{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

1488-Investigation of Composite Interphase 1990

1488-Investigation of Composite Interphase 1990 -...

Info icon This preview shows pages 1–9. Sign up to view the full content.

View Full Document Right Arrow Icon
Image of page 1

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 2
Image of page 3

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 4
Image of page 5

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 6
Image of page 7

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 8
Image of page 9
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Investigation of Composite lnterphase Using Dynamic Mechanical Analysis: Artifacts and Reality J. 1.. THOMASON Kontnkltjke/Shell-Laboratorium. Amsterdam Shell Research 8. V. Badhulsweg 3 1031 CM Amsterdam, The Netherlands The fiber-matrix interface is an important factor determining the overall me- chanical properties of composites. This interface is no longer regarded as a sharp boundary. but is now considered to be an “interphase.” |.e.. a region surrounding the fiber where properties differ from those of the bulk matrix. Although the concept of the lnterphase is rapidly gaining acceptance. its in situ detection and characterization remains a largely unsolved problem. Dynamic mechanical analy- sis (DMA) is a technique known for its sensitivity in the detection of inhomogeneity in polymer morphology. Recent publications have claimed that the inter phase in unidirectional fiber-reinforced composites is detectable using DMA. We have been evaluating a Du Pont 982 DMA on its uses in characterization of composites and their individual components. We have found that using heating rates higher than 2"C/min produces an artificial peak in the DMA loss spectrum of glass—fiber— reinforced epoxy composites at temperatures above the matrix glass—transition temperature 1T9). This peak was not present in the data from the unreinforced matrix nor in data from carbon—fiber-reinforced samples. This artifact could be interpreted as evidence of an interphase. However. our investigations revealed that it is in fact due to a complex interaction of the instrument. the thermal conductivity of the sample. the heating rate. and the sample modulus above 'l‘g. Despite this artifact. the high sensitivity .of the Du Pont 982 DMA enables detection of inhomogeneities in the composite matrix that are attributable to an interphase. INTRODUCTION It is well known that the properties of fiber-rein- forced polymer composites depend on the proper- ties of the two main constituents, l.e.. the fiber and matrix. Nevertheless. it is increasingly recognized that the interface between the fiber and matrix is important in defining these final composite proper- ties (1—6]. A more fundamental understanding of the interface region and the magnitude of its influence is therefore a prerequisite for the prediction and con‘ trol of composite properties. it has been known for some time that the level of fiber-matrix adhesion at the interface directly affects the mechanical properties of glass-fiber-rt'iiif()n‘t:tl epoxy resin composites (1‘10). However. there are a number of other possible properties of the interface region that could affect composite properties. ()llt' such possibility. which is currently being high- lighted in the literature. is that fibers may be sur- rounded by an “interphase”; i.c.. the region around the fiber may have different properties from the bulk matrix (7— 10). in many cases. the interphasc is delib- erately built into the composite by coating the fibers POLYMER COMPOSITES, APRIL 1990, Vol. 11, No. 2 with a layer of material. such as rubber. before inclu- sion in the composite. However. there are many cases where an interphase may be unexpectedly formed because of the nature of the materials used. in fiber- reinforced epoxy composites. these possibilities in- clude the following: 0 A rigid polymer layer next to the fiber surface caused by restricted molecular motion due to fiber-polymer interactions (1 1. l2). 0 A region next to the surface that is different in composition from the bulk matrix because of preferential adsorption of one of the resin for- mulation components (18]. I A region next to the surface that is different in i'()iii|)i)silioii because of the incomplete dissolu- tion of the fiber surface coating [present on most t‘tllllilll‘l‘t‘lill reinforcing fibers) in the bulk matrix (2| 0 /\ layer ”Ext to the surface with different nic- t‘llillllt‘ill properties resulting from residual ther- nml stresses. adsorbed water. or the presence of voids (i4. l5). E; "Iil theoretical studies on the influence of all) inimpiiasc on composite properties have been pub 105 ‘ J. i,. 'l‘Iioinuson Iished [8. 9). and these predict a signllicant cllccl in some cases. We have been investigating the interior eial region in fiberareinforeed polymer composites. in particular. we have been examining whether an in- terphase can be. tailored to enhance the desired com— posite properties. Furthermore. we feel it is important to determine whether the unexpected formation of an interphase can lead to a deterioration of those composite properties. A primary requirement in such an investigation is the ability to detect and Characterize the lnterphase in composites. This is complicated by the fact that the properties of the individual components (l.e.. the fiber and matrix) are usually measured simultane- ously. and these properties must be held constant if differences in measured values are to be accurately correlated with differences In the interphasc. We have been evaluating a number of relatively new techniques for the in situ characterization of the lnterphase. One of these is dynamic mechanical analysis (DMA). which has recently been reported as being able to detect the prcscncc of an tntcrphase in giass-fiber- and earbon-i'iber-relnforeed epoxy com- posites (16. 17). We report here the results of our investigation of DMA as a method for detecting the interphasc in such composites. EXPERIMENTAL I)M/\ data were obtained using a l)u l’ont 982 dynamic mechanical analyzer operated by a 9900 computer. For a full understanding of this work. it is necessary to describe the instrument anti its opera~ tion. In l)M/\. viscoeiastic materials are generally subjected to a low—strain sinusoidal deformation. and their responses to the deformation are measured. The two resultant measurements obtained are the mod— ulus, which is related to the stiffness of the material. and the damping. which determines the viscous loss properties. in the Du l’ont 982 DMA. the sample is clamped between two pivoted parallel metal arms (Fig. 1.) When one arm is slightly deflected the sam— pic is ficxcd. and when the deflection force is re— moved the arms oscillate at the resonant frcqucncy of the system. determined by the elastic properties of the sample. The 982 system measures this resonant frequency and the force input necessary to maintain a constant oscillation amplitude at the resonant fre- quency. which is a measure of the energy dissipation in the sample. The sample chamber incorporates a radiant heater and coolant distribution system. which enables these properties to be measured over a temperature range from — i 50°C to 500°C at heating rates quoted between 0.1 and 5t)"tT/mln. Unless oth- erwise Stated. DMA scans were made using a heating rate of 5°C/min. The sample temperature is recorded by a thermocouple placed nearby, Data are collected by a dedicated computer. which is also used for fur- thcr data analysis and plotting. From the measured data. the sample dimensions, and the instrument calibration parameters. the values of the storage and loss moduli of the sample are calculated I l 8]. 106 7 The composite samples for this work were ina- t'hiliCtI from unidirectional fiber-reinforced El’iKO'l‘E epoxy resin rods or laminates whose preparation has been described (1 ). Table 1 gives the components of the composite samples studied. RESULTS AND DISCUSSION Typical results from DMA measurements on a sam- ple of matrix A are shown in Fig. 2. The curves presented are the flcxural storage and loss moduli plotted as a function of temperature. The results clearly show the step-like drop in storage modulus and the peak in the loss modulus that are character- istic of the glass-rubber transition of the material. We have defined 1}, as the temperature of the maxi- mum of the flexural loss peak. Figure 3 shows a similar plot of data obtained from a sample of com- posite A. The storage modulus is much higher. both below and above the matrix 7}, because of the pres- ence of the high concentration of reinforcement. Ex- amination of the loss modulus curve reveals that the major peak. attributed to the matrix I"... has shifted SURROUNIMNG nAO-ANV NEAIEI ,CLAM"S \ , \ msvnuuz N! "(hue - [nun r “_‘PIVO‘5'— l i I | ‘L umnvco l ) LfiKA — l l I urcéoumur he I only” I L. __________ .1 . Fig. l. Schematic diagram of the Du Pout 982 UMA. Table 1. Details at Composite Samples Studied. Fiber Reference Content Matrix Code Fiber Type °/ovoi. Type' A Glass, Fiberglass FGRE 20/63 65 A 8 Glass, Pilkington 3 65 A C Glass. Bayer RV 7722 65 A D Carbon, Couriaulds XAS 65 A E1 Glass. PPG 1062-TNT 35 E E2 Glass. PPG iOBZ-TNT 45 8 E3 Glass. PPG 1062-TNT 55 B F Glass. Vetrotex R099 65 C (3 Glass. Owens Corning ABGZH 65 A _________-_—_.______—_—..—————————— ' Maid: A = EPIKOTE BZBEL + EPIKURE 113 Malrix B = EPIKOYE HEEL + Molhylhu-hydlnyhthalic nnhydrldo Malrix C = EPIKOTE “53/185 POLYMER COMPOSITES, APRIL 1990. Vol. 11, No. 2 We” Artifacts and Reality [——-] E‘ (6P0) [ ————— ] E" (MPal 20 250 200 m: 11‘: 15 150 1U 100 05 50 0 0 50 «00 150 200 2 50 YEMPERM’URE ('Cl Fig.2. Typical DMA results for unreinjorced 882851,] E113 matrix (A). [—] e' (GPO) [ ————— ]E' lGPn) Q0 20 . nu c" 35 l‘\ 30 15 25 20 10 15 1O 05 (l 0 0 50 100 150 200 250 300 YEMPERAl URE ('C) Fig. 3. DMA resultsfor Composite .4. to higher temperature. and a second peak has ap~ peared at even higher temperature. There are a number of examples in the literature of the appearance of a loss peak above Tg in the DMA of composites. The three principal reasons put for- ward for this second peak are given below: i. incomplete cure of the bulk matrix. which then undergoes further reaction when raised above its ’l‘.,(l9-21): 2. drying of the sample during heating (22): 3. the presence around the fiber of an interphasc region with a higher 7‘” (l6. l7]. Point 1 can be eliminated in our case. since the epoxy matrices were fully cured [evidenced by no increase in T” with heat treatment (23—24)). Neither is point 2 the explanation for our samples. because the second peak was still present in the DMA ‘05:; curves of samples that had been thoroughly dried. To clarify whether this second loss peak was indeed characteristic of the presence of an interphaso we ir-- stigated the influence of a number of paratnc tars. [For convenience. through the rest of this paper POLYMER COMPOSITES, APRIL 1990, Vol. 11, No. 2 the main loss peak due to the Tg of the matrix will be referred to as Peak 1 , and the second loss peak above T” as Peak 2.) Effect of Fiber Type Figure 4 shows the loss modulus curves obtained from different composites made using the same ma- trix system but different reinforcing fibers it can be seen that in glass-fiber-reinforced composites (A. B. C]. the relative height [and peak temperature] of Peak 2 appears to depend on the type of glass. in the ease of carbon fiber (1)). there is no evidence of a second loss peak; however, it is noted that Peak l is shifted to even higher temperature than in the glass-fiber— reinforced composites. This lack of Peak 2 from the carbon-fiber composites would be unexpected if Peak 2 were due to a higher '1‘” interphase because of restricted molecular motion due to interaction with the fiber. since it is generally accepted that this in— teraction is greater for carbon fibers than for glass fibers. The large difference in the E” peak height between the matrix and the composites is of course due to the fact that the complex modulus (B') of the composites is much higher. while the phase angle is of the same order. Effect of Fiber Content Figure 5 presents results from composites E143 made with different glass content. it shows that the relative height of Peak 2 is proportional to the amount of glass in the composite. Since the amount of inter— phase in a composite will also increase with the amount of fiber. this could be interpreted as evidence that Peak 2 is indeed characteristic of the presence of an interphase. Effect of Fiber Orientation Angle in all results presented to this point. the unidirec- tional composites had been positioned in the DMA with the fibers spanning the space between the t‘|:ll1]]\‘<, Lu, the 0 degree position. Figure 6 shows the influence of varying the fiber direction in come fLEXURAL LOSS NEW!) 51 ', /r\unnu \ ,s v, .>.,. Wadi, “news” “fiffj 150 200 250 YLMI)[QI\YUQE ('5) Fig] »i ":ff‘lf(5’1iff'l'”)‘§[!l'lflff’\'rfgfl’l’fl'l llMA loss t'til'l‘i'. 107 J. I,. 'l'lmmason FLEXURAL L055 (5P0) 35 30 25 2O V. Gl‘SS 55 15 10 O5 00 100 150 200 250 vcmrenmunr (1) My. 5. [allow ij'lbcr (‘mtlvui on I)MA loss rurmr/m' ('1)!!! post it.’ E. ‘0 05 100 ‘50 200 250 YEMPERMURE i'C) Fig. 6. lifleci ajfiber orientation angle on DMA loss curve for composite F. posite F between 0 and 90 degrees. It can be seen that the presence of the second loss peak is depend- ent on the fiber orientation with respect to the clamp— ing direction. Peak 2 disappears as the angle ls changed from 0 degrees (fiber-dominated properties] to 90 degrees (matrix-dominated properties). This is in agreement with the results of Reed (16) and has been proposed as evidence that the second peak is related to an lntcrphase. Reed argued that the pres- ence of an interphase around the fibers will be best detected when measuring fiber-dominated properties at 0 degrees. Effect of Heating Rate Since the heating rate can Influence the results obtained from thermal analysis (l l. 25]. this param- eter was also systematically varied. Figure 7 shows the effect of heating rate on the DMA loss curve of glass-relnforced-eomposlte A. It can be seen that the heating rate has little effect on the temperature of Peak 1. However. the temperature of Peak 2 is strong}; dependent on heating rate; furthr-mmre. at an ”,3 lower heating rates Peak 2 is not evident. This is further indicated in Fig. 8. which shows the temper- ature of the two peaks as a function of heating rate. This phenomenon was investigated in glass~rein' forced—composites A. F. and G. and the same trend was found in each case. Peak 1 being virtually inde- pendent of heating rate. and Peak 2 being linearly dependent on heating rate. The loss curves of a carboii-fibcr-reinforced sam- ple were also measured up to heating rates of 15°C/ min. and as can be seen in Fig. 9, no second peak was found. However. the temperature of Peak 1 was dependent on the heating rate. as shown in Fig. l0. Figure it shows the dependence of the Ta (Peak 1) of a sample of unrcinforccd matrix A on heating rate. Here too. it was noted that up to a heating rate of i5°C/min. there was no Sign of a second loss peak above T”. and that the temperature of Peak 1 was dependent on heating rate. As previously mentioned. the effect of heating rate on T“ seen in Figs. 10 and 11 for the unrelnforced and the carbon-fiber-rein- forced matrix was expected and can be attributed to NORMALISED FLEXURAL L055 10 05 13 Hznvmo am ('C/Mml 0 150 200 250 300 TEMPERATURE ('C) Fig. 7. Effect of treating rate on DMA loss curvefor com- posite A‘ PEAK Y EMPERATURE l'C) 280 0 win 2 o Pill I 8 O 260 2‘0 «)0 220 200 180 i60 , __1~__.L___L___J 00 2 5 5 O 7 5 10 0 t2 5 ‘50 NEAY'NG RATE (’C/min) Fig. 8. Effect of heating rate on loss peak temperature for composite A POLYMER COMPOSITES, APRIL 1990, Vol. 11, No. 2 Artifacts and Reality —— E'(GPul ---— Uterus 7O 0 so IOO 1 so zoo zoo sou TEMPERATURE ('Ci Fig. 9. DMA resultsfor composite D at 15°C/mtn. PEAK YEMPERAYURE I'C) 240 . 3 U 220 Q C 200 U 150 ' ‘50 _l__i._~_1 1 u I, oo 25 so 75 100 125 150 HEATING RATE ('c/mm) Fig. 10. Eflect of heating rate on loss peak temperature for compostte D. PEAK TE MPERAT URE (‘Cl 200 Q 1m.» INICK 6 1mm mien 190 O O o (7 0 mo 8 O I D u we ' 5 ”’0 ,AL.._WL t_ .1“. i ii 00 25 so 7‘) I00 125 i510 HFAYING RA” t’fimin) Fig. i l_ 1-;[fect of hunting; I‘(tit’ on loss peak It’litlil’lulill't‘ ffi' nmlrtx A. POLYMER COMPOSITES, APRIL 1990, Vol. 11, No. 2 thermal lag in the sample; e.g.. the temperature at point A in Fig. i is lower than the actual temperature recorded by the thermocouple near the sample sur- face. point TC in Fig. 1. Because of the time required to conduct heat into the sample. this thermal lag will Increase as the heating rate Increases. However. it Is not clear why the heating rate dependence of the ’l‘,, In the ca rbon—tibcr-reinroreed composites Is so much higher than that of the unreinforced matrix. indeed. if this was solely due to thermal lag one would expect the opposite. since addition of carbon fiber raises the thermal conductivity of the composite. Moreover. in the case of the glass-flbcr-reinforeed samples. the apparent independence of Peak 1 from the heating rate coupled with a strong linear dependence of Peak 2 on the heating rate cannot be explained by a simple thermal lag effect. At this point. we can summarize our findings on this phenomenon as loliows: O Unreinforced matrix: ——singlc peak = Ty ~—’i‘,, dependent on heating rate 0 Carbon-fiber~reinforced sample: —-single peak = Ty matrix —'t‘., at higher temperature than that of unrein- forced matrix ———'i}, dependent on heating rate 0 Glass-fiber-reinforced samples: ———doubie peak. main peak (Peak 1) = T” matrix ——Peak 1 at higher temperature than that of unreinforced matrix. but not as high as that of carbon—flber~reinforced samples —’i‘,, peak virtually independent of heating rate ——height of Peak 2 dependent on type of glass fiber -—height of Peak 2 dependent on volume fraction of glass fiber ——i’eak 2 disappears as sample fiber direction relative to Clamping direction varies from 0 to 90 degrees. ——posliion of Peak 2 linearly dependent on heat- ing rate. it is clear from this list that we are dealing with a complex phenomenon for which simple explanations such as thermal lag or the presence of an interphase do not suffice. From the foregoing observations. we can give a number of reasons why it is unlikely that the second peak in the loss curve is due to an inter- phase surrounding the reinforcing fibers: o The probable increase in '1'“ of an interphase layer due to fiber-matrix interaction is already sccn in the increase in '1‘” (Peak 1] in the com- posites compared with the unrcinforvctl matrix. t-‘urt hcrmorc. it Peak 2 was (tuc to an tntcrphasc whcic I'iiiciwnati‘ix interactions lead to a higher 't‘.,, om~ would expect it to he more prevalent in the (‘le'lflilldilit‘r-l't‘ill|‘(ii'(‘(‘(l composites where their is generally a higher level of t'ihenmatrlx interaction. iiowcvcr. Peak 2 is not found in the llMA mum-s: 1mm the caromi-t‘iher-reinforced Siliilplt‘h 109 J. I“ Thomason o The dependence of Peak 1. in the unrctnforecd samples. on the instrumental heating rate is probably due to thermal lag in the sample. which is dependent on the thermal conductivity. The glass-reinforced composite has a higher thermal conductivity than the unrelnforced matrix and. therefore. should have less thermal lag; however. Peak 2 in the composite is much more dependent on the heating rate than Peak l in the untein- forced matrix. The dependence of Peak 2 on the heating rate also takes a different form. This implies that Peak 2 is not a matrix-interphase related phenomenon. 0 Although the height of the second loss peak in- creases when the volume of the reinforcement. and thus the interphase. is Increased. the height of the main loss peak for the rest of the matrix does not decrease. as might be expected from the reduced volume of matrix. Sample Temperature Distribution For the above reasons we came to suspect that the appearance of the second damping peak was due to some complex artifact. Because of the strong depend! ence of Peak 2 on the heating rate. we reasoned that the temperature distribution within the sample played a role. We. therefore. attempted to n1casurc the actual temperature within the sample during heating in the DMA. To do this. we drilled fine holes in samples parallel to the...
View Full Document

{[ snackBarMessage ]}

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern