374 G.A.Schoeppner,G.P.Tandon and K.V.Pochiraju the fiber axis direction.The mechanical (stiffness and strength)and thermal (coefficient of thermal expansion [52]and thermal conductivity [51]) properties of these oriented PAN-based fibers are highly anisotropic in nature.Typical density values of these fibers range between 1.75 and 1.86g cmand a porosity of approximately 15%with most of the porosity appearing in needle-shaped forms with axes of the needles parallel to the fiber axis [77].Therefore,it follows that the diffusivity of the fibers will likely be anisotropic in nature.Unfortunately,experimental measures of carbon fiber diffusivity have not been obtained. Numerous researchers have attributed the anisotropic diffusivity of unidirectional laminates to the preferential oxidation along the fiber-matrix interface or along the interphase region.However,the interphase region is a very small-volume percentage of the composite and a very small area fraction of the surface area.It is questionable that the diffusion of oxygen through and subsequent oxidation of the interphase region by itself can have such a significant effect on the composite oxidation process.More likely,the rapid oxidation rate along the fiber length indicates that,in addition to diffusion of oxygen from the specimen surface to the oxidation front,the interphase region has a supplemental oxygen path or source.Two possible mechanisms or scenarios for the supplemental oxygen source to the interphase are being investigated.Firstly,if oxidation in the interphase region leads to early fiber-matrix interface debonds,the debonds provide a pathway for oxygen to penetrate deeper into the composite.Knowing that there are residual curing stresses at the fiber-matrix interface,it is likely that such a scenario would entail the interface debond propagating along with the oxidation front.The second scenario is that the oxygen diffuses into the composite along the fiber and to the fiber-matrix interface at a rate much greater than through the neat resin or interphase.Based on the aniso- tropic oxidation behavior of unidirectional composites,the axial diffusivity of the fibers would have to be much greater than the diffusivity of the resin.Unfortunately,quantitative measures of the diffusivity of carbon fibers are lacking.Although one might anticipate that the bulk diffusivity of carbon is representative of that of the PAN fibers,this would only be the case if the morphology of the fibers and bulk carbon match,which in general is not the case. For unit cell modeling of the diffusion behavior of unidirectional com- posites,the diffusivity of the resin,the interphase,and the fiber constituents must be defined.Only the diffusivity of the resin phase can be defined with any certainty due to lack of experimental data on the diffusivity of the
the fiber axis direction. The mechanical (stiffness and strength) and thermal (coefficient of thermal expansion [52] and thermal conductivity [51]) properties of these oriented PAN-based fibers are highly anisotropic in nature. Typical density values of these fibers range between 1.75 and 1.86 g cm−3 and a porosity of approximately 15% with most of the porosity appearing in needle-shaped forms with axes of the needles parallel to the fiber axis [77]. Therefore, it follows that the diffusivity of the fibers will likely be anisotropic in nature. Unfortunately, experimental measures of carbon fiber diffusivity have not been obtained. Numerous researchers have attributed the anisotropic diffusivity of unidirectional laminates to the preferential oxidation along the fiber–matrix interface or along the interphase region. However, the interphase region is a very small-volume percentage of the composite and a very small area fraction of the surface area. It is questionable that the diffusion of oxygen through and subsequent oxidation of the interphase region by itself can have such a significant effect on the composite oxidation process. More likely, the rapid oxidation rate along the fiber length indicates that, in addition to diffusion of oxygen from the specimen surface to the oxidation front, the interphase region has a supplemental oxygen path or source. Two possible mechanisms or scenarios for the supplemental oxygen source to the interphase are being investigated. Firstly, if oxidation in the interphase region leads to early fiber–matrix interface debonds, the debonds provide a pathway for oxygen to penetrate deeper into the composite. Knowing that there are residual curing stresses at the fiber–matrix interface, it is likely that such a scenario would entail the interface debond propagating along with the oxidation front. The second scenario is that the oxygen diffuses into the composite along the fiber and to the fiber–matrix interface at a rate much greater than through the neat resin or interphase. Based on the anisotropic oxidation behavior of unidirectional composites, the axial diffusivity of the fibers would have to be much greater than the diffusivity of the resin. Unfortunately, quantitative measures of the diffusivity of carbon fibers are lacking. Although one might anticipate that the bulk diffusivity of carbon is representative of that of the PAN fibers, this would only be the case if the morphology of the fibers and bulk carbon match, which in general is not the case. For unit cell modeling of the diffusion behavior of unidirectional composites, the diffusivity of the resin, the interphase, and the fiber constituents must be defined. Only the diffusivity of the resin phase can be defined with any certainty due to lack of experimental data on the diffusivity of the 374 G.A. Schoeppner, G.P. Tandon and K.V. Pochiraju
Chapter 9:Predicting Thermooxidative Degradation 375 fiber and interphase.Therefore,the role and the relative contribution of the fiber and the interphase to the effective diffusivity cannot be rigorously determined. 9.2.2 Neat Resin Behavior The attributes of the oxidation process of high-temperature polymers and polymer composites vary depending on the chemistry of the polymers being tested.The high-temperature aging behavior of several neat polymer resins and their composites has been studied [1,10,13,14,26,37,38,53. 64,65,95,113].In PMR-15,the oxidized material near the specimen's free surfaces is observed to have a darker appearance than the unoxidized interior using bright-field light microscopy Recent studies [19,28,63,78, 90,119]document the growth of the thermooxidative layer and the changes in elastic moduli and chemical composition resulting from isothermal aging.It has been observed for PMR-15 that the polymer degradation occurs mainly within a thin surface layer that develops and grows during thermal aging.Ripberger et al.[78]reported an oxidation layer thickness of approximately 55 um after 50 h aging and a layer thickness of between 107 and 129 um after 342 h of aging for PMR-15.Additionally,the interior core of the specimens (specimens greater than 2 mm thick)is generally protected from oxidative degradation during thermal aging and is relatively unchanged after aging. In this work,for both the neat polymer and composite specimens that are used to monitor the propagation of oxidation,samples are dry-sectioned from aged larger specimens,as shown in Fig.9.7,potted in 828-D230 epoxy resin,and cured at room temperature for 3 days.This sectioning procedure allows monitoring of four of the exposed free surfaces of the large specimen.The large specimens are subsequently placed back in the oven to continue the aging process.The diamond blade used to section these Aged specimen center (nonoxidized) sectioned Edge (oxidized) Cut Edge (Nonoxidized Interior)Oxide Layer Fig.9.7.Oxidation measurement procedures
9.2.2 Neat Resin Behavior The attributes of the oxidation process of high-temperature polymers and polymer composites vary depending on the chemistry of the polymers being tested. The high-temperature aging behavior of several neat polymer resins and their composites has been studied [1, 10, 13, 14, 26, 37, 38, 53, 64, 65, 95, 113]. In PMR-15, the oxidized material near the specimen’s free surfaces is observed to have a darker appearance than the unoxidized interior using bright-field light microscopy Recent studies [19, 28, 63, 78, 90, 119] document the growth of the thermooxidative layer and the changes in elastic moduli and chemical composition resulting from isothermal aging. It has been observed for PMR-15 that the polymer degradation occurs mainly within a thin surface layer that develops and grows during thermal aging. Ripberger et al. [78] reported an oxidation layer thickness of approximately 55 µm after 50 h aging and a layer thickness of between 107 and 129 µm after 342 h of aging for PMR-15. Additionally, the interior core of the specimens (specimens greater than 2 mm thick) is generally protected from oxidative degradation during thermal aging and is relatively unchanged after aging. In this work, for both the neat polymer and composite specimens that are used to monitor the propagation of oxidation, samples are dry-sectioned from aged larger specimens, as shown in Fig. 9.7, potted in 828-D230 epoxy resin, and cured at room temperature for 3 days. This sectioning procedure allows monitoring of four of the exposed free surfaces of the large specimen. The large specimens are subsequently placed back in the oven to continue the aging process. The diamond blade used to section these Fig. 9.7. Oxidation measurement procedures Oxide Layer Edge (oxidized) Cut Edge (Nonoxidized Interior) Aged specimen center (nonoxidized) sectioned fiber and interphase. Therefore, the role and the relative contribution of the fiber and the interphase to the effective diffusivity cannot be rigorously determined. Chapter 9: Predicting Thermooxidative Degradation 375
376 G.A.Schoeppner,G.P.Tandon and K.V.Pochiraju samples is washed with acetone and wiped clean with paper towels prior to cutting to minimize the amount of contamination from the cutting wheel. The specimens are then wet-sanded with 600-grit sandpaper and distilled water and polished using a 0.3-um alumina polishing media.Since the oxidized layer forms on all exposed free surfaces and propagates to the interior of the sample,the oxidized region is easily seen in the cross section.These potted specimens are used for both nanoindentation testing (MTS Nanoindenter XP)and optical and dark-field microscopy (Nikon Microphot-FXL,Model F84006)measurements.The MTS Nanoindenter has an X-Y positional stage with a resolution of 0.5 um and a force resolution of 50.0 nN.The reported transitional resolution for the Nikon positioning stage is 1.0 um.A Philips XL30 ESEM with energy dispersive spectroscopy(EDS)analysis capabilities is also used to examine the damage and the chemical state of the aged composite specimens. Figure 9.8 shows a photomicrograph of a PMR-15 neat resin specimen isothermally aged in ambient air at 343C for a period of 196 h.The figure clearly shows the oxidized region(much like a picture frame)on the two adjacent exposed free surfaces of the specimen.Between the outer oxidized layer and interior unoxidized region is a transition region,which is the active "reaction"or "process"zone.These observations allow easy measuring and characterization of the oxidized layer and the transition Oxidized Laye Unoxidized Interior Transition Region 94.60m107.6pm Potting Resin Fig.9.8.Photomicrograph of PMR-15 resin after 196 h of aging at 343C showing oxide layer formation,transition region,and unoxidized interior
section. These potted specimens are used for both nanoindentation testing (MTS Nanoindenter XP) and optical and dark-field microscopy (Nikon Microphot-FXL, Model F84006) measurements. The MTS Nanoindenter has an X–Y positional stage with a resolution of 0.5 µm and a force resolution of 50.0 nN. The reported transitional resolution for the Nikon positioning stage is 1.0 µm. A Philips XL30 ESEM with energy dispersive spectroscopy (EDS) analysis capabilities is also used to examine the damage and the chemical state of the aged composite specimens. Figure 9.8 shows a photomicrograph of a PMR-15 neat resin specimen isothermally aged in ambient air at 343°C for a period of 196 h. The figure clearly shows the oxidized region (much like a picture frame) on the two adjacent exposed free surfaces of the specimen. Between the outer oxidized layer and interior unoxidized region is a transition region, which is the active “reaction” or “process” zone. These observations allow easy measuring and characterization of the oxidized layer and the transition Fig. 9.8. Photomicrograph of PMR-15 resin after 196 h of aging at 343°C showing oxide layer formation, transition region, and unoxidized interior water and polished using a 0.3-µm alumina polishing media. Since the oxidized layer forms on all exposed free surfaces and propagates to the interior of the sample, the oxidized region is easily seen in the cross samples is washed with acetone and wiped clean with paper towels prior to cutting to minimize the amount of contamination from the cutting wheel. The specimens are then wet-sanded with 600-grit sandpaper and distilled 376 G.A. Schoeppner, G.P. Tandon and K.V. Pochiraju
Chapter 9:Predicting Thermooxidative Degradation 377 regions as a function of aging temperature and time.Figure 9.9 shows the evolution of the oxidized layer and the active "reaction"zone thicknesses as a function of aging time.Within the first hour,an oxidized layer 11.0 um thick forms on the exposed specimen surfaces.The thickness of the oxi- dized layer is seen to approach a plateau value as the oxidation growth rate reduces considerably for longer aging time periods,whereas the thickness of the active "reaction"zone remains nearly constant for the aging times considered. PMR-15 samples aged in air were also evaluated using a nanoindenter by scanning across the oxidized layer,through the active "process"or "reaction"zone,and into the unoxidized central area of the specimen.The indentations were spaced so as to give a minimum of three data points within each region.These measurements were repeated at several locations around the perimeter of the oxidized sample. Figure 9.10 shows the variation of elastic moduli from three individual scans as a function of distance from the specimen's edge for a specimen aged in air for 196 h at 343C.The data are normalized to the modulus of the unaged material as determined by nanoindentation.Scatter exists in the data as a result of sampling through different regions of the specimen.This figure illustrates the spatial variability and the heterogeneous nature of the oxidation formation.In general,the modulus of the material in the oxidized region is higher toward the outer exposed surface than it is near the inner reaction zone.The modulus increase in the oxidation zone is consistent with embrittlement of the oxidized region. 140 Oxidation layer 120 343℃ 100 0 60 40 20 Transition Region 20040060080010001200 Aging Time (hrs) Fig.9.9.Evolution of oxidation layer and transition region thickness with aging time
Fig. 9.9. Evolution of oxidation layer and transition region thickness with aging time PMR-15 samples aged in air were also evaluated using a nanoindenter by scanning across the oxidized layer, through the active “process” or “reaction” zone, and into the unoxidized central area of the specimen. The indentations were spaced so as to give a minimum of three data points Figure 9.10 shows the variation of elastic moduli from three individual scans as a function of distance from the specimen’s edge for a specimen aged in air for 196 h at 343°C. The data are normalized to the modulus of the unaged material as determined by nanoindentation. Scatter exists in the data as a result of sampling through different regions of the specimen. This oxidized region is higher toward the outer exposed surface than it is near the inner reaction zone. The modulus increase in the oxidation zone is consistent with embrittlement of the oxidized region. 0 20 40 60 80 100 120 140 0 200 400 600 800 1000 1200 Oxidation Thickness (µm) Aging Time (hrs) Oxidation layer Transition Region 343o C within each region. These measurements were repeated at several locations around the perimeter of the oxidized sample. thick forms on the exposed specimen surfaces. The thickness of the oxidized layer is seen to approach a plateau value as the oxidation growth rate reduces considerably for longer aging time periods, whereas the thickness of the active “reaction” zone remains nearly constant for the aging times considered. regions as a function of aging temperature and time. Figure 9.9 shows the evolution of the oxidized layer and the active “reaction” zone thicknesses as a function of aging time. Within the first hour, an oxidized layer 11.0 µm Chapter 9: Predicting Thermooxidative Degradation figure illustrates the spatial variability and the heterogeneous nature of the oxidation formation. In general, the modulus of the material in the 377
378 G.A.Schoeppner,G.P.Tandon and K.V.Pochiraju 1.5 1.4 Scan 1 Scan 2 Scan 3 sninpoW pazileuoN 1.3 12 PMR-15 1.1 Unaged modulus: 0.9 0 50 100 150200 250 Distance from specimen edge (um) Fig.9.10.Normalized modulus of PMR-15 resin after 196 h of aging at 343C Based on optical observations,the three distinct specimen regions are identified in Fig.9.10 as the higher modulus oxidized surface layer,the lower modulus unoxidized interior,and the reaction zone or transition region in which the modulus reduces from the oxidized to unoxidized values.The average thickness measurements of oxidized zones and the transition zone, as obtained from optical measurements,are plotted as dotted vertical lines in Fig.9.10.It is observed that the average thickness of the oxidation layer and active "reactive"zone measured by optical methods is in good agree- ment with the boundaries of three regions suggested by the nanoindentation data.Similar observations were made by Johnson et al.[50]using atomic force microscopy (AFM).They summarized that the outer"plateau"region is a homogeneous oxidized layer,which is a result of a zero-order reaction. The transition reaction zone is a diffusion-controlled oxidation zone,which is a result of a first-order reaction,and the third region in the specimen interior is the unoxidized PMR-15. Figure 9.11 shows the average elastic moduli of the oxidized region and the unoxidized interior of PMR-15 specimens aged at 343C as a function of aging time [78].In addition,the average elastic moduli of a PMR-15 specimen thermally aged in a nonoxidizing environment is shown in the figure.The reported average values of all the moduli measurements made within their respective regions include data from multiple scans. There is only a marginal increase in the modulus of the oxidized layer with aging time,as seen in Fig.9.11.Thus,once the material oxidizes,little change in material modulus occurs with aging time
Fig. 9.10. Normalized modulus of PMR-15 resin after 196 h of aging at 343°C Based on optical observations, the three distinct specimen regions are identified in Fig. 9.10 as the higher modulus oxidized surface layer, the lower modulus unoxidized interior, and the reaction zone or transition region in which the modulus reduces from the oxidized to unoxidized values. The average thickness measurements of oxidized zones and the transition zone, as obtained from optical measurements, are plotted as dotted vertical lines in Fig. 9.10. It is observed that the average thickness of the oxidation layer and active “reactive” zone measured by optical methods is in good agreement with the boundaries of three regions suggested by the nanoindentation data. Similar observations were made by Johnson et al. [50] using atomic force microscopy (AFM). They summarized that the outer “plateau” region is a homogeneous oxidized layer, which is a result of a zero-order reaction. The transition reaction zone is a diffusion-controlled oxidation zone, which is a result of a first-order reaction, and the third region in the specimen interior is the unoxidized PMR-15. Figure 9.11 shows the average elastic moduli of the oxidized region and the unoxidized interior of PMR-15 specimens aged at 343°C as a function of aging time [78]. In addition, the average elastic moduli of a PMR-15 specimen thermally aged in a nonoxidizing environment is shown in the figure. The reported average values of all the moduli measurements made within their respective regions include data from multiple scans. There is only a marginal increase in the modulus of the oxidized layer with aging time, as seen in Fig. 9.11. Thus, once the material oxidizes, little change in material modulus occurs with aging time. 0.9 1 1.1 1.2 1.3 1.4 1.5 0 50 100 150 200 250 Scan 1 Scan 2 Scan 3 Normalized Modulus Distance from specimen edge ( µm) Unaged modulus PMR-15 378 G.A. Schoeppner, G.P. Tandon and K.V. Pochiraju