MetajL«ife. STRESS N60 W15088 Bobolink Ave. Menomonee Falls, WI 53051 414-252-3804 Toll Free 800-366-1973 FAX 414-252-3956 Compliments of WebSite - www.badgermetal.com Improving Tool Life for Aluminum Die Casting Dies Drs. David Schwam and John F. Wallace Case Western Reserve University Cleveland, Ohio The dies employed in aluminum die casting are a very important and significant cost factor in the process. The steel diss currently account for about 10% of the cost of pro- ducing die castings, if normal die life of epproximately 100,000 shots par dio set is atiained. When problems are encountered with the diss and prematuro die lailure rosulls, the-ecet can greaily exosed this 10% value, In addi tion, the lost production time when replacing and/or repatt- ‘ing dies can add nase costs, Dis cas stjpments in North accordingly, the anmual cost of des is estimated to be at least $1 billion. Improvement in die life constitutes a worth while objective to enhance the competitive advantage of this industry. Die Failure ‘A nuabor of diforont mechanisms act reduce the Ife of dio casling dies. Those include heat checking (also termed fire cracking or thermal fatigue failure); gross cleavage fracture ccross the die segments erosion damage, o wast. ing of the die surface by molten metal ove; and choraical attack or convsion, Thermal fatigue cracking isthe most re: squeal cause of die failure Its caused by the fucuation of the eurlaco temperature of the die between the casting and Jubsicant application periods In the eycle. These fuctua Hons produce dimensional changes and mechanic sitganea tthe die surtace, ‘Ghose cracking, although it usually starts at the de sur: face. is not primary a die suriace problem bul relates ta the-cbilty of the dle material ‘crackin he presence of tensile stresses, The primary cause is the streseos produced by marked temperature difierences within the body af the die. Erosion of the die surface is produced primarily near the ingates into the die cavity where the die metal is worn ‘away... .or damaged, at times, by cavitation produced by the colldpse of gas bubbles entrained in the entering metal streams. Solution of the die material in the molten casting alloy algo can contribute to this wear or may produce sal dering of the casting to the die, Corrosion and chemical ‘tack result from the oxidation of the die surface by itsenvi- ronment. This produces axides that fil and expand the ther- ral fatigue cracks and decarburize and weaken steel die surlaces. The die lubricant employed should not be corto: sive to the die material: damage irom the orygen of the atmosphere is unavoidable. Die Materials Modern diemaking technology employs premium grade H 13 sieel as the primary die building material, The wide usage ofthis steel is based on ils generally satisfactory per formance under the relatively widespread types of environ- zmenis lo Which dies are subjected. Some use of other mate. slols—such as maraging steels, tungsten, molybdenum and copper alloys—oceurs, but H-13 eteel is now the primary material for aluminum die casting dies. Extensive studios conducted at Case Westem Reserve University (CWRU) under the direction of the NADCA Die Material Commitee, hhas optimized the characteristics of H-13 steel that are pre: ferred for dies, Tis stee!—known as premium grade H-13—is described, in a detailed specication published by NADCA. Compar ing it with commercial H-13 steel, it has closer chemical ‘analysis requirements and a low maximum permiselale sulfur content. This low sulfur has been determined to be ‘essential for the combination of thermal fatigue resistance ‘and resistance fo greas cracking. Although the specification tdoes not require that this premium grade H-13 stoel be pro- duced by VAR or ESR processes, the olher requirements have made these process o necessary pat of tho so! “Another [important feature of this specication is close ‘control of the nonmetallic inclusions to minimize fracture paths through the steel. This clean steel requires not only the low sulfur, but also ¢ low oxygen level in the steel. The specification also assures a sotind steel by ils ullzosonic requirement, and it tests the capability of the sical for cbiaining good toughness levels. Other necessary require. ‘ments are a relatively fine grain size, the absence of a poor ‘annealed microstructure, and particularly, the avoidance of massive carbidoa, This steel is now available from several tool and die steel producers. The NADCA committee that prepared the spec- ilication had represenictives of steel companies as well a8 die casters, assuring the availability ofthis stool. While pre- ‘lum grade H-13 sieel costs somewhat more than the com- mercial grade, it is well worth the added cost in terms of improved dio lite. Te coat ofa finished die involves finishing ccels that ara considerably higher than the price ofthe die steel. By using the premium grade, the percentage increase in die coats is small, and the improvement in die life more than compen- sates for this added cost, Dice are a relatively expensive par! of dle casting, but their life, except for prototype dies, 50 Important, that relatvaly minor costs in die materials ot in shaping the die are secondary considerations. ‘Measuring Properties ot Die Steels While the ultimate test fora die is how it performs during dio casting, the large number of variables—and fluctuation of those variables—make it different and frequently unfeasi bile to conduct a dotalled study on die life using dice, die inserts, or cores produced from different materials and DIE CASTING ENGINEER REPRINT MARYAPR 1905processed in different ways, For this reason, « laboratory type ol'est that can be conducted under controlled condi- tions is tne beet procedice for technical evaluation of die mmuteriais This testing hos been conducted using a spe- Gally developed thermal fatigue test for determining sus- coptibility to heat cracking, and a Charpy V-notch test to measure the telative resistance to gross cracking, ‘A thermal fatigue specimen has been used successfully for 28 years for this evaluation.’ This test has evaluated @ ange number of potential metals as die materials for ahi- minum die castings. The results obtained have accurately predicted the relative behavior of the materials in service. hisa2x2x7-in. rectangular parallepiped specimen with @ 154n, dia, hole in the center for internal water ccoling. “The tee! produces considerable constraint and high thermat fluctuations during immersion and removal from «380 alu- minum alloy bath, The temperature at the specimen cor- ners fluctuates from nearly 100° F to 200° F through this ‘immersion and cooling cycle, The four comers have a con- ‘stant 0.010-in sadius that Intensifies the prodominanly uni- ‘atl stress at this location. ‘The outer surface of the specimen is sprayed with a.com- mercial water base lubricant just before t enters the molten ‘luminum bath. Water flows through the central hole at a ‘constant rote, and the molten bath is maintained at 1350" F ‘The specimen is immersed for a constart period and removed from the bath for another constant period to pro- duce the thermal cycle at the corners. The eracis obtained ‘are measured periedioally in ¢ section 3 in, long on the cor- ners equidistant from each end. The more severe tne crack pattem, the lower the thermal fotigue resistance of the tested material. ‘The toughness (resistance to gross cracking) is evaluated. by tho Champy V-notch test. This test~conducted at room temperature—has proved to provide an accurate evalua- tion ofthis propery. Inc Ph.D. thesis by Richard Bertolo con- ducted ct CWRU, the results of the Charpy Venotch transi- tion curve were compared with the reeults of both the ‘compact tension fracture toughness test and the dynamic toughness test conducted with « pre-cracked Charpy test. It ‘was determined that the room temperature Charpy V-notch test results provided « good relative evaluation of the toughness of die steels. Since this is a relatively inexpen- sive, easily conducted feet itis being used for this relative evaluation, Optimization of Heat Treatment While the selection of die stool is important in obtaining Jong die life, the processing of these steels prior to their use is. even more important. The H-19 dies are typically placed Into service at about 46 Rockwall C, yet the surace ofthe iter thermal fetigue fesure is eventually softened to the low 30 Rockwell C range. The lower strength at this hard- ness increases the occurrence of thermal fatigue failure. ‘The softening vecurs because of tempering and carbide coalescence during this process. It occurs even though the surface of the dio does not reach the tempering tempera. tures used in heat treatment. The simultaneous exposure of the die surface to elevated temperatures and the presence of surface stresses produce this sottening effec. Observation of the crack paths in failed dies indicates, that the paths proceed irom one carbide particle o another, of from one hard inclusion to another Since these carbides sre frequentiy located at the grain boundaries, the thermal fatigue cracks tend to follow grain boundaries. Examination i the fatled dies reveais that the presence of large alloy carbides and inclusions lowers thermal fatigue resistance, ‘This behavior lea to the conclusion that dia steels should be processed to provide a microstructure as uniform, tem perresétant, hard, and strong as feasible. The hardness of i che is determined by the tempering operation. This value [nv ornerer ness ———— é ,, ton cownmovecr coo.eo secemens i 2 ‘paint |g Ee i e ze Rg [SOOLING RATE'™ dog. per min, aa Fig. 1. Summary of Impact results for continuously cooled ‘specimens are shown in the graph. is adjusted to some degree depending on the size of the die ‘and the severity of the application. ‘The preferred austenitizing temperature for H-13 has been determined to be 1875-1900" F to provide as much alloy solution as feasible without grein growth. Any signifi cant grain growth sharply reduces the toughness and ‘makes the dies eusceptible ta gtoss cracking. Rapid cooling from the custenitizing temperature is important. It should be 198 rapid as can be tolerated by distortion considerations, butt least 50° Plmin from 1750 to $50° F This heat treatment ‘s designed to quench to martensite and avoid the precipl- tation of damaging amounts of grain boundary carbides. ‘The details of this structure and the toughness properties care deveribed: Fig. | has been taken from this reference. ‘The value of the fracture energy of the stoel at room tem [perature indicates the relative resistance to crack propaga tionotiained at different rates of cooling from the austenitz. ing temperature. These steels were all double tempered 10 46-47 Rockwell C before testing, The toughness values coblained with cooling rates that preduced a bainitic etruc- ture were only about 3 ftlb. below the fracture energy at ae s0F/min & 160F/min Boo: 7 g 2000F/min & a Soo % B00 Number of Cycles Fig. 2. Etfect of cooling rate on thermal fatigue of H-13.rapid rates of cooling to martensite. Since rapid rates of cooling increase the distortion of the dio and can cause cracking, the use of cooling rates of 50° Fimin are consid: cred in many quarters o be a reasonaible compromise, par ticularly for the larger dies. ‘The cooling rate from the custenitizing temperature is cso important in determining the thermal fatigue resis: {ance of premium grade H-13 steel. At slower cooling rates, the alloy carbides precipitate at the grain boundaries and provide a path of weakness for thermal fatigue cracking. The effect cf cooling rate in * F/min through the 1750 to $80° F temperature range on the thermal fatigue properties is shownin Fig. 2,qs measured with the thermal fatigue spec: men discussed previously. ‘The influence of these same cooling rates on the dynamic fracture toughness transition curve is not as marked, as illustrated in Fig. 3. twos noted that the area cracked by thermal fatigue in this immersion specimen is some five times higher for 60° Fimin cooling rate vs. 2000" Fimin cooling rate, whereas the differences in toughness are less. ‘Another consideration in evaluating die steels for opti- ‘mum diel is the presence ol alloy segregation or banding in the steel. While some microsogregation has to be ex- pected during solidification of the steal prior to working, severe banding will reduce the thermal fatigue properties considerably ... and the toughness to a lesser extent. The two elements that segregate to a greater degree are molyb- enum and chromium. Both of these are present in H-13 stool. The control ofthe solidilication process abiained in the VAR or ESR processing and homogenization during pro- cessing reduces banding and minimizes its effec. ‘The tempering temperature employed for optimum prop- erties also is significant consideration. Alloy steels such as TL13 produce significant amounts of retained austenite dur- ing quenching; this austenite transforms to untempered martensite in ccoling from the lirst temper. Atleast a double fomper is needed to avoid the britleness that can result {rom the untempered martensite. Studies conducted on HI3 ‘at CWRU have shown that higher hardnesses alter temper- {ng Increase thermal fatigue resistance, because a higher veld etvength ol the stee! is oblained at operating tempera: urea, However, toughness generally decreases with increasing hardness, and gross cracking can become a serious prob Jem, The usual hardness of F-13 diee before service is 46-47 Rockwell C. Small dies can be placed into service at hard- nesses up fo 50 Rockwell C—and the largest dies at 43-45 Rockwell C levels—cepending on the tendency othe die to bbe subject to gross cracking. Higher hardnesses also pro- vide greater resistance to eresion. Cooling from tempering temperature can be in a, oll, or water, While temper ‘embritlement might be « problem, the eflect of this cooling rate is very small with the molybdenum present and low ‘Phosphorous content of H-13. ‘The carbon content at the surface of the steel determines the achievable hardness level, and has @ marked ellect on the thermal fatigue lif. Any changes in the local chemistry uring heat treatment, either through carburization or Gecarburization, can ailect the thermal fatigue behavior. Decarburization is detrimental and causes premature heat checks because the lower carbon layer has lower strength fond lower resistance to thermal fatigue and ercelon. The ‘conditions that produce decarburization also can reduce die properties by corresion. Carbutization may also be Getrimental; it exhibits lower checking resistance because the eurface layer does not have enough toughness. How- ‘ever, sight carburization can increase strength and ther- mal fatigue life. ‘Although hardness is «significant factor in de life, as dis- cussed, morely attaining the correct hardness range for an 1-19 die doss not assure good die life. The hardness level in the entire 42-80 Rockwell C range discussed can be accom- plished by several routes without necosearly obtaining the structures required for good die life. This makea it necessary to apocily processing of the die through hea treatment, act simply to require a given hardness level fa 2000F/min 20F/min ce 20 /ria Fracture Toughness (Ksi Vin) 2 38 Test Temperature (“F) Fig. 3. Effect of cooling rate on the dynamic toughness of H- 13 shown in ths graph Preheating Dies ‘Toughness of H-13 is significantly affected by its tempera: ture, as illustrated in Fig. 3. These results demonstrate that a higher toughness is oblained at temperatures of 30°F cand higher. The I-13 dies are considerably more suscepti- bile to gress cracking when not preheated to 300°F temper- ature or higher, In addition, the thermal stresses that are produced in a die surlace depend on the temperature dif ference, normally referred to os delta T, between the die surlace and the sub-surface of die. Proper preheating of dle reduces this delta T and ie important in extending die life. While various means are available for accomplishing this preheat, costing a number of preliminary die castings to heat the die is not recommended. Circulation of het oll through the die cooling lines is one of the better ways of preheating, but infrared heaters— given sulfcient tme and capacity—also can provide the needed preheat Die Surtace Treatments Considerable work also has been conducted on die coat- ings to extond dio lil. This has been succeeeful only when the coating is diffused into the die or core surface. Nitriding ‘and carbonitrding (such as Melonizing) extend die life. A diffused chromium layer also extends die hfe, but this latter process is expensive, time-consuming, and requires ter- [perature higher than 2000° F. Heat treatment has to follow this exposure, Types of coatings not diffused into the surface [have nat bean effective, The use of shot pening of the dio] surlace—known as the Metalife treatment—also extends| die life, because it replaces the residual tensile stresses that [develop on the die surface with compressive stresses. The| die is shot-poened initially and then pericdically (every| 10,000 to 20,600 ie cycles), therectter. ‘Thermal stress relieving the used die at S0° F under the| highest tempering temperature alo is used. Studies using] the thermal fatigue immersion specimen have demon: Jstrated that Metalife and thermal stress relieving have|reduced thermal fatigue cracking. The efecive use of hess hol pening or thermal stress relieving processes is to Japply these provedures during the life of the die, The die ia removed from service periodically, processed to reduce the! tension residual sess at the surface, and returned to use, ‘The application of die lubricants is neceseary to aid in| stripping, o reduce soldering, and aid in building up an xide-degenerate lubricant layer on the die surface, This is particularly necessary in the case of aluminum alloys of low tron composition, These low iron alloys may solder to the die unless the lubricant forms a barrier. The lubricant should not be corresive to the die surface. The procedure has been to use waterbased die lubricants diluted at 10-40 ‘6 | waterte-lubricant ratios. ‘The use of graphite or white graphite in the lubricant has been discouraged by environmental considerations, 60 present lubricants are primary hydrocarbons. The proper lubricant aids ia achieving longer die life, It may also be preferable to use an insulating lubricant to reduce heat ‘ans tothe dies. Application of the lubricant, particularly in large quantities, produces an abrupt drop in sie surface temperature. Ifthe quantity of lubricant applied is large, this can aocentuate thermal fatigue cracking. Effect of Electro Discharge Machining ‘At the present time, two procedures are employed widely for heat treatment of H-13 eel dies for aluminum dle cast ‘ngs. The lirst procedure is to machine the block of dio steel close fo final dimensions in the annealed condition when the stee] is relatively sek (about 26 Rockwell C hardness) ‘Then, the die is heat treated by qustenitiing in a vacwurn furace, cooled as rapidly as feasible by nitrogen gas blown into the vacuum furnace at pressures that are sev. ‘eral ines aimospheric, and then tempered at least twice 10 the hardness levels previously discussed. The distortion thet occuts during this heat treatment is corrected by final ‘machining. ‘Since these final machining operations are conducted on hardened steel, they are slow and expensive. tis desired 10 keep these operations toa minimum by predicting and con- trolling the distortion. The amount and causes ofthis distor tion when heat treating and cooling these diee ina vacuum, Samane are being studied in a NADCAVDOE-sponsored investigation currently underway ot CWRU. The project is designed to determine the extent of this distortien invari us shaped dies at different rates of gas cooling, including 50° Fimnin from 1750 to 890° F. This will permit the machining ofthe die in tho annealed state to as clove as feasible to final dimensions, thereby reducing final machining operations. ‘The second method ol processing F-13 dies involves the use of EDM. This method is o block out the general dimen- sions of the die by machining the annealed block of steel. Generous allowanoes are made to compensate for distor tion atter heat treatment, since final finishing by EDM oan bbe accomplished by EDM os readlly in the hardened as in the annealed state. The blocked-out die con be heat treated ‘oils high hardness level in conventional atmospheric con- trolled furnaces. Sufficient additional finish has been cillowed to remove the dimensional variances from distot- tion and any changes in surface conditions. ‘The rough machined die, before EDM finishing, can be cooled considerably faster from the austenitzing tempord- ture than is possible by gas injected into @ vacuum furnace. ‘The use of oil quenching also can be employed. The ad tional finish on the die configuration allows for more distor. tion to be removed than is feasible with conventional machining proceduces employed for dies heat treated in vacuum furnaces. The EDM process, however, does have some effects on metal structure that require considerction ‘and adjustment. This process has been studied at CWRU’, ‘and in resent PhD. thesis." DEPTH HARONESS WH _5 iiesonsy (Re MELTED AND SOUDIFIED 56-58 LUNTEMPERED MARTENSITE 2540 50-54 ‘TEMPERED LAVER 4oas 3449 UNAFFECTED BASE METAL 46-40 Fig. 4. Various layers in an EDM-machined surtace of H-13. ‘steel are described. ‘The EDM process produces diferent layers in from the surface, as illustrated in Fig. 4. The melted and solidified layer—algo called the white layer—is extremely hard and cracked, It has to be removed by surface grinding or pol ishing, since it leads o rapid surface cracking and deterio- ration. However, it is usually only about 0.0004-in. deep ‘and ean be readily removed. A tempered martensite layer ‘about 0.0012-in. thick undeslies the white layer. This layer ‘also is hard and brite, so it has to he tempered by thermal, treatments before die service. This leaves a significant deeper layer about 0.0028 in. thick that has been soltened, by the heat produced in the EDM process to 42 Rockwell C {from the original 47 Rockwell C. This layer cannot be al- tered beiore die service. Th order to avoid reducing die Ife by thermal fatigue] mechanical removal of the white layer and tempering of [the martensitic layer are necessary. Also, subsequent sur- face hardening treatments are desirable to overcome the| ellects of the remaining layers. ‘Summary ‘This article has presented the present state of knowledge in. obiaining good die lile for aluminum die casting dies. Rec- ommendations for selection of the die steel, heat treating procedure, die surlace treatments, and vaciuim treatment ‘or EDM processing of the die are considered. This area is, ‘one of advancing technology, however, and itis hoped that further improvements in all aspects will be attained in the future, to further exiend die lle For the present, itis suggested that die casters compare their procedures with those described, and evaluate whether their methods utilize present technology. < References 1. JC, Benedyk, D. |, Menace, sad 1. Wollaco, “Thermal Fatigue Bohavior of Die Materials of Aluminum Die Casing.” (270 SDCE Transactions, paper 111 J.F Wallace, W. Roberts, and E, Hakalinin."Infenco of Coal- Ing Rate on Microstructure and Toughnass of Promium Grade Hg Die Stool,” (989 NADCA Transactons, paper G-T8.013 3. LL. Centa and LF, Wallace, “Etoct of Electro Discharge Machining on Surface and Subsurface Microstructure of #13 Steal,” 1688 NADCA Transactions, paper C-T80.032 4. HI Kim, PhD. thesis, CWRU, “Analysis of Mitigation of the Influence of EDM on the Thermal Fatigue Properties of -12Di= Steels” 192