Document qmeL01546NQq1LByVro98N01K

PLAINTIFF'S EXHIBIT :.- ' V^k.v ' s ' * v'- ^ ,'` , " "./* V *.* ~ -V >'. .:; - `*;' V : \ . :f"T Xv-.; \\ TX...'U ;3-'-W* ..i- ' -:. -X l v -Y> \y\.f\* 'X SOCIETY OF AUTOMOTIVE ENGINEERS, INC. Two Pennsylvania Pla/a, New York. N. Y. 10001 Chemical and Functional Responses to Brake Lining Cure Variations M. H. Weintraub and J. P. Bernard Scientific Research Staff, Ford Motor Co. SOCIETY OF AUTOMOTIVE ENGINEERS Mid-Year Meeting Detroit, Mich. May 20-24,1968 680416 SCF-FA-6110 DDOnTTcn RV FOPH 680416 4--' Chemical and Functional Responses to Brake Lining Cure Variations M. H. Weintraub and J. P. Bernard Scientific Research Staff, Ford Motor Co. FORD MOTOR CO. RELIES upon several test procedures to assure consistency of the friction materials used on Its vehides. Among these procedures are pyrolytic gas chromato graphy (PGC), a test to characterize the organic constituents of the friction material, and the friction assessment screen ing test (FAST), a test which characterizes the friction and wear behavior of the material. Detailed descriptions of these tests and Illustrations of their efficacy have been pub lished previously. (1. 2)* In addition to their usefulness as quality assurance tests, these procedures can serve as valuable tools for laboratory studies of more fundamental phenomena Infliction materials. The sensitivity of both tests to material compositional varia tions has been shown previously. It has been found that these tests are also sensitive to certain processing variables such as cure time, cure temperature, and cure agent concentration. This paper describes a study of friction material curing con ditions for a typical liquid, oil-modified, phenolic resin sys tem and the relationship of the PGC and FAST responses to changes of cure. 'Numbers In parentheses designate References at end of paper. SAMPLE DESCRIPTION The response of PGC and FAST to variations in cure con ditions was studied using two groups of samples. The first group of samples was used to study cure time and tempera turn variations. The second was used to study variations of the cure agent concentration at selected times and tempera tures. The first group contained a single set of samples from the same wet mix. This model brake lining system con sisted of asbestos, an oil-modified, phenolic liquid resin (a typical brake lining resin), hexamethylenetetramine (hexa) cure agent, and barium sulfate. The samples were cured at fourdlfferent temperatures, nominally 250, 300, 350, and 400 F. for cure times of 1, 2, 4. 8. and 16 hr at each cure temperature. The hexa concentration recommended by the manufacturer -- 15% of the "solids" content In die resin -- was used in dais group of samples. The second group of samples consisted of three sets of samples with variations in the cure agent concentration. These mixes had compositions identical to the previous group, except that the hexa concentration was varied from mix to mix. Hexa concentrations of 10. 15, and 20% of the ABSTRACT------------------------------------------------------------------------- A study was made of the responses of pyrolytic gas chro matography (PGC) and the friction assessment screening test (FAST) to variations In curing conditions for a liquid, oll-modlfled. phenolic resin system. Both PGC, which char acterizes the organic resin, and FAST, which characterizes the friction and wear properties, show systematic variations with changes In cure time and temperature. A linear re- lationshlp exists between the area of one PGC peak (phenol) and the wear as determined by the FAST procedure. A chemical kinetic model Is postulated which relates the concentration of phenol produced on pyrolysis to a func tion of cure time and temperature. An Index of cure is Introduced which defines the curing conditions In terms of a single parameter. The PGC phenol peak area. FAST fric tion. and FAST wear correlate well with this Index of cure. ROHE 3186 oonniTr*:r> t 2 resin "solids' *ef* fcf this Rudy. The samples were come heated to a high temperature, thereby pyrolyzing cured at four different temperatures (250, 300, 350, end the organic material in the sample. The products of pyroly 400 D far cure tisnes f 1*1/3. 2*2/3 end 4 hr et eech cure sis are swept by a stream of helium into the chromatographic tempertttrc- The effect of cure egent concentration on column. The column separates the pyrolysis products, allow cure wi> expected to be more predominant et eerly cure ing the individual components to elute from the column times; therefore, cure time* beyond 4 hr were not Investi at different times. The presence of an organic component gated. in the exit stream is sensed by a flame ionization detector, the output of which is recorded with time at a series of TEST PROCEDURE peaks. Each peak corresponds to a particular decomposition product, while the area under the peak corresponds to the PYROLYTIC GAS CHROMATOGRAPHY - The PGC test quantity of that product. involves the pyrolysis of e sample followed by the instrumen The rest conditions used are those specified by Ford for tal separation end the sensing of die products of decompo sition. When the pyrolysis conditions ere kept constant, the products of decomposition can be used to characterize the organic constituents in the original material. A schematic quality assurance of brake linings. A more detailed de scription of the equipment and test procedure is given in the literature. (1) FAST * The FAST machine provides a convenient labo diagram of die pyrolytic gas chromatograph is shown in Fig. 1. ' A small solid sample is Inserted in the loop formed by the platinum-rhodium ribbon at the tip of the pyrolysis probe. Current is applied to the probe tip, causing the tip to be- ratory method for characterizing the friction and wear be havior of friction material. Test procedures for this machine in a constant output (that is, constant friction force) mode of operation are particularly well suited for quality control of brake lining materials. The machine provides highly reproducible results which correlate well with vehicle per- formance. SWUO.T(CMT*MoarlrCtc* ro* The FAST machine with Its associated recording instru mentation is shown in Fig. 2. and a schematic diagram of the test geometry and load control system is shown in Fig. 3. As shown in the schematic diagram, the sample clamping load is applied by hydraulic pressure acting on a rolling dia phragm load cell. The friction force is transmitted directly to the friction valve spool. In the constant output mode of opera tion, any variation of the friction force from equilibrium results In a displacement of the clamping valve spool. This causes the clamping pressure to Increase or decrease, so as to provide the friction force that re-establishes the original equilibrium. In this manner the control system maintains the friction force constant within 0.1 U> through modulation of the clamping load. The magnitude of the friction force Fig. 1 - Schematic diagram of gas chromatograph is selected by means of the load consol screw. Fig. 2 - FAST machine with auxiliary {instrumentation 8002 217 oonnr'rpn tiv vnon 3 The test oed for this study is the one Ford specifies fr ** `tuMity **surance of brake linings. This procedure calls for a flat 1/2 in. square sample to be dragged for 90 minutes against a cast iron test disc at a surface ve* locity of 23 fps, while maintaining a friction load of 17.4 jb *t the sample. The cast iron test disc reaches a temper* - . ture of 560 F during the test because of the heat generated *t die friction Interface. Since the rate of power dissipation is held constant, the history of temperature versus time and the total work done is the same in all tests; therefore, wear results (total sample weight loss) can be compared directly. A complete description of the FAST machine and details of la sensitivity and reproducibility can be found in the literature. (2) RESPONSE OF PGC TO CURE CONDITION VARIATIONS of peak A increases, while the area of peak C decreases.** Numerous other peaks show either an increase or decrease in area as the cure temperature is varied; however, the magnitude of their changes is less than that for peaks A and C. Even though peak B is a major peak, the peak area is lea sensitive to changes in cure time and temperature. The area of peak C plotted against cure time with tem perature as a parameter Is shown in Fig. 5. This peak shows a systematic decrease in peak area as cure time is Increased. Fig. 6 is a plot of die area of peak A versus cure time with temperature as a parameter. The solid lines in this figure represent the expected values based on a chemical kinetic model, which will be discussed in a later section. Thepolna indicate the experimental values. The data show a system atic Increase of peak area with increasing cure time for each cure temperature, except at the initial portion of the 250 F cure, where Initiation of cure is believed to have been mar- CURE TIME AND TEMPERATURE EFFECTS - The effect of cure temperature on the chromatograms is illustrated in Fig. 4. The chromatograms shown are those of samples cured for 16 hr at the four different temperatures (250,287*, 350. 400 F). As the cure temperature is Increased, the area **The peak area percentage is obtained by normalize tion procedure. For this resin system, 21 peaks with a rea sonably large response were selected and their peak areas measured. Then the percent area of a particular peak is equal to lu area divided by the sum of the 21 peak areas. *The nominal 300 F was actually 287 F. MICTIM MISSUM I i atwiMmuM Fig. 3 Schematic diagram of test geometry and load con* trol system Fig. 4 Effect of variation of cure temperature on PGC -- cure time 16 hr 8002 2188 ponnurpri v prvon 4 ` glnal. This erratic behavior for the 250 F curve li attributed * additional curing of the tamplei In die In jection prt ^ cbtotnetograph. Becauae the area of peak A It used In the construction of the chemical kinetic model, tbs major decomposition product corresponding to peak A was trapped and identified ts phenol. Numerous techniques for chromatographic peak identification are well defined In die literature. (3) In this case, the outlet gas stream just before peak A emerged wu passed through a U-shaped capillary tube submerged In liquid nitrogen. The condensed simple (In the capillary) then was Identified as phenol by ultraviolet absorption tech niques. For this particular resin system. It is apparent that die PCC Is extremely sensitive to changes In cure temperature; however, it was found to be sensitive to cure time only In the early stages of cure. As shown in Fig. 0. the curves tend to level off beyond a 4 hr cure. CURE AGENT CONCENTRATION EFFECTS To describe qualitatively the influence of cure agent concentration on tbe'PGC response, only the extremes In hex* level will be discussed (10% and 20% of resin solids). Generally, the PGC response to the 20%hexa samples was similar to die 15% bexa samples. This Indicates that 15% hexa Is adequate to achieve sufficient cure, la addition, the PGC response to the 15% hexa samples in this group of samples was consistent with the results described In the previous section. Fig. 7 shows chromatograms of samples which received a 1-1/3 hr cure at 300 F for the high (20%) and low (10%) percentage of hexa. The top chromatogram results from the pyrolysis of hexa alone. The high hexa chromatogram shows a small ihoaldet on the leading edge of peak D (arrow), which corresponds to the major peak of pyrolyzed hexa. The shoulder Is evident In the high hexa chromatograms for all curing conditions tested, it does not sppetr In the particular chromatogram shown In Fig. 7 for low hexa. but It does sppetr In other low and normal (15%) hexa chromato grams. The random occurrence of this shoulder in the latter chromatograms Is attributed to local variations in the distri bution of hexa within the samples. The variation of the PGC results with hexa concentration it not yet understood. A change In the bexa concentration did not produce a systematic variation In the atea of any of the PGC peaks, at was observed with cure time and tem perature variations. In fact, the low and high hexa samples showed a reversal in the area of peak A as the temperature Is Increased. That Is, for cure temperatures of 250 and 300 F and cure times of 1-1/3, 2-2/3. and 4 hr, the areas of peak A for the 10% hexa samples generally were lower than the 20% hexa samples. At cure temperatures of 350 and 400 F, the area of peak A generally was higher for the 10% hexs samples; this behavior suggests a complex cure mechanism. CHEMICAL KINETIC MODEL OF RESIN CURE Fig. 5 - Variations in area of Peak C with cure conditions A chemical kinetic model Is postulated which relates the area of peak A (phenol) to the cure time and tempera- Fig. 6 - Variations in area of Peak A with cure conditions ft ' 5* Ffg. 7 - Effect of variation in hexa concentration on PGC 21H9 DOAr>Trcn ov crvr> turs. () The model U construetedby relating die rate of formation of crossllnked itnicture (x) to die concentration of noncroidInked prectinon (y). It U farther hypothesised tbat phenol Is a measure of the ctotilinked structure. The lmerreladonihipi between the precursors, crossllnked struc ture, and cure conditions can be expressed mathematically by the following equations: Concentration of precursors at any Instant; y y0 - (i) t * Core time T 11 Cure temperature 5 Combining Eqs. 1-4 gives the following differential equa Horn . -E/RT. R n dt V <y0 * ePA> (5) If n 11 and Is non tero, Eq. 6 can be integrated to give Chemical rate equation: % W* % ` V* - (" 11 .H = dt ky" Arrhenius equation; -E/RT k = k0e (2) (3) Relates peak area to concentration: x 3 PA (4) where: y 5 Concentration of noncrossllnked precursors to x x 3 Concentration of crossllnked structure -- phenol is a measure of x n 3 Order of reaction k 3 Chemical kinetic rate constant kg a Arrhenius constant E 3 Activation energy c = Constant yQ 3 Initial Concentration of y A = Relative area of phenol peak P 3 Constant for this resin system when tested using same apparatus and test procedure R 3 Universal gu constant Dividing by (cP)* : (n - 1) kQte -E/RT (cP) 1-n (7) The boundary conditions require that the peak area go from a minimum to a maximum, as cure time varies from zero to Infinity, Thus, at the lower bound (t 3 0) the peak area must be a minimum, and the difference between the two terms on the left-hand side of Eq. 7 must equal zero. There fore, at zero cure time the minimum area is equal to Afl (minimum A a Afl at t 0). At the upper bound (t E/RT finite), the peak area should be a maximum, and the Table 1 * Values of Constants from Chemical Kinetic Model aq 9.7% A 19.0% M . E 19,900 cal/mole n 2 B 6 X 10^/hr Table 2 - Calculated Area of Peak A from Model Equation (Deviation Of Test Data In Parentheses) Time, hr 1 2 4 8 16 Temperature, F 250 9.75 (-0.1) 9.80 (-2.3) 9.90 (-1.9) 10.09 (-1.2) 10.45 (-0.05) . 287 9.87 (+0.2) 10.04 (-0.1) 10.36 (-0.4) 10.93 (-1.6) 11.87 (-0.9) 350 10.73 (-1.4) 11.66 (+0.1) 12.79 (-0.6) 14.34 (-0.06) 15.89 (+0.3) 400 12.60 ( 0.1) 14.12 (-0.5) 15.70 (+1.5) 16.99 (+ 0.2) 17.87 (-0.3) , + * . R002 a 1 *+ 0 oo r\r>T trcn "ov coo 1# 6 left-hand tide of Eq. 7 should approach <aflnlty. Thlt ctn occur only If the maximum peek tree Is y^/cP (maximum 3 for this oil-modified, phenolic resin system. For example. '7t appears that a sample cured for 4 hr at 350 F has a similar A yQ/cP). Let A^ denote the peek eree for e maximum cure temple, end for convenience, let e new constent B (n - 1) k^/ 1-n cP . Substituting die above terms Into Eq. 7 gives the following: state of cure to one cured for 1 hr at 400 F. Preliminary comparisons between the values of cure index and FAST wear (total weight lots) and friction indicate a strong quali tative agreement for this resin system. This will be illus trated in a later section. - FUNCTIONAL RESPONSES TO CURE CONDITION CHANCES ..1-n a (AM * A> * (AM ' 0} - ,, -E/RT Bte ' (8) CURE TIME AND TEMPERATURE EFFECTS - The FAST The form of Eq. 8 is compatible with the observed PGC phenol peak area. A computer solution of Eq. 8. utilizing curve fitting techniques, resulted in the curves shown in Fig. 6. The PGC data can be expected to Include some variance due to sample, procedure, and Instrument effects. A normal distribution of these variances was assumed; hence the values of A., A .. n. B. and E were sought which mini* Ora 6 mixed the sum of the squares of their absolute'differences between model and test data. The best set of values ob* rained by the curve fitting is given in Table 1. The calculated values of peak area and the deviations from the experimental values are given in Table 2. The literature value for the activation energy for meth ylene bridge formation is 13.7 Kcal/g mole. (S) With cross linking, there would be multiple bond formation, thereby increasing the activation energy. Also, hexa reacts with phenolic resin to form aromatic amines which again would Influence the activation energy. Thus, a value of 19.9 Kcal/g mole for the activation energy is quite reasonable. INDEX OF CURE - An index of cure can be defined which reduces the time/temperature curing conditions to a single parameter. It is defined as: A * A0 lcsr-rrxloo` M0 (9) machine test procedure used was able to show significant changes in the friction and wear behavior due to the varying cure conditions studied. Fig. 8 shows the sample weight loss during the FAST test plotted versus cure time, with cure temperature as param eter. The effect of cure temperature on total sample wear increases nonllnearly with temperature in the same manner as the PGC response described previously. The cure time effect on wear is also similar to the PGC response: that is. the effect it gteatest for the low cure times and rapidly di minishes above the 4 hr cure time. The solid curves shown represent the best fit to the data of an equation of the same form as was used to generate the PGC peak areas from the mathematical model of the cure kinetics. Fig. 9 shows four of the friction traces from the FAST machine and Illustrates the type of friction differences found. There was generally a slight overall Increase in friction level with Increasing cure, especially at the higher test disc temperatures (> 500 F). A rational interpretation can be made of the friction variations during the beginning of the test. Variations which occur during the first three minutes of the test may be disregarded due to seating and surface conditioning effects. The variations during the remainder of the test can be attributed to the cure variations. Seyond 20 or 30 minutes, the cure effects diminish as the materials approach the same state of cure due to the heat generated at the friction interface. This equation is based on the assumption that samples with the same phenol peak area have a similar "state of cure." Calculated values of the Index of cure are given In Table Table 3 - Index of Cure Time, hr Temperature, F 250 287 350 400 1 0.5 2.0 n 31 2 '' 1 3.5 20 47.5 48 2 7 33 64.5 4 13 50 78.5 16 8 23.5 66.5 88 Fig. 8 - Variation In FAST weat with cute conditions 800? /-\rs f */"vr-\ r\ x x COEFFICIENT OF FRICTION COEFFICIENT OF FRICTION Ptg. 9 - Variations in FAST friction with changing index of cure -- complete test traces .30 250*F- 4hr. 287*F* Ihr. ^ *C.-* 40 .30 ___ l___ I_ _ i 11 1 .30 250*F* 16 hr. 350*F-1 hr. \ le,# le || .40 SO _ _ 1- 1 __ 1__ 111 .30 4 4O0*F - Ihr. . 350 F-4 hr. 400*F- 2 hr. 31 Xe 33 Ie 47.5 .40 SO _ 1 1 _ I__ _ iii .30 * 350*F-I6hr. 400 F* 8 hr. Sw^ie,&5 ^ Ie78.5 .40 .50 11 1 -i___ I___ I__ __'-L- 0 3 IS 20 0 S 10 IS 20 0 S 10 IS 20 0 S 10 IS 20 0 S 10 IS 20 TIME, MINUTES Fig. 10 - Variations in FAST friction with changing index of cure -- Initial 20 minutes *002 2192 . PRODUCED BY FORD $1 ' V ..s Increasing degree of cure. To demoornate foe relationship between foe chemical cure variations as detected by PGC and foe functional variations in foe form of FAST results, the area of peak A Is shown in Fig. 11 plotted against FAST total sample weight lost. The solid line represents the linear fit to the data which provides a correlation coefficient of 0.93. The mean error from the linear fit Is 3.23% with the maximum deviadonoflO%occuRlng (or the 4hr400 F cure. These results illustrate that chemical changes detectable by PGC ate functionally significant, at least, in the system investigated In this study. Fot this resin system, variations In foe curing conditions produced a corresponding variation In foe FAST wear. Thar is. Increasing the cure results In an Increase of both the phenol peak area and wear. This provides important support fot the use of these two methods fot both quality assurance and laboratory testing. SUMMARY AND CONCLUSIONS Fig. 11 * Correlation between area of Peak'A and FAST weight lost An oil-modified, phenolic resin-asbestos system was used to study the chemical and functional responses to variations in cure conditions. Five cute times at each of four cure The first 20 minutes of the test results are shown in Fig. 10 for all the cure conditions studied. These curvet are ar temperatures were investigated using pyrolytic gat chroma tography (PGC) and foe friction assessment screening test (FAST). ranged in order of the sample's "index of cure,' as defined in the preceding section. A systematic change in shape with increasing index of cure is readily apparent. The amount of The PGC results show a systematic variation with changes In cure time and tempeiituie. The FAST friction and wear results show a systematic change between cure conditions horn the stepllke shape of variation with cure time and temperature consistent with the 250 F cure ample friction Races to the sloping shape foe PGC response. of the 400 F cure sample friction Races Is qualitatively For this resin system, a linear relationship exists between consistent with the values of the index of cure noted on the area of foe phenol peak, as determined by PGC. and foe each curve. CURE AGENT CONCENTRATION EFFECTS - Friction wear determined by foe FAST procedure. A chemical kinetic model Is postulated which relates foe and wear variations due to varying the cure agent concen concentration of phenol produced on pyrolysis to a function tration in the material were small compared to the effects - of cure time and temperature. The model accurately fits of cure time and temperature. foe actual phenol concentration at determined by PGC. In general, those samples cured at 250. 300, or 350 F had An Index of cure it defined which reduces the curing con slightly decreased wear with excess hexa (20%) when com ditions to a single parameter consistent with the functional pared with the low hexa samples (10%). At the 400 F cure, test. the increased hexa bad a slightly detrimental effect, and Variations In cure agent coneenaatlon influence the PGC the wear rate was greater then for foe low hexa samples. and FAST responses, but In a more complex manner than foe The 15% hexa level wear results generally fell between the cure time and temperature variations studied. 10% and 20% results and were qualitatively consistent with the results of the first group of samples. The friction differences between foe low and high hexa samples were greater for foe low temperature cures where the high hexa (20%) samples had slightly lower friction than foe low hexa samples. This effect occurred for about foe first 30 minutes of die test, after which there were no ef fects clearly attributable to hexa level. These friction and wear results are consistent with known vehicle performances that Is, the cure effects upon friction rapidly diminish with use, while foe effects on wear affect the lining for a much longet time. INTERRELATIONSHIP BETWEEN FAST AND PGC REFERENCES 1. R. L. Getler and B. H. BIggefs. "New Chemical Test for foe Characterisation of Organic Brake Linings - Pyro lytic Gas Chromatography.' Paper No. 670080 presented at SAE Annual Meeting, Detroit. January 1967. 2. A. E. Anderson, S. Gratch, and H. P. Hayes, 'A New Laboratory Friction and Wear Test fot foe Characterisation of Brake Linings.' Paper No. 670079 presented at SAE An nual Meeting,,Detroit, January 1967. 3. Howard Purnell. "Gas Chromatography.' New York: John Wiley & Sons, Inc., 1962. 4. A. E. Anderson and R. L. Gealer. Private Communica tion. As previously shown, both foe FAST friction and wear 5. Robert W. Martin. "The Chemistry of Phenolic Resins." data and foe PGC peak A area tbow systematic changes with New Yorks John Wiley & Sons, Inc.. 1956. -- . w-r *. 7** w .--TT*y7rP, V'*^ 8002 21*3 mAmirtn rv pCtm 8002 3194 TIU ppot to nbjoct to nvtolooi Iuioomou o4 oplalooo ttoueti la pipon at IkaoM* on tbo wther'i tai mo bto owpawlbUUp. aa< tbo ioelocy'c tevo'ar, Cba papor boo Society ofAutomotive Engineers, Inc. boo* odltoA bp IAI for aallofm tpUag uO (oanaL OUcobIoo will bo priotaA with iba papat IIU It patoltobol U IAI Tiiouctlow. f pomwioa 10 pttolUb ibto popor la Mi at lo pan mum iho IAI Pablkoiioai DirUtoa lai tho luiboa. J2.P*|o booWo*. Ptmtod laU.I.A. pnnn(rrcrt nv Fonr>