* 1 2. - I OF I ORNL P 2361 NO . . is C . "; sellele, ... e to ..: ir tt 17 9 M36 940 .. . 11:25 | 14 ILS . MICROCOPY RESOLUTION TEST CHART NATIONAL BUREA I OF STANDARDS -1963 11 . ORAL-P-2361 Conf. 660578-8 CISTI PRICES SEP 2 2 1986 ac. $1.00; MN_57 RELEASED FOR ANNOUNCEMENT IN NUCLEAR SCIENCE ABSTRACTS MASTER WEITECT OP CURING STRESSES IN EPOXY RESINS* R. S. Crouse and R. J. Gray Oak Ridge National Laboratory Oak Ridge, Tennessee - '. . - - ABSTRACT - - Metallographic examination of some brazed "T" joints revealed cracking wher: other mechanical tests indicated a sound joint. The size and location of the cracks pointed to a failure occurring during curing of the specimen mount. Several mounting procedures were tried to confil'm this suspicion, and i'ially a sound joint was retained using epoxy resin and peiletizea alumina. The addition of pelletized alumina to the epoxy resin prevented cracking by absorbing some of the curing stresses. Nickel plating also eliminates cracking but introduces other difficulties when etching. INTRODUCTION When a metallograpcer is faced with the necessity of mounting a very fragile or easily damaged sample, he usually resorts to some castable resin, such as an epoxy. For instance, crystal bar zirconium and thin walled sort tubine or very brittle irregular shapes may be severely damaged when mounted in a pressure mount, but apparently suffers no damage when cest in a resin mount. Some recent results in examination of "T" joint test brazes have cast Some doubt on the "stress-free" qualities of castable resins. Since castable lesins are used more and more in metallographic mounting, this report pro- poses to look at some of the effects of curing stresses and suggest some ways to offset them. AR *Research sponsored čy the U. S. Atomic Energy Commission under contract with the Union Carbide Corporation. LEGAL NOTICE : This report was prepared as an account of Goverament sponsored work. Neither the United Suates, por the Commission, nor any person ucuing og tebalf of the Commission: A. Makes day warranty or representation, expressed or implied, with respect to the accu. racy, completeness, or wefulness of the information contained in this report, or that the use of way information, apparatus, melbod, or process disclosed lo this report may not infringe printely owned righus; or B. Assumes any liabilides with respect to the use of, or for damages resulung from the un of nay laformation, appart tus, motbod, or prscess disclosed in this report. As and in the above, "person acting on behall of the Commission" Includes way em- plogue or contractor of the Commission, or employee of such contractor, to be extent that soch employee or contractor uf the Commission, or employee of such cootractor prepares, dierenlaales, or provides access to, why Information pursuant to his employmeat or contact wie ebe Commission, or his employmeal wild sucd contractor. - 17 - - - - 2- PROCEDURE AND RESULTS AR The veiding and Brazing Group of the Metals and Ceramics Division of Cak Ridge National Laboratory (ORNL) produced some "T" joint test brazes and requested a metallographic examination. Eefore suomitting the samples they were each dye penetrant inspected for cracks. Some of them were then severely deformed by hammering the leg of the "T" flat. If they survived this treatment they were suomitted for metallographic examination. A stan- darci procesiure for mounting in Araldite 502* cpoxy resin was employed, and the specimen vas ground and then polished on a Syntron vibratory polisher. Figure 1 is a photomacrograph of the polished cross section showing an as- brazed and an us-brazcd and bent sample. The irst polished specimens examined were round to have cracks through the vrace joints as shown in Fig. 2. In view of the previous deiurination tests, the technician performing the metallography concluded the cracks must have occurred during the mounting operation. He reasoned that since the joints survived the "hammer test" and since the cracks appeared to be exten- sions of the separation between mount and sample, the shrinkage in the epoxy resin must have caused the cracks. Several other mounting materials were tried with results similar to those with Araldite. Figure 3 shows the type of crack that appeared in a Bakelite mount. Other mounting materials, which included Hysol. 2105, Guick Mount. Epoxybond putty, and Maraglás 655,8 cracked the oraze to sone decrei. An interesting observation was noted with the Epoxybond putty which haruens in about two hours. As soon as the mount hardened sufficiently to handle, it was polished and the joint appeared sound. However, about 16 hours later, without further polishing, the joint was found to be cracked. This material is known to continue to cure for about 48 hours; and this seemed to add evidence to the idea that curing stresses were causing the trouble. The worst cracks were i'ound in Maraglas (see Fig. ), although normally shrinkage is reportedly less in this epoxy resin than other materials. Pos- sibly the curing temperature was too high and that caused excess shrinkage. As shown in Fig. 1. the mount separation reproduces t indicatin; that hardening took place before cracking of the joint and the separation followed or occurred at the same time as the cracking. * *Ciba Proilucts, wunmit, iicw Jersey. Zlysol Corporation, Olean, liew York. Fultoa Metallurgical Products Co., 4/10 Ellsworth Ave., Pittsburgh, l'a. "Atlas Mineral Products Division, The Electric Storage Battery Co., Mertztown, Pa. O'The Marblette Corp., 37-21 30th St., Long Island City, New York. . - 3 - A specimen mount polished on both ends was viewed in transmittea iisht betecn crossed iicol filters to observe the stress patterns show iniz. 5. An area o concentratior. at one fillet is indicated by an arrow. At the top of the picture is a large stress loop which was caused by the set scroll used to holci the roount in the weight for polishing. [.! äri elicz to analyze the ätresses in the counts, J. II. width 0. thu causes to "q" joints and observed tbem as the resin curec.. He noted con:- a perioi ur compression near the beginning of the curing time, but they ievejed 01: airove the ücro poi.lt, uy the end oi' ll hours, at which tiile the totul strail was only about 200 microinches. Since no increase in stress var notcü, ilo readings were taken after lil hours. ivo cracles were i'ounu. after about 16 hours curing time, but the joints did eventually develop craclis. The presence of the strain gauges and loads may have rearranged the tresses sui'ficiently to delay damage to the joint. Although ic has lot been determined just how and wher. the stresies act, inears oi' preparini specimens of this type without cracks have bec! ioviscri. di'i. ure o how:; & sound joint that was nickel plateu beiore mounting. The nickel plate apparently "beet's up" the section and absorbs the stresses by separiztini i'rom the sample; verifyin:there is better achcrence, et exuerteri, between the nickel plate and the epoxy than the nickel plate and the specilien. This procedure is undesirable, however, because these separation could result ir staining problems upon etching, and the presence or the nickel might impede the etching oi the sample by electrolytic actio.i. The most promising procedure to date is to use pelletized alumina in epoxy. No has been reported,' addition of pelletized alumina to ar epoxy resin mount absorbs the shrinkage and results in a closely adhering mount. Figure 7 shows a joint mounted in pelletizeà alumina-Tilled Araldite epoxy resin. Although no cracks occurred in the joint oi an unbent specinen, cracks vere foured in the bent specimens even when mounted in fiiled epoxy. The crucks were observed to start at the compression side of the vend and proceed is time across the joint. Figure 8 shows the cracks developed in a beat "T" joint. Further stuciy will be made in an effort to better understand this prob- Lem and to devi:;c ways of preventing uamage to samples. CONCLUSIONIS The supposed stress-free character oi' some castabic resin: has been demonstrated to be not so. For a inejority of metallographic specimens and specilnveil shapes the imposed curing stresses are not troublesome, however, the possipility of incurring specimen damage during the curing stage must be remembered. Observable damage to "T" joint brazes can be dinimized by nickel plating prior to mounting or by mounting in epoxy plus pelletizeà alunina. . . AC!!/UWLEDGMENTS - - - - - - sit vi:;h to crutct'uliy aci.novicuse L. G. Bry. Oil of the icecinei 2012! lielaine 101:p for ili i.v.luaisle aid in preparing the braze joints and for his interest in the worl.; and 2, k. Boya of the letallography Group for first propoji!), the cause of the cracks and for preparing spe:imens. - - ... REFERENCE 1. . . Calabru, "Maraset ispoxy Resin and Coors Pelletizen Alumina co a letal.lographic Specimen Potting Mixture", Proceedings of the 13th lievili lography Group Hecting Held April 20-22, 1905 at Cax iidco rational Laboratory, Oak Ridge, Tennessee, CRNL-TM-110l, p. 29. LIST OF FIGURES Y-71575 liis. 1. 'acro Photo or Typical "T" Joint Test Bruzes. Shows 2:3-5 razca and beni cumples. o; % Y-71572 Fi;. 2. Crack in Braze Joint Caused by Curing Stresses in the zipoxy court, 100 ? Pix. 3. Cracn in Brauc Joint Mounted in Bukelite. 100 X V-771570 Y-71571 i'iij. t. "" Joint Mou:ited in Maraglas 65. Note ha surface is repiicite!! oy mountinis naterial. 100 1-71592 'ij. ;. treisio i'attern in Araldite Epoxy liesin llout as keycalce iri Polarized Light. - 2X Fico ü. Pickel Plated "q'" Joint Mounted in Araldite poxy Resi: 100 % Y-71564) Y-71500 Fig. 7. "T" Joint Mounted in Pelletized Alumina-Filled Aralaite Epoxy Resin. 100 % Y-71573 l'ię. . BC:lt "T" Joint That Cracked Although Mounted in Pelletiucu. Alumina-Filleci ilraldite Epoxy Resin. 100) Y. NATO .. .... .. . . .. : t S . - - . -- ! Sam : '1.'. . .? ! woj. A Fig. 1. Macro Photo of Typical "T" Joint Test Brazes. Shows as-brazed and bent samples. 4 x the Epoxy Mount Fig. 2. Crack in Braze Joint Caused by Curing Stresses in 100 X esimin 2- - - 22 - - --- - -. - - -.- r . .... c-.* V. vositas hes 3 . LL -3.135 !:-::- 1 1 Fig. 3. Crack in Braze Joint Mounted in Bakelite 100 X Lin ikarin-viointi . .. . mir # s wru 1 . XOC. - 53-5!.. SECO- 1 1 it 1 . K ISSN . Fig. 4. "T" Joint Mounted in Maraglas 655. Note how surface is replicated by mounting material. 100 X h - - • . Fig. 5. Stress Pattern in Araldite Epoxy Resin Mount as Revealeå in Polarized Light 2 X AUT M Muut . .. : } . > . . . -0.035 1:+5S - XOOini • On Super . . - - $ 2 Fig. 6. Nickel Plated "T" Joint Mounted in Araldite Epoxy Resin 100 X -,Sr -, t.cve ta { *33.is Fig. 7. "T" Joint Mounted in Pelletized Alumina-Filled Araldite Epoxy Mount 100 X - - - - - - . 1ml * T - ' - - * ':- - y . ' M. s - -- 1 - I .. • X i. . - S . Fig. 8. Bent "T" Joint That Cracked Although Mounted in Pelletized Alumina-Filled Araldite Epoxy Resin 100 % ' ... W W their WWLIMU W WW.LT ' ' l' www..' W . N . . .. . END care -o - - - DATE FILMED 10/21 / 66 M in .: . T - .