Document Rj7RL95LvLDmKdBZjyyRQLzVV

Factors By H: ; long been the aim eld of adhesion to predicting not only vhat, but for prel of the resultant lere is to consider s involved in such le at the outset to ion of an adhesive, pioneered in this hat there are two onds: mechanical anical joints form r dovetailing of the us surfaces, while hich can occur on s rough surfaces, om intermolecular fractions between ad. Probably both sent to some degree cBain then further id which wets both 1 can be solidified bond and is thereHe apparently conbond to form if the nd a tensile stress lowever small, ittributes three essive: it must be a t, and it must be ation. Equally init not stipulated, veight, composition, ties like coefficients reover, no specifics idherend surfaces are ce for this definition e and quantitative ag as both adherends .rge number of ma1 and low molecular y a substantial range, physical properties, idhesive was applied achieved usually by olvents were some- ion implies that wette to bond formation stion of what is not stting has no uni- 'N OF THIS PAPER IS publication or for the thor. Address all com . M. Headquarters, 1916 :a, 3, Pa. - or of Chemical Engr..-- .te of Technology, Cam- t Dewey A Almy Chemical , formerly with Massachu-- oology. ' then in parentheses refers ded to this paper. . . j^dhcsive Bon. d$ Formation and E. W. Merrill2 : dsnnition but is formed could be destroyed by cooling Teisted to the contact to 100 C. below the adhesive's melting : Jhanid and the solid temperature, due to stresses resulting wlL ^>e remembered from unequal coefficients of expansion. : ^sssaraed as the angle The successful bonds formed by mercury ir * droplet by the are particularly interesting, since this uite solid surface material forms some of the largest xrropiet rests and is known contact angles with ice, glass, m. The liquid. One styrene, etc. Bonds therefore formed considers wetting to here regardless of the magnitude of the sc- .angle is less than contact angles involved. s-etunc if it exceeds In these tests, no extensive control juroirr, considers com- of adherend surfaces was attempted, occnr if the contact though they were as fresh, clean and . xruv: wetting is in- . smooth as possible. Despite all reason asHte as the angle in- able precautions, adsorbed layers of m: materials, when gases and water vapor, and probably contact angles with traces of grease, were undoubtedly nasi and the cases present, since, no high vacuum tests ----- I\inBain fall largely were undertaken. Similarly, surface characteristics are influenced not only cssnner combinations by age but by methods of preparation, ir mgr contact angles a familiar example being the differences tot specific adhesion, in metal surfaces depending upon 33ve experiments were whether they were prepared by grinding, weight materials milling, polishing, etc. ''indeed, tests turner (4) and Wood- were deliberately undertaken on a given evaluate the possible system in which different Jdnds of sur sEHiK other than con- faces were prepared, and in which ~h nutations were also h surface contamination as by oxidation large and small dif- /jwas permitted. In all of these cases, rmr Mints, polarities, bonds formed. As further evidence, of expansion, etc. the every extensive search made among - smal. amount of ad- others by Loughborough and Haas (8) wa- placed upon the to find solid surfaces to which water u anc the combination will not stick upon freezing deserves annrr to keep tempera- mention. This is of critical importance zmiioni throughout the in de-icing of airplane wings during ' ~ above the adhes- flight. No case of nonadhesion is The system was reported in spite of the great number of armn 10 C. below this materials tested. --^urenucm formation of a While the evidence is not, of course, rveiy investigated by complete, the foregoing suggests that jrr c--rroiet with a spatula an adhesive bond will form if a low u Temperature. The molecular weight material at a tempera were tested, in- * ture above its melting point is brought eombinations in which into contact with a solid surface and yta: a higher melting is then allowed to freeze in place. iiuE, benzene, cellulose nes. ice, paraffin, poly- at7f-mii- nitrate, potassium -tt. sodium chloride, nine chloride; lead ass; n-octadecanol; e. and stearic acids; " dibutyl phthal- semasnone, nitrobenzene; diphenyl, naph- rnhenanthrene; glass, - .rtiimKiakablyin all oases, r ^course, the bonds so High Polymers as Adhesives The capacity of thermoplastic high polymers to stick to solid surfaces when molten and then cooled to solidification is well known, and in this respect their behavior closely parallels that of low molecular weight materials. The tacki ness often exhibited by molten poly mers reflects this capacity to stick, as well as the reluctance of these polymers to flow resulting from their high vis cosity. Since a polymer usually -does not show a well-defined: melting tempera- STLCOPCB4092373 ' will store from 6 to 10 months, as will rvanety of end'uses. Full data on the * to have serious economic disadvantages. melamine resin treated materials. * ` . physical -properties of this type of 'A new technique of injecting the resin Laminating pressures recommended molding or laminate are available from into the dry filler while it is in .a closed for the chemical glues when they were a variety of sources. and heated mold appears to have certain first introduced were in the 25 psi. -Because of the types of molds em advantages which may provide a solu range as compared to 100 to 250 psi. ployed, tooling costs were exception tion to the problem. for the low-pressure phenolic resins. ally low, and changes in the molds A very interesting technique of proc However, it has been established that could be made rapidly and at low cost. essing the chemical glue type of resin a marked improvement in properties This had a decided advantage in the is that of continuous lamination. In is noted when the chemical glues are war effort because of the many changes this process a series of webs of paper, processed at 100 psi. and excellent which were made m design of such-items cloth, or fiberglas fabric are individually properties are obtained when molded as plane parts as combat experience resin treated from separate treating at 1500 psi. The previous statement indicated a weakness in a structure or pans and then laminated into a single is not intended to bC facetious but is as engineering developments dictated ply by passing them simultaneously based on established facts and con a change. through squeeze rolls. At the squeeze stitutes the basis of the author's refer As a result of the acceptance of parts rolls, a layer of cellophane is imposed on ence to these materials as chemical molded in this manner during the war the top of the ply and another on the glues when added in excess in the resin- period, it was predicted that a large bottom and the composite web is passed filler ratio and processed at low pressures. peacetime industry would be estab into a heating oven. Longitudinal On molding or laminating the pheno- lished based on production of products pressure is produced on the web-- lic-treated materials, the treated sheets using the same materials and techniques while in the oven--by differential speeds are placed between corrosion-resistant as employed under war conditions. on squeeze rolls at opposite ends of . steel sheets to produce a flat laminate, This prediction failed to take into con the oven. Lateral pressure is produced or the die-cut pieces are loaded into the sideration that the costs of the war by contraction of the cellophane, while mold cavity. In some cases, the die- program were underwritten by the held under tension by tenter frames, cut parts are stapled together for ease United States Government and that from the oven heat. The cellophane in handling and loading the mold the high material costs for resin and acts as supporting web, pressure pad, cavity. Because the treated material, inorganic filler were generally disre- _ and, in addition, imparts a high finish is dry and the material essentially self- garded as long as- satisfactory perform to the laminate during the curing lubricating, a minimum of make-ready ance was obtained. Temporarily, at operation. This process has several is required which tends to offset the least, the usual economic yardstick was advantages over the more conventional longer curing time noted above when discarded and end-use performance be manner of laminating individual sheets compared with the chemical glues. came the only criterion of value. in a hydraulic press, but recently dis Molds are either of the metal-to-metal "When the normal economic laws once closed figures indicate that the more type or metal with an expandable bag again became part of our business life, modem hydraulic presses, when oper or pressure membrane on the other side. the high material costs could not be,; ated efficiently, produce laminated ma- The mold is usually heated by steam, disregarded. In addition, the inexpen-w terial at twice the footage per unit time direct gas flame, or electricity. After sive molds did not stand up under longp?1 compared to a continuous laminating molding, the finishing operations in production runs, and as a result of a&` machine. When a paper-polyester resin clude sawing, grinding, drilling, punch combination of these factors, the ex- * laminate of the type .produced in this ing, painting or any other required pected volume in this field failed to manner is compared to a melamine operation. materialize, and as a result production surfaced phenolic laminate, it is found The chemical glue treated material has been confined to prototypes, models, " to be lower in surface hardness, less is processed in much the same manner and a few specialty uses. Apparently stain-resistant, and lower in abrasion except that lubrication of the mold by the engineering and tooling ability of resistance. The resins usually em special lubricants is quite often re proponents of this type of production ployed have lower heat-distortion points quired, and the customary use of an did not match their enthusiasm and and because the molecular Bize does not air-excluding membrane requires ' ad optimistic outlook which failed* to in permit good fiber penetration the di ditional time for assembly. In view clude some very important and per mensional stability of the laminates is of the fact that the sticky material tinent economic considerations. rather low when compared to the requires more time for lay-up assembly, Some effort has been devoted to at melamine-phenolic type of material. this additional make-ready time tends . tempts to -employ cellulosic fillers and This type of laminate, because it can to offset the advantage of faster cure inorganic fiber mats to offset the high be produced in any desired length, may of this type of resin. Molds for this cost of woven inorganic filler. In the_ possibly find application in wall panels resin may be made of plaster, wood, case of cellulosic fibers, the molecular and similar applications. cist plastic, metal or any combination weight of the chemical glues does not As new materials are developed and of these materials. Heating of the permit fiber penetration, with the result improvements in processing techniques treated material may be accomplished that low dimensional stability results. . are worked out for presently available by infrared lamps, heated ovens, steam, In the case of mat materials, the tech -materials, the. field of low-pressure or electricity. Finishing operations are nique of handling is to load the dry filler laminates and molding materials will similar to those noted above. - . into the mold and then to add the resin undoubtedly find its proper place in During the war period, the contact as a pour charge over the dry mat, after the plastics industry. pressure ^type of resin was employed almost ^exclusively with' fillers of the , - inorganic type. Tins ! type of filler which a film of cellophane or other ma terial is placed- over the cavity charge and dieitt stod^pressure 'applied.'f^e-; -v. ,;FV.------- ----------------------- . j. .' " fe------------ !---------------------------------------------- -- j^-.wras wetted readily by the resin, and, cause of ^hp^Knessive make-ready^time. iqif 4he properties of the filler, exceptionally 4iigh lues 'were 'produced "Jabd required and in view Of the fact that.Wi iexcessqfresin ust be added to allow ''.i'MM idort itk>fhe inal^adB^PPlgl#3 tit TT f . i^ti8factorily4in,*',j^d6';j DSW 587039.01 STLCOPCB4092374 ture,"the questionarises as to whether a minimum temperature exists below which sticking becomes impossible. To explore this point, Judge (6) brought together various combinations of plas tics and adherend surfaces at controlled temperatures. This was done by pres sing together in a tightly fitting mold a thin foil.of the adherend upon the flat face of a disk of the plastic under test,. approximately in. in diameter and 0.2 in. thick. The entire apparatus was located in a thermostatic bath whose temperature was .controlled to 0.2 C. The system was of course first brought to temperature before pressure was applied, whereupon con stant temperature was maintained throughout the test. Occurrence of sticking was determined qualitatively by pulling the foil from the disk. If sticking developed at a given tempera ture, the run was repeated at a new tem perature approximately 2 C. lower until a minimum temperature was discovered at which sticking occurred but below -which it did not occur. Table I presents the results using tin as an adherend. The evidence indicates that each poly mer tested shows a unique temperature above which sticking develops but below which it rapidly becomes dif ficult and impossible to form a bond. For polystyrene, this temperature lies in the neighborhood of 80 C. and is evidently little affected by pressure. Above this temperature, the capacity for sticking develops rapidly, and in deed polystyrene becomes noticeably tacky to a light touch with a spatula at 110 C. Below 80 C., however, pelystyrene quickly loses its capacity to stick even at prolonged exposure at high pressure to an adherend surface. Other plastics were found to behave in a fashion analogous to polystyrene. While Table I reports only results for tin* as an adherend, entirely similar results were obtained for aluminum, cellulose, and styrene foil as adherends, whose minimum temperatures of bond formation were found to lie within 2 C. of that of tin. . Further inspection of Table I shows that the minimum sticking temperature for these polymers appears to coincide with their second order transition temperatures. It will be remembered that the second order transition tem perature, a term coined by Ehrenfest (5), is defined as the temperature at which curves of primary thermodynamic properties like volume and enthalpy plotted against temperature show a. dis continuity in slope. The values re ported in Table I were determined on the materials tested by the use of a dilatometer designed according to Jenc- kel and Ueberreiter (7), with 20 min. allowed -for each observation point. STABLE I--MINIMUM TEMPERATURES OF ADHESION TO TIN. , Adhesive Polystyrene*........................ .. > Second Order Transition, deg. Cent; 81 Polyvinyl butyralb.................... Polymethyl methacrylate*.... Polyvinyl alcohol^.................... 45 *63 58 * Dow Styron, R-1-K27, GA type. 6 duPont, V. F. 7100. c duPont Lucite, SED 3633. duPont 52-22, Type A. Time of -Pressure . Application 7Mhr. 1 hr. 6 rnin. 10 hr. 1 hr. 2 hr. 10 min. 1 hr. lOmin. 1 hr. 5 min. Pressure Applied, . psi. . 10 000 10 000 10 000 50 50 10 000 10 000 6 000 .6 000 5 000 5 000 . ' Minimum Temperature of Boud Formation, deg. Cent. .76.5 80 82 80 . 85 40 45 50 65 55 65 The curve for the polystyrene used here is shown in Fig. 1 and is typical. These transitions have been very extensively studied, among others, by Boyer and Spencer (2), and are con sidered due to a rapid increase in rotational and vibrational freedom of the molecular chain segments at these temperatures. Thermoplastic polymers often (but not always) exhibit second order rather than first order transitions. A polymer's second order transition temperature is relatively unaffected by pressure but is lowered by solvent or plasticizer addition; this temperature rises with increasing molecular weight, but soon reaches a maximum constant value which is then independent of further molecular weight Increases, providing no cross-linking occurs. Boyer and Spencer (3) showed that thedij second order transitions are not equili-.5 brium effects, since the discontinuity in the curve of Fig. 1 disappears if 60 hr. instead of 20 min. are allowed for each observation point. The coincidence of the minimum ad hesion temperature and the transition temperature appears consistent with these facts. Adhesion undoubtedly re quires a sufficient mobility of the atoms and molecular segments in the adhesive surface so that their force fields can key into the companion force fields'of the adherend surface. Without such mobility and orientation, no perceptible bond formation can occur. Mobility increases as the temperature rises above the transition temperature but is rapidly lost below it, which explains why poly styrene shows a rapidly declining capacity to stick below 80 C. (see Table I). If mobility at a lower tem perature is required, then this can be attained by reducing the polymer's transition temperature through solvent addition. Thus a nitrocellulose lacquer can be made to stick to metal surfaces far below the transition temperature of dry nitrocellulose, since the orientation required for bonding occurs before the solvent evaporates completely. On the other hand, mobility can be destroyed by raising the transition temperature through cross-linking. Thus, thermo setting resins used as adhesives are usually applied in partly polymerized form and then "set up" in place by heating, oxidation, etc. When applied in molten or solution form, orientation can clearly occur before extensive cross linking destroys mobility. When ap plied as a dry powder or film, the polymer is usually made to melt before setting-up, and so again has an oppor tunity to orient. In all these cases, of course, a poor bond results if a film of low cohesive strength separates the jI i | j/ \/ '/V/ Znm4t C. -^ to 40 40 40 100 no ~ 740 140 i.TM44**ATU#C , DCG CENT. - ' . Fig. 1.--Specific Volume Versus Temperature for Polystyrene (11). -- ASTM BULLET1N . . i DSW 587040 STLCOPCB4092375 'Wffnwpi adherend and ^adhesive, Such as 'the '.^witli contact angles than with polarity, mold lubricants and parting agents used It is not intended to imply here that TABLE IV.--ADHESION OF VARIOUS MA TERIALS TO STAINLESS STEEL. in molding. the polarity rule is necessarily incorrect. ,The foregoing, therefore, suggests These examples are cited to show that I- that a liquid will bond to a solid sur bond strength does not appear to de Melting Poipt, deg. Cent. Test Tempera ture, deg. Ceht. Shear Strength, psL 1 face with whieh it is brought isotherm ally into contact if the combination is then cooled to below the transition temperature (either first or second order) of the liquid. A bond will of course also form if the liquid's transition temperature is increased above the system's temperature, as by cross linking or solvent evaporation. pend solely upon one factor. Variables like coefficients of expansion and other physical properties must also be con sidered, since these may cause stress concentrations which weaken an other wise strong bond, as discussed below. Effect-of Bond Thickness: Naphthalene., Water............. fi - Methyl naphthalene Silver nitrate Wood's metal a - Methyl naphthalene Lead stearate Stearic acid Paraffin n-Octadecane. so 0 35 212 62 -22 116 69 47 28 66 -20 14 ieo 60 -40 (T) 62 43 29 10 90 78 63 48 37 32 30 12 7 2 It has been known for a long time Bond Strength . While it has been suggested here that any liquid can be made to bond to any solid surface, the strength of the bonds so formed, has not yet been con sidered. A great many factors have been recognized as affecting bond strength, and a few deserve brief dis cussion. It has, for example, been proposed that strong joints cannot be made to polar adherents with nonpolar adhesives, and vice versa. The experi mental evidence for this generaliza that bond rupture strength varies in versely and widely with thickness. A typical curve is shown in Fig. 2, showing results for a polymethyl methacrylate bond between two steel cylinders if in. in diameter, stuck together end. to end, and tested in tension at a loading rate of about 6500 psi. per sec. Each of the data points is the average of three tests. The bonds were formed isothermally at the various temperatures shown, by inserting a layer of molten polymer between the cylinders placed ficients of expansion of adhesive and adherend more closely into line, and so improves the thickness-strength be havior of the bond. The dips in the curves at small bond thicknesses .are presumably due to "starvation" or to stresses resulting from compression of the particles of filler. Further evidence. indicating the in fluence of thermal stresses upon bond strength is presented in Fig. 4. Here is shown a typical curve of bond strength versus temperature of test at constant bond thickness for a thermo tion is not entirely satisfactory.. Table II presents the test results of Lough borough and Haas (8), showing the TABLE III.--ADHESION OF WOOD'S METAL (MP 62 C.) TO VARIOUS ADHER- ENDS, AT 60 C. plastic polymer, in this case, poly methyl methacrylate. These tests, as before, were on metal cylinders stuck shearing force needed to cause a button of ice to break loose from the indicated Shear Strength, psi. together end to end and tested in ten sion at a load rate of 6500 psi. per sec. adherends, as determined in a eentri-fugal testing machine. Tables HI and IV show Woodruff's findings (13) on the force required to shear off a Glass.......... .................................. Lucite........................................... Bakelite....................................... Polystyrehe................................. Stainless steel.............................. 151 The coefficient of expansion of the ad 84 46 ' hesive is roughly ten-fold that of the 8 37 metal in-this case. As the temperature i^/alls, inspection shows that bond rup- button of the indicated adhesive 0.3 in. thick and 1.0 in. in diameter from a single adherend surface, by application of force parallel to this surface by means of a yoke which fits snugly around the adhesive button. The ratp of loading was approximately 50 psi. per min. Results should be regarded as of com parative value only. The findings . are evidently not consistent with the IV ture strength increases yntil the second end to end, after which the bonds were order temperature is reached. Here cooled to room temperature for test. the rupture strength decreases abruptly, All care was exercised to keep the possibly because the material is no cylinders properly aligned while allow longer capable of rapid flow to remove ing the adhesive freedom to contract strain inequalities. The strength again. as it wished during cooling. The re rises, reaches a maximum, and then sults show that bond strengths at room falls off to zero as the internal thermal temperature appear independent of stresses become more and more impor temperature of formation within the tant. Many cases of bond failure may pojarity criterion; for example, Table ranges investigated. be attributed to the development of II shows ice, which is highly, polar, to Various explanations of this thick such stresses. Figure 5 presents the stick as well to polyethylene as to glass, ness-strength behavior have been pro same test data for unfilled and filled which are respectively nonpolar and posed, and of these, four will .be men polystyrene. The benefits to be de polar. The equal strengths for Wood's tioned here. The first attributes this rived from adjustment of coefficients metal bonds to stainless steel and bake- behavior to the development of stresses are evident. lite in Table III or for naphthalene and resulting from differences in thermal A second explanation for the thick water to stainless steel in Table IV are coefficients of expansion between ad ness-strength curves as in Fig. 2 at again exceptions to the polarity rule. It hesive and adherend. This theory tributes this behavior to differences in is worth mentioning in passing that these , was well restated recently by Turner moduli, Poisson's ratio, etc., of adhe bond strengths correlate no better (12), who proposed overcoming the sive and adherend. It appears that, difficulty by adjustment of the coef oven when thermal expansioq coef TABLE II--ADHESION OF ICE-FILM TO MISCELLANEOUS MATERIALS, -26 C. TESTS BY LOUGHBOROUGH AND HASS, BY CENTRIFUGAL TECHNIQUE. ficient of the adhesive to correspond with that of the adherend by incorporation of suitable pigments and fillers. Figure ficients are perfectly matched, unequal moduli of adhesive and adherend would still result in a curve qualitativelysimilar . '- - Adherend Shear Strength, psi. 3 shows that unfilled polystyrene bonds 1 to Fig. 2. ,-,It willije noted that the curve thicker than 0.10 in. actually break forfilledpolystyi^iieinFig. 3 still shows 185 spontaneously ; upon cooling, even ST aluminum.'........ .............. 124 --220 ------ ,_though_good bonds form initially, due land vthifl, xa&y f$fcyiue to -differences . iUus (06% dlioa)...................... = 168 evidently to the development of high a;<(M^r ^;^;:^hffl?3^,<!oeffi<5ient8 or the MioroMope slide.................. - -422 ..-.US XFolyetliyiene.............................. 160 ^stresses upon cooling. Filling -the sty- ^pnoduli.i j J .. 'nRubber......,,.,....................... 170 'ytPBDB; brmgs .jthe mgIMIpjMjt W4M ------- STLCOPCB4092376 i 4000 f no<y TEMPERATURE OF BONO , DEG CENT. Fig. 2.--Bond Strength at 25 C. Versus Thickness (11). Adhesive: polymethyl methylcrylate. Adherend: lathe-turned steel. Fig. 4.--Bond Strength Versus Temperature (II). Adhesive: polymethyl methacrylate. Adherend: emery-ground steel* These bonds were formed at 240 C.t and were all 0.1 in. thick. third possible explanation of the thick thin bonds may far exceed the cohesive (6) M. B. Judge, "Minimum Adhesion ness-strength behavior of adhesives. This attributes the higher average strength of thin adhesive sections to the fact that these necessarily contain strength of the adhesive material in bulk. The desirability of further work to establish. the reasons for thicknessstrength behavior is clear, since cor Temperature of Thermoplastic Re sins," Massachusetts Institute of Technology Thesis (1947). (7) E. Jenckel and K. Ueberreiter, "On Polystyrene of Various Chain fewer structural flaws than thicker sections. Bikerman's (1) ingeniousanaly sis of tests of paraffin wax bonds be tween metal blocks showed that this theory goes a long way toward explain rective measures can be taken only after causes are known. References ; Lengths," Z. Phys. Chem., Vol. A182, p. 361 (1938). (8) D. L. Loughborough and E. G. Haas, "Reduction of Adhesion of Ice to De-icer Surfaces," Journal of the Aeronautical Sciences, Vol. 126, ing his experimental findings. (1) J. J. Bikerman, "Strength and Thin p. 13(1946). The fourth possible explanation of this thickness-strength behavior,. pro posed by McBain and Alexander (9), ness of Adhesive Joints," Journal Soc. Chemical Ind., Vol. (SOT, p. 23 (1941). (2) R. F. Boyer and R. S. Spencer, "Ad (9) J. W. McBain and J. Alexander, "Colloid Chemistry" (Alexander), Vol. Ill, Ch. 1, Interscienee Pub. Co. (1931). } suggests that orientation occurs in ad vances in Colloid Science" (Mark), (10) J. W. McBain and D. G. Hopkins, ; ; hesive films due to the influence of near-by solid adherend surfaces. McBain's X-ray study of glue films used Vol. II, Ch. 1, Interscience Pub. Co. (1946). (3) R. F. Boyer and R. S. Spencer, "Thermal Expansion and Second. ; "On Adhesive and Adhesive Action, Journal of Physical Chemistry Vol. 29, p. 188 (1925). (11) E. W. Merrill, "Certain Cohesive as wood adhesive, however, uncovered Order Transition Effects in High; and Adhesive Characteristics of | no supporting evidence. Polymers: Part III. Time Ef- ;r> Thermoplastic High Polymers." | 'f ! It will be noticed that these four possible explanations of thickness strength behavior are not necessarily fects," Journal of Applied Physics Vol. 17, p. 398 (1946). (4) N. H. Callner, "Adhesive Behavior of Low Molecular Weight Materials," Massachusetts Institute of Tech nology Thesis (1947). (12) P. S. Turner, "Thermal Expansion Stresses in Reinforced Plastics," l mutually exclusive, that is, all four Massachusetts Institute of Tech Journal of Research, Nat. Bureau i may simultaneously affect bond nology Thesis (1947). Standards, Vol. 37, p. 239 (1946). , i strength. Only the last two theories mentioned offer a reasonable explana- (5) P. Ehrenfest, "Phase Transitions," Communications from the Physical Lab. of the Univ. of Leiden, Supple (13) C. L. Woodworth, "The Adhesion of Low Molecular Weight Compounds " Massachusetts Institute of Tech ! tion for the fact that the strength of ment 75B, p. 8 (1933). nology Thesis (1947). i\ 0 O.Ot O.O* 0.04 0.04 0.10 0.12 O.!* 0.16 0J4 BONO THICKNESS, IN. I Fig. 3.--Bond Strength at 25 C. Versus Thickness (II), ' Curve A: polystyrene without filler (cubical coefficient of expansion: -. - . 24a x 10"* reciprocal degrees centigrade) bonded to lathe-turned steel sur faces (cubical coefficient: 33 X 10"*). r>. Curve B: polystyrene filled with 60 per cent by weight of asbestine (cubic * .^coefficient 114 X 10"*) bonded to emery ground aluminum surfaces (cubic coefficient 75 X 10"*). -so 'Ao - -go o to 4o 40 to too TEMPERA TORE OF bSnO , DEC, CENT, Fig. 5.--Bond Strength Versus Temperature (11). Curve A: polystyrene without filler (cubic.coefficient 240 X 10~*) bonded to emergy-ground aluminum surfaces (cubic coefficient 75 X 10"*). Curve B: polystyrene filled with 60 per cent by weight of asbestine (cubic coefficient 114 X 10"*), bonded to emery-ground aluminum (cubic coefficient 75 X 10"*). All bonds were formed at 220 C., and were 0.1 in. thick. M BULLETIN TPo g3 - DSW 587042 STLCOPCB4092377