ( D G E B A ; Epon 828; Shell Chemical Company) that had an average molecular weight of 380 g/mol. The rubber modifier was an amineterminated butadiene acrylonitrile, copolymer (ATBN) that contained 17% bound acrylonitrile, and had a number average molecular weight of 3400 g/mol. A T B N is sold under the trade name of Hycar (BFGoodrich Chemical Company). The curing agent was tris(dimethyl aminomethyl) phenol tri(2ethyl hexoate), an amine salt catalyst (Ancamine K61B; Pacific Anchor Chemical Company). Film formulations are given in Table I. When possible, the amount of curing agent was adjusted to ensure that the ratio of amine to epoxy groups was maintained at unity. All three components were dissolved in toluene to make a 30% solids solution, which was cast on Teflon-coated plates. The films stood for 12 h under ambient conditions to allow the solvent to evaporate, were dried under room temperature vacuum for 1 h, and cured in a vacuum oven (0.5 torr) for 1 h at 60 °C and 2 h at 120 °C. The thicknesses of the final films ranged from 12 to 20 μπι, but the thickness of any individual did not vary by more than 0.5 μπι. Studies of the infrared spectra of these formulations taken at various points in the cure cycle showed that the samples were fully reacted at the end of the cure. In all cases the epoxy absorption at 863 c m , measured with a Fourier transform infrared spectrometer (Cygnus model 100), completely disappeared. Furthermore, no changes in the spectra were observed after curing for longer times or at higher temperatures. (Most of the films were made simultaneously to minimize any effects that could arise from inconsis­ tencies in sample preparation.) At room temperature, the films were die cut into dog-bone-shaped specimens to measure their tensile strengths. Room temperature stress-strain curves to failure were recorded on a tensile tester (Instron model 1122) at a crosshead speed of 5 mm/min. Absorption experiments were performed by immersing samples in a liquid and weighing them at various times until they reached saturation. Several swelling media were used: (1) a 50/50 mixture of acetone and methyl ethyl ketone ( M E K ) ; (2) 100% toluene; and (3) solutions of the 50/50 acetone-MEK mixture that contained varying amounts of - 1

Table I. Film Formulations Component Epoxy ATBN(X16) Curing agent Amine-Epoxy ratio (A/E)

Formulae Based on 100 pbw Epoxy 100 0 10 1.0

100 50 9.0

100 100 7.9

100 200 5.8

100 300 3.7

100 400 1.6

1.0

1.0

1.0

1.0

1.0

100 475 0 1.0

100 550 0 1.16

Riew and Kinloch; Toughened Plastics I Advances in Chemistry; American Chemical Society: Washington, DC, 1993.

100 600 0 1.26

308

RUBBER-TOUGHENED PLASTICS

Downloaded by UNIV OF CALIFORNIA SANTA BARBARA on March 20, 2016 | http://pubs.acs.org Publication Date: May 5, 1993 | doi: 10.1021/ba-1993-0233.ch012

catalyzed epoxy resin in dissolution (Table II). This last series of solutions, (3), was used to impregnate the films with increasing amounts of epoxy and curing agent to observe the degree of absorption that could be expected from the pure catalyzed resin and what effect this could have on the physical properties of the films. Dog-bone specimens that had been infused with liquid resin and catalyst were cured a second time following the same procedure described earlier and also were tested on the tensile tester.

Results The stress-strain curves for the initial cast films containing between 50 and 600 parts by weight (pbw) of the A T B N modifier are shown in Figure 1. The behavior ranges from a leathery plastic to a soft rubber, depending on Table II. Composition of Swelling Media (By Weight Percent) Solvent 1 Toluene (100%) Acetone (50%) Acetone (45%) Acetone (35%) Acetone (20%)

Solvent 2 MEK MEK MEK MEK

(50%) (45%) (35%) (20%)

Solute

Catalyzed resin" (10%) Catalyzed resin* (30%) Catalyzed resin* (60%)

" Catalyzed resin = 91% DGEBA liquid epoxy (Epon 828), Ancamine K61 B.

5000

STRAIN (%) Figure 1. Stress-strain curves forfilmsof various ATBN content.

Riew and Kinloch; Toughened Plastics I Advances in Chemistry; American Chemical Society: Washington, DC, 1993.

Downloaded by UNIV OF CALIFORNIA SANTA BARBARA on March 20, 2016 | http://pubs.acs.org Publication Date: May 5, 1993 | doi: 10.1021/ba-1993-0233.ch012

12.

MCGARRY AND ROSNER

309

Epoxy-Rubher Interactions

composition, but in all cases a substantial amount of ductility is present. Plots of individual parameters, such as modulus, ultimate strain, and ultimate tensile strength (UTS), as functions of the A T B N content convey the relation­ ship between composition and properties and are shown in Figures 2, 3, and 4, respectively. Three regions of behavior are observed. Transitions occur at A T B N concentrations of 150 and 475 pbw. In the first region, where the A T B N concentrations are less than 150 pbw, the film acts like a flexibilized plastic: both modulus and strength are relatively high, but the samples are tough and

ATBN Content (pbw) Figure 2. Modulus at 75-100% strain as a function of ATBN rubber content. 500

I ΙΪΙ Kl



400

Ji

C

.2

a Β χ

300

y

W

ε

1

•a

D

β

0 0

100

300

500

700

A T B N Content (pbw) Figure 3. Ultimate extension as a function of the ATBN rubber content.

Riew and Kinloch; Toughened Plastics I Advances in Chemistry; American Chemical Society: Washington, DC, 1993.

Downloaded by UNIV OF CALIFORNIA SANTA BARBARA on March 20, 2016 | http://pubs.acs.org Publication Date: May 5, 1993 | doi: 10.1021/ba-1993-0233.ch012

310

RUBBER-TOUGHENED PLASTICS

Figure 4. Ultimate tensile strength as a function of ATBN rubber content.

leathery. Other studies have reported similar properties when the rubber modifier content ranges from 20 to 100 pbw (13). Films with intermediate A T B N content, between 150 and 475 pbw, behave like a rubber. The stress-strain curves have the familiar shape of an elastomer, and the modulus and UTS decrease linearly with increasing A T B N content whereas the ultimate extension increases slightly. The third region starts at A T B N concentrations > 475 pbw, when the ratio of amine-to-epoxy units in the formulation exceeds unity. Under these conditions the modulus and UTS level off whereas the ultimate extension undergoes a large increase. One explanation for this behavior is the nonstoichiometric nature of the formulations used: As the amine-to-epoxy ratio departs from unity, more of the functional end groups of the A T B N chains remain unreacted after all the epoxy is consumed. The unreacted chains are not chemically incorporated into the cross-linked network of the film and thus do not contribute to the film strength. Further, it is probable that the unreacted chains act as plasticizers and promote intermolecular slippage, which results in the higher strains. Returning now to the two-phase systems of rubber-toughened epoxy, if the elastomer particles in the glassy matrix have compositions similar to the corresponding films, then it is possible to estimate the elastomer properties by the data on the films. Although the exact composition of the particles is not known, the rubber within them has been estimated to range between 30 and 80% by weight (as mentioned previously). This estimate corresponds to A T B N contents of approximately 45 and 400 pbw in the homogenous films. Thus, the particle properties probably correspond to regions 1 and 2 in the film behavior: a transition from a flexibilized glass to a typical rubber. In all

Riew and Kinloch; Toughened Plastics I Advances in Chemistry; American Chemical Society: Washington, DC, 1993.

Downloaded by UNIV OF CALIFORNIA SANTA BARBARA on March 20, 2016 | http://pubs.acs.org Publication Date: May 5, 1993 | doi: 10.1021/ba-1993-0233.ch012

12.

MCGARRY AND ROSNER

Epoxy-Rubber Interactions

311

cases, considerable ductility seems to exist; the particles would be highly deformable. Now consider the rubber particle-epoxy interface. How probable is a sharp compositional discontinuity here? Some idea of the thermodynamic interaction may be inferred from the swelling behavior of the films in various solvents. The ability of a liquid to act as a solvent for the rubber in the particle is shown by the amount of that liquid absorbed by the cross-linked film. The better the solvent, the greater the degree of absorption expected. Cured homogenous films containing 300 pbw of A T B N in two organic solvents swelled rapidly and significantly. Figure 5 shows that room tempera­ ture immersion in a 50/50 mixture of acetone and M E K caused a doubling in weight; immersion in toluene produced a 200% increase. This weight gain demonstrates that both mixtures are excellent solvents for the D G E B A - A T B N rubber. Immersion of the samples in hot liquid epoxy (uncatalyzed D G E B A at 105 °C) produced different but no less dramatic results. Unlike the organic solvents where the films reached saturation within minutes, the samples in the epoxy gained weight more slowly without approaching saturation (see Figure 6). The films continued to absorb for up to 12 h before finally disintegrating: At elevated temperatures, the resin is a good solvent. The temperature dependence is shown in Figure 6. If the swelling of the film by the hot epoxy is considered as a kinetic process, then the slope of the plot of rate vs. inverse temperature can be used with the Arrhenius equation to determine the activation energy. The data give a value of 20 kj/mol, which is lower than the value for a chemical reaction (14), but reasonable for a thermodynamic interaction.

300 Toluene

c

•a α g

-i

200

1Acetone/MEK

100

J

,

100

200

Time (Min) Figure 5. Weight gain offilmscontaining 300-phw ATRN immersed in organic solvents at 25 °C.

Riew and Kinloch; Toughened Plastics I Advances in Chemistry; American Chemical Society: Washington, DC, 1993.

312

RUBBER-TOUGHENED PLASTICS

Downloaded by UNIV OF CALIFORNIA SANTA BARBARA on March 20, 2016 | http://pubs.acs.org Publication Date: May 5, 1993 | doi: 10.1021/ba-1993-0233.ch012

Previous studies of the phase separation of solutions of liquid epoxy and C T B N indicated that the acrylonitrile content of the rubber is an important factor (10). Absorption experiments performed with cured films that contain various amounts of acrylonitrile confirm these findings. Figure 7 shows the liquid epoxy uptake at 105 °C of three films that each contain 300-pbw A T B N with 0-, 10- and 17-wt % acrylonitrile. Despite the differences in thickness,

500

400 Time (Min) Figure 6. Weight gain offilmscontaining 300-pbw ATBN immersed in DGEBA (Epon 828) at various temperatures as a junction of time.

200

* 17% A N , 450 μ • 10% A N , 400 μ • 0%AN, 350 μ

Time (Min) Figure 7. Weight gain offilmscontaining 300-pbw ATBN immersed in DGEBA (Epon 828) at 105 °C as a function of time. Acrylonitrile content of the ATBN andfilmthickness are given in the legend.

Riew and Kinloch; Toughened Plastics I Advances in Chemistry; American Chemical Society: Washington, DC, 1993.

MCGARRY AND ROSNER

Downloaded by UNIV OF CALIFORNIA SANTA BARBARA on March 20, 2016 | http://pubs.acs.org Publication Date: May 5, 1993 | doi: 10.1021/ba-1993-0233.ch012

12.

313

Epoxy-Rubber Interactions

the data show that the acrylonitrile content is the dominant factor in the swelling behavior and, therefore, it is also influential on the solubility of the rubber in liquid epoxy. From these results it is apparent that, during the curing of a rubber-toughened epoxy, a considerable amount of local mixing and interdiffusion between the rubber and the epoxy matrix is probable before gelation. This means that the boundary that develops between the two phases in the rubber-modified formulation will be broad on a molecular scale and the surrounding matrix will contain a finite amount of the rubber in solid solution (15). Such a situation could explain the large local shear deforma­ tions in the epoxy when a crack passes through the region: The rubber-plasticized epoxy yields and flows, and absorbs substantial work that is manifest as increased fracture toughness. The situation also explains why the cavity in the epoxy, which is usually much larger than the rubber particle resident in the epoxy, retracts and collapses when the material is heated above its (locally reduced) glass-transition temperature. In Figures 1-4, the properties of epoxy—rubber formulations cured to a single-phase structure were presented and discussed. Figures 6 and 7 show that similar, cured-rubber films can absorb substantial amounts of liquid epoxy. In Figure 8, the results of another set of absorption experiments with a rubber film are presented. The film contained 300-pbw A T B N , 100-pbw D G E B A (Epon 828), and 3.7-pbw amine salt catalyst (Ancamine) curing agent. The film was dried and cured in the manner already described, and then it was immersed at room temperature in various solutions of a e e t o n e - M E K - l i q u i d D G E B A resin catalyzed by the amine salt (Ancamine), as listed in Table II. The solution uptake is shown by the upper curve in

200 Q Swelling in Solution

c

•a α

A Permanent Swelling

'53

40

60

80

100

Resin Concentration in Solution (%) Figure 8. Weight gain of afilmcontaining 300-pbw ATBN rubber in solution and after drying

Riew and Kinloch; Toughened Plastics I Advances in Chemistry; American Chemical Society: Washington, DC, 1993.

Downloaded by UNIV OF CALIFORNIA SANTA BARBARA on March 20, 2016 | http://pubs.acs.org Publication Date: May 5, 1993 | doi: 10.1021/ba-1993-0233.ch012

314

RUBBER-TOUGHENED PLASTICS

Figure 8. Then the samples were removed from the solvent solution and the acetone-Mek solvent was desorbed and evaporated. The residual permanent weight gain was due to the epoxy-curing agent and it was linearly propor­ tional to the concentration of epoxy-curing agent in the immersion solution. (Extrapolating the two curves shows an expected uptake of 160% weight gain in 100% liquid epoxy at 25 °C.) The new composition of each film specimen can be determined from the amount of resin and curing agent absorbed. For example, i f a film originally composed of 300-pbw A T B N and 100-pbw D G E B A showed a permanent weight increase of 52% due to the addition of epoxy and curing agent, the new composition would be 300-pbw A T B N and 291-pbw D G E B A . If the formulation is again normalized by the epoxy content, the new composition becomes 103-pbw A T B N and 100-pbw epoxy; the A T B N content has been effectively lowered by nearly two-thirds. After a second vacuum heat treatment to cure the newly absorbed epoxy that had diffused into the film, the system appears to form a structure very similar to those produced for Figures 1-4. The samples were tested in the same way as the original tensile specimens. Figure 9 compares the moduli of the original films from Figure 2 with the moduli of films modified with the imbibed catalyzed resin. Both samples pass through the flexibilized plastic-elastomer transition at about the same point, but the original films appear to be somewhat stiffer. Despite this difference, the results suggest that fully cured ATBN-based elastomers are capable of absorbing large quantities of epoxy and curing agent, which can then be subsequently incorporated into the network of the film by a second cure. It is tempting to

3000

20



a

α»



^

\

2000 cp

^ \ \ \\\

J 10 "8

\\ \\

2

0

100

200

h 1000

300

400

ι

Q Original Films

A Films Swollen with Resin

500

A T B N Content (pbw) Figure 9. Comparison of modulus as a function of ATBN rubber content for originalfilmsandfilmsswollen by resin and catalyst.

Riew and Kinloch; Toughened Plastics I Advances in Chemistry; American Chemical Society: Washington, DC, 1993.

12.

MCGARRY AND ROSNER

Epoxy-Rubber Interactions

315

speculate that an interpenetrating network is produced in such a manner, but that remains to be proven.

Downloaded by UNIV OF CALIFORNIA SANTA BARBARA on March 20, 2016 | http://pubs.acs.org Publication Date: May 5, 1993 | doi: 10.1021/ba-1993-0233.ch012

Conclusions The solubility of reactive liquid rubbers in liquid D G E B A is widely recog­ nized. This solubility can be exploited to produce two-phase cured solid mixtures in which the rubber-rich particles increase the fracture toughness of the glassy epoxy matrix by causing plastic shear deformation around the particles. Even if the particles contain substantial amounts of epoxy, they can still exhibit rubbery properties: low tensile modulus and considerable extensi­ bility before fracture. The solubility of liquid epoxy in cured films of rubber-epoxy also is substantial. Liquid epoxy can be cured i n situ to produce films with properties consistent with the overall proportions of epoxy and rubber present. Because of such solubility, it is postulated that the epoxy immediately surrounding a particle i n the two-phase toughened mixture contains rubber in solid solution or reacted with the epoxy, enhancing its ability to plastically deform. Solubility considerations strongly suggest the existence of an interphase rather than a sharp boundary i n the rubber particle-glassy epoxy systems.

References 1. McGarry, F. J. Proc. R. Soc. London A 1970, 319, 59. 2. Rubber-Toughened Plastics; Riew, C. K., Ed.; Advances in Chemistry 222; Ameri­ can Chemical Society: Washington, D C , 1989. 3. Rowe, E. H . ; Riew, C. K. Plast. Eng. 1975, 31, 45. 4. Bascom, W. D.; Ting, R. Y.; Moulton, R. J.; Riew, C. K.; Siebert, A. R.J.Mater. Sci. 1981, 16, 2657. 5. Sultan, J.; McGarry, F. J. Polym. Eng. Sci. 1973, 13(1), 29. 6. Glad, M. D. Ph.D. Thesis, Cornell University, 1986. 7. Kinloch, A. J.; Shaw, S. J.; Tod, D. Α.; Hunston, D. L. Polymer 1983, 24, 1341. 8. Kinloch, A. J.; Shaw, S. J.; Hunston, D. L. Polymer 1983, 24, 1355.

9. Kalfoglou, N. K.; Williams, H . L. J. Appl. Polym. Sci. 1973, 17, 1377. 10. Wang, T. T.; Zupko, Η. M. J. Appl. Polym. Sci. 1981, 26, 2391.

11. Sayre, J. Α.; Assink, R. Α.; Lagasse, R. R. Polymer 1981, 22, 87. 12. Manzione, L. T.; Gillham, J. K.; McPherson, C. A. J. Appl. Polym. Sci. 1981, 889.

26,

13. Riew, C. K. Rubber Chem. Technol. 1981, 54, 374.

14. Barrows, G. M. Physical Chemistry; 4th ed.; McGraw-Hill: New York, 1979; p 683. 15. Bascom, W. D.; Cottinger, R. L.; Jones, R. L.; Peyser, P. J. Appl. Polym. Sci. 1975, 19, 2545.

RECEIVED

for review March 6, 1991.

ACCEPTED

revised manuscript June 17,

1992.

Riew and Kinloch; Toughened Plastics I Advances in Chemistry; American Chemical Society: Washington, DC, 1993.