Document NEa0XGEw07REyV1qpQ7QoMpzQ

<H3i?S DESIGN OF 56SH NAILS FOR ASBESTOS SIDING Memoranda written on June 18th and November 22nd, 1937, give reports of some experiments in attempting to pro duce an aluminum alloy nail equivalent to the caclmium or tin plated brass nail now in use. Mr. Jones of the Chicago Office had done some work on this problem and had felt that it would be possible for us to use a nail which had a larger outside diameter than .093'', which, we had previously been advised, was the maximum. The present series of tests were run to show the effect of various types of grooves and various root diameters on the holding properties and driving properties. These tests were conducted in accordance with the tests out lined by Johns Manville Company specifications. OBJECT OF THE INVESTIGATION: The object of this Investigation was, first, to de termine what type of groove marking on the shank would be most suitable for the use of aluminum nails to obtain a satisfactory holding power, and, second, what size wire and what root di ameter of the grooves would be required to obtain a nail with driving properties equivalent to brass nails. MATERIAL USED: The material used was 17S .095" diameter for the pre liminary shape of the groove tests, and 56SH wire stabilized * for 4 hours at 250F. The sizes of 56SH wire were .083", .087", .091", and .095" diameter. The mechanical properties of the 56SH stabilized wire were as follows: ATC 0034869 2- - Diameter Tenstile Strength .083" .087" . 091" . 095" 60,050 58,150 lbfts., 60,180 l! 58,200 II oNer It rf sqif . It tt inIt It It The original tests were made on yellow pine blocks taken from the Carpenter Shop. Due to non-uniform results of thi3 material, some seasoned yellow pine was obtained for later tests. PROCEDURE: A series of 17ST nails 1" long were made up with the follow ing types of grooves or indentations on the shank: (1) Straight knurl 32 pitch (2) Straight knurl 40 pitch (3) Diamond knurl 40 pitch (both sharp and flat) (4) A drive screw thread (5) A drive screw thread with a diamond knurl (lower half drive screw thread, upper half diamond knurl) (6) 40 pitch standard thread (both sharp and flat threads) (7) 30 pitch annular grooves similar to the brass nails (both sharp and flat) (8) Plain nail (9) Nail caustic dipped 1-1/2 minutes to roughen surface These nails were driven in the unseasoned yellow pine for l/2" and the pressure required to push them and pull them from the wood was obtained from the testing machine in the Laboratory. The 56SH nails were made with annular grooves and rolled to various root diameters on the various sizes of nails. The sides of the nails were in some case3 filed flat to approximate a barbed nail. In some cases the nails were made with approximately 1/4" near the point plain without grooves, to start them into the wood more easily. Fur ther study was made of the effect of the shape of the point upon driv ing by filing the point by hand to different angles ranging from 80 to 140 included angles. A round point, instead of the usual square point, was included for comparison. ATC 0034870 -3- The nails were made l-l/2" and 2" long, as the 2" nail would be worse for driving qualities, while the 1-1/211 long nail was specified to be used in Johns Manville test specifications. A copy of these specifications is included in the Appendix. The 56SH nails were tested by hand driving in the yellow pine, by measuring the load required to push, the nails l" into the wood and the load to pull the nails out of the wood, on the testing machine in the Laboratory, and were tested for bending strength on a special jig made up according to Johns Manville specifications. RESULTS: Figures given in Table I show the results obtained upon driving and pulling the 17ST nails of various types. The only type which had holding power equivalent to the annular groove nail was the 4u pitch thread which gave a smaller root diameter, and therefore would be more prone to bend in driving than the standard nail with annular grooves. Flattening the top of the annular groove nail by opening up the dies did not seriously affect the holding power, and it also gave a somewhat larger root diameter. Results on the 56SH nails are somewhat confusing, as in the unseasoned yellow pine a good many of the nails bent on being pushed into the wood, even on the Laboratory testing machine. Table II shows the results of tests on these nails as compared with the standard brass nails. The holding power of the aluminum nail with the full grooves seems fully equal to that of the brass nail. Tables Ila and lib show tests made at different times, so that changes in the wood may have taken place. The results are not directly comparable to those in the maid table. Table III shows the results of holding power tests in sea- ATC 0034871 -4- soned yellow pine of .091" diameter nails with the grooves formed to shallow depths to improve the driving characteristics. Results on full grooves with the sides ground off to approximate a barbed condition are included. The bending tests on the stabilized nails given in Table IV indicate it will be necessary to use at least .091" diameter wire to obtain satisfactory results. Even at this size, the results are somewhat toward the low side of the specification. The difference \ between stabilized and unstabilized nails is very considerable as 1 shown by the test, the unstabilized nails having a bending load 1 about 12% greater than the unstabilized nail. Table V shows the effect of varying the point angle on the driving characteristics of the nails. The values given are the average depth to which the nails could be driven by hand into un seasoned yellow pine. It is apparent that there is not much poss ibility of improving the driving characteristics by change in this direction. Driving tests made by hand hammering the nails into either seasoned or unseasoned yellow pine were so variable as to be worth less as a basis of comparison. Even the standard brass nails usu ally bent before driving completely. DISCUSSION OF RESULTS: Prom the results It is apparent that nothing but the annu lar type of groove is suitable. Neither diamond or straight knurls gave a holding power comparable to that obtained with the grooves. A drive screw thread tended to turn when pulled from the wood, with very little pressure being required. In addition, this nail was not 0034872 -5- stiff enough and bent during driving on the Laboratory machine. The use of annular grooves which had been filed off, producing a condition similar to that which would be obtained by barbing the wire on rolls, did not have holding power greater than 75.^ of that obtained on the standard brass nail. The shallow grooves , which were in effect little more than lines in the surface of the nail, seemed to have a holding power very close to that of the full depth of groove. Even with the penetration of the dies giving a root diameter only about .002" lower than the out side diameter of the nail, the holding power was somewhat over 75?s of the holding power of the brass nail driven into the same wood. The use of such a shallow groove would indicate that a burr on the surface was quite important in improving the holding power. The possibilities of using this type of grooving are very attractive because of the larger root diameter which would give a stiffer nail, and also because the outside diameter of the nail would not be greatly increased. The customer's dies for forming their asbestos shingles are made with a hole size of between .093" and .103". A nail larger in outs_ide_ciiameter j than .093" would necessitate their redesigning their dies to make a larger hole. This would increase the differential in cost between thev aluminum and the brass nail which they would have to have before they would consider changing to an aluminum nail. From the results of jhe bending test it is apparent that a wire diameter of .091" or .095" will be required to produce a nail to meet their specification. Both of these sizes are on the borderline of meeting the bending requirement. This bending test was put in to insure a nail with satisfactory driving characteristics. Since the individual driving the nails will have a marked effect on their ten dency to bend, it is felt that the only satisfactory way of seeing ATC 0034873 6- - whether the nails can be driven is for the customer themselves to run tests. All of the tests on 56SHnails with the various types of grooves were made on stabilized material. This wbuldhave a yield strength of around 48,000# per sq. in. as compared to around 53,00055,000# per sq. in. on unstabilized material. It is fejj^Jihah---if-- these nails are on the borderline of being acceptable to the cus tomer, that an unstabilized nail should be satisfactory. Reduction in the length of the groove from about 1" to 3/4" did not affect the holding power of the nail appreciably. The point of the nail was made smooth for about l/4-3/8tt to pro vide a pilot for starting the nail through the asbestos shingle. The holding power of these nails was very little different from that of the nails which had the grooves clear up to the point. However, the nail with grooves extending along the shank only l/4" from the point did not give holding power equal to the fully grooved nail, so that it would not be possible to leave the larger section of the shank plain, with grooves just at the point to increase the holding power. CONCLUSIONS: The results of the test in this report show that no other type of grooving will give a satisfactory holding power as the annu lar type of grooves specified by the customer. The diameter of wire for an aluminum nail to obtain equal results to that of the brass nail must be at least .091". Unsta- bilized material is superior to stabilized material so far as the driving characteristics of the nail are concerned. Superior driving characteristics can be obtained by back- ATC 0034874 -7 ing the dies off, so that the depth of groove is only about .001" to .002" compared to .007" or .008" on the brass nail. The holding power of such a nail i3 equivalent to about QQfo of the holding pow er of the standard brass nail. RECOMMENDATIONS: It is recommended that samples of the stabilized nails of .091" diameter and .095" diameter with the shallow grooves be sent to the customer for their tests. If these are on the border line of being satisfactory, some unstabilized nails which have aged at least a month since the fabrication of the wire should be prepared and sent for further trial. May 5, 1938. A. HARTWELL ATC 0034875 TABLE I HOLDING POWER OF .093" DIAMETER 17ST NAILS Description Push IN Range "Average Pull OUT Range Average Ratio ofPull to Push Average Plain 67-75# 72# 26-39# 33# .46 Dipped 63-87# 72# 43-61# 50# .69 Grooves Sharp Grooves Intermediate 110-112# HI# 86-92# 89# 64-111# 76-84# 89# 79# .80 .89 Grooves Flat 67-84# 76# 84-88# 86# 1.13 40 Threads Sharp 103-126# 111# 58-99# 77# .69 40 Threads Intermediate 78-i6# 92# 40-83# 66# .72 40 Threads Flat 89-124# 103# 48-91# 64# .61 Straight 40 Tooth Knurl 61-85# Straight 32 Tooth Knurl 77-120# 74# 93# 31-34# 34-58# 33# 47# .45 .50 32 Tooth Diamond Knurl Sharp 76-98# 32 " n " Flat 86-99# 89# 92# 51-58# 43-54# 55# 49# .62 .53 Sharp indicates dies were pressed into full depth to form metal up to a sharp point. Flat indicates dies were backed off to leave top of knurl or groove flat. Intermediate indicates dies were backed off somewhat from sharp grooves but not as much as on the flat ones. ATC 0034876 DATE ON HOLDING POWER OF 56SH S TA B ILIZE D GROOVED NAILS - UNSEASONED YELLOW PINE 8 2 of <x> bO a) G <D > < H bC rH c G a) P-, CC d o *d si i--t O rH * g 03 PU 03 i--1 03 i--1 tO lO 03 03 p OP P CD c o g to 33 -P o> 03 03 JO > O cO i--1 JO 03 P P rH 03 <D rH mCG Xi P 0 O JO t--( c Si rH t--1 -P rH 3 rH o X >> rH .G ,G P p rH C rH PP -P P G c o c C > oP P C 03 G to aJ t- 03 to JO jo 03 JO JO rH oo O i--1 03 to to to JO to rH O i--( 03 CO 02 1 1 1 1 03 C- 1 1 rH rH i--i rH rH rH o rH _rH to i--1 ^H i--1 03 to 02 1 f--1 i--i O rH rH 03 -c<PD < < 03 < < rH to rH rH < 03 < to to 1--1 rH o0>2 JO to bO G o to to lO c- to o " 03 t}< 03 CO 03 CO > rH CO CO 03 rH rH rH 03 03 oi--I 1--1 r--1 rH rH 03 to rH to 03 03 to 03 03 03 rH 03 O 03 t- bC o C CO 1 CO 1--1 P o Go o p fc 0J # p p o CO r> o to 03 o 03 00 03 o o03 m t> C- I--i IO CO 03 <* to 03 to i--! O tO 03 to 03 rH 03 <# rH to to to to C- CO t~ i--1 rH 1--1 03 1--1 rH 03 03 03 03 03 03 03 03 03 11 1 03 03 to 111 CO 03 to to 11 to C- I 11 1 111 oC- rH to CO 03 1 orH rH 03 rH to CO to to 03 rH 03 1--I CO 03 03 to C- 03 rH rH rH 1--i rH i--1 l--i 03 03 03 03 03 rH i--( rH i--! 00 rH to to to t- CO 00 CO CO oo oo o c- to 03 CO CO CO ooo to o 00 03 oo to o to to to t- to 03 03 03 03 03 03 CD oo ooooo o o 03 03 rH o o 03 03 03 rH rH o O o 03 to o O 03 rH rH c G o JO 3 O.P o to 03 rH CD 03 o 0to CO rH CO rH o o CO rH to <3 t- OQ 03 CO 03 03 03 C\ \ rf* \ r+# o rH CO rH CO ft. p CO rH CO D-, p CO CO x. 11 1 p rH 03 03 rH 03 03 03 03 03 rH 03 03 03 03 03 03 03 03 M4 rH X X X x X X X X 1 X X X X X X X X X O to CO to t> c- 0- 0- c- rH rH rH rH rH rH to to to 03 00 ao CO CD CD co 00 CO CO 03 03 03 03 03 03 03 O 33 o o o o o o o o o o o o O o O o o O Q ATC 0034877  60 0) B 03 CO SO lO > r-i H C to Cco ffl, Ca> Cc- rH 6C r-i rH rH C 1 1 Z to O' PL, OS 03 CO i--1 1--1 to to 1 CD 05 i--1 lO to > C'03 LO rH Oi +3 G o -p p .Q cE 05 <D rH SO SO rH 1 P i--1 rH 00 CO rH rH O 1--1 < < s * to 03 to rH to o 05 03 03 rH lO to lO i--i o to rH CO to CO 03 03 1 11 C- to 05 05 CO 1--1 CO rH to rH a B o >d POP Po OB X) SO 60 E B O p si o p <4 oe - oo .-+* e n tip t>> rH rH E E P rl rH O B >4 H PO *d O W B ti ao a > >e EO B O eo o PBB O tiO COP 11 1 CQ co CQ 60 d Pi -- -p <> EO o C:>O CCO5 i--i rH 1 H M to CO o W 3 03 < en sai 3 E 6C C OS rH1 0i3 tto- cCOrH rH 05 to -o 03 rH CO 03 to CO to to to rH rH C<* co O- 03 03 rH rH CM 03 rH i--1 | t 1 lO rCHO 013 rH o |H In to to W 3 CQ E<h rH rtHo 1 to COOJ tr1Ho CO > rH co to > CM 1 a> to 1*4 |H w CO 3 CO CQ 03 <1 -< 05 CO 05 to CO rH rH to to 03 1 05 > 03 03 03 rH rH rH rH rH rH i--1 p o bo as -p e Po O o o B ,Q S3 E43 e o P P E *H b o n CQ co 0o 03 05 O 01 rH P B dd b N Pd rH E P 4* co CQ SO p P CO rH 03 03 BP 3 55 K X Oh to io lO a 05 05 05 ooo B CQ 01 rH ^rl ctf & tn oE 01 aJ bB CQ cq O m 03 03 03 u HXKd to to to 05 05 05 c cj o o o -p CO B 3 Ss < rH P W B d B <t, N rt m rH O rH B as BX3 CQ <M 03 P a B K B rH 3 d B<--1 E'-' B 05 c p co CO n > o OE BO 60 P P EE BO Bp E P >E B> E rH <4 O 3 d> o B ErH OO o P > o XI B E o p > o SO P P rH B C <4 3 r-i P <4 E B OH 3 de *d o B B Js! OO B pep <4 so O 3P 3 O O <4 O xS fcOB 60 P CQ P P EP E PPPP > rH *5 r-i 1 1 d d P rH 33 d Bid E > E E O> oo BO BP Bp xs si P 60 P 60 tiO B 60 EE p P XS _ >E> E P XS E P B P B OO 3tOP CBQ 3 P QB 3 P E O rH n BE i 1 l I lO B co as e CQ Ph B o ATC 0034878 a barbed n a il. same, the approxim ate appeared th e y to o f groove diam eter b u t bottom ro o t SEASONED YELLOW PIN E BLOCKS USED to to to ) < u cc 05 to E> 03 to > CD to to to o rH o rH > 05 to 02 <D 03 03 03 03 03 03 03 to 03 rH to < to to to to to iO CD to to GO 05 o to to rH o O CO 05 o <* to to to iH to 03 03 to 03 03 to 03 tjt to 31 1 1 1 ( | 1 1 rH 1 Ph CO 05 << 05 to to to to -O 03 o CD to c- o 03 03 03 03 o rH to 05 03 1 05 03 03 03 CO 03 03 03 03 rH <D rH to to CO to to to aJ CO 03 to > LO 03 03 03 03 ?H to O to rH to 03 rH O (D 03 03 03 03 03 03 03 03 03 03 03 > <5 to to to to O O iH to 03 'i' IO to << CO 03 X! to 05 CD to to to IO 03 O 03 03 03 03 03 03 03 03 03 03 03 3 l | 1 | i I 1 111 1 toPH to CO O to o to 03 to IO IO rH to o to 05 05 03 03 03 03 03 rH 03 03 rH rH to to 05 03 03 rH aJ to to toto 05 O 03 05 to o o o o oQ 05 05 o 05 o CO o CO o 05 05 CO O CO CO -P o o PC tD to to oQ E- to c- 05 o o o o o o o4 05 O 05 05 o 05 05 05 05 05 o rH rH O rH c CO o =t= CO 00 o CD CO CO o > CO pi > > CO Hi cc CQ CO at -H U -p 03 03 03 03 03 03 03 03 03 03 CQ p. rl X X K X K X X K X X u o LO to to rH rH rH rH rH to rH e c o c o o o om 05 C 05 o 05 05 05 05 05 05 05 o 05 as P Q CO o > o o u bO o rH rH CO a) <o Si XJ wo +5 >> aJ U ft. o > CO 11 CO o > CO to the o ff in . d .0 0 3 " ra te ly e l i f . of epa d s a e le s id e s no h iflfies rteendc ith w it w s Rough grooves Smooth groove ad were is y es. There so th e in " - -L s#l R- SB L V ATC 0034879 TABLE IV BENDING FORCE TO PRODUCE .100 DEFLECTION Description Specification O.D. Root Dla. Force in Lbs. to deflect. 100" Range Average 30 minimum Brass Nail .092" .078" 34.7-36. 35.7 . 083" S .093" .078" 22-22.5 22.3 .083" R .087" S . 090" .095" .081" .083" 25-23. 5 24-24.5 23.2 24.3 .087" R .093" .085" 24.5~26.5 25.3 . 087" D .086" 27.5-28. 27.7 .091" S .100" . 08o " 30-55. 31.6 .091" R .099" . 089" 29.5-30.5 30.0 .091" p .092" 33. 3-54. 33. 6 .091" D .091" 32.5-33. 32.6 .095" S .095" R . 104" .102" .086" 32.7-33.5 .090" 34.7-41.5 33.1 39.0 .095" P .095" 37. 5-38.5 37.9 .095" P Unstabilized 41.5-45.5 43.3 . 095" DrC 33-36. 34.8 . 095" DrF 34-34. 34.0 .09l"Am.St.&Wire(Unstabilized) 32.5-36.5 34.6 .095" VS .0975" .093" 35.5-36.5 36.1 .095" Sc .0965" .095" 38-39. 38.3 . 095" L .091" VS . 091" Vs#l .091" Sc .091" L .0975" . 094" .094" .094" . 094" . 094" 37.5-38. .086" 32-33. .089" 33.2-34. .090" .092" 37.7 32.6 34.6 33.5 33.5 ATC 0034880 TABLE V EFFECT OF POINT ANGLE ON DEPTH OF DRIVING WITH HAMMER Shape Included Angle of Point Standard Square Point 58 Round Point 66 Standard Square Point it it it 66 oo o> it it it tl If It 100 110 The chisel point 66 Points blunter than 140 Average Depth-Inches 4/8 5/8 3/6 3/8 3/8 7/16 2/8 1/8 ATC 0034881 SPECIFICATION NUMBERS 65024-1 M*-W-A-P'-G 65025-1 M-V/-A-P-G 65026-1 H-.T-a-P-G 65027-1 li-Vf-A-F-G 65029-1 M-V/-A-F-G 1 1/4" x 14 GAUGE 1" x 14 ft 1 3/4" x #14 ft 1 1/2" x 14 ft 2" x 14 Fage 1 f 3 pp JOHNS-MAIN'ILLE CORPORATION PURCHASING SPECIFICATIONS FOR TIN PLATED BRONZE CASING NAILS APFROVED DECEMBER 10, 1937* I, Composition i The nails shall be manufactured from Bronze wire. II. Dimensions; (a) Lbnr;th: The lengths as measured from under side of head to end of point shall be: (b) Gage: 1", 1 1/4", 1 1/2", 1 3/4", 2" The nails shall be 14 Gage; Stubs' (-0.083"). Measurement shall be made prior to plating the nails. ir 6, ` .OO (c) Shape of Point; Diamond point - Standard. (d) Shape of Head: Flat button, 5/32" diameter, l/64* thick (Min.) Variations from the above values shall not be of greater magnitude them that allowed by the U. S. Government, Federal Specification FF-N-101 for Standard Brass Nails of similar sizes. (e) Count per pound: Nall length 1" l 1/4" 1 1/2" 1 3/4" 2" Count /lb. 550 Min 470 400 320 280 III. Finish (a) Fbrmation: Barbing of the annular ring type 30-2 threads/inch, shall extend along the shank of the nail, 1/8* from the head end of the diamond point formation, to 1/4* from the head. (b) Colori Dull Tin ATC 0034882 SPECIFICATION NUMBERS 65024-5-6-7 and 9-1 M-U-a-F-G Page 2 of 3 pp. (c) Coating? The bronze nails shall be Tin plated either prior to or following the barbing operation. No subsequent acid treatment or coating shall be used to increase the holding power. The coating thickness shall be *0003* min. IV. Workmanship; Nails shall be true to shape, '.veil, finished and reasonably free from imperfectly formed or misshapen pieces, and shall not vary from the specified tolerances. The nails shall be free from corrosion, reason ably smooth, free from either excessive deposits of Tin or uncoated spots. V. Performance Requirements; (a) Bendingi Nails less than 1 l/2* long shall not be subjected to bending test. Nails shall be supported on 3/8* diameter rollers, placed 1.125* apart (center distances). A bending load shall be applied at the center of the span through a 1/4* diameter rod or bar. During the load application, the center deflection shall be measured to the nearest one-thousandth of an inch (0.001*).- The nails shall meet the following require ments : 1. Minimum ultimate load for any one nail - 30 18. 2. Average ultimate load - 35 18s. minimum. 3* Maximum center deflection at 30 18. load - 0.1". (8) Holding Power: Nails shall be driven through a hole in a tin metal plate into a 2* x 2 1/2* x 3 1/2* seasoned yellow pine block. The hole in the metal plate shall be approximately 0.005 inches larger in diameter than the diameter of the barbing on the nail shank. Following the nailing operation, both metal plate and wood block shall be gripped in suitable holders and mounted in a tensile testing machine. Load shall be applied at the rate of 150 18. per minute in such a manner as to exert a direct tensile pull on the nails. Nails are to be tested for holding power when driven across the grain, perpendicular to the annular rings of seasoned yellow pine. The nails to be tested shall be 1 1/2* long and shall show the following characteristics: ATC 0034883 I SPECIFICATION NUMBERS 65024-5-6-7 and 9-1 m-W-A-P-G _____________________________ Page 3 of 3 pp. 1. The head of the nail shall in no care break free from the shank. 2. The nail shall have at least 75% of the holding power of the standard nail.* The average of not less than five tests with the standard nail* shall be construed as the holding power of the length of wood chosen for test use. * Standard Nail - 1 1/2* Hassall Tin plated Bronze Casing Nails - /,14 gage - Ring barbing as described by the manufacturer Is "threaded with 30"V annular rings for one inch." VI Inpsaction; Four per cent (1 box out of every 25 boxes) of the nails received shall be inspected for the following characteristics: a* Type b. Class c. Material d. Size e. Finish f. Workmanship Net './eight h* Count All specifications covering this particular material will be issued under serial numbers indicated above with a designating number following. The designating numbers will be changed with each revision of the specifications, while the specification numbers will remain the same. Reference to these specifications on requisitions and in correspondence must always include the designating number, else the reference is incomplete and the letter or requisition should be referred back for completion. ATC 0034884 KEY TC NAILS PHOTOGRAPH 1. Brass Nail. 2. Smooth Grooved Nail. 3. Rough Grooved Nail. 4. Shallow-grooved with Pilot Point. 5. Shallow-grooved with Pilot Point- Large root diameter. 6. Smooth Grooves for l/4n near End. 7. Scratches on Surface of Nail. S. Lines on Surface of Nail. 9. Drive Screw Thread and Diamond Knurl. 10. Dipped 17ST Nail. 11. Sharp 40 Pitch Thread. 12. Sharp Grooves. 13. 40 Pitch Knurl. 14. 32 Pitch Knurl. 15. Flat Diamond Knurl. 16. Intermediate 40 Pitch Thread. 17. Flat 40 Pitch Thread. 18. Intermediate Grooves. 19. Sharp Diamond Knurl. 20. Flat Grooves. ATC 0034885 ATC 0034886 INTERNAL CORRESPONDENCE August 23, 1938. . From JOHN W. HOOD I METALLURGICAL DEPT. 'h--S----'-f--a---;---cEhDGEWATER WORKS AUGZ'5'38 To MR. R. L. TEHPLIN NEW KENSINGTON n) SEP 2 1938 HJ, j I RE: DESIGN OF 56SH NAILS FOR ASBESTOS SIDING faf/tfauG 26 1338 jC/f1 flUQ 10This is to thank you for your letter of August 15th suggesting yu-~ tji1e use of 14S nails for this application. You may be interested in the reasons why we have not previously tried this alloy for this purpose. When the question as to the development of these nails first came up, we discussed the matter with Mr. Bossert as to the alloy to be used, including 14S, 27S, and 56S. At that time, Mr. Bossert felt that if 56S had satisfactory upsetting characteristics, it would be the better alloy from the corrosion standpoint, and we accordingly did no work on 14S. This question of corrosion continues to be im portant as the customer desires to have the nails stay light colored and not turn dark as long as possible. We believe that probably the high copper alloy would turn considerably darker than 56S. The second factor which held us from using the 14S nails was v the conclusion reached in Research Report PT34-11, of February 14th,'v-^ 1934. 14ST nails were compared with 17ST and 27ST at that time, and the conclusion was reached that these nails were too brittle for gen eral use. The mechanical properties quoted in that report on 14S were: Yield Strength 56,300#/sq.in., Tensile Strength 69,080#/sq.In. It is probably the Yield Strength which is the deciding factor In the bending of the nail during driving. We were obtaining a Yield Strength of around 56,000#/sq. in. before stabilizing in 56SH, and on the last lot of material we obtained a Yield Strength of 61,700#/sq.in. without stabilizing. Room temperature aging would drop the first material to NK 4)0)1 MM-3-37 POINTCO IN U. S-A. ATC 0034887 2TO. ...Date. 3-2.5~.58. ___Sheet No. around 51,000#/sq.in. and the second to about 5b,000-57,000#/sq.in. The material has been sent to the customer for test. With this ma terial there would apparently be no appreciable advantage from the use of 14S, as the yield strengths would be about the same. This question is a cost proposition throughout, and the use of a heat-treatable alloy would add about 5/ a pound to the price of the nails, particularly if special equipment, such as a salt bath, were required for the 14S material. There would also be the further difficulty of the possible bending and distortion of the nails dur ing heat treating. Although neither 56SH nor 14S are standard al loys for wire products in the .091" diameter size involved, it is probable from the comparison of 17S and 5bS^H prices, that there will be a further cent or two a pound difference in the material cost. We do not have a sufficient margin to enable us to pay -these extra b/ a pound and still obtain the business, from present indica tions. You may be interested to know that while the customer had stated that the material described in our report was about 40% under in bend ing tests, they ran some preliminary tests on unstabilized material which came through quite well, much to their surprise. They have sent these samples to their Development Division, but do not expect to be able to give us a report for at least two months, due to the vacation problem in the summer. DICTATED BY A.HARTWELL:S CC: Mr. E. H. Grotefend Mr. T. W. Bossert HOOD NK tet 100M-I0-J6 PWtNTCO IN U. . A. ATC 0034888 NO' oiur*"f--t'; JU*/ ^,l- ALUMINUM COMPANY OF AMERICA Aluminum Research Laboratories Chemical Metallurgy Division New Kensington, Pa. 4C04-7 No. XF-289 October 10, 1949 PERFORMANCE OF ALUMINUM aLLUY Nr.ILS IN aSBESTQS CEMENT SIDING INTRODUCTION: Several years ago, when aluminum alloy nails were introduced to the building trades, tests were started to eval uate the merits of these nails for specific applications. One of these involved lathers nails for plaster and another pertained to nails for asbestos cement siding. In both cases, the trade was concerned over the ability of the nails to resist the alkaline materials. ^In this connection, there is an erroneous belief that all alkaline materials are particularly corrosive to aluminum alloys. This is certainly true for caustic solutions and certain other alkaline materials, such as sodium carbonate or trisodium phosphate, but there are important exceptions to this.^For ex ample, aluminum alloys are resistant to ammonium hydroxide, calcium carbonate, and may have adequate resistance to other alkaline materials under the conditions encountered in service. This is apparent from the satisfactory use of aluminum alloys in contact with alkaline materials such as plaster and concrete. It is known that alkaline solutions of many compounds, such as sodium carbonate or trisodium phosphate, are not corrosive to aluminum alloys when sodium silicate is present (Figure l). Asbestos cement siding also contains silicates and these might favorably influence the chemical properties of this material. ATC 0034889 XF-289 Page 2 OBJECT: To present laboratory tests and outdoor exposure data on the performance of aluminum alloy nails when used in asbestos cement siding; and to augment these v/ith information on the re sistance of aluminum alloys to other alkaline materials. PROCEDURE AND RESULTS: Commercial lots of asbestos cement siding represent ing six different manufacturers were secured on the open market for these tests. Chemical analysis revealed that all of these materials were alkaline in nature. They contained from 0.08$ . to 0.13$ free lime(l) and a hot water leach of them had a pH of about 11.0 (at 25C.). The tests employed 61S nails conform ing to Federal Specification Q,Q-A-325. Accelerated Laboratory Corrosion Tests A number of tests were made in the laboratory. To simulate service conditions, most of the specimens used 61S siding nails driven through the asbestos siding and into yellow pine. Some of these assemblies were continuously immersed and others were alternately immersed (1-1/2 minute cycles) in distilled water and tap water. A set of nails removed after 10 days' exposure were found to be in good condition. The nail heads and the shanks in contact with the siding were only slightly corroded. This is a relatively severe test because XTT Method by Snelhand, Bii'fens,, "Commercial Methods of Anaiyisis' Chapter 13, P.277: hot alcohol ana glycerol solution with titration using an alcoholic solution of ammonium acetate. ATC 0034890 IF-289 Page 3 any water soluble alkali in the 2n x 6" panels of siding used were allowed to concentrate in the solutions, whereas, in service there is free drainage of water. Other assemblies were exposed in a humidity chamber where continuous condensation occurs at a temperature of 125F. In this test, the asbestos siding and tne pine wood backing were thoroughly saturated with water. After 10 days in this test, the 613 siding nails were not corroded significantly. In an earlier test using roofing nails of 61S, no corrosion occurred even after 3 months' exposure. In this same test were galvanized steel siding nails. These suffered complete failure of the gal vanized coating and rusting of the nails. Cadmium plated bronze nails endured this test as well as did the 61S nails. A test panel containing a number of 61S siding nails was supported in an almost vertical position and tap water was allowed to flow continuously, but at a slow rate over the sur face of the siding and, of course, the nail heads. Several nails were removed after 10 days' exposure and were found to be in good state of preservation. The surfaces in contact with the siding were only superficially etched. Atmospheric Exposures The encouraging results obtained with the preceding accelerated corrosion tests are being confirmed by outdoor exposures in progress on 61S nails driven in the six lots of asbestos cement siding. ATC 0034891 XF-289 Page 4 The nails were driven through the siding and into yellow pine. One set was exposed to tne industrial atmosphere at New Kensington, Pa., and a duplicate set was exposed to the seacoast atmosphere just 300 feet from the ocean at Point Judith, R.I. The high rates of attack shown in Figure 8 for mild steel demonstrate the relatively severe corrosive conditions existing at both of these locations. The assemblies were exposed at an angle of 45 which allowed rainfall anc dew to flow over the asoestos cement siding and the nails. a visual inspection after about two years' exposure showed that the 61S nails were in excellent condition at both test sites. The attack was too slight to be visually discern ible. The surfaces exposed to the atmosphere showed only mild weathering and the under sur face of the heads and the shank con tacting the asbestos cement siding appeared to be free from attack. This is apparent from photographs in Figures 2, 3 and 4. Ilicroscopic examination was also made of longitudinal sections of these nails. The sections in Figure 5, and 6 are typical. These illustrate the excellent condition of the nails, especially of the heads and the shanks which had been adjacent to the asbestos cement siding. The depth of attack measure ments made at 100 diameters are as follows: ATC 0034892 XF-289 Page 5 Alloy ______ Exposure 613 Industrial Atmos. 615 Seacoast Atmos. Death of Attack Head and Shank Head Exposed Contacting to the WeatherSiding Period Avg.Max. r.vg. Max. 2 yr. 0.0009" 0.0017" 0.0003" 0.0005" 2 yr. 0.0010 0.0021 0.0004 0.0008 It is apparent from a comparison of the above data with that in Figure 10 which were obtained on freely exposed aluminum alloys, that the asbestos cement siding has certainly not had any harmful effects. Also note from Figure 10 that the rate of weathering of aluminum alloys decreases to a very low value after about two years. This results from film formation. It is unlikely, therefore, that continued exposure will cause any real changes to occur in the already low rate of weathering of the 61S nails in asbestos cement siding. In support of these findings is the satisfactory per formance of 61S siding nails which were used to attach asbestos cement siding in six to eight large housing developments in the San Francisco area. The installations were made about 1-1/2 years ago and the aluminum alloy nails are reportedly in excellent condition. Some Fundamental Considerations In a dense material, such as asbestos cement siding, the rate at which alkali can be leached by water from the siding is of more importance than the total amount of alkali present in the product. The alkali v/ould be expected to be leached more ATC 0034893 XF-289 Page 6 rapidly from the surface layers than from the interior of the siding. The depleted surface layers would provide a diffusion barrier through which the alkali leach would have to diffuse before it could reach the surface of the siding and the nail heads. Hence, the total amount of alkali present is no\, of as much importance as the rate at which the alkali can be dis solved from the siding. Two of the more important materials likely to be pre sent in leaches of asbestos cement siding are lime (Ca(GH)2), and calcium carbonate (CaCOa). Each of these compounds -was added to distilled water to form a saturated solution and pro vide an excess as a fine powder. Specimens of commercially pure aluminum (2S-H14) were exposed to these saturated solutions, vith the results given in the following table. Tor purposes of comparison, the rate of attack by sodium hydroxide (caustic) solution of about the same pH as the above solutions are included. Solution Sat. Ca(0H)2 4 g/l NaOH Sat. CaC03 0.016 g/l NaOH M_ 12.7 12.8 9.3 - 9.5 9.5 Depth of Attack inches per year 0.0105 1.6700 0.0017 0.0040 The corrosion was uniform and the rates of attack in inches penetration per year were calculated from weight loss data. The rate of attack in the saturated solutions of the materials likely to be leached from the siding were low, being 0.0105 inches per year for saturated lime and 0.0017 inches per year for saturated solutions of calcium carbonate. ATC 0034894 IF-289 Page 7 These rates are much lower than that in sodium hydroxide solu tions having the same pH values. The significance of these data are at least two fold: (l) the alkaline nature of a solution as measured by pH is not an accurate criterion of the corros iveness of a solution to aluminum alloys; and (2) the rate of attack in a saturated lime solution, although measureable,is not marked, evidently because of the formation of a protective film over the metal. Furthermore, lime solutions may have their pH reduced by absorption of carbon dioxide from the air with the formation of calcium carbonate. Consequently, part of the leach- able lime on the surface of the siding may be converted to cal cium carbonate. If this conversion occurs, the leach would be even less corrosive than a lime solution as shown by the above data. A number of tests were made on a single lot of commer cial asbestos siding to indicate the effect of repeated leaching on the alkalinity. This material had a free lime content of 0.05$, and a water leach of it had a pH of about 12.0. A 2" x 6" panel of the siding was coated with wax, except for the face intended for exposure to the weather. The panel was immersed in 150 ml of distilled water, and the pH of the solution was measured at various intervals from one minute to 22 hours. This water was not changed, but at the different periods, the panel was temporarily removed, rinsed, and immersed for one minute in fresh distilled water and the pH of this solution was measured. These data are given in Table I. ATC 0034895 XF-289 Page 8 The data reveal that a fresh surface of asbestos cement siding provides enough alkali to increase the pH to about 11 dur ing the first minute and continued leaching gradually increasing the pH to a value of about 12.2 in four hours. The intermediate immersions for one minute in fresh water, on the other hand, showed a gradual decrease in the pH to that of a neutral solution (pH 7.0), indicating a reduction in readily available alkali on the surface of the siding. This was especially true when the siding had been leached for four hours (Table I, Run l). After 22 hours' immersion, the panel was transferred to fresh dis tilled water and allowed to remain until a pH of 10.0 was ob tained. This required 40 minutes, whereas, only one minute was needed initially. This relatively slow rate of leaching was demonstrated in another test in w'hich water slowly flowed over a 12" high panel of the siding. Tap water with a pH of 7.5 was used. After two minutes, the drainage had a pH of about 10.0; after five minutes, the pH was 6.6. It then dropped to 7.8 in 90 minutes; and thereafter a pH of 7.5 was maintained, this being the same as that of the incoming tap water. The immersion test and the flowing water test do demonstrate that the alkali on the skin of the asbestos siding is likely to leach away after a few rains, etc., and that ^yc 0034896 thereafter water draining over the surface of the siding is not likely to pick up any significant amount of alkali. The density of the asbestos cement siding and the expected slow XF-289 Page 9 rate of diffusion of alkali in this hard material probably accounts for this favorable condition. In view of the preceding results,it was reasoned that since diffusion of the alkali is evidently sluggish in this sid ing, it is unlikely that the shanks of the nails would be affected by more than the amount of alkali present in a narrow annular ring about the nail. It is apparent from the various corrosion tests that this has had very little affect. To illustrate this more quantitatively, the following test was conducted. The material (1.891 g.) from a l/2n circle of the siding was powdered and mixed with 25 ml of water. Nails of 61S were then exposed to a number of these prepared solutions (pH 12.0). Two nails were removed after 4, 8, 24, 48, 72, and 'J6 hours' exposure, cleaned, and loss in weight determined. The results are given in Figure 7. The weight loss data in Figure 7 show that the cor rosion virtually stopped after about 48 hours' exposure, in spite of the fact that the solutions remained decidedly alkaline. Evidently, a protective film was formed on the 61S. Visual in spection disclosed that the nails had remained in very good condition, the corrosion being only superficial. This was con firmed by weight loss data showing that the corrosion had penetrated an insignificant 0.00046". Aluminum Alloys Versus Other Alkaline Building Materials ATC 0034897 Aluminum alloys are also resistant to the action of other alkaline building materials, such as plaster, concrete, IF-289 Page 10 magnesium oxychloride, magnesia block insulation. Some t est data are briefly given to illustrate this. Plaster Laboratory tests have revealed that neither lime-base plaster or gypsum-base plaster had -any harmful effect on 61S aluminum alloy nails. These plasters caused only superficial etching of the aluminum alloy nails and this served to increase the adhesion of the plaster to the nails. The nails had been driven in plasterboard, then covered with 3/3n thick layer of the above plasters, and a finishing coat containing 1 part plas ter cf Paris and 2 parts lime. One set was cured for 10 to 21 days anc examined; a duplicate set was subsequently exposed to 100$ relative humidity at 125F for 7 days, at which time the plaster separated from the plasterboard. The corrosion in all cases was of a superficial etching type, the maximum attack being no greater than 0.0002". It might be added that large quantities of 61S lathers nails have been used for this application during the past two years and no corrosion problems have been reported. Concrete ATC 0034898 Freshly prepared concrete causes some uniform mild surface attack of aluminum alloys but this corrosion does not seem to continue after the concrete has hardened and cured. The data in Tables II, III and IY illustrate this. It is obvious from Table II that 3S-0 was only corroded to a depth of 0.0005" during a 4-1/2 day setting period. The oil coating on alloys 4S-0 and 52S-H38 was of no benefit, but again the corrosion was XF-239 Page 11 superficial in nature. A second test in which concrete was poured into boxes nude of 52S-H32 and the assemblies exposed to humid conditions or immersed in water showed (Table III) that the corrosion was slight after 6 months. The corrosion was not appreciably greater than that expected from the initial action of the wet concrete. Another test utilized ln diameter rods of 53S-T6 imbedded in blocks of concrete. Some of these were stored indoors up to 8 years and others were exposed out doors for 8 years to the industrial atmosphere at New Kensington, Pa. The data in Table IV disclose that the bond strength im proved with tine of curing and that the exposure did not have any corrosive effect that would be reflected in a decrease in strength. Sections of the rods were examined microscopically at 100 diameters and the depth of corrosion measured. These data in Table IV illustrate that the corrosion wras shallow. Note that the metal imbedded in concrete was less corroded than those portions of the bars which had been freely exposed to the weather. Magnesium Oxychloride Cement Magnesium oxychloride cement prepared from magnesium oxide and magiesiun chloride is alkaline in nature and a water leach of it had a pH of about 9.5. laboratory tests have shown that aluminum alloys, including 61S-T6 have an inherently high resistance to corrosion by this material and one superior to galvanized steel or plain steel. Like concrete and plaster, this product causes superficial etching of aluminum during the casting and setting period, but subsequent exposure to alternately0034899 IE-289 Page 12 wetting and drying conditions does not cause any further corros ion of significance. Corrugated Asbestos Cement Roofing A sample of commercial grade corrugated asbestos cement roofing about 1/4" thick was placed-in contact with 61S-T6 and exposed for 3 months to 100$ relative humidity at 125F. This material contained 3.8$ free lime and a water leach of it had a pH of 11.6. Although the material was alkaline in nature, it did not cause any corrosion of 61S-T6 in the three month period, even though the specimens were continually wet during this time. 85^ Magnesia Insulation Small blocks of 85$ magnesia insulation were sandwiched between panels of several aluminum alloys, including 3S-HL4 and 4S-KL4 and exposed for 6 months in the 100$ humidity chamber at 125F. The samples were continuously wet and visual inspection revealed that the panels were in excellent condition and free from corrosion. The magnesia insulation was alkaline, a water leach of it having a pH of 9.9. Resistance to Atmospheric Weathering of Aluminum Alloys The data in Figures 8, 9, and 10 are included to illustrate the higji inherent resistance to corrosion of aluminum alloys when exposed to a wide variety of weathering conditions, including seacoast, industrial, urban, rural, and tropical. There is very little difference in the resistance to weathering of aluminum alloys 2S, 3S, 4S, 52S, 53S, and 61S; consequently, the data for one alloy may be used to indicate the expectedATC 0034900 XF-289 Page 13 performance of the other alloy. It is apparent from the data that aluminum alloys weather at a slow rate and that the rate decreases to a very low value after about two years. Obviously, data obtained within two years can give a good measure of the ultimate performance expected of these aluminum alloys with re gard to their resistance to weathering. The aluminum alloys were, of course, outstandingly superior to steel (Figure 8), and comparable to copper (Figure 9). In some strongly indus trial atmospheres rich in sulphur bearing products, aluminum alloys would be expected to be superior to copper. The surface appearance of aluminum alloys will vary with the nature of the environment. Aluminum will generally weather to a light gray, which should blend well with asbestos cement siding. In industrial atmospheres, the accumulation of industrial soil will cause some darkening of the nail heads; this is true of not only aluminum nails but nails of any other commercial alloy. The siding may also be darkened somewhat so that the nail heads should not be conspicuous. 0NCLD5I0MS: Since we understand that the six lots of asbestos cement siding are representative of commercial production, it is believed that the following conclusions are justified. (l) Accelerated laboratory corrosion tests and exposures for two years to relatively severe industrial and seacoast ATC 0034901 atmospheres show that aluminum alloy (61S) nails are suitable for use in asbestos cement siding. The 61S nails revealed a |C-J XF-289 Page 14 high inherent resistance to weathering, and contact with the siding of six different manufacturers did not have any corrosive effect on the nails. (2) Furthermore, 61S nails employed 1-1/2 years ago in several large asbestos cement siding installations in San Francisco are reportedly giving excellent service. (3) Laboratory tests show that this favorable performance might be attributed to the following: the dense nature and stability of asbestos cement siding actually resists the leach' ing of appreciable amounts of alkali from the siding; and, in addition, 613 nails are resistant to any alkali leached from the siding under service conditions, because of a protective film which the metal develops. CJW:mcc ATC 0034902 IF-289 Page 15 TABLE I EFFECT OF LEACHING THE FACE OF ASBESTOS CEMENT SIDING ON THE RESIDUAL ATXATINTT7 OF THE SURFACE Distilled Water Not Changed^ Cumulative Panel Temporarily Immersed^) Exposure Run Period In Fresh Distilled Water pH Exposure pH 1 1 min. Uoi 5 min. 3 10 min. 4 15 min. 5 30 min. 6 1 hr. 7 2 hr. 8 4 hr. 9 6 hr. 10 22 hr. 11 12 13 14 15 16 11.0 11.6 11.4 11.5 11.7 11.7 11.5 12.2 12.2 11.5 1 min. 1 min. 1 min. 1 min. 1 min. 1 min. 1 min. 1 min. 1 min.. 1 min.(3) 5 min.(3) 10 min.(3) 15 min.(3J 25 min.(3) 30 min.(3) 40 min.(3) 10.0 9.9 9.7 9.7 8.7 9.6 9.2 . 7.0 7.2 7.0 7.0 8.6 9.1 9.8 9.9 10.1 (1) Panels 2" x 6" entirely waxed except for the face which is intended for exposure to the weather, immersed in 150 ml of distilled water, the initial pH of which was 5.5. The gradual increase in pH was caused by the leaching of alkali from the siding. (2) Panel was temporarily removed from the solution (1) above, rinsed, and immersed for 1 minute in fresh distilled water, and the pH measured. (3) After 22 hours of leaching in solution ,(l), the panel was placed in fresh distilled water and allowed to remain to determine the time required to again produce a pH of about 10.0. CJW:mcc ATC 0034903 XF-289-Page 16 TABLE II RESISTANCE OF ALUMINUM ALLOYS TO CORROSION BY CONCRETE DURING THE SETTING PERIOD Alloy 3S-0 4S-0 52S-H38 Surface Oiled No Yes Yes Setting Period 4.5 da. 4.5 da. 3.5 da. __ Peplh of Attack Average Maximum 0.0005" 0.0005" 0.0008 0.0008 0.0003 0.0004 Note: (l) Specimens were 0.064" x 4" x 6". (2) Portlant cement: Si02 21.7$, A1203 4.9$, Fe203 4.2$, CaO 61.4$, MgO 4.4$, S03 1.6$, Ignition Loss 0.73$. (3) Mix used: 1 part cement, 2 parts sand, 4 parts gravel. CJW:mcc ATC 0034904 XE- 289 Page 17 TABLE III RESISTANCE TO CORROSION OF 52S-H32 ALUMINUM ALLOY IN CONTACT WITH CONCRETE AND EXPOSED UNDER MOIST OR WET CONDITIONS FOR SIX MONTHS AT lov Exposure Depth of Attaok(l) Average Maximum Appearance 52S-H32 100$ relative humidity at 30 C 0.0011" 0.0014" Uniform etching 52S-H32 Alternately to 100$ humidity 0.0008 at 52C and to refrigerator at 4C 0.0019 Uniform etching 52S-H32 Half immersed in tap water 0.0005 0.0009 Uniform etching Note: (l) Corrosion of surface in contact with the concrete,, (2) Bubble slag concrete had been poured into shallow boxes formed from 52S-H32 sheet 0.032" thick. CJW:mcc ATC 0034905 TABLE IV 53S-T6 ALUMINUM ALLOY RODS (1" DIA.) IMBEDDED IN CONCRETE! 1) AND CJW:mco 8 t> CO 8 CO r--1 CO to CT> -sH 03 CO O C\J CO aos CO cot ooo CO P P '>>>> CO i--ICO o 03 Er* ;; CO < oPO X co o co o cO O o oO oo0 o=( o o o 03 n<TdOD O CO :> cr> r-t o o 03 o eg c0 o O 3 o O o 0 o 0 o oo -P -P Si M fC-D TOO3 O --1 03 X in 03 r--1 o oO o oO ooe OCO pe (=<D T033 P > o oo D- CO CO P CO H <--I o O I o o o o o o CO o o o oo H <m a o3 p o rH PP O O -rH O -P 1-4 o co) -p p o p oo p 1--( 0) o CO <pM ---- O a> -P m -p p H po -aop P< Po rH o o -p 1 oa -P p Cm 03 PO O t P4 P o H o o 03 p 03 p o c o o Cm o 03 M O O rH XI P< a) CD nd n 00 X o ' rH Td ao p H -P 03 03 03 od PO P 03 -P 03 CO rH Td t i--i *X3 03 -P O -p o p p. p to __ --1 '--' -P o a P P 03 P 03 Cm Cm M id 03 X! p P cc 03 P O -i--* p C3 03 03 m o p o Cm O P o H P CO P rH a CO X C3 H P< o C3 03 o po rH a t> X nd nd-p 03 P-H M i--1 a p pp o QN 03 XT'-289 Page 18 ATC 0034906 Figure 1 Tests were made on aluminum alloy 2S-H14. The solu tions were alkaline; the pH of the sodium carbonate ranged from 9.0 to 11.4. The pH of the sodium carbonate - sodium silicate solutions ranged from 10.3 to 11.3. The straight sodium carbonate solution was corrosive, whereas, the sodium carbonate - silicate solutions were not significantly oorrosive even though alkaline. ATC 0034907 ALUMINUM COMPANY OF AMERICA ALUMINUM RESEARCH LABORATORIES :PORT NO DATE FIGURE NO JL XF-289 Page 19 j 4UO-47 "mnunot'" h<> t i*J - 750 700 650 600 550 500 450 400 350 300 250 200 150 50 5.0 4.0 3.0 2.0 l.C 0.1 0.2 0.3 % Concentration ATC 0034908 Neg. No. 56410C Figure 2 Photograph showing condition of nails and asbestos cement siding after two years' exposure to either the seacoast atmosphere at Poiht Judith, R.I., or the industrial atmosphere at New Kensington, Penna. ATC 0034909 ATC 0034910 Nego No, 56445A Figure 3 Photograph showing good condition of 61S nails as removed from asbestos cement siding.after two years' exposure to .the industrial atmosphere at New Kensington, Prana. ATC 0034911 ATC 0034912 Heg. Ho. 56344A Figure 4 Photograph showing good condition of 61S nails as removed from asbestos cement siding after two years1 ex posure to the seacoast atmosphere at Point Judith, R. I. ATC 0034913 ATC 0034914 Beg. Ho. 60309 Mag. about 8X Figure 5 Photomicrograph of longitudinal section of head and shank of 61S nail (roofing) driven in asbestos cement siding and exposed two years to industrial atmosphere at Hew Kensington, Penna. No significant corrosion had occurred. Sliver on top, center of head and ridges and valleys of shank are mechanical in origin. Beg. Ho. 60310 Mag. about 81 Figure 6 Photomicrograph of longitudinal section of head and shank of 61S nail (roofing) driven in asbestos cement siding and exposed two years to seaeoast atmosphere at Point Judith, R.I. Ho significant attack was evident. Ridges and valleys of shank were mechanical. ATC 0034915 ATC 0034916 ALUMINUM COMPANY OF AMERICA XF-289 Page 24 ALUMINUM RESEARCH LABORATORIES REPORT NO_______ _____DATE FIGURE NO_____Z____ otc o Exposure Tine - Hours ATC 0034917 Figure 8 Data shows the relative severity of the test sites at New Kensington and Point Judith. The severity of them is indicated by their effect on mild steel. Note the high inherent resistance to weathering of the aluminum alloys, which include 61S. Also, that the rate of weathering de creases with time. Loss in tensile strength is used as a critical measure of extent of corrosion. ATC 0034918 C a> PCdD <d D0**OH ch E 0033 > -rd 03 oa*i up-'pd rCD Ca>d 03 oCd P C oa CD '-< CD 43 -rd Ch P 03 CD O c S+> cd o -n Cd rH P <H O d) cd oxi ai+3 CO -P 03 Cd_q 03 3 03 <H OH-P Cd -H <H ) B <1> 0 0) -P 03 >; co t>> 0) o Si -P Ce -rH H J3^~v CO Cl> S -P ^ Ih o . o CD 03 C = -rH Si >-h -c^x: aour<ad poe co -p H a)rH<eo^ t co H (DHO PCI o . <c-4dP si O CD >,P o . Cd 43 CuO 03 cH CH CD CD TO 03 Cd 0) 43 03 Cd--------- -p 03 -P 03 00 S3 CO CO 0/0 C CO G) OH H (3 rH I--( -rH -P O *H rH E Cd-r-t 03 cd CdP C CQ P Cd 03 s 03 03 -P P 3 o 3 03 GxJ <--i a *h (--I 03 H E CO *H Si 3 cd 3 03 rH 43 03 03 cd Cd 0,0 CQ 03 Cd CU 03 Q3rrJ f-t X 043 H U 03 rH -d O 100 /vjC 0034919 Figure 9 Data obtained from A.S.T.M. tests showing that aluminum alloys (2S-H14 [l/2H] and 3S-EL4 [l/2H]) weather at a slow rate in a variety of atmospheres and were equal to copper. (He real differences exist between 61S-T6 and either 2S or 3S in atmos pheric exposures.) ATC 0034920 XF-289 Page 26 LOSS IN TENSILE STRENGTH PERCENT note: ALL MATERIAL EXPOSED AS MACHINED TENSION SPECIMENS O.OJ5" THICK PITTSBURGH, PA. (INDUSTRIAL) ALTOONA, PA. (INDUSTRIAL) A NEW YORK CITY (INDUSTRIAL) A ROCHESTER, N.Y. (INDUSTRIAL) SANOY HOOK, NJ. (SEACOAST) KEY WEST, FLA. (SEACOAST) LA JOLLA, CALIF. (SEACOAST) O STATE COLLEGE, PA. (RURAL) A PHOENIX, ARIZ. (RURAL) Fig. 9 ATC 0034921 Figure 10 The curves show the rate of attack of aluminum alloys over a period of 10 years to the seacoast atmosphere at Point Judith, R.I., and over 14 years to the industrial atmosphere at New Kensington, Penna. These curves are extrapolated to a period of 42 years. Note that the attack is shallow and that the rate of weathering of aluminum alloys decreases markedly after about two years; also, that aluminum alloys erposad at many other localities in this country generally were not appreciably affected. ATC 0034922 XF-289 Page 27 o > cr Z O O _o az 2 O 6 if 2 jsr : -J o -- ?o 2u 2-- 5 to o CO o? $z*oU"I* >^ . to <r ^ a _ w 2 No iuj z d O J < > X d!3< <io isiljo o lego o <r a: ?"S t ? - o jwoi* ^ <o o to to w X ' <*0> ItOn z o Q. O CC i X o' a UJ ( </> ^ UJ cUJ aUIJ uj a. o <n n OK UJ UrJ 52 CL Oto zo5 CO tfOo" tro>"_ 5< < to* to csi to -- Oo UJ UJ to to oo <0 (0 ** to* to a: uj x Q_ to O 2 <5 <9 h- _J < < h(T CO K to E Z 2 < 3k2 q. < O m i it = 55 -l "i* - J to to* to" to* to* io o, r *o to* to to to* cj cu fj cj <D U C3J> rH lu UoJ UoJ c<r a<: (/) z << o o -- to co < o o < 0 O8 zO O u.<7-i . z to it to a) X* . 2 << J CD u. o SsXSi 385|E 5uaj: z<5 dS*5 -5 V o to to to jto ic <0 to ttoo wo itNou^>eJ(OOtoJMto w -->JC> ATC 0034923 E. H. 3*2rTW-t2-30 Prrnud m U.S.A. no ntozJR. A. R. L. - NE'" lINSIN^TCN V ALUMINUM COMPANY OF AMERICA t / 8 HOTED 79<tg ALUMINUM RESEARCH LABORATORIES PROGRESS REPORT JUN 2 71942Page 1 , ____ ___ E. H. D., JS. Problem no CL-64 - XNTSSTIGATION 0? THE PERMANENT MOLD PROCSSS 1803452 CONCLUSIONS SUGGESTIONS Operator A. Dean date 5-20-42 EXPERIMENTS REMARKS (Cast Iron and Aluminum Molds Cast to Finished Dimensions) Ashestos-C waent Mold Synopsis Attention has been given so materials for pc rmanent molds. One phase of thin problem is to pro luce east iron pernanent molds quicker and chea]>er than the presen; method. Another is to produce molds from materials other than eas; iron. In the present report an attempt was made to cast cast iron inold halves to fini shed dimensions, Surfaces were smooth and of satisfactory quality for a production mold but the castings were warped and hal to be reworked tc a greater extent than was desired. A better molding sand, one of which is now under consideration, might eliminate this warping. The mold halves were finished to print dimensions and hare been used to pro duce commercial castings. The se qond phase of this problem was Invest ijgated by prepar- ing molds from ti e same pattern equ.pment as the cast iron mold but of aluminum and of dn asbestos cement fixture. Thus materaials of higher and of considersMy lower thermal e mduetivity than ca st iron were available. It aI peered from these preliminary experisje nts that alumi- num alloy molds sight be siade to fi: lal dimensions and used in production provided the same freedom from warp ng encountered in the cast iron mold, can be achl eved. After some preparatory experia^e nts a set of mold halves of an asbc stos-camant mixturi was made. The s cracked during these experiment and the mold halve s were also warped After a thermal treatment to exp4l moisture it appealred that castings |Qf satisfactory external appearedce also could be p;foduced from suoh aj mold. A miore scopic examination of castings from the three molds revealed that t$ one of asbestos-coment produced castings of internal appearance similar to a sand casting Solidification of castings in this mold was much slower than in e metallic mold. fhere was no significant differeence in micro appearance of the castings produced in the uncoated ejast iron and alumiir'am alloy molds. A light spray of wash on the metal lio molds reduced ihe chill to small tmt noticeable extent in each case. These experiments indicate the need for furt ler attention to the problem of csating permanent moEds to finished dimjaneions whether from cast iron, elluminum or other materials. Introduction ATC 0034924 The pre sent report is a scussion of & phas)b of an investigation of material s for permanent mdlds. The occasion for the particula 3427 iU`12-39 Printed in US.A. ALUMINUM COMPANY OF AMERICA Rev. S-1927 ALUMINUM RESEARCH LABORATORIES PROGRESS REPORT Page 2 Problem no CL-64 - INVESTIGATION OF THE P2BUAHEKT MOLD PROCESS Operator A. Dean datf $-20-42 CONCLUSIONS SUGGESTIONS EXPERIMENTS REMARKS experiments desoribed in this report was a request fr Ur. G. Kohl of the permanent mold plant that we attempt to cast mold halves to dimensions or at least so close to the final contour olf the desired mold that finish!ng operations would he relatively slaple and inexpenslve. The parti cjular job was a sbklJ. bracket for Yougjht Sikorsky Aircraft Corp., two of which were to be poured from a central gate. Three sets of molid halves were prepared from the pattiem equipment, namely, a set of cast iron, one of .l132 alloy and one of a cementasbestos mixture, A microscopic study has been made of brackets cast in each of the ttree molds. Experimental Several attempts to make -he iron mold halves in green sand some of which had been obtained fro& the Cleveland Coojperative Stove Co., were not sucjcessful. The iron {scrap Meehanlte lted in H Plant) washed and cut ttye green sand. A d:^- sand mold was ustod with some success but the bore sand mold was ot hard enough for the Iron. The mold halves were warped slightly an< had to be reworked in the machine shop. The Alj 2 alloy mold halved were cast in greem sand and were warped to a great or extent than the iron. Better results would undoubtedly be obt sjlned by casting thd aluminum mold hal res In dry sand, A132 was selected because of its loir coefficient of exjpanslon and because of its high resistance to abtrtis ion. Although the metal was fluxed with chlodine the gas contend of the castings was high. The use of aluminum alloy molds in prodnotion would requirje greater attention than Is givejn to iron molds. ]'or example, while arming up the mold, the tempera tune must be kept >elow 950F to prevjent melting. Any wash applied to tjhe surface should >e easy to remove sLnce scraping or the use of a wire brush will rapidl; impair the sharpnjs so of contour, One of the salt binder type washes Soluble in water ml ghfc be used. An attefmpt was made to el: mlnate the warp frbm the A132 mold halves by heating them in^as fired furnace under a 1oad of 100 pounds, Some improvement |was effected but t*h' e warp was not completely eliminated The A132 alloy mcld was not machine finished to the customer's print and was not used in production. The third mold, one of ceuent-asbestos was made only after a preliminary invesVtAiggaaVtAioWnU wHlhLAicUhil wWAilAl.,. bfcrew described AInU vs ame detail now. Interest in mold materials of relatively low thermal cbnductivity has been expressed fr am time to time by different individuals familiar with the permanent mo!3d process. For exemple Transits has been suggested as a material which was hard, very smooth and perhaps could be used witfcout the appllcat;. on of a wash. Ur. L. W. Kempf discussed with ATC 0034925 3427 31^-13-3$ Printed in U.S.A. ALUMINUM COMPANY OF AMERICA R*. 5-1927 ALUMINUM RESEARCH LABORATORIES PROGRESS REPORT PROBLEM Nr. CL 6A - IHYBSTIGATIO'ri OF THE pgRMAJTgflT MOLD PROCESS Page 3 Operator W. A. Dean date 5-20-A2 CONCLUSIONS SUGGESTIONS EXPERIMENTS REMARKS Kr. R. H. Sell anil of the Mellon Inslfc itute, the nse of molding mixtures similar to Trans:.te. In a letter f rom Mr. G. H. !?agniT to Mr. I. If. Eempf (November \0, 1941 Re: Kixturea for Molding T:rt nsite) a mixture was suggested consisting of 15# asbjo stos fibre and 85^ Portland oement by weight. LumnJ.te cement, because of its hl^ier thejfmal stability was reported to be preferable to Portingd cement. Direct] ons also were given for the preparation of such a mixture. A supp;_y of Lumnlte cemen|t was obtained and also three samples of asbestos recommended by fir. Fleck of Johns Manvllle Co. These were identified as follows AH - Shingle fibre (0.0-5.0-10.0 -1.0) - $73.00/ton 6D Waste. Stlucco or Plaster fi 5re (O.O-O.O-7.O-9. 0)-$30.00/ton 7D Refuse or Short fibre (0.0-D .0-5.0-11.0)-$30. CO/ton All price i1 s F.O.B. the mine The designation of the different grhdes in parentheslc represents the ounce8 of materia,1 from a 16 ounce sample left on thre e screens and collected in the bottom of the pan AH therefore. 1s coarsest and 70 the finest. The first investigation wis carried out in e sub-press used In the Amsler maqhlne which producep a cylindrical spc<olmen of 0*50 square inches in area. A sample of each grade of asbestos was soaked in water over night and at the start of the experiment Lumnlte cement was added to eael to make a 15# asbestos-85# cement (ty weight) mixture, Excess water was decanted, the sub-tress cavity filled with the mixture and specimens prcd*u"ced by applying pressures of 1000, u2000, 5000, 10,000 and 20 .OCX little differenee in the appearance of specimens made from the different grades of asbe9tola except that those from AH we 19 ^Lightly rougher ihan the others. Ajll were readily scratched with a pen knife. Somewhat harder surfaces would have de- veloped had the specimens been tempered with water during the aging period following their formation, jin attempt was made to Brlnell the specimens. Many of them cracked bug in some instances readings were obtained. Figure 1 shows the speciiiens, the pressures at which they were formed and tjhe Brine11 readingu where failure did not occur. The AH (long fibre) specimen were strongest and as a group the 7D specimens were next. As wicjuld be expected, resistance to failurje increased with pressure at which the specimens were formed. An att empt was made to de 1 .ermine the best ra tio of cement to asbestos. Grade 6D was examined aiice it offered grea ter room for improvement than ei ther AH or 7D. At two pressures of 2 300 p.s.i. and 10 000 p.s.i. mixttures containing Id , 15, 25, 35 and 53 per cent asbestos by we igbt were formed into specimens. These specimens were ATC 0034926 3427 5W-12-3 Printed m U.S.A. ALUMINUM COMPANY OF AMERICA Rev. S-lft27 ALUMINUM RESEARCH LABORATORIES PROGRESS REPORT ?&ge 4 Problem no CL-64 - XSV33TIGATICK 0? T55 rZaittiigST MOLD rHOCSSS Operator A Dsan date 5-20-42 CONCLUSIONS SUGGESTIONS EXPERIMENTS REMARKS tempered with wa ter 4 hours after thay were formed anc received three additional appliRations of water at hourly intervals, The results, shown in Figure indicate the opt iferom concentration f asbestos fear wo-rl mnB hardness is in the neighbor}*Lood of 15 to 25 pcr cent. From these experiment^ it was concluded that desirable eon 4itions for the proposed cement--4sbestos mold would be a mixture of 1! to 25 per cent of long fibre asljestos and a high fyralng pressure (10 ,000 p.s.i. or greater). A prel minary experiment showed that reduct i on in volume of a 16-2/3^ asb4 stos-84-1/3^ cemenjt mixture at a prei;sure of 10,000 p.s.i. was appro;ima tely 4 to 1. *L th this informat i('n a boiler plate rectangular frasr^ was made. The si ies were of 1" plat e, the ends of 1/2" stock and the frame was held together with 1-3/8 threaded bolts (drawing CD3-D-1(>636). The bottom yf the box was clojied with a hard maple board 1" thick. Pressure in the press was to b< transmitted through a built hard maple block that fit into the boiler plate frame with 1/3" Clearance on the sii es and 1/4" on thu ends. A view of the frame with bottom maple boari and maple block shre shown in Figure 3. A sido view of the equl nt is shown in F gure 4. The inside dimensions of the box were runghly 7" deep, 14 long and 4-1/2" wide. Plaster casts were taken Tram the patterns from which the cast Iron and AV 2 mold halves alreja4y described were made. These plasters were us4d as patterns to e a pair of dies also of A132 alloy, which werf to be used subseq lently to form the asbestos-oement mixture. The dl4a were machined to establish a 1/16" clearance when fitted in the boi. Figure 5 shows these dies after they were used to make the asbestos-cement mold halve s. The gate, both risers and the two bracket cast ngs are apparent. Both dies cracked during the experiment. The 1500 ton press at the forge shop was us>d for these experiments. Thn boiler plate form with 1" hard mapld bottom was placed on the larjrer die of the press. The framo was filled to within 2" of the top wi a mixture of I6-2/35C asbestos and q3-l/3 Luamlte cement (by weight } Then one of tha two A132 alloy d es was plaoed on top of the mitture. The built up maple block was placed on top of the aluminum die and pressure applied from the upper Me of the press. At first a light pressure was applisd to the assembly and the mixture squirted out at uhe top. It was apparent that the in 1 entional clearance in the equipment were not enough to permit the water to squeeze out. The press die wat then very slowly irifted into contat t with the Re sambly and allowfd to drift until tie top of the boilir plete frame was contacted. The assembly was raised on two blocks and the contents ATC 0034927 M27 5M-12-3C Printed tn U.S.A. ALUMINUM COMPANY OF AMERICA Rev. 5-IW7 ALUMINUM RESEARCH LABORATORIES PROGRESS REPORT Pag 5 PROBLEM Nr> CL-64 - INVESTIGATION 0? THS PERMANENT MOLD PROCSSS operator W A. Dean hatf 5-20-49 CONCLUSIONS SUGGESTIONS EXPERIMENTS REMARKS of the box slowly pressed out of tho bottom. Light ta pping permitted the A132 alloy dije to be drawn from the asbestos-oemenjt. mixture. The die had cracked, It was thought, dujflng the initial ap plication of pressure. Both pie blocks were wirped somewhat. When t he other half of the asbestos-cement sold was made the upper die of the press was drifted into contact wi h the assembly and this time non b of the mixture scuirted out. On r amoving the alumlnui die it was found to have cracked also. Both asbestos- jement mold halves were warped becaujse of the cracked lies. Figure 6 sho its the asbestoscement mold halve $. These experiments indicated the n<ed far more substantial base and follow-up block tlan hard maple. Alt o dies with greater ductility than A132 alloy aia needed. With soiie experience in seating the dies Ln the asbestos-cenent mixture (a vibrra tor might be useful) perhaps c racking of the dies and the resulting warping of the mold halves can b prevented. The experiments showed the need for the slowest possible tpplioation of pressure to permit the excess water to be squeezed from ;he mixture. The mold halves were tern;pjered with water, 4 11 and 24 hours after they were pressed. The halves were dressed with a file to reduce the extent of the warp. The mixture may be filed read ly provided no fine particles of steel or iron are encountered. These probably are present as impurit ies in the cement. To fur ;her investigate these mold materials castings of B195 alloy were made In each mold. The pouring temperature' was about 1350QF. This alloy is use<[ to produce the Voight Sikorsky braclet commercially. For the first exp4riment the cast ir an and aluminum mo] da were used without a wash anc the asbestos-eerie it mold had been ;ed only at room temperature. All three were heated ;o about 250F aftar which a series of about 15 east! were made. The iron and aluminum molds were then sprayed, lightly wi th Silocel and in >he meantime the as bestos-cement mold was aged for 3-1/2 hours at 300 *F to drive out moisture, which before this treetie:t, had bubbled I ito the gate of the casting giving it the appearance of having been mad ) in green sand of too high a moisture content, This thermal trea aaent effectively stopped the evolu- tion of moisture ^hen eastings were >oured into the mold subsequently, Castings made in tjhe asbestos-cement]mold solidified fair slower than those in the metal lie molds. As mentioned previously only the cast iron zriold was finished to the print for the casting. Both the A132 and asbestos-cement molds produced castings of considerably thicker section than required. The only seotlons of clomparable thioknesj in eastings made in the three molds were the rils. These were used for micro examine tion. Castings selected for examination were those loured after it was thought the mold A70 0034928 3427 31M2-39 Printed in U-S-A. ALUMINUM COMPANY OF AMERICA Rrr. 5-1927 ALUMINUM RESEARCH LABORATORIES PROGRESS REPORT 6 Problem No. CI.-6A - INTCSTIGATIOM QF 1HS PBjM&HFiaT 'J0T.T1 PBOCgSS Operator W. A. Dean date 5-20-L2 CONCLUSIONS SUGGESTIONS EXPERIMENTS REMARKS was at operating lemperature Figure^ 7 and 8 are mi crotsactions of B195 a1Toy castings made in the asbes ios-cement mold be re and after the aging treatment at 300F. There s little different in the appearanc^ of the micro- constituents of t]Ke two eastings or in their dendritic patterns. Both are quite coarse and are of similar appearance to a sahd casting, Figures 9 and 10 uhow the microstruc|tures of oastings 1 iade in the cast iron mold used wit h and without a cJ0 3Lting of mold wash, The constituent size and dendritic patterns are much finer than those produced in the asbestos-cement HU> Id. Of interest i observation t hat the light spray of wash decgreased the chill, trueture in Fi,gure 10 is alight- ly coarser than t: at in Pigure 9* es 11 and 12 si ow the structures of oastings made n the A132 alloy moId. Castings madd in the iron and aluminum mold A which had not bee a sprayed with was were of very similar micro app4 arance. The effee t of the light wash on the aluminum mold reduced chil to a greater erte at than a similar tfash on the iron mold as may be seAn from a comparison of Figures 10 an(. 12. This dif- Terence in the eff ectiveness of a wi h on the iron and aluminum molds should not be re gnrdeed as slgniflcanjt in view of the 1: mited information since the wash on the aluminum mold nay have been sligltly heavier than that applied to t] Le iron mold. It i 3 well known that ight layers of wash considerably reduce chill but s ibsequent applicati ons of the same thickness do not bring about a corre spondingly further reduction in chill. MscussionandCoi Lelusions 1. An attempt lias been made to ea st oast Iron mold halves directly from patterns of the finished casting. The mold h&lve^ warped and had to be worked in tljte machine shop to 1 greater extent than was desired before they could be used to produoe commercial castings. It is thought that the dry sand mixture was not pr :>per for this job. Another sand has been recommended l>y Ur. O. Gardner o f the G.T.L. and tlis will be used when an experiment of this kind is t cried again. The strface of the iron mold halves t .a cast were very a sooth and of a qua! ity that could have been used wit hout further finis ling had the molds not warped. It is the opinion at this time that molls cast to dlmsnsicns can be made successfully if wurping can be overcome 2. A set of mold halves were also made of A132 allojf fresi the same equipment. These too warped badly akd were never finis hed for the produetion foundry. Their surfaces wars also very smooth and it is thought that molds of a suitable aluminum alloy might well be tised in the foundry particularly for obs where the numbpr of oastings reqtjired ia limited. Such muuoj.ludos,| aos ixs pwointeud o.ut i..n1 t.h.e..e bwoudjyr ow.f t.h..i.sk. rxepo: v, would require more than normal care during preheating and cleaning alter use to prevent ATC 0034929 3427 3M-13JS Printed in U.S.A. ALUMINUM COMPANY OF AMERICA Rct. S-IM7 ALUMINUM RESEARCH LABORATORIES PROGRESS REPORT PaQ 7 Problem No CL-oA - INSTIGATION 0? TFT PERMANENT MOLD PROCESS Operator >7. A. Dean Date 5-20-42 CONCLUSIONS SUGGESTIONS EXPERIMENTS REMARKS malting In the t i:rst case and to preserve their sharpn< i ss in the second. Further irorlc on this phase of the permanent mo Id problem appears warranted, 3. An investigu tion ha 8 been made of a relatively poor heat conducting material, nam* ly an asbestos-ceaeint mixture. To obtain mold halves which would produu e castings similar to those made in bhe east iron and aluminum molds aip eady described, AA3 2 alloy die halves a boiler plate frame and other e uipment was prapara d to press mold hi.Ives from a 16-2/3# asbestos-^3-1/3# cement mixt.lire in the 1500 toi. press at the Forge Shop. Duri&g the experiments the A132 alloy diesi cracked and the mold halves were ijrarped. The mold haIves were dressed so that castings could be produced The surfaces of :hese mold halves 1 ere also smooth. To drive moisture from the mold a traataent at 300? vt s found desirable 4. Solidification of the casting Ln the asbestos-oei te:nt mold was much slower than :m the metallic mol is. Before the as Cestou^ement was given a thermal treatment moisture bubbled from the mold/contaminated the casting. A microscopic examinat Lon of corresponding sections from easting*? made in ;ha three molds shoWed the structure <f the ones made in the aabestos-coment mold had the appearance of a saz d casting. Those made in the two metallic molds were if fine structure Characteristic of the permanent mold process. There appeared to be no di fferenoe in the structures of castings made in the uiooated iron and aj umtnuffl molds, The effect of a light layer of wash m both of the methllic molds was apparent in the roduetion of chill wiich the wash prodineed. The wash on the aluminum a< ild was more effect Lva but that on th< aluminum mold may have been slightly thicker than fhe layer sprayed Qn. the cast Iran mold. 5 While these experiments do indL cate that molds oi simple design can be made from isbestos-ceaent mixpures the structure of castings made in such a mold wo;ld be expected to >e more like that qf castings made in sand rather t:hah. ln a permanent >Id. cc-Hr.E .H. Dix ,,JX. Mr.F. jarfline Mr.G.Kohl Mr.T.D.Stay ATC 0034930 'je ATC 0034931 i* 's ATC 0034932 10 ' '4e ATC 0034933 AFC 0034934 12 A 1 C 0034935 13: 'V 3 ATC 0034936 Fig. 7 Asbestos-Cement Mold-Dried at room tempera ture. Microstructure of 13th casting at 100X etched with 0.5$ EF. Negative A3222. Fig. 3 Asbestos-Cement Mold-heated 3~i/2 hours at 300?. Microstructure of 11th casting at 100X9tched with 0.5$ HF. Negative A3220. ATC 0034937 Fig. 9 Cast iron mold - no wash - microstructure of 8th casting at 100X - etched with 0.5$ HF Negative A3218. Fig. 10 Cast iron mold light wash - microstructure of 14-th casting at 100X - etched with 0.5$ HF Negative A32A1. ATC 0034938 Fig. 11 A132 Alloy Mold - No wash - Microstructure of 11th casting at 100X - etched with 0.5# HF Negative A3221. c. vV vWt:--. >>./ .y- .t . ,, ^ * -C*^ . w % Fig. 12 A132 Alloy Mold - light wash - Microstructure of 11th casting at 100X - etched with 0.5# HF Negative A321 9 ATC 0034939 R. 2. JL 3427 5M-12-38 Printed in U.S.A. ALUMNUii SISE^RCr: LAIN KZv, iKENSINGTON R^t. 5-1227 ALUMINUM COMPANY OF AMERICA T WS7& 1809462 ALUMINUM RESEARCH LABORATORIES IIK rogress report ,,o7 KL 23 1940 >-. -v ^ p. i PROBLEM No.CUT--5791, Cl.C.T.-287--TFTWRWTWftTIQB- OF 5TIITA3TMTY OF SBAPBITE- TREATED ASBESTOS CASKET TOR OSE WITH HAC*PSIPW 6LI.OTS- 6-4-40 CONCLUSIONS SUGGESTIONS EXPERIMENTS REMARKS Rf: (1) Letter T. Wood to K. W. Daugherty 5/28/40 Re: Deter- mina of Suitability o Graphite-Treated Asbestos Casket for i With Magnesian .oys (2) Book 8, p. 34, 6/5/40 (3) Pro, Report 10/25/35, jd-54, C1.C.T.109, Determination of the ctrolytio Corrosion Characteristic^ of Magnesium Alley* in Contaot with Fe and Bon-Ferro os Alloys. (4) Progre Report 5/1S/S7, BX-S728, Cl.C.T 1, Evaluation of Qe Motors Cork an Fabric Insulating Strip and DuPont nsulating Tape. (5) Progre Beport 2A0/39, diS-4989, Cl.C.T. 2|B, fellutex Gasket terial for M&gnerlum Base Alloys. A gra te-treated asbestc s gasket was ob tailed from the Bendlx Aviation rp. to test its eq:rrosive effect on Magnesium. The gasket la used a a earburetor throtitle body gasket. a*id is .032* thick. Two si ar piece*, each ilth a single surfabe area of about one square ineh, here out from the gasket, superimposed and clamped firmly between tWi machined pieces cf AH265-C alloy, Each piece of the alloy was S/l|l x 5* x 2".and he d not bean coated, They were clean- ed with easing h id gasoline oefore assembly. To &voi$ metallic eontaets the dampish was dene by neaaa of rubber bands. The ela sped assembly was continuously immersed in a vertical position in tap v ter at room temperatore for 8 weeks, with a daily change of water, ninee previous simitlar tests had been I made in that manner. After thi 8 week exposure the AM26S alloy surfaces which had contacted the gasket material were cleaned with chromin acid and photo- graphed (Figure 1 As shows i in Figure 1 each specimen was sever^ly pitted where in contact with tie gasket, partioul sxly along the ga alet edges. The maximum depth of bitting was approxijaately .03 inches, This indicated aleotrolytio action probably between the graphite and 4agneeinm. Cornparison of this result with previous resalts with other gasket materials, as reported under ICl.C.T. lOt, 178, ind 29, indicates that the graphitetreated asbestos is a decidedly unsaltable gasket mateilal for use in contact with magnesium. StBHAHY Pieoes of a graphlte-treatfd asbestos gasket were damped between machined surfaoes of AME05-0 alloy, and continuously immersed in tap water for 8 weeks. After exposure the magnesias surfaces which ATC 0034940 3427 5M-13-39 Prated in U.S.A. ALUMINUM COMPANY OF AMERICA Rrr. 6-IM7 0056475 FROM R. A. CHRISTINI CHEMICAL PRODUCTS & PROCESSES ALCOA TECHNICAL CENTER - B T0 VITO CEDRO CHEMICAL PRODUCTS & PROCESSES ALCOA TECHNICAL CENTER - C ALCOA PRIVATE INFORMATION 1995-11-07 09471 RE: EVALUATION OF CAMERON-FIAT ITALIAN Mg PROJECT* 1 Letter Report No. 09-95-39 Abstract The Cameron-Fiat Italian Mg project proposes to use serpentine (asbestos) as part of the Mg ore feed. Mass and energy balances were calculated and used to calculate costs to compare to Northwest Alloys' costs. The major conclusions follow. 1. Cameron's process using serpentine (asbestos) as a partial MgO feed source does not produce a cost advantage even with no costs attributed to the serpentine. 2. Due to CaO requirements in the slag, dolime and/or lime must be fed, thus reducing the serpentine portion to only 10-18% of the total oxide feed. 3. Larger slag volumes are required which will reduce productivity. These slag compositions have little sales value due to the high SiC>2 content. 4. Five raw materials must be fed which requires more capital for the feed system. (NWA has only 3 feed tanks per furnace.) 5. Higher power requirements will increase costs by $0.01-.03/lb Mg due to increased amounts of raw materials and slightly higher operating temperatures. Recommendations 1. Continue to follow Cameron's process to assess the technical advantages of a plasma atmospheric process. 2. Do not participate at this point unless Cameron can show a way to improve the economics. <2? . R. A. CHRISTINI ms/3065 cc: ED-D P. R. Bridenbaugh, ATC-D-CEA (w/o attachments) T. V. Pierce, ATC-C-STVA E. L. Sandman, Northwest Alloys J. C. Sever, Northwest Alloys M. D. Ballain, Northwest Alloys ALCOA ATC 0034943 s*.S6j 'y Introduction As pan of our on-going magnesium processes evaluation study, Cameron's plasma Magnetherm process^1) using asbestos for a feedstock was evaluated. Cameron has proposed that his high temperature atmospheric pressure version of Magnetherm using a plasma arc could utilize asbestos as a feedstock. Asbestos or serpentine is a Mg0-Si02 ore containing 45-50% MgO on a dried basis. It is actually slighdy richer than dolime which is about 41% MgO. However, the high concentrations of SiC>2 (45-50%) are a detriment to a normal Magnetherm operation feeding FeSi since slag chemistry control is criucal for good operation. Cameron has never published detailed material balances for his proposed process. It was decided that mass balances would be calculated for various raw material feeds and several slag compositions: (1) normal NWA composition, (2) a mid-level MgO slag composition from Cameron's patents, and (3) a high level MgO slag composition from Cameron's patents. Calculations Simplified raw material compositions were assumed by ignoring low level impurities. Thus serpentine was assumed to be 50% MgO and 50% Si02- Lime and magnesite were assumed to be 100% CaO and 100% MgO, respectively. Mass balances were calculated for four cases: (1) A standard Magnetherm case was done for comparison, (2) serpentine was used to produce a normal NWA slag since this is a process that definitely would operate, (3) and (4) two optimized balances producing Cameron slags were calculated (see Table I). All four were normalized to the same crude Mg production of 30,000 lbs which is the current upper limit for one cycle production at NWA. A refine recovery of 91.7% was assumed for recovery of good Mg (final product) from crude Mg. This is fairly typical for NWA. Standard NWA raw material costs were assumed for calculating Table II. Since serpentine (asbestos) tailings have zero value, it was assumed the Cameron-Fiat plant would have very little cost involved in that raw material. ms/3065 -1- ALCOA PRIVATE INFORMATION ATC 0034944 Raw material costs include calcination costs and transportation costs for magnesite, lime, dolime, and bauxite. Raw material costs: Serpentine Magnesite Lime Dolime Bauxite A1 Ferrosilicon (75%) Mb 0 0.10 0.05 0.05 0.10 0.70 0.40 Energy balances were also calculated based on the mass balances of Table I. -Dolime was assumed to enter the furnaces hot at 1000C. All other raw materials are at room temperature. Both the Magnetherm and AN-6 case (Magnetherm slag using serpentine) were assumed to operate at 1550C. From Cameron's patent, the ACL-4 mid-level MgO slag (13% MgO) operates at 1700C. The ACH-8 high level MgO slag (25% MgO) operates at 1900C. Results Various combinations of raw materials were tried for each final slag composition. Due to the high concentrations of CaO needed in the slag, either dolime or lime must be added to the mass balance in large quantities. Since serpentine brings significant quantities of nonreactive Si02 that hinders the FeSi reaction, large quantities of serpentine cannot be used. Thus, serpentine only amounted to 10-18% of the total oxide feed. The second significant aspect of the mass balances is the slag weight. The slag is 26-40% larger for the three serpentine cases compared to Magnetherm. Since slag volume in the reduction furnace can be a capacity constraint, these large slag volumes could reduce productivity proportionally. Since none of these slags has a sales value, the larger volumes are just more material that needs to be landfilled. Finally, the three new mass balances all require 5 different feed materials to balance the slag composition. NWA has just 3 feed tanks per furnace and feeds 4 materials by preblending two of them. Five materials would either require even more blending (with potential inaccurate feed rates) or more feed tanks which means higher capital. Note that none of the mass balances use bauxite instead of A1 as the AI2O3 source for the slag. Several mass balances were calculated using bauxite but the slag volume increase was substantial and the FeSi requirement was very large. The costs for all the bauxite mass balances were over $ 1.00/lb Mg and were not included here. The costs at the bottom of Table II show that both AN-6 and ACL-4 are comparable to Magnetherm. However, as mentioned above, all three serpentine cases have larger slag ms/3065 -2- ALCOA PRIVATE INFORMATION ATC 0034945 weights and thus lower furnace throughput. This lower productivity will lead to higher costs due to poorer asset utilization and the distribution of fixed costs over fewer pounds. Energy usage was examined for two reasons: (1) the higher slag weights means more raw materials have to be heated to high temperature and (2) both of the Cameron processes run at higher temperatures: ACL^l at 1700C and ACH-8 at 1950C. The energy increases are significant from 2.94 kwh/lb Mg for Magnetherm up to 4.24 kwh/lb Mg for ACH-8. At NWA's current power rate of $0.026/kwh, the incremental power cost is $0.01 to $0.03/lb Mg. The power cost will not be the deciding factor but it is important to consider. Discussion The one arena in which the Cameron-Fiat process is important occurs if the Italian government is willing to subsidize the process to use it to dispose of asbestos tailings or to provide jobs. In a recent conversation, Cameron said the Italian government would provide half the capital to build the plant to provide jobs since the closing of the asbestos mine had hurt that area. Reference 1. A. M. Cameron, "Magnesium Production," U.S. Patent 5,090,996, 1992 February 25. ms/3065 -3- ALCOA PRIVATE INFORMATION ATC 0034946 Table I. Mass Balances Serpentine (Asbestos) Feed Material In - Serpentine - MgO Si02 Magnesite - MgO Lime - CaO Dolime - CaO MgO A1 Maenetherm 10.9 62.8 44.4 6.8 AN-6 NWA $lag 15.4 15.4 30.8 58.5 42.3 9.8 ACL-4 Cameron Slag Mid MgO 15.7 15.7 7.8 75.4 54.6 7.5 ACH-8 Cameron Slag High MpO 6.8 6.8 36.7 -- 56.2 40.7 10.1 Si 13.4 10.9 12.9 10.7 Fe 4.4 _JL2 4.2 4.0 142.7 186.8 193.8 172.0 Slag Out - CaO - MgO - A1203 - Si02 Crude Mg Good Mg 62.8 5.4 13.0 27.1 108.3 58% 5% 12% 25% 89.2 7.7 18.8 152.0 58.7% 5.1% 12.4% 23.9% 83.2 20.4 14.1 AQJ. 158.4 52.5% 12.9% 8.9% 25.7% 56.2 34.2 19.2 27.4 137.0 30.0 ~ _ 27.5 30.0 -- -- 27.5 30.0 -- _ 27.5 30.0 -- 41.0% 25.0% 14.0% 20.0% -- 27.5 Resid Si Fe 1.3 4.4 144.0 1.1 JZJ. 186.8 1.2 4.2 193.8 1.1 _AQ 172.1 ms/3065 -4- ALCOA PRIVATE INFORMATION ATC 0034947 Table II. Costs Raw Material Serpentine' Magnesite lime Dolime Bauxite A1 FeSi Slag Mg Residual Serpentine/Mg Magnesite/Mg Lime/Mg Dolime/Mg Bauxite/Mg Al/Mg FeSi/Mg Total Maenetherm 10.9 108.0 -- 6.8 17.9 AN-6 NWA Slag Composition 30.8 30.8 100.8 -- 9.8 14.6 ACL-4 Cameron Slag Mid MgO 31.4 - 7.8 130.0 -- 7.5 17.1 ACH-8 Cameron Slab High MsO 13.7 36.7 -- 96.8 -- 10.1 14.6 108.0 27.5 5.7 152.0 27.5 4.8 158.0 27.5 5.5 137.0 27.5 5.1 lb/lb -- .40 -- 3.93 -- .25 .65 $/lb -- .04 -- .20 -- .18 .26 lb/lb 1.12 1.12 3.67 -- .36 .53 S/lb 0 - .05 .18 -- .25 .21 lb/lb 1.14 - .28 4.73 - .27 .62 $/lb 0 - .01 .24 .19 .25 Ib/lb .50 1.33 -- 3.52 -- .37 .53 S/lb 0 .13 - .18 -- .26 .21 .68 .69 .69 .78 ms/3065 -5- ALCOA PRIVATE INFORMATION ATC 0034948 Table III. Energy Balances - Serpentine MgO SiCb Magnesite Lime Dolime CaO MgO A1 Si Fe Masnetherm lbs kwh -- -- -- -- 10.9 5.29 -- 62.8 15.64 44.4 14.96 6.8 1.82 13.4 5.45 4.4 1.66 AN-6 lbs kwh 15.4 7.48 15.4 3.62 -- 30.8 11.15 58.5 14.57 42.3 14.26 9.8 2.62 10.9 4.44 3.7 .62 ACL-4 M kwh 15.7 8.37 15.7 4.05 ---- 7.8 3.09 75.4 20.59 54.6 20.18 7.5 2.19 12.9 5.76 4.2 .77 ACH-8 lbs kwh 6.8 4.15 6.8 2.01 36.7 22.44 ---- 56.2 17.60 40.7 17.26 10.1 3.40 10.7 5.47 4.0 .84 A1 reaction Si reaction (.9) 8.17 27.95 80.93 11.76 22.73 93.25 9.87 29.50 104.37 15.25 28.07 116.49 Good Mg 27.5 27.5 27.5 27.5 kwh/lb Mg 2.94 3.39 3.79 4.24 ms/3065 -6- ALCOA PRIVATE INFORMATION ATC 0034949 TITLE. MR^rESiHf^-BIl, JR. nv Aluminum Company of America,-*--------(1--1-9-)- - ALUMINUM RESEARCH LABORATORIES R^K^Ort/cE Library PROGRESS REPORT SPECIFIC METHODS OF REDUCING THE RISER VOLUME OF ALUMINUM AND MAGNESIUM CASTINGS OPERATOR- E. E. Stonebrook CL--291 .PROBLEM NO- -DATE 6-9-49 1803-449 SEVENTH REPORT: RISER REDUCTION ON NADLER FOUNDRY AND MACHINE CO. CASTING NO. M-6-A BY ASBESTOS--SILOCSL TYPE INSULATION 'J0[ -i ' 7 '% SYNOPSIS Previous riser reduction tests2,3.6) bad estab lished the effectiveness of Hi-Temp, an asbestos-silocel type material, when used as insulation around blind risers. This study has now been extended to include riser reduction tests on another production casting with Hi--Temp insulation. It was found that the weight of the two risers used on Nadler Foundry and Machine Co., part No. M-6-A, could be reduced from 10.0 to 4.2 pounds, a reduction of without any sacrifice in casting quality. A cost analysis, however, indicated that it would not be economical to use this method of riser reduction on a casting having such relatively small risers, but that insulation of considerably larger risers(3) should offer some cost advantage. Another application for such riser insulation might be for the production of castings where the problem is more one of providing adequate feeding rather than riser reduction. REFERENCES (1) Progress Report by V. E. Sicha and E. E. Stonebrook, CL-291, 9-12-45# "First Report, "Riser Reduction by Heating or Insulation". (2) Progress Report by E. E. Stonebrook and V. E. Sicha, CL-291# 3-15-46# Second Report, "Riser Reduction of Wedge Castings by Heating or Insulation". (3) Progress Report by E. E. Stonebrook and W. E. Sicha, CL-291, 5-13-46, Third Report, "Riser Reduction by Insulation as Applied to A132 Alloy Production Castings". 3427 1014--10-46 Printed in U. 8. A. ATC 0034950 Aluminum Company of America (120) ALUMINUM RESEARCH LABORATORIES PROGRESS REPORT SPECIFIC METHODS OF REDUCING THE RISER VOLUME OF ALUMINUM TITLE- AND MAGNESIUM CASTINGS OPERATOR- E. E. Stonebrook CL-291 -PROBLEM NO- 6-9-49 -DATE- (4) Progress Report by E. E. Stonebrook and W. E. Sicha, CL-291, 7-29-46# Fourth Report, "Riser Reduction Methods Applied to Large 195 Alloy Step Castings". (5) Progress Report by E. E. Stonebrook and W. E. Sicha, CL-291, 3-29-46, Fifth Report, "Breakdown Tests as a Gauge of the Quality of 195-T4 Alloy Castings Produced with Insulated Risers". (6) Progress Report by E. E. Stonebrook and W. E Sicha, CL--291, 3-7-47* Sixth Report, "Riser Reduction on Wedge Castings by Heating or Insulation". (7) Notebook #5243, pages 1, 2 and 3. OBJECT It was desired to make riser reduction tests on a production casting using asbestos-silocel type insulators. PROCEDURE AND RESULTS Pattern equipment for two castings was secured from the Cleveland Sand Foundry for riser reduction tests. Both were Nadler parts. Part No. 15-C-2 is a cover plate casting. It was deemed unsuitable, however, for riser re duction tests using asbestos-silocel type insulated risers because of the irregular shape of certain portions of the casting adjoining the risers. These relatively thick-walled insulators are not readily adaptable to use around top risers unless the casting surfaces upon which the Insulators rest are flat. Furthermore, it is often undesirable to have the bottom of the insulators in contact with a casting because the rate of solidification of the portions of the casting thus contacted will be reduced considerably. Three views of the other Nadler casting, desig nated as No. M-6--A, are shown in Figure 31* Form risers are located on two sides of the casting and it is poured two-up as indicated in the photographs. The small letter U17 I0U--10-4& Printed in U. S. A. ATC 0034951 TITLE- Aluminum Company of America ALUMINUM RESEARCH LABORATORIES PROGRESS REPORT SPECIFIC METHODS OF REDUCING THE RISER VOLUME OF ALUMINUM AND MAGNESIUM CASTINGS (121) OPERATOR- E. E. Stonebrook CL--291 -PROBLEM NO- 6-9-49 -DATE markings on the Nadler casting photographs indicate the type and location of chills originally used in producing this part. It was found later that the casting could be produced without any chills, and their use was discontinued. One of these cast ings was poured at 1340*F with 356 alloy from heat C5007 fol lowing the practices used by the Cleveland Sand Foundry. The different parts of the casting had the following weights: Gating Riser A Riser B Trimmed casting 5.5 5.7 4.3 -j&Z lbs n a a Poured casting a Radiographs showed that the casting contained general fine to medium gas porosity and areas of small to medium mot tling. A few small gas hole3 were noted in one area, but the casting was considered to be of acceptable quality. A blind riser form was made for the top side riser (B) by molding loose asbestos-silocel cement (Hi-Temp) and baking to drive off the added water. The approximate size of the original open riser was 2-7/8 in. x 2-1/8 in. at the base and 3-7/8 in. x 2-1/2 in. at the top with a total height of 4-1/2 in. above the riser pad. The blind riser had the same dimensions at the base but was only 3 in. high with suitable draft. A squeezer was used to ram the asbestos-silocel cement in an aluminum tube. Some difficulty was experienced in re moving the formed insulator from the tube ana it was necessary to use the squeezer in conjunction with a ram in ejecting the insulator. A rectangular-shaped riser form was used for riser A and the resulting riser cavity was lined on four sides with 1 in. thick blocks of Hi-Temp. The top of the riser was left open. These insulated risers are shown in the A sketches of Figure 32. Another casting was poured from heat 05078 using the insulators Just described. The weights of the different parts of the casting were as follows: 3427 10U--10-43 Printed in O. & A. ATC 0034952 Aluminum Company of America (122) ALUMINUM RESEARCH LABORATORIES PROGRESS REPORT SPECIFIC METHODS OF REDUCING THE RISER VOLUME OF ALUMINUM AND MAGNESIUM CASTINGS TITLE- OPERATOR- E. E. Stonebrook CL-291 -PROBLEM NO- .DATE 6-9-49 Gating Riser A Riser B Trimmed casting Poured casting 5.5 2.1 lbs 2.4 16.9 n w 6.9 0 Radiographs shoved a localised area of small to medium shrinkage porosity on one side of the bottom vail ven der riser A, and the casting vas considered to be of unsatis factory quality. The insulated B riser used in this test appeared to adequately feed the one side of the casting and was retained in subsequent tests. The next test (heat C5121) involved the use of two insulated blind risers located at the parting line of the mold in place of the single A riser used in the previous test. Each of the two insulators used for the A risers vas the same size as the insulator used for riser B. Figure 32B shows sketches of the locations of the three risers used. Weights of the different parts of the casting from heat C5121 were as follows: Gating Two A risers Riser B Trimmed casting Poured casting 5.6 lbs 4.4 2.4 17.1 m it a 29.5 n Radiographs shoved that this casting vas of satis factory quality but the weight saving was only 3.1 lbs. which would not begin to offset the cost or Insulating the risers. Accordingly, one additional casting was poured from heat C5241 using only one insulated A riser at the parting line instead of the two used in the previous test. Radio graphs indicated that the part was of acceptable quality. The weights obtained with this practice were as follows: 3427 10M--10-4* PrmUKlinU. S. A. ATC 0034953 TITLE, Aluminum Company of America ALUMINUM RESEARCH LABORATORIES PROGRESS REPORT SPECIFIC METHODS OF REDUCING THE RISER VOLUME OF ALUMINUM AND MAGNESIUM CASTINGS (123) OPERATOR- E. E. Stonebrook .PROBLEM NO- CL--291 -DATE. 6-9-49 Gating Riser A Riser B Trimmed casting Poured casting 4.9 lbs 2.0 tv 2.2 It 17.2 R 26.3 If This is a total weight saving of 6.3 lbs. compared with the casting produced with standard, uninsulated risers. It was thought that this casting represented nearly the maximum riser reduction possible with insulation, and no further tests were made. DISCUSSION If the cost of melting and pouring one pound of aluminum is 3 cents, then the total saving resulting from the riser reduction achieved with the casting poured from heat C5241 would be about 19 cents. There would also be some saving in trimming cost since the large side riser was entirely eliminated from the experimental casting. These savings should be balanced against cost of insulators and a possible increase in molding time resulting from placing the insulators and favoring them during the ramming opera tion. Although accurate values for insulator costs could not be obtained from the small scale method of fabrication employed, it is doubtful whether the saving with this cast ing would justify the cost even if the insulators could be used several times. , .In the case of castings having relatively large risers, U1 riser insulation should offer some cost advantage, particularly if the insulators can be salvaged and reused several times. There are also jobs where only a few cast ings are to be produced and the cost of scrapping one or more castings would far outweigh the expense of providing insulators. If added assurance is desired that the castings will be adequately fed, then insulated risers would offer one solution to the problem. J2T tOM--10-48 Printed ia U. S. A. ATC 0034954 TITLE- Aluminum Company of America ALUMINUM RESEARCH LABORATORIES PROGRESS REPORT SPECIFIC METHODS OF REDUCING THE RISER VOLUME OF ALUMINUM AND MAGNESIUM CASTINGS (124) OPERATOR- E. E. Stonebrook CL--291 -PROBLEM NO.. 6-9-49 -DATE- CONCLUSIONS (1) Use of asbestos-silocel type Insulated risers permitted the poured weight of Nadler casting No. M-6-A to be re duced from 32.6 to 26.3 lbs. without any sacrifice in casting quality. (2) The cost of using insulated risers on this particular casting would be greater than the savings effected by the 6.3--lb. reduction in poured weight. (3) From an economic standpoint, use of asbestos-silocel insulators should be restricted to castings having relatively large, bulky risers or to short runs of rela tively expensive castings for which satisfactory gating practices have not yet been developed. cc Hr. M. V. Daugherty, ARL, Cleveland Mr. E. H. Dix, Jr., ARL, New Kensington Mr. W. E. Sicha, ARL, Cleveland Mr. H. J. Rowe, Pittsburgh Mr. E. V. Blackmun, Cleveland 3427 10U--10-43 Printed in U. S. A. ATC 0034955 Three W hioh V ie Was ws Us o ed f f oN ra dl R er is e C r a s tin Red g uc No. tio n M -6-A T e s ts w ith Hi-Tem p In s u la tio n . ATC 0034956 tc .annoid to (12*) & ATC 0034957 (126) Figure 32 Heat C5121 Sketches showing types of insulated risers used in producing two Nadler M-6-A castings. ATC 0034958 R. A. MARRA CHEMICAL SYSTEMS DIVISION ALCOA TECHNICAL CENTER-E T0 MEMORANDUM 0053420 S9V6081 1992 March 12 RE: LETTERPJEPORT 6.92.28; CHARACTERIZATION OF PABCO METALTEMP Background Calcium silicate board materials are used as both headers and feedboards at Warrick Operations for the Horizontal Direct Chill (HDC) casting of aluminum ingots. The HDC process was originally designed to use Header Grade Molten Metal Marinite (HGMMM) for the header and feedboard. HGMMM is an asbestos fiber-reinforced calcium silicate board material. Due to the health hazards associated with this asbestos containing material, replacement materials for both the header and feedboard are being evaluated. An acceptable header replacement should have the characteristics shown in Table L The feedboard should have similar characteristics, although the resistance in molten metal attack is less important Since a large material area of the feedboard is exposed to molten metal contact, the moisture and carbonate content is extremely important Table 1 Desired Characteristics for HDS Header Material Characteristic Importance Molten Metal Resistance. (Low Wettability) Low Thermal Conductivity High Compressive Strength Low Moisture/Carbonate Content High Abrasion Resistance Low Porosity/Permeability Adequate Strength (MOR) Low Thermal Expansion (Good Thermal Stability) - Prevent metal/header reactions which degrade die header surface and result in ingot defects. - Prevent molten metal solidification at the header surface. - Prevent crushing ofheader in the mold package. - Prevent risk ofexplosion when contacted with molten metal and avoid outgassing which degrades ingot quality - Prevent degradation of header surface during casting. - Minimize lubricant absorption. - Allow machining and handling of header. - Minimize change in header/oil ring overhang and stress build-up in mold package. PABCO Metaltemp and BNZ Materials Mctalform calcium silicate boards have been evaluated for feedboard and header replacement These materials are boards produced by pressing slurries of lime (CaO), silica (SiOr, typically added as a colloidal material), wollastonite (CaSi03>, and alkali resistant glass fibers. The pressed boards are autoclaved ALCOA ATC 0034959 S*.4s3 |BV. MEMORANDUM 1992 March 12 Page 3 marinite headers for casting 5xxx series ingot There have been severe quality problems with the BNZ material in terms of outgassing (bubbling) of feedboards. This not only results in poor ingot quality but is a serious safety issue during casting. Feedboard breakage during or after machining has also been a major problem. The qualification of the Pabco Metaltemp material for feedboards (for 5xxx and 3xxx series alloys) and headers (for 3xxx series alloys) should help eliminate these quality issues as well as having a favorable cost impact since material is currendy available in Warrick's inventory. Characterization Techniques DENSITY: The density was calculated from dimensional and weight measurements on l"xl"x6" bars machined from the boards. The bars were dried at 230F for a minimum of four hours prior to measurement HARDNESS: The material hardness was measured using a Durometer conforming to the ASTM 2240-75 test method. The hardness was measured on both the die side and the screen side of the pressed boards. MOLTEN METAL RESISTANCE: The molten metal resistance was evaluated by both metal immersion testing and direct examination of used header surfaces. In the immersion testing, 3"x3"xl" specimens machined from the boards were dried at 230F and submerged 1.5 inches in a 5182 aluminum alloy at 1400F for 3-10 minutes. The changes in specimen thickness and weight were measured immediately after removal from the molten metal and after the specimens had cooled for 15 minutes. COMPRESSIVE (COLD CRUSH! STRENGTH: The compression strength was measured by loading r'xl"x6'' specimens perpendicular to the pressing direction. The testing was performed at Westmoreland Testing Laboratories. LUBRICANT ABSORPTION: The material's propensity for absorption of the lubricant was determined by hot oil immersion tests. In the hot oil immersion test, 2"x2"xl" specimens were immersed in 400F oil (the same composition used during casting). The weight change was measured after 0.5,1, and 3 hour exposure times. MODULUS OF RUPTURE: The modulus of rupture was measured on l"xl"x6" specimens. The testing was performed at Westmoreland Testing Laboratories. MICROSTRUCTURAL ANALYSIS: Scanning electron microscopy (SEM) was used to examine the microstructural features. Elemental dispersive X-ray analysis (EDAX) was further used to characterize the materials microstructural and chemical make-up. ATC 0034961 MEMORANDUM 1992 March 12 Page 5 discussions, Willie Lansdale has stated that the more recent Pabco boards do not appear to contain these white spots. A memorandum from R. A. Marra (1991 May 03) discussed the microstructural and chemical analysis of Pabco Metaltemp header surface used to cast 3082 alloy at Warrick Operations. The analysis showed that magnesium and to a lesser extent aluminum attacks the calcium silicate bond phase at the header surface. It has also been shown the additions of minerals containing FeO (such as red mud and verariculite) are beneficial in preventing this attack. Therefore, it is recommended that the metaltemp boards be used as header material only for the casting of low magnesium containing alloys and the modified composition containing red mud and vermiculite be used for the higher magnesium containing alloys. Conclusions and Recommendations The Pabco Metaltemp board materials have physical and chemical characteristics similar to the header-grade molten metal marinite (with the exception of containing no asbestos). In comparison with the BNZ Metalform material, the Pabco material's cold crush strength and modulus of rapture are significantly higher. The higher density Pabco boards also showed much less oil absorption than the BNZ material. Pabco's material should be considered for use as feedboards for HDC casting of 3xxx and 5xxx series alloys and for headers for casting the lower magnesium containing alloys. The modified mold clamp design should be used to optimize performance. The higher density boards (50 pcf) should perform better than the lower density materials because of their higher strength (more easily machined and handling), higher crashing strength (less likely to fail in the mold package), and lower propensity for oil absorption. The 600F heat treatment appears adequate for these applications. In fact, the higher heat treatment (1100F) may result in increased porosity and oil absorption and a decrease in the material strength. The material should be monitored for density, strength, hardness, moisture and carbonate content (TGA/Mass Spec), and hot oil absorption on a regular basis. The microstructure should also be examined to ensure that large unreacted CaO is not present in the board and especially along the machined header surface. It appears Pabco has done a much betterjob than BNZ in controlling their lime supply and screening the raw material to remove any large particulates. ATC 0034963 Table II Pabco Meteltemp Board Characteristics Material HG/MMM BNZ Metalform 104-6-2 104-8-2 112-16-1 112-16-1 AA HT 1100F 112-16-1 XG Dry 230FF HT 1100F Test Density (pcf) Hardness Molten Metal Immersion Hot Oil Immersion Compressive Strength Modulus of Rupture Units pcf 45.811.4 Durometer 69.613.2 % weight gain % dimensional change % weight gain 0.5 hr 20.3 1 hr 31.3 3 hr 44.8 psi 19601160 psi 15201160 48.1 64 34.7 41.6 57.5 1450 670 51.410.3 71.410.9 1.4910.04 0.410.2 52.610.3 61.011.3 NM NM 49.811.4 70.710.7 0.3 0.2310.21 25.512.6 24.910.9 30.414.1 ' 30.411.5 35.911.1 37.013.8 22101340 25601250 10801460 12801260 23 26.2 32.9 21001160 9501140 48.710.5 68.910.7 0.33 0.2310.21 34.6 37.5 48.9 1810160 9701210 51.90.2 71.80.9 NM NM NM NM NM 19201140 9901130 ATC 0034965 Table II Pabco Meteltemp Board Characteristics vTO 8^ ccnn Hac VO cn C4 04 04 04 04 & o 25 cn 2 04 ^ a uo * 8 EX Ov 04 a C O I ^- 2 ? +5 ^ 2 s o t-~ r~ -- n tt in I 1 fO"' ooov Or*' ^o oS 00 . ^ o o 5 51 5 2 ^O' mv> J OH n-o Cncn4 \p 2 oQ 44.4444 44 44 T1 00 -- 04 o \0 oo cn cn \o h O' 2 'O P <N - S3$2 <=> n & i--/\ oirot " c. n 04 00 O; S O n ri p - 2 n44 o4o4 n44 644 gT] 'r oo o 9 cn cn cn ~ ov O " n 2 HS 04. --. O^' t** -- 04 Tt ? cn $ -- si 04 3 o $ o ?3 8 20--\ 20044 vO SSss ^.^22 oo r** tt VO o s 2 s 2 s 2 5o1 Ov 5 04 m r- m Ov 04 S s? VO 00 O 2 VO o ocn ocn 00 cn cn vO TT cn cn wn 44 OOO 44 Ocn VO cn oo 44 55 0o^0or r So* oo ci -- ^ -- cri cn ^ in ^ p"i ro o\ o oo eJ = in `S o a Soci 6-mc0 _sg 6.we0 .*b u fe jr -- --- g 60.o eo n ^ g. 55 5 "5 'C c> ^ | S > ts .*oi ^ ob O 1 Co Co .CO s O Osv I23 2 O o 2 ba4E> o Xo e oE E 0o>0 ,035 '55 O 3S s.-i 6i 12 ATC 0034966 Table II Pabco Meteltemp Board Characteristics Material 123-16-1 XQ1 Dry 230FF HT 1100F 124-8-2 124-8-2 AA HT 1100F 124-8-2 XQ4 Dry 230FF HT 1100F 124-14-1 124-18-1 Test Density (pcf) Hardness Molten Metal Immersion Hot Oil Immersion Compressive Strength Modulus of Rupture Units pcf Durometer % weight gain % dimensional change % weight gain 0.5 hr 1 hr 3 hr psi psi o0 Z Z Z S.g *2 2 2 44.310.5 59.610.9 1.710.0 0.1710.08 53.3 54.4 78.2 12501120 5701170 43.410.2 56.710.4 0.57 0 59.8 66 72.5 910140 400190 43.810.1 62.010.6 NM NM NM NM NM 1150150 670150 44.610.5 61.312.0 1.310.1 0.3210.08 39.50.9 62.411.2 1.310.1 0.3210.19 58.2 79.5 84.1 1241110 7901270 70 83.4 85.7 9501100 4401120 ATC 0034967 GA 31646.' Microstructure of a Typical Pabco Metaltemp Board Figure 1 ATC 0034968 GA r*4*2 Pabco Metaltemp Board Showing Unreacted CaO Particle. (Top Photo: SEM Image, Middle: Ca Map, Bottom: Si Map) Figure 2 ATC 0034969 ALUMINUM COMPANY OF AMERICA ALCOA LABORATORIES T /> / ELIMINATE ASBESTOS IN ALL FURNACE APPLICATIONS R. G. LaBAR March 13, 1974 "CONFIDENTIAL" INGOT CASTING DIVISION REPORT NO. 11-74-12B103 Copy No. Reported by R. G. LaBar Group Leader Noted by L G. K. Turnbull Section Head ATC 0034970 .9a/ ALUMINUM COMPANY OF AMERICA ALCOA LABORATORIES t/> ///^ ELIMINATE ASBESTOS IN ALL FURNACE APPLICATIONS R. G. LaBAR March 13, 197*4 "CONFIDENTIAL" INGOT CASTING DIVISION REPORT NO. ll-7*4-12B103 Copy No. Reported by R. G. LaBar Group Leader Noted by G. K. Turnbull Section Head ATC 0034971 -i- SYNOPSIS This report summarizes the program to replace asbestos containing products in applications other than HDC and FDC headers. A broad number of candidates (Table I) and test data (Tables II-VI) were narrowed to a specific group of recommended materials. Appendix A contains the recommendations along with updated procedures for installation and use. Whereas the project is closed out, assistance and advice will be provided as required. Further, candidate materials for insulation, trough linings and packing, basin and FDC pot linings, etc. will be evaluated in Project 06W053H} Refractory Evaluation. ATC 0034972 -ii- TABLE OF CONTENTS SYNOPSIS .................................................................................... INTRODUCTION .......................................................................... CANDIDATE MATERIALS ...................................................... EVALUATION AND TEST PROCEDURE .............................. RESULTS ............................................. .... .................................. RECOMMENDATIONS ................................................................ CONCLUSIONS .......................................................................... PAGE NO. i 1 1 2 2 4 4 LIST OF TABLES TABLE I.................................................................................... Major Candidate Materials by Category TABLE II.................................................................................... Laboratory Tests on Castables TABLE III......................... ...................................................... Laboratory Tests on Moldables TABLE IV.................................................................................... Laboratory Tests on Board Products TABLE V.................................................................................... In Service Evaluation Results TABLE VI.................................................................................... Physical.and Thermal Properties 6 9 10 11 12 13 LIST OF FIGURES FIGURE I.................................................................................... Linear Firing Shrinkage Insulating Moldables 14 APPENDIX APPENDIX A............................................................................... Recommendations for Preparation and Installation 15 ATC 0034973 INTRODUCTION Overwhelming evidence in recent years links human exposure to asbestos dust with several forms of cancer. One result of this is the U. S. Asbestos Standard* which greatly restricts the exposure limit (no. particles/volume air/ unit time) effective July 7, 1972 with a further reduction effec tive July 1, 1976. The restrictions are so severe, that the dust handling equipment, monitoring equipment and additional personnel required to monitor concentrations will make the use of asbestos containing materials impractical as well as uneconomical. The major thrust towards solving the problem is the elimination of asbestos containing products in all applications. However, the unique properties of asbestos and many asbestos contain ing products are not readily matched by any one substitute for all applications. In fact, many different materials had to be qualified and specified for use in various applications; hence, the purpose of the project and this report. Preliminary to this report was a recommendation issued by D. R. Barch and the author in April 1973. The appendix of this report is an update of the original recommendations based on laboratory and in-service results. CANDIDATE MATERIALS The number of replacement materials is great, as is the variety of trade names and manufacturer's claims for them. To minimize the confusion. Table I lists the materials according to the following categories: 1. Castables - supplied in bags or drums to be mixed with water and poured, tamped or troweled in place. 2. Moldables - supplied in bags or drums to be used as is (if premixed) or mixed with water and tamped, troweled and/or molded by hand. 3. Board - in most cases to be cut, installed and used as is. However, some products require coatings to improve erosion and corrosion resistance when used in metal contact. il. Miscellaneous - includes packing materials, gasketing, cements, etc. 5. Preformed Special Shapes - includes drop-in trough liners and other special shapes (FDC pots, etc.). As a list of major candidates. Table I is neither a list of recommendations, nor an all inclusive compilation of alterna tives. Recommendations are detailed later in Appendix A. Materials intentionally excluded from Table I include inadequate candidates plus other materials which were evaluated and for one *Federal Register, Vol. 37, No. 110, pp. 11320-11322 (6/7/72). ATC 0034974 -2- reason or another failed our laboratory tests, but will appear elsewhere in this report for comparisons. EVALUATION AND TEST PROCEDURE The methods for testing asbestos replacement materials vary with the Intended use. The standard cup and immersion tests were used to evaluate those materials which could be prepared in the required sample configurations and which would be used in metal contact in transfer systems. Because the severity of metal corrosion at 1500 and 1400 F is significantly greater than that at 1350 F and lower, many materials (for applications downstream from the holder) were tested in contact with metal at 1350 F. In several cases, materials were evaluated in service and found to be suitable prior to any laboratory tests. Laboratory, in-service evaluations were conducted on nearly all materials in various forms to determine those performance characteristics not attainable in standard test procedures, e.g. thermal stress cracking, erosion resistance, handling characteristics, etc. RESULTS Table II contains the results of standard tests on a number of castables. Of particular interest is the rather high silicon pickup from most of the fused-silica castables and yet their actual performance, as noted in Table V, is quite satisfactory. Conversely, the low silicon pickup from Alcoa C-l is somewhat counteracted by its tendency to crack from differential thermal stresses. Tables III and IV contain the results of tests run on moldables and board products. The comments under thermal stability refer to degradation due to any decomposition of the material during the test or subsequently to its exposure in air and normal levels of humidity which can cause rehydration of some cementitious phases, i.e. dicalcium silicate, calcium sulfate, etc. The conflicting results between cup tests and in-service evalua tions can be attributed to lower metal temperatures and, in most cases - lfess corrosive alloys than 7075- The low temperature tests produce better correlation with inservice evaluations. Further, the use of coatings, e.g. Lumnite, Glasrock Sealer, J-M Ceracote, Zirconite, whiting, etc. on moldables and board products upgrades their ability to resist erosion and wetting (metal adherency) so that their in-service performance is significantly different from that predicted from cup and immersion tests. ATC 0034975 -3- Results of plant trials at Warrick using Thermo 340 (R--30) and a developmental mix designated R-616 (contains equal amounts of fused-silica and plaster with somewhat less calcium aluminate, CA-25) were less than satisfactory. The materials were not sufficiently durable to survive beyond 2 1/2 mos. (Thermo 340) and 5 mos. (R-6l6) respectively. These results correlate with our observations in the laboratory. We have found that with materials containing considerable amounts of plaster (CaSOjj) partial rehydration occurs during cyclical use. This phenomenon produces swelling and dusting which weakens the material. The A-641 product was developed with this in mind so that it only contains. 1JE plaster which is added to accelerate the hydraulic set of the CA-25 cement so that higher productivity of cast pieces is possible (1? plaster addition halves the time required to develop handling strength). One unique feature of the R-340 material is that it expands during curing more than it shrinks during firing. ' We have utilized this property to successfully line pump piping and FDC pot shells. The linings are tight, quite durable and easy to install as the R-340 mix is fine-grained and thus flows easily into small annular spaces. As mentioned above, we have developed a new product called A-641. This material, in a preformed and fired shape (board or otherwise) can be cut to the required size with standard metal cutting tools and bolted in place. However, because the material is brittle, a back-up cushion of ceramic fiber blanket or ceramic fiber paper is recommended. A companion product now under evaluation is A-646. This is a slight modification incorporating a corrosion inhibiting zinc borosilicate frit which all but eliminates wetting and metal attack (see data in Table IV). Another new product development is A-647. The concept of this material was derived from J. P. Austin's Mandoseal (See Table III) which contains Portland Cement (unstable in cyclical use). The objective was a moldable mix using expanded vermiculite blended with various other materials to obtain a product which could be wet mixed more easily than mineral wool containing moldables, would have low drying and firing shrinkage and could be used as a substitute for asbestos shorts/Lumnite. The need for a suitable moldable in many applications is critical. One of the major problems with their use is cracking due to shrinkage. In order to quantify shrinkage versus temperature and product, a series of tests was conducted. Samples of each material were formed into bars 1 in. x 1 in. x 10 in. and then measured after curing, air drying, oven drying to 230 F, firing to 600 F and then to 1350 F. The data are shown in Figure 1 and Table VI. Of particular interest is the extremely low firing shrinkage of Consolidated's Moldable,Panelex Super (still being evaluated in the laboratory). Variform and A-647. The JM-375 L and asbestos/shorts Lumnite seem to increase their rate of shrinkage ATC 0034976 -4- with temperature which is attributed to irreversible phase changes in their fibrous components. The Panelex Super is a new product containing some mica (probably vermiculite) which would tend to counteract shrinkage of the mineral wool constituent. As noted in Table VI, the total shrinkage of the Consolidated Moldable is 0.0?. However, as seen in Figure 1, shrinkage occurred during air and oven drying to 230 F. Reports from the field in dicate that this product does not shrink at all. This is under standable in view of the fact that in our test, the bars were allowed to air and oven dry lying on one side. Thus the bottom did not dry at the same rate as the top surface, causing the bars to warp (compared to the other materials, this product dries slowly). Finally, a comment is warranted about moisture content. As reported for JM-375 L, its shrinkage was only 1.4?. However, the mix was prepared using 50? water. When 75? water was used, the shrinkage nearly tripled. This type of sensivity led us to embark on the development of A-647 and simultaneously begin to search for other products such as Plisulate and Panelex Super. The prices in Table VI are not precisely accurate but were based on recent data to provide a comparative base. 'The figures for A-647 and A-641 were estimated from available data on raw materials and trial manufacturing (costs to us by outside vendors for preparing experimental materials). The range of prices for fused-silica castables is the difference between the least expen sive (Glasrock 840) and the most expensive (Silfrax ARC). RECOMMENDATIONS In keeping with our overall effort, previous recommendations have been amended to reflect the results of our work at the Laboratories and to include the results of various and suc cessful plant trials. These recommendations appear in Appendix A. While the reader may at first be somewhat confused by the diversity of materials, it is recognized that local availability and delivery will at times necessitate the selection and use of alternate products within a specific category. Recommendations on the use and source of supply for A-641 and A-647 will be forthcoming as more extensive in-service trials are completed and/or when agreement is reached with a vendor(s) for their manufacture. CONCLUSIONS 1. Molten Metal Marinite and asbestos insulation can be replaced by asbestos free materials in troughs, basins and for back-up insulation. 2. Asbestos shorts/Lumnite and asbestos rope can be replaced by ATC 0034977 -5asbestos free materials for troughs, pot linings and miscellaneous packing. 3. The recommended substitute materials are described herein. They are numerous and, in many cases, suited for a limited number of applications dependent on thermal conductivity, erosion resistance, corrosion resistance, bulk density (as this pertains to inclusions and metal quality), workability and cost. lj. In many instances, asbestos free substitutes for Molten Metal Marinite are less expensive, more refractory but in no instance as uniquely versatile or flexible. 5. More care in handling and use will be required when using other materials where Molten Metal Marinite has been the standard. ATC 0034978 -6- MAJOR C A N D ID A TE M A TE R IA LS BY CATEGORY u 03 U PG O O G ~ o * O G= Z M GM M CO rH a; o <H; co o tv s -rH o o o3 * o *H Pi O B <D = o o = s 3 -P o O o G <U 0) -p o H -P Xo 3 at >9 03 tc tt> tH 03 0) aS (h O O co G M co 4) P N o oo G 03 W oo G M rH g X 3 u 03 Oo co G H u o G iH o o co <0 <c rH o p rH H V g 'G a> > o 9p G Pi rH oP G o 'G G ,Q 03 O 03 Cm G G a3 2 -AS o GS 3 g 03 Pi O PQ Vi 4) i o gO o5 ~PS u CO = z o fi ,Q p^ z pq Pq a) O rH u o 03 Pi G rH rH o3 J3 tH rOH ACO >9 u o O H CO < < 3= pq 3 'G G G 2 i >9 0> 3 q rH rH to H o o hH G <c C x> CO O Pq 2 *H Ta o G 03 G O O rH O? O O c * a; rH O cn a5 pt Tj r--"* i PQ rH pq O O CO *G s O cn o G oo -=t* CM co co O CM rH PEH g p H 3 4) 3 Oh G< a> p fi*G G o3 Pi PQ AS O AS O AS O O < X rH *rl CO 1--l H CO rH O Pi co O Pi a3 Pi i--t G CO H Cw o g h O *H CO H a3 a3 CO rH 0) <D CO rH rH G H G G *H O O HH CQ tH rH Pi 1 1P o CO G o a3 1 aj O p OX CO o pq 1 aj rH l"D o X < CO hH CO H X ITS 4) g G G G H o rH o co C*- rH Cm O u m 03 tH CO p3* G i G u G VO rH 3 aj O 1 o *3 pq > u c CO pg G G rH c5 OH aa) oG tIo 0) CO >: Pi 0)o_, CO taC or rH p oG s p to Oa3 03 O co gc CbO <U rl +3 P aj raHJ *CH Gg co 3 G rH M a3 0O3 03 G O G -rH co o rH p GG G G iH CO g Gp QG CM cn pG Gc H O rH G O 03 P H P Pq >> Pi o co a? rH 8 r-H O Z o rH G o o G P rH G GG Pi H 03 G g H H rH 2G # *H G P ' aC3 to o H a3 *d p05 a3 rH ^. G ca3 o a O -P rH aj 'O G a3 G p* g 0) -P iH GO iH <HQ 4) P rH rQ <M Cm tH G G rH P VO tH H c3 3e OG G x: g 3 s <> 3 -H H tH gO pq to q. G rH C rH o G G *H G G x: O O CO > O pq # CM cn * -S' m u\ ATC 0034979 7- CE R e fr a c to r ie s , In c Babcock & W ilcox Foseco, In c. 03 aS r-4 CO -p to u o o G a 2 0) o M CO G H G X E O CO o 0) OJ G 0) o a> rl u o oa 03 rH aS Xo o a> O U Ph It G --1 -P G o -p rH rH rH CO 3 V0) u t--1 G G 05 H *rl H G>G *H O G g -P 3 G aS a Sh M S-. C0 0) O aS -P G rH gG 3> co g aS 3 rH P c u G CO a G 8 p CQ aS G 08 G G C3 o aS o 3 G as c .X eg S 3 cj Cm G a X oS M ii CO 05 05 >> G <1) 3 H -X o a) o o pg o o G rH o H p aS CO CO O Og H 3 ,X z 3 CO o oi o CO o GD oC U 2 0 23 0) jz Sh u & c aS O G ) u ,P p& X o > o OS go G O w aS as o o PQ 3 s G O rH C <3 aS o aS rl O AG OE x <D ^ CQ UO <D o O-VO G rH CO P O Ph < CU O u0) Pi X) c a! l< 3 1 2 0m O E-* -P Pc* OJ CO o GG rH O rl rH X 5C O 0) 6 0 * >> 3<d U H -p aS oOo E m CQ ,X CO O G O o u G Jh rH <3J 0) 0) >> Ph U G3 CO O a) -p l o G Jh aS CQ GaS VD Ks > O < O -rf CO o E E <D G 3X 43 X3 G aS u 1--i i--l u a) <M o a> Cm CO a> O > aH] Uu 0) ) G Ocd a> H G -Eh w o Eh Eh Therm otect A M - Board K alm in 5000 Ceramic fib e r , <D pi aS o M GCO e5 CO GCaSO yj o rH aS Oo CO oCO G G3 eSh^} aS <D u>3 oS' -<pu G0gc) G Ei <om0) G aS o GaJ<o-et --c0<ou)- r--s 0) -p G aS G a) E o cm g aS GG g G3 aS GG GGo O a V G^ ^G3 G<11 Po .caXs p G oaS > g S-4 p G Ga aS 0) -CpO Gas GP G^3 <M G_ GO G3 a Ph S <u G aS aS aS 3 a> U Ph Gp <H GO 3E pa; a * OJ O o G GGGaCO o pG M CO aS a> * pc S Ga0>3 EG G0CO) EG CO ir\ GHa GCO G<D aSS co g PE Tco- *co* vo GOac> 0) o O Sx GtGo oaS Ph G G3 G <u o GhGQ -0P) CXO aS O OJ T<JU -htcU<HOou E ss <oM wp P<udh Eouh ATC 0034980 8 ATC 0034981 -9- c o 0) nHP a> aj X O rHH P a? 0) ><D -- s -- - -- 0) G .G rH bO ,oH o rH 2S CO < p CO 0) tr\ (X, . bO <u 2S o t- 43 CO <V <u tx in a P 0 0 XO o H Ph nH u CO p o\ c~- cn oooo os o\ CO OS Ei~i HOS CO rl otos os rH to o o \D O -oS** cvi os rH rH r--i CrHO CO o OJ rHI a oS o o rH o< CO OJ a! C- 41) 4pe1) oo lf\ 3 CO EH rH Xo o CO oo a- CM 00 CO <x) c 3 M a o X V O u pa uCQ u to 03 a rH i--( oo s CQ o ncH 3 u B rH nH CO -rH > OJ H pH nH CQ O E 0) 8 rH n0H9 O ce 0) o EC c X S3 <u*H rtHt CO 403) nH hH I Oi 43 CQ X rH 1 O aS Oo rH < 43 CQ a5 O Ph CO LABORATORY TESTS ON CASTABLES ATC 0034982 10 LABORATORY TESTS ON MOLDABLES P4 H0p) oOO cOO c o 4.3 >: o3 O H 4J rH a> < a> LTN Ph o r- -p to o d> &4 -i a P- g u o iH p< -H U CO 3 5 X CVJ p. 0) H a> n o s OJ o rH o p* o tr\ on H 0) O S OJ o o o . 4 c 0 0 t/\ f--1 O rH a> ,0 nH O to 0) 5E rH a) 0) *4 0) CaO > 0) <u X CO > d> CO cn cn o\ 0 O OJ 0 j* ITS ITS 0 O - < 0 CVJ -d* rH C*h P`4 c O b & C O tr\ ir\ O ir\ cn cn trv cn rH rH rH rH -P <D a> 0) 0 4-3 xs 0 d) rH <u gO'"** rH P rl x) < -P > ,0 "d rH rH aJ 3 O a) aS TJ xi ecgj C3 d) O rH c CO H rH rH Jh ITv CO 'H 4-> 0 rH O OO 0- O *4 <M cn tj E O ft u aJ 02 CO S3 ,0 u 3 >1 d) PH u vO rH s a 1 OO CO S) O0 > < CO O vJ ^ ATC 0034983 -u -Cp raH) o u Ooor (h a] &H 11 i1 1 r0"*) woK o o > *H 03 P*< -11- m 3EvQH o g < o > PQ M 25 sO pHq CO E-t EH CO 13 >H a 8 3 o 2 <D o 25 1 i <aU> 0) a> C>0Q) C>aO) ca PQ OoaJ 0) SO5 O pO bo 0) -p* E-* 1--1 1iti b& oo -sr -3- rH rH 1 1 O H -P >: g5 O rH -P Ha < 0) in Pn t- o t- 1 <D O IS I { -p CO <D 0) E- P [Q P P PO O H PQ *H CO Po n: CO 1o ! t I--1 o Il 1 1 1 CVJ P< o H CQ is i in CO rH | \ 0) rH 1 43 ,0 H bO | . tjO V rH *H bO rH O 0) CO 25 z | cvj CVJ VO 1 o * o o - 1 a\ ON rH 1 VD rH -p' OJ (Q 1 | & in o in oo CO CO in rH rH rH < P o CVJ O H o <D Jp OO o* rH P 3 c OO *O 3 BO rH PQ rH VO PT H n PQ <D >J <u 1 VO VO 4> 6 aJ S i1 S<< <L> > H -P H *d aJ 43 H <m aJ 43 H > rH -pVO 1 < O O rH 43 ao) H <H 3 o s a} m H VO -3VO i < * ATC 0034984 12 IN SERVICE EVALUATION RESULTS >1 p c nH CP rH H .g 0) cp X w *d o o= = = e> pp cc *d OP 0) 'd o rH r--1 o o r-1 Q o 0) ^ <o id V. id u u O rl O -H p o o o X 5 O OS X O o o a) t> w pt, O w o fl< o > 0) cp *3 < OP rH C 03 O P CP 2 OP OP CP rH rH rH ,a & ,> b0nH H H P bO P P P t*) P to.C rH nH ^ J3 bO (U H rH bO 0) <D U a) rl M OP nH S = s = = G bO bO H C -H C *H C bO H S3 rH O OP op H O H O H O fl) rH O CQ S5 K S W s CO s CQ s z CQ S5 d) cp P CO CO CP P as G CP bf rH c H r0H) oa) CP HX o 'odo o= = = > p pP p c cg c CP CP CP cp rH rH rH rH H d Ph U *rH pa4) Hgas fc iOod O ood u rH COP wX ri OO o rH CP CP WX rH CP CP wX rH CaP u nH wX at p *CHO C0O) pc CP C . o CP nwCOUHO CP w 'd o O= s r o p G *d CP o rH o rH o CP o X >W P c OP rH rH id h v id h T) oo oo o OO 0 0 0 X0 o o o *d o o o * iood >o OP a rH o Q ,c 2d 0P CO p nH H Jh C a3 X 3o o 1 -3* CM COCO o cD CM t CO CO a3 rH o CP Pd p OP tJ < CQ CP pP O H P< rH < H CP rH o3 CM O CP rS g 3 u CQ P a) O 2O ^ *! u o CO CO nH o CO nH rH g H o 3 CQ -Q P 1 X 1 nH O 3o g 1 g 0$ Oh p 9 O O n rH CP o OO hQ C oo H PU ITv o rH t"- nH t*" Jn O CO nH -d* CtJ CO O g c o CO d rl Ch J- o B U & gS o d rH O u CP H > #H CO hH o nH d rH CQ < 1 G VO <~l C CP CP rH VO CP at 1 O O > < CQ O ATC 0034985 PHYSICAL AND THERMAL PROPERTIES 0) .p 2 -H c -H 2 *4 as s 0 1o 0) St gn VO O* st I 1 OO OO UN OO CM tr\ u\ ON o O VO C--Pf 1 o o \D t l OI UN OJ 1 C-- UN * >5*. UN CO r--l * o rH HI -=t * t~ * -d- VOI I <1 *o J- CTV rH oOo Ov Orooo c\ n co i/v -p o 0) -p o J* OOOJ JU-N * t- UN o \D VO i1 oo rH rH * XX O cnst rH rH OJ * UN 1 1 11 OO f*--I on un iI UN II II o CM CM O VO CHO oUnN -13- o 4a-5s r--HI O CO OCO t-dvo < E o **H aS > t-3 un bon i st VO rH *O CO VO H co CD UN UN 1 1 OJ UN t-- St O CO o UN 1 1 st O vo st * 1i ii OH 1 OO 1 rH ON 1 .1 11 o UN O II II -3* UN * UN rH *V VO OJ CO -O 1 1 1 1 pC*O * st st rH CO 1 I oo P-I co -3 a5 GO O iO O oo -p a> 03 -P a) -h tS HI aShs rH O O CO CO a> rH i aS i v o <5 <U -H -P CO rH 09 3 H aS & CO o O P3* UN rH OJ O CO I OVD 1 OV UN VO co on rH o I0TN- CO rH OJ OHO O CM O E' en o rH OO o on o ooo CO UN OJ iH rl C*-- UN OJ o rH o rH O rH rH VO rH f-- 1 O rH rH rH o UN t- t- Q OJ O UN OJ o co UN UN vo o UN 1 1 rH 1 1 ON vo 1 1 11 OJ O ii ii cr\ ON UN rH vo VO 1 rH u\ 1 1 UN OJ -3* CO o OJ vo rH oo o UN 1 00 rH T( a) *p aS g -P CO <D <D aS 0a3> tao> &+5 cSoh: as P< a -p CO a) o 4 ) O<S -pHH >5 -P H r-s COOJ vCp-Or4O"J- -P <H H CO 0 O X <D o o -P O Ad m Ac hOS'-^O'-' 33- c4) ti <tJ fH f0 fH P '-- -p o o PQ CO o o H P CO OO -p <H I -69- oa> S4 0. ATC 0034986 -14- JEAR FIRING SHRINKAGE INSULATING MOLDABLES FIGURE 1 ATC 0034987 APPENDIX A -15- Recommendations for Preparation and Installation (Supercedes Memo by D. R. Barch and R. G. LaBar of April 27, 1973). A. Melter to Holder Castables: Alcoa C-l, SJP Cast, Kastolite 30, fused-silica castables (See Table I in text). 1. Recommended minimum wall thickness is 2 inches. 2. Follow manufacturer's recommended mixing instructions for a pourable mix using minimum water. 3. Coat forms with oil, grease or polyethylene sheet (Stapled to forms). 4. Tack weld expanded metal mesh lathing to inside of trough 1/2 - 1 inch from steel, work. 5. Post-casting procedure: a. Cover casting with plastic sheet for 24 hours. b. Dry casting in air for 24 hours. c. Castables should be heated slowly to 600 F prior to use, preferably overnight. Caution: Avoid flame impingement and rapid heatup. 6. The recommended coatings, except for Silfrax ARC which requires none, are, Lumnite, Plistix 900 or 900F- Moldables: Solar 80 Plastic 1. Recommended minimum wall thickness is 1 1/2 inches. 2. Follow manufacturer's recommended procedure for installa tion paying particular attention to leaving a roughened surface - do not smooth or overwork the surface! 3. One half to one inch of mineral wool or fiber glass board back-up insulation is recommended. It can be glued to the steel lining with sodium silicate. 4. Post-molding procedure: a. While some air drying is permissible slow heating of this material is recommended after installation is complete. Apply heat uniformly across the entire section. b. Heat slowly to 230 F and hold for 1/2 day per inch thickness. c. Raise temperature to 600 F and hold for 1/2 day per inch thickness prior to use. ATC 0034988 -16- Caution: Avoid flame impingement during initial heatup and use. 5. Coatings are not required. Brick: Coral P or Chas. Taylor 12054 a. Use R & I Super 3000 mortar. b. Standard practice for refractory brick installation. B. Holder to Molten Metal Treatment Process: Castables: Fused-silicas or Kast-O-Lite 30. 1. Follow recommendations in Section A for castables. Moldables: JM-375L, Variform B, Consolidated Moldable. 1. Use 50 lb. water to 100 lb. JM-375L for power mixing. Maximum addition should not exceed 75 lb. water to ICO lb. JM-375L! 2. The same restrictions stated in 1 above apply to variform. 3. Thorough mixing will permit the use of minimum water and produce more uniform, workable materials. 4. Consolidated Moldable is ready to use as received. 5. Tack weld expanded metal mesh lathing to inside of trough 1/2 in. from steel work for long trough sections, unless the trough is equipped with an overhanging lip. 6. May be hand molded, tamped and/or troweled in place. 7. Do not overwork the surface! 8. Post-molding procedure: a. Cover'with plastic sheet for 24 hours. b. Air dry for at least 24 hours, (longer for Consolidated Moldable). c. Heat each section (between joints) uniformly and slowly to 230 F overnight. d. Continue heating to at least 600 F prior to use. Caution: Avoid flame impingement, non-uniform heating, rapid heatup and scraping with sharp edged tools. 9. The recommended coatings are Lumnite, Plistix 900 and 900 F, J-M Cerakote, whiting (apply whiting by rubbing it into the surface after step 8 b. above and brush away all the excess). ATC 0034989 -17- Board and Preformed Shapes: Vacuum Formed Ceramic Fiber (see Table I in text), Glasrock Foam (50 pcf). 1. These materials may be installed using the same procedure as with Molten Metal Marinite or follow manufacturer's instructions (especially in the case of Glasrock Foam). 2. Recommended coatings are Lumnite, Glasrock Sealer (with Glasrock Foam), J-M Cerakote, Plistix 900 and 900 F, whiting (as per instructions under moldables above), or coating supplied with Kalmin iFoseco). Caution: Avoid flame impingement and scraping with sharp edged tools. C. Downstream from Molten Metal Treatment Process Moldables: Same as B above. Board: Same as B above except do not use Lumnite or Plistix coatings. D. Packing and Miscellaneous Ceramic fiber bulk, blanket and rope may be used in place of asbestos rope .for packing, etc. Thermo 12, Kaylo 10, or Ceramic fiber board products may be used for insulation in place of Thermobe.stos, Superex, Molten Metal Marinite, etc. General In Service Care and Cautions 1. Avoid excessive preheating. If preheating is necessary, apply uniform heat, avoiding hot spots and direct flame impingement. 2. Avoid mechanical abuse. a. Install protective caps over linings near tap spouts. b. Lift skulls by hand, minimize scraping and/or gouging. 3. Extra care is required when installing dams in drop-in liners, to avoid breakage. Recommend use of ceramic fiber paper (Fiberfrax, Cerapaper, etc.) gasketing around dams and j oints. 4. A program of scheduled inspection and maintenance should be implemented. It is recognized that individual plant differences exist in trough usage and design, but it has been concluded that these differences can be taken care of if made known and discussed. ATC 0034990 -18- The ingot plant metallurgists are requested to inform M. L. Redhair of contacts with various vendors and receipt of sample materials. In addition, brief feedback on plant experiences, either pro or con, with the recommended materials are requested. New materials and/or revised procedures for installation and use will be added to the list as they develop. ATC 0034991 FROM G. K. TURNBULL INGOT CASTING DIVISION ALCOA TECHNICAL CENTER - B TO MR. J. E. HATCH INGOT & POWDER PRODUCTS DIVISION PITTSBURGH OFFICE - 23 "CONFIDENTIAL August 8, 1974 RE: ELIMINATE ASBESTOS IN ALL FURNACE APPLICATIONS ' Ingot Casting Division Report No. 11-74-12B103 Asbestos has become thoroughly engrained in ingot casting practices through generations of us^. We'vd learned where and how to use it, and have even adapted our practices to take maximum advantage of its characteristics. Yet we must curtail such use and do so over a fairly short transition period. We could hardly expect to have a shelf item that is the ideal substitute: superior to asbestos in every way. Yet in the series of engineering tradeoffs inherent in such substitution we are fortunate to have a series of alternatives adequate to undertake nearly every aspect of substitution. Some preexisted and have been evaluated by plants and Alooa Laboratories. Others were recently developed at the Laboratories. Major candidates are compared in the attached report by Dr, R. G. LaBar. Preferred candidates have been singled out into the iAppendix section and paired up with detailed installation instructions. Preferred technique for use of the report is to use Appendicized materials and limit them to the applications and procedures described. If local conditions, refractory availability or other considerations override this approach, some insight into the short comings of other candidates are described in 'the body of the report. It is strongly recommended, however, that use of the not-recommended materials be coordinated through Dr. R. G. LaBar or Mr. D. R. Barch to provide maximum opportunity for their success and to extend the extremely important "In Service Evaluation Results" of Table V. A separate report will be issued covering parallel developments on asbestos elimination in headers. Mr. Sartschev plans to expand narrow initial distribution by submitting to you a list of plant recipients. v G. K. TURNBULL Nf GKT:sh r Attachment cc: R. Speax/TID J T. R. Gauthier/J. H. Dunn - Pgh K. J. Brondyke/R. T. Teeter - ATC R. G. LaBar - ATC Div. File T. B. Bonney/R. P. Carter - Pgh 7 a R. C. Cook - Pgh ALCOA 7-1B6304 ATC 0034992 One copy routed to SMIC: A. J. Sartschev - Pgh E. D. Quade - ATC E. V. Blackmun - Pgh M. C. Schoetz - Pgh R. W. Wrenn - Pgh E. L. Rooy - Tenn M9fl (<trv. !*) ALCOA TECHNICAL CENTER 100 TECHNICAL DRIVE ALCOA CENTER, PA 15069-0001 HDC Header and Feedboard Material Development and Production Qualification 5rrrfC-'. ; ^Advanced Ceramics Center r ^ AlcoaTechnlcajl Centepfr Mr J. Dunlay, W. Lansdale,/ Ml C. Lukens, B. W> Siebe. Alcoa, WarridkOperations .'*W ' Nfv Alcoa Private Information ATC 0034993 0265431 0054363 Distribution List for Division Report No. 06-93-85 Alcoa Private Information cc: ID - D ADCC T&M - ATC-C-ADCC (Route) R. Bachowski / R. T. Richter - ATC-B-MMP J. S. Benjamin - ATC-C-ADM R. A. Bonewitz - ATC-C-ADM M. J. Bruno - ATC-B-MMP V. Cedro - ATC-C-CSD M. G. Chu / D. A. Granger - ATC-B-MMP G. J. Hildeman - ATC-B-MMPR. A. Kois - ATC-B-MMP (R. A. Marra - ATC-C-ADCC) A. J. Sartschev - ATC-C-MMP A. Pearson - Knoxville M. Scherbak - Knoxville P. Thomas - Tennessee J. A. Bloomer - Warrick (M. J. Dunlay - Warrick) (W. Lansdale - Warrick) (M. C. Lukens - Warrick) B. C. Owens - Warrick F. M. Packer - Warrick (B. W. Siebe - Warrick) T. E. Leary - Wenatchee D. M. Scheumann - Wenatchee CSD File ATC 0034994 0054963 HDC HEADER AND FEEDBOARD MATERIAL DEVELOPMENT AND PRODUCTION QUALIFICATION R. A. Marra Advanced Ceramics Center Alcoa Technical Center M. J. Dunlay W. Lansdale M. C. Lukens B. W. Siebe Alcoa Warrick Operations IT SHOULD BE NOTED THAT THE MATERIAL FORMULATIONS DESCRIBED IN THIS REPORT ARE CONSIDERED HIGHLY PROPRIETARY AND SHOULD NOT BE DISCUSSED OR DISCLOSED WITH ANY NON-ALCOANS WITHOUT THE CONSENT OF WARRICK PERSONNEL. Prepared by: Date: > // R. A. Marra Noted by: Date: /fyjr M; J. Bruno _____ Noted by: H Vfr -tbec. 1 Date: '/M V. Cedro Pfi r it- CS/3273P CSD Report No. 06-93-85 1993 December 14 ATC 0034995 EXECUTIVE SUMMARY Calcium silicate board materials are used as headers and feedboards at Warrick Operations for the Horizontal Direct Chill (HDC) casting of aluminum ingots. Header Grade Molten Metal Marinite (HGMMM), an asbestos-reinforced board material, has historically been used in this application. Due to the health hazards associated with this asbestos containing material, replacement materials for both the headers and feedboards needed to be identified for Warrick Operations to meet its ingot production demands. This report describes the activities which led to the development and production qualification of Pabco PMT and PMTHG calcium silicate materials for feedboards and headers. During the program, an improved understanding of the effects of the header material and mold package design on-the casting performance was developed. This understanding was critical to the success of the program. An excellent working relationship was established between ATC, Warrick (MPE-Ingot and Production) and Pabco (materials vendor) personnel. The development of standard operating procedures by Warrick mold preparation and ingot casting personnel greatly aided the evaluation of potential replacement materials. The use of designed experiments and statistical analysis provided the basis for downselecting the potential material formulations. Statistical methods were further utilized to optimize the material formulation, mold package design, and casting process variables. Proposed activities for further material refinements and HDC process optimization which should result in improvements in casting life, ingot recovery, and ingot quality are discussed. CS/3273P i ATC 0034996 TABLE OF CONTENTS Page No. EXECUTIVE SUMMARY .................................................................... LIST OF FIGURES ..................................................................................... I. INTRODUCTION .................................................................................... II. PROBLEM STATEMENT ..................................................................... III. SUMMARY OF ACCOMPLISHMENTS ............................................... IV. PROGRAM RESULTS - SIGNIFICANT EVENTS .......................... V. MOLD PACKAGE UNDERSTANDING ................................................ A. Header Material Physical Characteristics .................................................. B. Molten Metal / Header Chemical Interactions ........................................ C. CHI Absorption............................................................................................ D. Header Failure Mechanisms .................................................................... E. Header Thermal Expansion ..................................................................... VI. FUTURE IMPROVEMENTS (Current, Planned, and Proposed Activities) .................................................. A. Casting Process Optimization .................................................................... B. Materials Development ............................................................................. 1. Red Mud Level ................................................................................. 2. Calcium Silicate Board: Xonotlite Structure.................................... 3. Red Mud/Vermiculite Ratio............................................................. C. Feedboard Cracking ................................................................................. D. Header Overhang ............................................. E. Mold Shop Consistency ........................................................................... VII. CONCLUSIONS ...................................................................................... i in 1 3 4 5 11 11 12 14 15 16 18 18 18 18 19 19 19 20 20 21 CS/3273P u ATC 0034997 LIST OF FIGURES Figure 1 Figure 2 Figure 3 Figure 4 Header material cold crushing strength. Header material modulus of rupture. Header material oil absorption. Header material thermal expansion. CS/3273P in ATC 0034998 I. INTRODUCTION Calcium silicate board materials are used as headers and feedboards at Warrick Operations for the Horizontal Direct Chill (HDC) casting of aluminum ingots. The HDC process was originally designed to use Header Grade Molten Metal Marinite (HGMMM) for the header and feedboard. HGMMM is an asbestos fiber reinforced calcium silicate board material. Due to the health hazards associated with this asbestos containing material, replacement materials for both HDC headers and feedboards are required for Warrick Operations to meet its ingot production demands. An acceptable header replacement should have the characteristics shown in Table I. The feedboard should have similar characteristics, although die resistance in molten metal attack is less important The moisture and carbonate content of the feedboard is extremely important since a large area is exposed to molten metal contact Table 1 Desired Characteristics for HDC Header Materials Characteristic Molten Metal Resistance (Low Wettability) Importance - Prevent metal / header reactions which degrade the header surface and result in ingot defects. Low Thermal Conductivity High Compressive Strength - Prevent molten metal solidification at the header surface. Prevent crushing of header in the mold package. Low Moisture / Carbonate Content High Abrasion Resistance Low Porosity / Permeability Adequate Strength (MOR) Low Thermal Expansion (Good Thermal Stability) - Prevent risk of explosion when contacted with molten metal and avoid outgassing which degrades ingot quality. Prevent degradation of header surface during casting. - Minimize lubricant absorption. - Allow machining and handling of header. - Minimize change in header / oil ring overhang and stress build-up in mold package. CS/3273P 1 ATC 0034999 Pabco Metaltemp (PMT) and BNZ Materials MetALform calcium silicate boards have been evaluated for feedboard and header replacement These materials are boards produced by pressing slurries of lime (CaO), silica (SiC>2; typically added as diatomaceous earth and a colloidal material), wollastonite (CaSiO}), and alkali resistant glass fibers. The pressed boards are autoclaved to convert the lime and silica into a calcium silicate compound. The autoclaved boards are further heat treated to remove physically and chemically bound water. Recent work on identifying a header replacement has shown that chemical reactions between the metal and header material are important in determining the header life, Le., the life that the header can be used before ingot defects and excessive scrap results in casting termination. It is believed that the magnesium and aluminum in the molten metal attack the calcium silicate bond phase which results in defects on the header surface. The molten metal can solidify in these areas. The solidified material continues to grow until it is trapped in the solidification front and pulled out by the moving ingot. This process results in ingot strawberries, tears, or surface cracks which increase the ingot scrap and eventually lead to termination of the casting run. This report details the header and feedboard program activities and accomplishments since the first quarter of 1991. The format of the report describes the progress (where we are in the program) and the chronological steps that were taken to develop and implement the replacement materials. Details on the understanding of the mold package functionality are also discussed. The success is finding alternative HDC mold package materials and improving the mold package design can be largely attributed to the emphasis that had been placed on understanding the mold package and the header / molten metal interactions during casting. Finally, planned and proposed activities for future improvements are presented. CS/3273P 2 ATC 0035000 II. PROBLEM STATEMENT In 1990 October, the following project objective was established: New products for marinite replacement must meet the following criteria: 1) The average mold life must equal or exceed the average life with header grade molten metal marinite for all alloys; and 2) The casting recovery must be greater than or equal to 88% due to header related problems. In 1991 October, this objective was modified to reflect the changes in ingot production schedules to reduce "flowtime." With the emphasis of reducing the material "flowtime" through the plant, achieving maximum mold life was no longer the primary goal. The new objective was: Develop a replacement for the header grade molten metal marinite header to be used in 5XXX alloy casts which meets production goals. The header life objective is 28 hr. The casting recovery must be greater than or equal to 88% due to header related problems. To further reflect the desire to link header performance with production goals and flowtime reduction, the problem statement was recently changed to: Develop a replacement for the header grade molten metal marinite header (HGMMM) to be used in 5XXX alloy casts which meets production goals. The header life should be sufficient to ensure that less than 10% of all casting terminations are not caused by defects related to the mold package. The header life objective is 28 hr. The casting recovery must be greater than or equal to 88% due to header related problems. CS/3273P 3 ATC 0035001 III. SUMMARY OF ACCOMPLISHMENTS The following data summarizes the accomplishments of the HDC Header Materials Development Team in developing and implementing a replacement for HGMMM. The average HDC mold life on 5XXX series alloy castings has increased from a 1991 BNZ MetALform header average of 12.4 hr to a 1993 (January-June) average of 24.0 hr with Pabco PMTHG2.5 headers - a 94% increase. Average HDC recovery on 5XXX series alloy castings has increased from 83.9% in 1991 using MetALform headers to 88.0% in 1993 (January-June) using PMTHG2.5 headers. The percentage of molds shut down for nonheader related reasons (3XXX and 5XXX series alloys) has increased from 49% in 1992 to 64.2% in 1993 correlating with the 100% conversion to Pabco materials on all 3XXX and 5XXX series alloy casts. In addition to a higher quality header and feedboard material, flowtime considerations would drive this percentage. A complete conversion from the BNZ MetALform material was accomplished in 1993 January. During 1991 and 1992, a considerable amount of resources (dollars, time, utilization, recovery, etc.) were expended while utilizing BNZ MetALform materials for HDC headers and feedboards. Average HDC mold life on 3XXX series alloy castings has increased from a 1991 average of 41.7 hr to a 1993 average of 52.0 hr. Recovery has also increased from 86.3% in 1991 to 92.0% in 1993. A direct comparison of 3XXX series alloy mold life during 1992 shows that the Pabco PMT material has an average life 30% higher than the BNZ MetALform material - PMT average of 56.8 hr versus MetALform average of 43.8 hr. Also, the 3XXX series alloy cast with PMT headers had a higher recovery than those with MetALform headers - 92.5% versus 91.9%. CS/3273P 4 ATC 0035002 IV. PROGRAM RESULTS - SIGNIFICANT EVENTS This section describes the significant events that led to the implementation of Pabco PMT and PMTHGXJC materials for feedboards and headers in the mold packages for Warrick's HDC casting. The events are documented in chronological order. Only the highlights of these activities are provided; references which present more detailed descriptions are provided. 1991 February-March: R. A. Marra published (1991-03-07) an ATC data letter "Header / Metal Surface Reactions - Magnesium Reduction" which postulated the important role of header surface oxide component reduction on the header life and ingot quality during HDC casting of 5XXX series aluminum alloys. It was determined that magnesium, a highly reducing element, will reduce both calcium silicate and iron oxide components. If iron oxide is present, it should be preferentially reduced, thereby preventing (or, at least, delaying) the detrimental reduction of the calcium silicate bond phase. HGMMM contains a large amount of amosite-type asbestos fibers (MgFe6[(OH)Si40n]2) which have a high content of iron oxide. This may contribute to the success in using this material as a header for HDC casting. Based on these results, it was decided to evaluate calcium silicate boards with additives to control the board chemistry and reduction reactions. 1991 April: Based on their laboratory capabilities and willingness to develop new material compositions, Pabco was selected as the preferred company to work with on developing modified calcium silicate boards with iron-containing additives. During a 1991-03-21 visit to Pabco [Ref: Warrick Memorandum: D. C. Sikora to HDC Marinite Replacement and Process Improvement Team, "Progress Report -1991-03-16 to 1991-03-22"], Alcoa discussed the desire for developing these modified materials. It was determined that boards containing vermiculite, red mud, iron oxide (FeO), and combinations of these materials would be evaluated for potential use as headers for 5XXX series alloy casting. 1991 May-July: Pabco performed lab trials to determine that levels of additives which could be processed into a calcium silicate board product Coupons of these materials were evaluated for molten metal resistance at ATC Based on these trials, a materials screening DOX was developed to evaluate the potential substitutes for HGMMM and BNZ MetALform. The materials consisted of: Pabco modified PMT with 13% vermiculite; Pabco modified PMT with 2% iron oxide (FeO); CS/3273P 5 ATC 0035003  Pabco modified PMT with 10% red mud; Pabco modified PMT with 5% vermiculite + 5% red mud; Pabco modified PMT with 10% vermiculite + 2% iron oxide (FeO); BNZ MetALform as a control; and Header grade molten metal marinite (HGMMM) as a control The screening experiment was designed to generate header material specimens for analytical examination that have had similar molten metal contact in a high magnesium alloy casting process (800 ingot inches or 3 hr cast time of alloy 5082 on HDC strand 1A). Each material was tested in three different casts. The test was designed such that the machining, mold shop, and casting variables were blocked. [Ref: Warrick Memorandum: M. J. Dunlay to G. R. Swanberg, "Screening of Marinite Replacement (Header) Materials," 1991-07-12]. Materials testing was performed on all materials before and after casting at several mold positions. The material testing included approximately 25 elemental, structural and physical analytical techniques performed at ATC, Warrick, Pabco, and an outside laboratory [Refs: Warrick Memoranda: M. J. Dunlay to G. R. Swanberg, "Update on Testing of Board Materials used in Header Screening Evaluation," 1991-06-20, M. C. Lukens to M. J. Dunlay, `Testing of Header Materials," 1991-07-23 and M. C. Lukens to M. J. Dunlay, `Testing of Header Materials-Revised," 1991-11-13]. 1991 November: The material testing and statistical analysis of the 1991 July materials screening tests were completed. The following material criterion measures were found to be significant: Cold crushing strength; Modulus of rupture; Thermal expansion (perpendicular to the pressing direction); Leco carbon analysis at mold positions 2,5, and at positions relative to the oil ring; Oil absorption on unused materials and correlation of oil absorption and Leco carbon analyses; and SEM with EDAX analysis of magnesium, aluminum, iron, calcium, and silicon on used header surface. Based on these results, the calcium silicate board material with 5% vermiculite and 5% red mud additions was selected to be used for confirmation evaluations. CS/3273P 6 ATC 0035004 1991 November-1992 January: Experiments were conducted to qualify PMT feedboards and PMT headers for 3XXX series alloy casting. During these tests, problems were observed with header crushing at the area in contact with the oil ring. Separate testing using pressure loading paper and strain bolts confirmed the need for redesign of the mold clamp design [Ref: Warrick Memorandum: W. Lansdale to M. J. Dunlay, `Teedboard - Header Clamp Evaluations," 1991-11-26]. A new clamp was designed which provided a more uniform distribution of the compressive forces. Experiments with the old clamp and new clamp showed that 80% of the explained variability (55%) was a two way interaction between the clamp and material, [Refs: Warrick Memoranda: W. Lansdale to M. J. Dunlay, "Feedboard / Header Material and Header Clamp Evaluation Results," 1992-01-09 and W. Lansdale to J. A. Bloomer and T. D. Plassmeier, "Feedboard/Header Material and Header Clamp Confirmation DOX," 1992-01-13]. Based on these findings, the clamp design and feedboard type were incorporated in the confirmation trail of the Pabco 5% vermiculite + 5% red mud (PVRM) material on 43" 5082 castings (Ref: Warrick Memorandum: M. J. Dunlay to G. R. Swanberg and P. D. Thomas, "Upcoming Marinite Replacement Materials Experiments," 1992-01-17]. 1992 January: The PVRM confirmation trial was completed during the week of 1991-01-20. This trial Evaluated seven possible combinations of header, feedboard, and clamp design. A total of 11 casts of 43" 5082 alloys were completed on HDC Complex No. 5. The data obtained strongly supported the conclusion that an alternative mold assembly package utilizing headers machined from Pabco calcium silicate boards with 5% vermiculite and 5% red mud additions, Pabco Metaltemp feedboards and the new clamp design results in superior mold life as compared to the mold package comprised of BNZ MetALform headers, BNZ MetALform feedboards, and the old clamp design. [Ref: Warrick Memorandum: M. J. Dunlay, W. Lansdale, M. C. Lukens, B. W. Siebe, and R. A. Marra to G. R. Swanberg and P. D. Thomas, "Marinite Replacement Material Screen (1992-01-20) Update," 1992-01-29]. Casts using the alternative mold package resulted in two to three times the life of the typical mold assembly averages. Other findings of the trial included: A mold assembly combination of PVRM header, MetALform (MF) feedboards, and new clamp design may also provide a competitive mold package for casting of high magnesium alloys. The potential of using both PMT and MF feedboards would offer the long term advantage in supplier / inventory flexibility. No mold assembly combination using MF headers resulted in an improvement over the historical MF header mold life of 12 hr. CS/3273P 7 ATC 0035005  The thermal dimension changes (material thermal expansion) have a significant impact on ingot quality and therefore on mold life. Ingots cast with the PVRM header showed a more prominent (improved) lap pattern as casting time progressed. More than 24 of the ingots cast with the PVRM header were successfully hot rolled. 1992 March: The 1992 January trial highlighted the importance of not cracking or crushing the feedboard during clamping. Based on these results, new clamps were ordered for all mold sizes. 1992 January-June: During the first half of 1992, Pabco encountered difficulties in producing boards with the 5% vermiculite and 5% red mud additions. The problems appeared to be related to difficulties in dewatering and drying the boards as the red mud level was increased. Alcoa provided assistance to Pabco in characterizing the red mud to determine the cause of the production problems. [Ref: ATC Letter Report 06-92-54: R. A. Marra to Memorandum, "Pabco Red Mud Characterization," 1992-07-10.] It was found that the red mud raw material (J. R. Goslee Company, Powder Division, R-20 Product) contained large fraction of fine particles. This may effect the pore structure in the board and contribute to the difficulties in dewatering and drying the material during production. Also, the presence of silica, calcite (CaCC>3), and the de-silication product (sodium aluminosilicate) may have an effect on the slurry gelation. Permeability measurements (ASTM Test C-577-87) of boards produced with varying red mud levels showed a decrease in the permeability with increasing red mud content. This supports the postulation that the red mud is effecting the pore structure. 1992 May: Pabco and Alcoa agreed to manufacture boards with red mud levels increasing in 0.5% increments; Warrick would evaluate the casting performance of these materials. During the week of 1992-05-26, Pabco boards containing 5% vermiculite and 0.5% red mud were evaluated for casting performance of 43" 5082 ingots. [Ref: Warrick Memorandum: M. J. Dunlay, W. Lansdale, M. C. Lukens, B. W. Siebe, and R. A. Marra, "Marinite Replacement Materials Evaluation with Pabco Vermiculite (5%) and Red Mud (0.5%) Header Board," 1992-06-04.] The effect of board heat treatment temperature (600F and 950F) was also studied during this trial. The results strongly supported the conclusion that a PVRM header with 0.5% red mud additions will provide superior mold life versus that offered by the MetALform materials. This screening trial did not generate casting data which would indicate a concern with the lower level of red mud (0.5%) versus the 1992 January trial with 5.0% red mud levels. However, subsequent materials testing provided data to suggest that higher red mud levels would provide improved performance. This is discussed in more detail in the Section V. There was not a significant difference in casting performance at the different heat treatment temperature. Previous CS/3273P 8 ATC 0035006 characterization of PMT materials [Ref: ATC Letter Report No. 06-92-28, R. A. Maria to Memorandum, "Characterization of Pabco Metaltemp," 1992-03-12] showed that higher heat treatment temperatures resulted in increased porosity and oil absorption and decreased strength. Therefore, it was decided that a standard heat treatment of 600F would be used to produce all production boards. 1992 June: M. J. Dunlay, W. Lansdale, M. C. Lukens, and R. A. Marra visited Pabco in Fruita, Colorado, to conduct a technical review of the marinite replacement project status and observe the manufacturing of additional vermiculite - red mud boards [Ref: Warrick Memorandum: M. J. Dunlay, W. Lansdale, M. C. Lukens, and R. A. Marra, "Pabco Visit of 1992-06-08 to 06-10," 1992-07-15]. The following items were discussed: Management / business items; Vermiculite and red raed formulation issues; Vermiculite and red med process issues; Pabco fabrication processes; and Pabco laboratory and testing capabilities. The visit was very informative and provided an opportunity to begin a joint effort to resolve PVRM fabrication issues. Management issues that Pabco considers important were identified. The close working relationship that was established between Alcoa and Pabco was a key element to the success of this program. 1992 August: Pabco PMT boards and boards with vermiculite and red mud additions continued to show significant potential for use feedboards and headers. Material specifications [Ref: Warrick Memorandum: M. J. Dunlay to D. C. Sikora, 'Tabco Material Specifications," 1992-08-13] were developed to provide consistent quality of production materials. 1992 September: PMT qualification trials (10 boards) were completed and PMT feedboards (used with the new clamp design) were qualified for casting of all alloys and all sizes. [Ref: Warrick Memorandum: W. Lansdale to M. J. Dunlay, "Qualification of Pabco's Metaltemp Material - Feedboard Applications," 1992-06-04.] The PMT header was also qualified for all sizes of 3XXX (low magnesium) alloy castings. CS/3273P 9 ATC 0035007 1992 November: Secrecy and purchasing agreements were signed between Pabco and Alcoa for the production and use of boards containing red mud and vermiculite additions. These agreements included 10 year confidentiality and exclusive use clauses. As a measure to protect the confidentiality of this material formulation, a new material designation was adopted. This designation, PMTHGXJC, represents Header Grade Pabco Metaltemp material with XJC percent level red mud content IT SHOULD BE NOTED THAT THIS FORMULATION IS CONSIDERED HIGHLY PROPRIETARY AND SHOULD NOT BE DISCUSSED OR DISCLOSED WITH ANY NON-ALCOANS WITHOUT THE CONSENT OF WARRICK PERSONNEL. 1992 June-1993 January: Production quantities of PVRM with 0.5%, 1.0%, 1.5%, 2%, and 2.5% red mud content (5% vermiculite in all materials) were successfully used in the casting of 5XXX series (high magnesium) alloys. Based on the casting results, materials testing, and Pabco's ability consistently to produce board with 2.5% red mud additions, trials to qualify the PVRM material containing 5% vermiculite and 2.5% red mud were completed. In 1993 January, this material was qualified for 5XXX series castings on all ingot sizes. 1993 January to Present: Pabco PMT and PMTHGX.X materials have been used almost exclusively (>90% of casts) for casting of 3XXX and 5XXX series alloys, respectively. As summarized in Section HI (Summary of Accomplishments) substantial increases in casting life and ingot recovery have been realized. There have been several 5XXX casts which have exceeded 100 hr and several 3XXX casts which have exceeded 200 hr indicating that significant potential for process improvements exist with these mold packages. [Ref: Warrick Memorandum: M. J. Dunlay, W. Lansdale, M. C. Lukens, R. A. Marra, and B. C. Siebe to G. R. Swanberg and D. C. Ashley, "Pabco Header / Feedboard Update," 1993-07-26.] CS/3273P 10 ATC 0035008 V. MOLD PACKAGE UNDERSTANDING During the course of this program, an improved understanding of the effects of the header material and mold package design on the casting performance has been developed. This understanding has been critical to the success of the program. This section briefly discusses the impact of header material characteristics and mold package design on casting performance. A. Header Material Physical Characteristics The physical characteristics of the header material have a large impact on the casting performance of the mold package. This discussion will concentrate on analyses conducted on the Pabco vermiculite + red mud material (PMTHGXJC) and the comparison with Header Grade Molten Metal Marinite, BNZ MetALform, and Pabco Metaltemp. Correlations between the material characteristics and casting performance for other materials were examined in the 1991 July materials screening trial [Ref: Warrick Memorandum: M. J. Dunlay to G. R. Swanberg, "Update on Testing of Board Materials used in Header Screening Evaluation," 1991-06-20 and Warrick Memorandum: M. C. Lukens to M. J. Dunlay, `Testing of Header Materials," 1991-07-23]. PMTHG materials with 5% vermiculite and red mud levels of 0.5% to 5% were characterized with respect to cold crush strength, modulus of rupture, oil absorption (% weight gain in coupons submerged in oil for 3 hr) and thermal expansion. The results are shown in Figures 1, 2, 3 and 4, respectively. Similar analyses on Header Grade Molten Metal Marinite (HGM), BNZ MetALform (MF), and Pabco Metaltemp (MT) are shown for comparison. The following conclusions can be made: Red mud additions of 2.0% and higher result in increased cold crushing strength [Ref: Warrick Memorandum: M. C. Lukens to M. J. Dunlay, W. Lansdale, R. A. Marra, and B. W. Siebe, "Cold Crush/M.O.R. Update," 1993-04-22]. The strength of these materials is significantly higher than the BNZ MetALform strength and at least as high as that measured for the marinite material. A similar trend was observed in the material's modulus of rupture (MOR) [Ref: Warrick Memorandum: M. C. Lukens to M. J. Dunlay, W. Lansdale, R. A. Marra, and B. W. Siebe, "Cold Crush/M.O.R. Update," 1993-04-22]. Red mud additions of 2.0% and higher resulted in significandy higher strength than PMTHG materials with lower CS/3273P 11 ATC 0035009 levels of red mud and the BNZ MetALform materiaL Molten metal marinite had the highest modulus of rupture although the strength was not significantly higher (at the 0.1 level) than the PMT material or PMTHG materials with red mud contents of 2.0% and higher. The asbestos fibers in the HGMMM material tend to strengthen the material. Oil absorption has been shown to have a significant effect on header performance. All of the Pabco materials (PMT and PMTHGXJQ show significantly less oil absorption than the HGMMM and BNZ MetALform materials. PMTHG materials with intermediate level of red mud additions (1.0%, 1.5%, and 2.0%) showed the lowest oil absorption. [Ref: Warrick Memorandum: M. C. Lukens to M. J. Dunlay, W. Lansdale, R. A. Marta, and B. W. Siebe, "Oil Absorption Update " 1993-04-22], The header material's thermal expansion characteristics also play an important role in the mold package performance. This is discussed in more detail in Section V.4. Materials with negative expansion coefficient or low positive values of expansion coefficient will maintain the header overhang which results in a more prominent lap pattern. The PMT, MetALform, and PMTHG0.5 have a significantly higher coefficient of thermal expansion than the other materials. [Ref: Warrick Memorandum: M. C. Lukens to M. J. Dunlay, W. Lansdale, R. A. Marra, and B. W. Siebe, "Thermal Expansion Update," 1993-04-31]. It is clear that the performance of the header is related to these physical characteristics. Optimum performance can be obtained by developing a material which has high crushing strength and modulus of rupture, low oil absorption, and a negative or low positive coefficient of thermal expansion. The Pabco PMTHG2.5 material (now qualified and being used almost exclusively for HDC casting of 5XXX alloys) has a favorable combination of these characteristics. The importance of header's chemical composition and its impact on reactivity with the molten metal is discussed in the next section. B. Molten Metal f Header Chemical Interactions Chemical surface analysis of used headers along with thermodynamic considerations showed that both magnesium and aluminum (to a lesser extent) will reduce the oxide components in the header board [Ref: ATC Memorandum: R. A. Marra, "Header/Metal Surface Reactions Magnesium Reduction," 1991-03-07]. Analysis of the marinite material showed the importance of iron oxide in controlling the magnesium reduction reactions. Thermodynamic calculations CS/3273P 12 ATC 0035010 showed that magnesium will reduce both calcium silicate and iron oxide. If iron oxide is present, it should be preferentially reduced. SEM / EDAX analysis confirmed the reduction of the header surface by both magnesium and aluminum. The much higher content of Mg (especially in comparison with the alloy compositions) indicate that the reduction by Mg is much more severe. This is consistent with the observed difficulties in casting SXXX series alloys. The EDAX analysis suggested that the role of iron oxide in the asbestos containing marinite boards was important in preventing the attack of the calcium silicate bond phase and thus the degradation of the mechanical and physical properties. Since the iron oxide components are discrete phases (can be thought of as fillers), their attack is much less detrimental to the integrity of the board than the attack of the calcium silicate matrix phase. The iron oxide component can be thought of as a sacrificial phase which prevents degradation of the matrix. A similar analysis was performed on the Pabco PMT header material [Ref: ATC Memorandum: R. A. Marra to Memorandum, "Pabco Metaltemp: Header/Metal Surface Reactions," 1991-05-03]. The PMT material is comprised of a calcium silicate bond phase with wolastonite and alkali resistant glass fiber fillers. There is no detectable iron oxide content in the material. SEM / EDAX analysis of a used header (5182 alloy cast) showed that magnesium and, to a lesser extent, aluminum reduced the calcium silicate bond phase during casting. The wolastonite fibers were quite resistant to chemical attack. Based on these observations, it was recommended that iron oxide-containing additives such as vermiculite, red mud, and iron oxide be added to the PMT material in an attempt be prevent the attack of the bond phase and the deterioration of the material's physical and mechanical properties. Pabco agreed to produce boards with these additives which were evaluated in the 1991 July materials screening test (See Section IV). SEM / EDAX analysis of the headers used during this screening evaluation confirmed the importance of iron oxide in controlling the molten roetal/header chemical reactions. The analysis of a defect (strawberry) on a 5XXX alloy ingot showed that fragments of the header material were present below the ingot surface. This supports the hypothesis that chemical reactions between the metal and the header leads to the attachment (freezing) of metal on the header surface. The solidified metal gets trapped in the solidification front and is pulled out by the moving solidified ingot. CS/3273P 13 ATC 0035011 C. Oil Absorption There is evidence that the amount of oil absorbed in the boards has a significant effect on the generation of ingot defects and thus the header life. The effect is not completely understood although various observations support its importance. The amount of oil absorption is much greater at the bottom and sides of the header than on the top. This was determined quantitatively during the detailed analysis of the header materials used in the 1991 July materials screening evaluation. It is much more common to observe defects on the sides and bottom of the ingots than on the top surface. In 1991 September, a BNZ MetALform header was used for casting of 5182 alloy. The header life was 38 hr, this is much longer than the typical life (1991 average life of 12.4 hr) of the BNZ material for casting 5XXX alloys. [Ref: Warrick Memorandum: M. C. Lukens to R. A. Marra, "Analysis of BNZ Header# 1109005-1," 1991-10-03]. Visual inspection of the header suggested that oil absorption was extremely low. This was confirmed by Leco carbon analysis of the used header. SEM / EDAX analysis of the used header showed that there was substantial chemical interactions between the molten metal and header in the region near the oil ring (surface cation composition: Mg-8.5 wt%, Al-5.6 wt%, Si-30 wt%, Ca-56 wt%) whereas there was no chemical attack in the region where there was no oil absorption (surface cation composition: Si-30 wt%, Ca-70 wt%). It appears that the oil absorption and the products of the pyrolysis effect the calcium silicate's wetting characteristics thus allowing the molten metal to react with the header. In 1975, R. LaBar [Ref: ATC Memorandum, R. G. LaBar, "Johns-Manville/Alcoa Meeting, Warrick Operations - 1975-06-17," 1975-07-22] postulated that the oil and oil vapor is intruded into the pore structure of the marinite board and is pyrolyzed to produce carbon, hydrocarbons, hydrogen, carbon dioxide, water vapor, and possibly some fatty acids. These pyrolyzed products can react with the components (iron oxide impurities, free lime, lime-rich silicates, and silica-rich phases) in the header material. The net result of these reactions is a weakening or destruction of the bond phase and an increase in thermal conductivity due to carbon deposition. He suggested that the increased difficulty in casting higher magnesium may not be related to corrosion by the alloy but rather to the additional lubricant required to cast these alloys. It is obvious that the oil absorption plays an important role in the header degradation. Further studies are needed to optimize the lubrication rates and determine the impact on casting performance. CS/3273P 14 ATC 0035012 D. Header Failure Mechanisms Various header failure mechanisms have been proposed (Ref: R. LaBar, "Johns-Manville / Alcoa Meeting, Warrick Operations - 1975-06-17," 1975-07-22 and ATC Memorandum: M. G. Chu / G. J. Hildeman to R. Bachowski, D. A. Granger, J. E. Jacoby, R. A. Marra, R. J. Milauskas, and R. A. Raraser, "Proposed Mechanisms for HDC Header Failure and Coarse Dendritic Structure Formation", 1991-05-22). As discussed above LaBar emphasized the important role of oil absorption and pyrolysis. The work by Marra, see Section V.l.b, showed the importance'of magnesium reduction of the calcium silicate bond phase. A report by M. Kulak [Ref: ATC Report 52-85-09, M. Kulak, "The Prediction of Temperatures and Stresses in Molten Metal Marinite Header During HDC Casting at Warrick," 1985) described the development of finite element models to predict the temperature and stresses in the molten metal marinite material during casting. The results agreed well with experimental temperatures and strains measured during an actual production run. It was concluded that: "the failure of the MMM header cannot be explained solely on the stresses predicted by this analysis. This is consistent with the life expectancy of the header, which may vary anywhere from hours to days. The life of the header is dependent on maintaining the integrity at the comer overhang region. Failure in this region is related to both the thermomechanical stresses and the material degradation with time because of oil vapor penetration and chemical attack at the hot face. Any clamping damage in this critical overhang region will accelerate failure." The proposed failure mechanisms can be classified into three primary categories: physical / mechanical failure, chemical attack, and failure resulting from thermal profile changes. Physical failure is believed to occur when excessive oil soaks into the header and weakens the header to an extent that it cannot withstand the thermal stress caused by the nonuniform thermal field. Fragments of the header break away from the header overhang region. Molten aluminum freezes and sticks on the back side of the oil ring which results in tears and strawberries on the bottom and side surfaces of the ingot. Chemical failure results from chemical reactions of the magnesium and, to a lesser extent, aluminum with the calcium silicate bond phase in the header. This reaction gradually corrodes the header behind the oil ring. Molten aluminum freezes and sticks on the back surface of the oil ring or corroded header which results in tears and strawberries in the cast ingot CS/3273P 15 ATC 0035013 Thermal failure results from excessive oil absorption into the header. The oil pyrolyzes leaving a carbonaceous deposit in the header which increases the thermal conductivity. Molten aluminum freezes and sticks on the header due to a lower surface temperature. This results in tears and strawberries on the cast ingot. It is likely that all of the proposed mechanisms play some role in the header failure. A hypothesis for the failure mechanism was proposed in 1991 November (Ref: ATC Memorandum: M. G. Chu to G. R. Swanberg, "Possible Mechanism for HDC Header Failure," 1991-11-04). This hypothesis was developed based on the results of analyses of the content and distribution of carbon (oil and pyrolysis products) in the used headers, chemical composition at the header surface, and analysis of the subsurface structure and chemical composition of the ingot defects. It was proposed that as excessive oil soaks into the header, the oil pyrolyzes and changes the molten metal wetting characteristics of the calcium silicate material. The header is wetted by the molten metal and chemical reactions between the metal and header components occur. The header surface is corroded and the material is structurally weakened. Defects (microcracks, voids, surface roughening) form on the header near the header / oil ring overhang. The metal begins to freeze and stick in the defective regions. The solidified metal gets trapped in the solidification front and is pulled out by the moving solidified ingot This results in the commonly observed strawberries, tears and surface cracks. It is possible that fragments of the header are also pulled out by the moving ingot This results in larger surface defects and ultimately to the failure of the header. E. Header Thermal Expansion: Effect on Mold Package Clamping and Header Overhang The thermal expansion of the header material during casting can result in crushing the header in the mold package clamping region. This problem was more prevalent when dissimilar materials are used for the header and feedboard or when stronger materials such as the PMT board are used. A new clamp was designed which distributes the clamping forces more evenly on the header surface. This clamp also results in a closer alignment of the center points of loading between the clamp / feedboard contact surfaces and the oil ring / header contact surfaces. This clamp design resulted in the ability to utilize the stronger Pabco PMT and PMTHG materials by eliminating crack initiation sites and the crushing at the header / oil ring interface. CS/3273P 16 ATC The thermal expansion of the board material in the direction perpendicular to the pressing direction (during board production) is important since it effects the header / oil ring overhang. A low positive or negative thermal expansion coefficient results in the retention of the overhang as casting progresses. This behavior was observed with marinite and PVRM (PMTHG) headers. It is possible that the slight negative thermal expansion of these materials results in an increase in overhang due to material shrinkage and expansion of the metal components in the mold package. It was noted that the BNZ MetALform material which had a high positive expansion coefficient had the lowest level of lap pattern and the PVRM (PMTHG) compared most favorably with marinite which had the most prominent lap pattern. The lap pattern has been shown to correlate with increased mold life and recovery. CS/3273P 17 ATC 0035015 VI. FUTURE IMPROVEMENTS (CURRENT, PLANNED, AND PROPOSED ACTIVITIES! A. Casting Process Optimization PMTHG2.5 has been qualified as a marinite replacement material for HDC casting of 5XXX series alloys. Further improvement in mold life and ingot recovery may be obtained through optimization of the casting process parameters. As discussed in Section V, it is believed that the lubrication rate plays an important role in the casting process. The casting speed and water mold spray are also believed to have a significant impact on the casting performance. A DOX has been designed to evaluate the impact of these casting variable on the PMTHG2.5 mold life and recovery. The results of this trial will be used to optimize the HDC process parameters. The following matrix shows the proposed experimental design: Oil Flow (ml/min) H2O Mold Spray (g/min) Speed (inVmin) Low (-) 1 300 4.75 Midpoint 2.5 350 5.0 High (+) 4 400 5.25 The experiment will be designed to block the following variables: mold package, casting strand, temperature, bottom sprays, alloy type (43" 5082), filtration, skimming, metal head, start-up procedures, and people. The measurements will include casting life, recovery, and reliability. B. Materials Development 1. Red Mud Level While the PMTHG2.5 material is providing acceptable casting performance, it is not known what the optimum level of red mud should be. Pabco has fabricated the following tests boards:* * Pabco Metaltemp with 5% vermiculite and 2.5% red mud; * Pabco Metaltemp with 5% vermiculite and 5.0% red mud; and * Pabco Metaltemp with 5% vermiculite and 8.0% red mud. CS/3273P 18 ATC 0035016 ATC has conducted material evaluation tests (density, crushing strength, TGA, thermal expansion, and oil absorption). The results showed that these materials had similar cold crushing strengths and densities. There was a slight reduction in oil absorption with increasing red mud content Thermal analyses showed the materials had similar weight losses and small negative expansions from RT to 700C. HDC casting trials will be performed on materials which show the promise (based on material evaluation and comparison with existing materials database) for improved casting performance. 2. ' Calcium Silicate Board: Xonotlite Structure The primary phases in the Pabco PMT and PMTHG boards is the tobermorite and wolastonite forms of calcium silicate. The wolastonite is added to increase the hardness and strength of the board; it is a very stable form of calcium silicate. Tobermorite forms the board's matrix phase. This phase is a less stable form of calcium silicate. The xonotlite form of calcium silicate is a more stable phase and is expected to be more resistant to attack by the high magnesium alloys. This material may also have expansion characteristics that are closer to the molten metal marinite material. Pabco has agreed to produce boards with the xonotlite phase with the additives (5% vermiculite and 2.5% red mud) in the present PMTHG2.5 material. The initial production run was unacceptable due to a high amount of white spots (unreacted lime). Material characterization and casting performance trials on these boards will be conducted. 3. Red Mud / Vermiculite Ratio The vermiculite content has remained constant (5%) in all of the PMTHG boards that have been evaluated. The effects of varying the vermiculite content and the red mud / vermiculite ratio on the material characteristics and casting performance should be examined. C. Feedboard Cracking Feedboard cracking during and after machining has been an significant issue with the BNZ MetALform material. Although this has been less of a problem with the Pabco PMT material, some cracking was observed in this material during 1992 December and 1993 January, [Ref: Warrick Memorandum: M. J. Dunlay to Memorandum, "PMT Feedboard Cracking," 1993-01-15]. Tests were conducted at Warrick and ATC to determine the effects of water absorption (humidity changes) on the board's weight gain and dimensional stability. It was concluded that the PMT material required about a 7 day time period to stabilize to different CS/3273P 19 ATC humidity / temperature conditions [Ref: Warrick Memorandum: W. Lansdale to M. J. Dunlay, "HDC Feedboard Cracking (Metaltemp) Update," 1993-02-12]. It was also found that the cracking could be reduced by decreasing the machining feed rate by 30%. Pabco will continue to investigate their manufacturing process in an effort to determine the root cause of the cracking problem. There was been extensive works at Apex to develop an in control and capable machining process. The specific activities include: Establishment of control charts on critical dimensions; Resolving issues with the machine manufacturer on laser calibration of arms, board tuning to appropriate frequencies, establishment of circle-diamond-square test, etc.; Program changes and adjustments of machine settings to improve "out-of-control" comer conditions; A 43" mold template has been supplied to Apex by Alcoa which will be used as a standard for the establishment of a tracking system; Testing to determine if improved machining accuracies can be obtained by using a "coordinate" based program instead of the "incremental" based program; Continuing to develop measurement capability; and Establishing SOP's for all header, feedboards, and key sizes. D. Pepdgr_gy.srhang The header-oil ring overhang plays an important role in the casting process. It has been proposed to further examine the effects of the overhang dimension on the casting performance. E. Mold Shop Consistency Programs recently initiated in the mold shop have had a positive impact on the consistency of mold packages delivered to the casting complex. These include tracking mold inside diameters, input actual data versus average data, and performing audits of the mold buildup process. CS/3273P 20 ATC 0035018 VII. CONCLUSIONS A replacement material for header grade molten metal marinite has been developed, evaluated, and qualified for the HDC casting of 5XXX alloys. The key elements and activities which contributed to the success of the program were: Warrick / ATC Teaming: An excellent cooperative team was established with ATC and Warrick personnel. Well Defined Project Objective: A technical objective which was consistent with Warrick production requirements was established to provide a measure of project success. Analysis of Header Materials: The use of analytical techniques to determine the header degradation mechanisms and role of the header material chemistry on casting performance led to the development of potential material formulations for the replacement material. Close Vendor Working Relationship: An excellent working relationship was established with Pabco. The program would not have been successful without the commitment of Pabco to assist in the development of the replacement material. Materials Screening: With the cooperation of Warrick's ingot production, a materials screening test was performed on the potential replacement materials. Through the combined efforts of Warrick and ATC personnel, the analysis of the materials screening test led to the selection of the PMTHG material. Process Modification / Optimization: The development of standard operating procedures by Warrick personnel gready aided the evaluation of die potential replacement materials. The understanding of the mold package functionability which led to process modifications such as a redesigned clamp was critical to the success of the program. Experimental Design and Statistical Analysis: The use of designed experiments and statistical analysis provided the basis for downselecting the potential material formulations. Statistical methods were further utilized to optimize the material formulation, mold package design, and casting process variables. The systematic approach used to evaluate and qualify the replacement materials provided confidence that the materials could be successfully implemented in production. CS/3273P 21 ATC 0035019 While the PMTHG material has been qualified as a header for the casting of 5XXX alloys, material refinements and HDC process optimization may lead to further improvements in casting life, ingot recovery, and ingot quality. Planned and proposed activities include: Optimization of the red mud content; Optimization of the red mud / vermiculite ratio; Evaluation of boards with xonotlite as the matrix phase; Optimization of HDC process variables Qubrication rate, mold spray, casting speed); Evaluation of the effect of header overhang on casting performance; and Resolution of the feedboard cracking problems. CS/3273P 22 ATC 0035020 (iSd) Ml6uaiis Msmo p|oo Cold Crushing Strength 95 % Confidence for Mean Figure 1 Header material cold crushing strength. ATC 0035021 (iSd) aJnjdny jo sninpoyy Modulus of Rupture 95% Confidence for Mean Figure 2 Header material modulus of rupture. ATC 0035022 UIBO U|B!0M % 95% Confidence for Mean Figure 3 Header material oil absorption. ATC 0035023 O e> co (Do/9"0 !) luajojweoo uoisuedxg Thermal Expansion to 700 95% Confidence for Mean Figure 4 Header material thermal expansion. ATC 0035024 CONFIDENTIAL EVALUATION OF CANDIDATE MATERIALS FOR HDC INGOT CASTING HEADER APPLICATION AT WARRICK OPERATIONS By M. W. Vance Ceramics Division Alcoa Technical Center Alcoa Center, PA 15069 1985 August 27 Report No. 03-85-47 Copy No. / CONFIDENTIAL Executive Abstract This report summarizes a characterization program intended to improve understanding and develop a data base for selection of new header materials for the HDC casting process. These efforts have been guided towards determining key properties of currently used asbestos containing Molten Metal Marinite, and in conjunction with using finite element analysis techniques to define regions and magnitudes of stress in the header, provide rationales for selection of substitute materials. These inputs were shared on a routine basis with Warrick Operations personnel to insure that testing and design study results were of immediate value and then trial results at Warrick used as a feedback to guide further testing and analysis. To date, key properties and design parameters have been successfully identified and two new calcium silicate bonded candidate materials recommended for field trials. Future ATC technical service activities will include limited testing or test recommendations, post mortem analysis and coordination with HDC operations personnel, suppliers. Purchasing and Engineering. Active participation in the HDC header task force activities will continue until substitute materials are successfully introduced into the HDC processes. MWV:PAN:8 ATC 0035026 TABLE OF CONTENTS Executive Abstract Page No. 1.0 INTRODUCTION ...................................................................................................... 2.0 MATERIALS DESCRIPTION .................................................................................... 3.0 TEST DESCRIPTION .............................................................................................. 4.0 DISCUSSION OF RESULTS ................................................................................... 4.1 Bulk Density and Strength .................................................................. 4.2 Bulk Density, Thermal Properties, and Oil Absorption..................... 4.3 Thermal Dilation and Stability ................................................................. 4.4 Differential Thermal Analysis(DTA)............................................................ 4.5 Load Relaxation Studies.................................................................................. 4.6 Molten Metal ImmersionStudies ................................................................... 5.0 CORRELATION OF TEST RESULTS AND DESIGN ANALYSIS ............................ 6.0 PETROGRAPHIC EXAMINATION OF NEW AND USED HEADERMATERIALS ............. 7.0 DESCRIPTION OF WARRICK TRIAL RESULTS ..................................................... 8.0 OVERALL SUMMARY AND RECOMMENDATIONS ....................................................... REFERENCES ........................................................................................................... TABLES .................................................................................................................. FIGURES ............................................... APPENDIX- SEMS of New and Used Header Materials................................. 1-3 3-7 7-10, 10 10-11 12-14 14-16 16 16-17 18 18-22 23-24 24-25 25-28 30-33 I-IX 1-18 A-l-20 MWV:PAN:8 ATC 0035027 1.0 INTRODUCTION During 1984, a decline in the quality and threatened availability of asbestos containing Molten Metal Marinite (trade name for Manville Corporation's calcium silicate board) used for HOC headers prompted a renewed ATC program**" to evaluate new materials for that application. The primary goals of this program were to initially perform materials characterization and design studies to understand how Molten Metal Marinite functions successfully, then to characterize alternate materials and conduct design and field studies to fit them into the HDC process. Hopefully this will lead to successfully finding more than one non-asbestos substitute to minimize sole-source dependency on one header material supply. Additional benefits of this program are the potential applications of these new materials for headers in HDC and FDC casting processes at other plants and for improved headers used in FDC casting aluminum- lithium alloys. If acceptable for the more rigorous header requirements, candidate materials could also qualify for use in molten metal containment applications such as troughs, floats, basin linings, and dams. This report describes the materials testing program that has been underway during 1984-1985 and Its relationship to the design studies performed on the instrumented header/assembly at Warrick.*7 ' 8 The understanding of how header grade Molten Metal Marinite performs in this application can then be applied to more effectively substitute alternate materials. Many discussions with Warrick personnel and field observations conducted along with ATC studies have resulted in a checklist of important parameters that must be considered in g relation to specific header requirements. These parameters are: * See references. MWV:PAN:8 ATC 0035028 1. Available in reproducible large board sheets which are free of warpage, cracks, voids and delaminations. The internal structure of the material i must also be homogeneous without excessive concentration of reinforcing materials or soft zones (white spots). 2. Good machinability to form a smooth, defect-free, as-machined surface having acceptable chip and crack resistance. ; 3. Adequate strength for handling and installation. 4. Adequate strain capability to deflect when mating with the mold. Through-thickness compressive strength must be capable of withstanding clamping loads. 5. Minimum permeability to, and solubility in casting lubricants. .6 Acceptably low thermal conductivity and thermal diffusivity to minimize chill during start-up. 7. Acceptable thermal shock resistance. Acceptable resistance to cracking and chipping in the overhang region. 8. Relatively predictable and stable long-term dilation behavior (low permanent shrinkage In all directions) as a result of elevated temperature exposure. Low thermal expansion coefficient to minimize stresses. MWV:PAN:8 2 ATC 0035029 9. No harmful outgassing when contacted by molten metal. 10. Adequate corrosion resistance and nonsticking in magnesium containing molten metal. By combining efforts with Warrick personnel, all of these parameters have been q addressed. The materials specification for MMM-HG,* given in Table I, addresses a few of the aforementioned parameters, but needs more elaboration from an engineering viewpoint. Therefore, it is also a major goal of this work to produce more complete guidelines for HOC header materials. 2.0 MATERIALS DESCRIPTION The following materials have been included in our evaluation program. Supplier/Material Description Manville/Molten Metal Marinlte Regular Grade (MMM-RG) Asbestiform fibers (amosite) bonded with a hydrated calcium silicate phase (amorphous tobermorite). * See references MWV:PAN:8 (continued on next page) 3 ATc 0035030 List of header materials evaluated - continued - Supplier/Material Description Manville/Molten Metal Marinite Header Grade (MMM-HG) Increased fiber content over MMM-RG bonded with the same hydrated calcium silicate chemistry. Heat treated at 1100F (593C). Manville/Metal Mover (Old) Manville/Metal Mover (New) (A4.3-1 and A3.1-4) Alkali resistant glass and wollastonite (CaSiOg) fibers bonded with a hydrated calcium silicate phase (amorphous tobermorite). Heat treated i at 975F (524C). Alkali resistant glass and wollastonite (CaSIOj) fibers bonded with a hydrated calcium silicate phase (amorphous tobermorite). Heat treated, at 975F. Manville/XM-4 and XM-6 Alkali resistant glass and wollastonite fibers bonded with a hydrated calcium silicate phase (amorphous xonotlite). Heat treated at 975F. MWV:PAN:8 (continued on next page) 4 ATC 0035031 List of header materials evaluated - continued - __________ Supplier/Material Description ________________________ Nichias/Pyrotek/N-14 Wollastonlte and E-glass f.lbers bonded with an amorphous hydrated calcium silicate binder (xonotlite). Not heat treated. Nichias/Pyrotek/L-16, L-17, L-18, and KX Wollastonlte and carbon fibers bonded with an amorphous hydrated calcium silicate binder (xonotlite). Density is decreased from L-16 to L-17 to L-18 to KX. Not heat treated. Pabco/32 Fiberglass bonded with an amorphous hydrated calcium silicate (xonotlite), bentonite and Portland cement. Autoclaved at 400F (204C). Not heat treated. Alcoa Technical Products Division/ Pertnatech Sigma (R-680) Calcium aluminate cement bonded fused silica castable with a nonwetting additive. Fired at ~1500F (816C). MWV:PAN:8 5 ATC 0035032 The reinforcement phase of Molten Metal Marinite (MMM) consists of amosite fibers which are chain-silicates (amphiboles)*1^ of the composition (Mg,Fe+3)^ S1g022(0H)2 (more correctly, these fibers can be designated Cummingtonite which is the magnesium-rich composition). Amosite fibers are basically ash-grey in color, with a coating of Iron oxides that give the fibers a reddish-brown color after the board Is heat treated at 975F. The hydrated calcium silicate binder phase is not composed of fully crystalline tobermorite*1^ (plate crystals of the composition 5-6CaO*651027* 5H20) but an amorphous gel phase (l.S-Z.OCaO.SiO^.nHgO).*11'1^ MMM is formed by filter-pressing an aqueous slurry of asbestos fibers, lime, and silica and then autoclaving at 100 psig/330F (166C) to form,'the pseudo tobermorlte hydrated calcium silicate phase* 13 which rigidlzes the board. The asbestos fibers, if too short, make the resultant board excessively brittle. Long asbestos fibers improve board toughness but result in homogeneity problems (laminations) if fiber length becomes excessive. Asbestos fibers in excessively large, unbroken bundles (crudy fiber or pencils) are considered defects. The ability of asbestos bundles to be broken down into minute 1 urn fibers gives the soft, as-formed board acceptable durability and after autoclaving, develops a lightweight, tough, machinable material. Post-drying*14 or heat-treating at 1100F (593C) is performed to remove excess water. Because the amosite fibers dehydrate at 1200-1355F (650-735C), they Impart reasonable elevated temperature stability to the board over a wide range of temperatures. Most of the shrinkage noted in MMM board products at temperatures above 1400*F (760C) is attributed to substantial dehydration of tobermorite gels. XonotlIte-type calcium silicate * See references MWV:PAN:8 6 ATC 0035033 binders tend to shrink less because of the lower amount of hydrated water present in the binder structure (SCaCNSSiC^^O). A number of reinforcement materials have been considered as substitutes for amoslte; for example. E-glass, mica, cellulose (kraft), alumino-si 11 cate ceramic, and wollastonite (calcium silicate) fibers. None of these materials either by themselves or combined, when mixed with hydrated calcium silicate binder would have the same combined green and fired strength, fired toughness, and homogeneity as MMM for header applications. The highly anisotropic properties of these board materials are related to the preferred orientation of the longitudinal axes of high-aspect ratio (length to diameter) fibers or fiber bundles parallel to the flat board surfaces. Therefore, special attention to a candidate material's anisotropy was given during the sample preparation, testing, design analysis and evaluation of data. 3.0 TEST DESCRIPTION The various characterization and evaluation tests and rationales for using them were previously described.*4 ' 5 * 6 An update of these tests is given below: Test Parameters___________ ___________ Rationales o Density, Density Gradient, and Durometer (Type D) Correlate density and material homogeneity with firmness (Durometer). Density may also be correlated with heat transfer properties and strength. MWV:PAN:8 (continued on next page) 7 /: V ATC 0035034 Test Parameters Rationales o Modulus of Rupture, Strain to Failure and Modulus of Elasticity at Room and Elevated Temperatures (1000F-538C) Determine; strength levels at room temperature and at elevated temperature using a three-point bending method. Modulus of elasticity may be used to calculate maximum deflection allowable for clamping headers onto the mold package. o Compressive Load Relaxation Test at Room Temperature, 1100F (593C) and 1400F (760C) Determine the extent of stress relief for different materials obtained when the;header is clamped (prestressed) on the mold and is heated during the process. o Overhanging Ledge Compressive Load Test at Room Temperature Investigate the chipping and cracking resistance of as-received and treated materials by simulating the stress-riser effect in the header's overhanging lip region. Treated specimens were subjected to oil soaking and coking at 1100 and (continued on next page) MWV:PAN:8 8 ATC 0035035 Test Parameters Rationales 1400F. MgC^ was added to the coke to determine the effects of Mg or Cl2 on header strength. o Modified "B" Immersion in 5182 at 1400F (760C) Determine wetting, corrosion resistance, dimensional stability, and crack resistance of candidate materials. This is employed as an initial screening test for unknown specimens. o Dilation Measurements as a Function of Increasing Temperature Determine the thermal stability of specimens as a function of anisotropy (measured parallel and normal to the board-pi ane). o Hot Oil Tests and Coking Tests Determine the extent of oil saturation and effects of oil and coking on the hardness and compressive strength of a material using the compressive ledge test previously described.. (continued on next page) MWV:PAN:8 9 ATC 0035036 o Differential Thermal Analysis (DTA) combined with Weight Loss Measurements Determine the temperature of dehydration reactions and weight losses occurring as a function of increasing temperature. Relate this information to possible outgassing in service. o Microscopic Examination of New and Post-Mortem of Used Header Materials Use the scanning electron microscope (SEM) to identify morphological structure of new and used materials and relate properties and performance to changes in the micrestructure. 4.0 DISCUSSION OF RESULTS 4.1 Bulk Density and Strength The overall HOC header grade board requirements are given in Table I. Adequate strength and strain to failure are needed for handling resistance, integrity In clamping to form a seal, and resistance to chipping at the overhang. See comparative values. Table II. The design of the header and mold assembly are shown In the cross-section view. Figure 1. During installation, the clamp rotates on its vertical appendage and pinches the header against the oil ring (Figure 1). On the other hand, excessive density accompanied by Increasing strength is counterproductive if thermal conductivity Increases occur. Elevated levels of open porosity (lower density) would increase the amount of oil absorption. MWV:PAN:8 10 ATC 0035037 Generally, it has been observed that excessive chipping occurs for header grade material at the overhang if the ratio of amosite fiber to C-S binder becomes too low or the length of fiber bundles becomes too short. This could be attributed to loss of toughness which would be coexistent with a prevalence of C-S binder. The morphology of the amosite flbers/C-S binder indicates a relatively high fiber/hinder ratio. A high strain-to-fallure is also desirable when clamping the header to conform to a mold which may have warped during or after service. For example, a flexural deformation of the order of 1/8 to 3/16 inch/foot of length is obtainable with header grade MMM before cracks are initiated during installation. Strength increases were noted at 1000F (538C) for all the C-S bonded materials except the 1-16, L-17, L-18, and KX materials. This may be associated with the oxidation of carbon fibers. The KX and Pabco/32 materials were dropped from further consideration because of relatively low strength and crushing resistance. The Permatech Sigma exhibited a significant increase in strength but decreased maximum deflection at 1000F. This would indicate a greater sensitivity to cracking if distortions occurred In the mold or temperature gradients become too severe during service. Past results of Permatech Sigma field trials in HOC casters at Warrick and at the Arnold Research facility have indicated that chances of immediate success using the board-type materials would be greater at this time. Further HDC trials using Permatech Sigma have been deferred at Warrick but are currently being pursued at Alcoa's Tennessee operations using a redesigned clamping system. MWV:PAN:8 11 ATC 0035038 4.2 Bulk Density, Thermal Properties, and Oil Absorption Generally, heat transfer properties of the various materials are related to composition, bulk density, porosity, and pore size and morphology. Values for thermal conductivity and diffusivity for various candidate header materials are given in Table III.*13 ' 15 Materials having high thermal conductivity and diffusivity are more likely to chill molten metal. Comparing those values at 600 (316C) and 1000F (538C), thermal conductivity usually increases with temperature for the C-S bonded materials. If pores of the header were saturated with casting oil (AA Standard), which is mostly composed of castor oil having a relatively high density (60.1 Ibs/ft ) and heat capacity (0.434 Btu/lb-F) the density and thermal conductivity of the header would markedly increase and substantially Increase the possibility of chilling molten metal. The relative absorption of lubricating oil** by the various candidates is shown in Figure 2. The lowest weight pick-up for a header material saturated with warm AA standard lubricant was exhibited by MMM-HG. The ratings for the various materials according to lowest to highest degree of pick-up are: MMM-HG, MM, SFX, XM, N-14, L-18 and L-16. Because a definite relationship exists between the extent of oil pick-up of the header material and casting performance, a number of coatings and impregnants were evaluated in prior work*17 for purposes of decreasing corrosion, oil pick-up and moisture pick-up. Silicone resin coatings and impregnants were found to be the most * See references. MWV:PAN:8 12 ATC 0035039 promising in decreasing oil absorption and Al^O^, TIOj and BN pigments were added to the silicones and applied to the header for improved molten metal resistance. However, during casting trials, exterior coatings were not effective in decreasing oil absorption or corrosion. The most promising materials easily impregnated the header material and provided an "in-depth" oil repulsion. Recent studies at ATC involved the impregnation of header materials with Dow Corning 805 which effectively decreased oil absorption. However, this effect declines when the header is heated up to molten metal temperatures (1200-1300F). Some protection may be provided in lower temperature zones through the cross-section of the header shown in the computer plot. Figure 3. At temperatures above 800F in air, silicones tend to break down to form silica and smoking may become a problem area. For most candidate materials, no reproducible change in maximum compressive load resulted from oil saturation, although the compressive stress (psl) as a function of strain (inch/inch) plots (Figures 4 and 5) show a pronounced change in slope. This could be attributed to an initial collapse of the pore structure at the onset of loading followed by continued deflection to final failure. Only the compressive strength of L-16 seemed to be affected by oil saturation. The data given In the next table lists strength values determined for the two conditions: MWV:PAN:8 13 w ATC 0035040 Material As-Received Initial Stress Crushing Stress* 011-Soaked Initial Stress Crushing Stress MMM-HG MM(A4.3-1) XM-4 N-14 L-16 L-18 1300 1200 1000 1425 750 950 2475 2130 1475 2363 3325 2400 1400 1250 1200 1400 400 1100 3150 2100 1600 2050 2100 1850 * Stress-psi .Coking the oil at 1400F resulted in significant decreases in strength for header grade, MM, and XM materials. The 1-16 and N-14 materials retained a major portion of original strength. The compressive loading was performed on specimens with a 0.040 inch overhang to simulate the header/oil ring design. The as-received and oil soaked MMM-HG, N-14, L-16, and L-18 materials all withstood loading to failure (crushing) without chipping of this overhanging portion even when coked at 1000F. The MM and XM chipped after 1000F exposure. None of the materials had suitable chip resistance or strength when coked at 1400F or retained compressive strength when exposed to the coke/MgCl2 mixture at 1000 and 1400F. This loss of bond strength was attributed to chlorine pick-up. 4.3 Thermal Dilation and Stability Thermal dilation measurements 18 ' 19 were performed on candidate header materials to characterize their behavior at elevated temperatures. As MWV:PAN:8 14 ATC 0035041 previously described, header grade MMM has composite-like properties because of the degree of preferred orientation of asbestos fiber bundles parallel to the board plane (upper and screen-side surfaces of the board).. Therefore, two separate dilation measurements were made on specimens whose orientations were parallel and normal (perpendicular) to these surfaces. A typical plot of % linear change as a function of temperature for header grade MMM is shown in Figure 6. In-plane expansion occurs from 32-1292F (0-700C) followed by a dramatic shrinkage (2.OX) from 1292-1472F (700-800C). This marked shrinkage also occurs at a direction normal to the board plane and there is very little net expansion. This behavior seems to be related to most of the calcium silicate (C-S) bonded materials, which is shown in dilation plots (Figures 7 through 11). The tobermorite-type materials (MMM-HG and MM) exhibited greater shrinkage than most of the xonotlite-type materials (N-14, L-16, L-17, L-18, and KX). Anomalies are the non-C-S bonded Zlrcar and Permatech Sigma materials which are processed at temperatures above the molten metal use temperature (ZIRCAR) or composed of a calcium aluminate cement bonded aggregate (Sigma) respectively which exhibit a continued increase In thermal expansion. Typical values for average thermal expansion coefficients are given in Table IV and are useful for comparing the anisotropic nature of the candidate materials. For design studies, incremental values of dialation as a function of temperature must be considered if non-linear expansion behavior occurs. During the casting process, the shrinkages exhibited by the C-S bonded materials will tend to offset (relieve) stresses caused by thermal expansion and clamping. However, excessive shrinkage normal to the board plane could completely break the desired seal between the oil ring and header directly causing defects in the as-cast Ingot. \. MWV:PAN:8 15 ATC 0035042 Understanding material characteristics such as dilation is important because bolt-torques and clamping movement required to maintain effective sealing through the use of Belleville washers may change with new materials. Dilation effects are modified considerably when the material is subjected to compressive stresses induced by clamping loads and relaxation occurs. 4.4 Differential Thermal Analysis (DTA) Because of the tendency for C-S bonded board materials to pick up moisture in humid atmospheres and contain chemically bonded water associated with the calcium silicate and asbestos, DTA and weight loss versus temperature measurements were performed*20 to determine the temperatures where these reactions occur (see Table V). Most of the materials contained up to 3% mechanical water (adsorbed) and when heated at temperatures of 1292-1472F (700-800C), significant amounts of chemically bonded water continued to evolve. Reactions observed for L-16 and L-18 were attributed to the oxidation of carbon reinforcing fibers. Weight losses were detected for samples heated up to 2012F (1100C), which is well beyond the process temperatures encountered. 4.5 Load Relaxation Studies To determine the extent of compressive load relaxation for various conditions, various specimens of candidate header materials were subjected to two different load levels (400 and 800 psi) at room temperature, 1100F (593C) and 1400F (760C). The results of these tests are summarized in Table VI. MWV:PAN:8 16 ATC 0035043 At room temperature, all the candidates relaxed less than 101 even at higher loads (see Figure 12). Increasing test temperature to 1100F (below the C-S dehydration temperature) the MMM exhibited lower levels of relaxation than the other candidates at 400 psi (see Figure 13). Increasing the load to 800 psi increased the relaxation for all candidates except L-18. At this point, it is suspect that some materials may exhibit crushing or initial failure and therefore relax less for the higher load (800 psi). Comparing test results for the 400 psi loads, tobermorite bonded materials relaxed substantially at 1400F which is in the range of the dehydration temperature 1298-1472F (700-800C). This is further evidenced by an almost total relaxation of load for MMM-HG and MM (A4.3-1). It was surprising that the Manville xonotlite materials (XM) also relaxed almost completely at 1400F (see Figure 14). This correlates with the dilation behavior and shows that excessive temperatures for these materials may also result in loss of seal between oil ring and header. The Nichias/Pyrotek materials did not relax as readily at 1400F which is attributed to lower shrinkage and maintenance of load carrying ability at that temperature. These materials tend to relax more than MMM-HG at 1100F, and the continued use of Belleville washers and 150 in-lb bolt torques is a reasonable requirement for Initial installation trials. The relatively high relaxation at temperatures of 1400F for MMM, suggests that the more stable xonotlite bonded Nlchlas candidates would offer better sealing for high temperature swings in the process, but may require slightly greater torques for sealing at nominal temperatures. MWV:PAN:8 17 ATC 0035044 4.6 Molten Metal Immersion Studies Bars of candidate header materials were immersed in molten 5182 alloy at 1400F (760C) for 3.5 and 24 hours*^ to determine the effects of intimate molten metal contact on their stability and corrosion resistance (see test results. Table VII). Of recent materials considered, L-18 had superior resistance to shrinkage after exposure to 1400F. The XM board showed a marked improvement over MMM-HG. Except for the color change of the MMM-HG, none of the materials described in Table VII were wetted or reacted by the 5182 alloy. This test by itself was not sufficient cause to warrant acceptance or rejection of a material because of its excessive temperature over that commonly experienced in ingot casting (1220-1360F) (660-738C) (see Figure 15) and because of the isothermal exposure in lieu of hot-face exposure as actually experienced by the header. However, the sensitivity of MMM to elevated process temperature swings is shown by these tests. Overall, results of this testing closely agrees with diiation data formerly described because corrosion did not seem to be an obvious factor. 5.0 CORRELATION OF TEST RESULTS ANO DESIGN ANALYSIS Prior to conducting finite element analysis to determine the levels of temperature and stress in a Warrick HOC header*7 ' o * 22 , a header and clamping mechanism was instrumented with thermocouples and strain gauges to measure ten^eratures, mechanical strains and thermal-mechanical strains during phases of: installation of the header on the mold, assembly of the mold and basin. MWV:PAN:8 18 ATC 0035045 preheat, and start-up of casting and steady-state operations. Typical temperatures at a given location in the header is shown in Figure 16. To summarize this work, the following conclusions are related to results from materials tests on MMM: 1. Consistent with room temperature relaxation tests, very little clamping load relaxation occurred when the header was applied to the mold. 2. The tendency of the hot face of the MMM header is to contract when exposed to metal contact. This is attributed to the low or negative through-thickness expansion. 3. Without the compressive clamping loads, the thermal gradients in the header during start-up and operation induce zones of tension at the hot face (see Figures 17 and 18). This too is related to the negative through-thickness expansion. 4. Surface flaws oh the hot face tend to open during start-up in the absence of clamping compressive loads. The negative through-thickness expansion influences this. Flaws are related to the laminar planes of fiber orientation. 5. Clamping loads Induce compressive stresses in the header at the hot face and header/oil ring contact region. This compressive load opposes tension (from thermal effects) in the header. (These loads relax with excessive temperature.) MWV:PAN:8 19 3 ATC 0035046 6. Nonuniform clamping loads Introduce differences in clamping stresses and axial restraint around the header circumference. Excessive clamping loads in the critical overhang region (a stress-riser effect induces 3-4X average stresses) may cause localized damage. This may be in the form of cracks or crushing. Damage can be minimized by distributing clamping loads over a greater area In lieu of the point contact clamp design currently used. The use of the resilient Fiberfrax gasket between header and back-up board also decreases stresses. 7. Clamping loads are expected to decrease with time because of thermal expansion in the bolts and MMM header contraction in the clamping direction. Use of a material having positive through thickness expansion may result in higher stresses and cracking if torqued at present levels. 8. Extremely high thermocouple measured temperatures at the top corners and sides of the header indicate nonuniformity of cooling or combustion of the lubricant around the perimeter of the header and explains the localized damage to the experimental header during start-up. Higher thermal expansion materials would be more vunerable to damage than lower expansion materials. 9. Fluctuation In metal temperature with time cycles stresses in the header material, accelerating material damage (through crack propagation) accumulated through clamping and start-up conditions. MWV:PAN:8 20 ATC 0035047 10. Thermal expansion tests have shown severe dimensional contraction in the MMM at temperatures above 1300F. Recorded metal temperatures fluctuating above 1360F as previously shown could cause tensile cracks to occur in the hot face. 11. Combined effects of corrosion at the hot face, material shrinkage, and oil penetration and temperature effects in the header in the overhang region may cause material degradation and accelerate failure. This has been observed to be more of a problem area associated with magnesium containing aluminum alloys (5182). Chlorine pick-up also causes deterioration of material strength. Using these findings to determine the selection of materials, the following comparisons can be made: 1. Tobermorite-type C-S bonded materials (MWi-HG and MM) tend to exhibit negative expansion coefficients in the through-thickness direction and therefore form tensile stresses on heat-up. Materials having positive expansion coefficients are In compression on heat-up. XM, L-16, L-17, L-18, Zlrcar, and Permatech Sigma all have positive through-thickness expansion coefficients. The marked shrinkage in MMM-HG and MM observed at 1292-1472F (700-800C) versus lesser shrinkage means more load relaxation and higher temperature stress relief than the other candidates. At lower mean temperature (1100F), MMM exhibited the lowest relaxation of any material tested. The higher expansion coefficient materials such as Zircar could develop excessive stresses in the header application. MWV:PAN:8 21 ATC 0035048 2. Excessive positive through-thickness expansion may result in chipping of the header for materials where shear or compressive failure results from excessive clamping loads. Cracks will not tend to open on the working face during heat-up. 3. The XM did not exhibit a marked through-thickness and in-plane shrinkage at 1292-1472F (700-800C) to the same degree as MMM-HG and MM. L-16, L-17, and L-18 showed a marked improvement in stability contributing to a decreased tendency for cracking and loss of seal between header and oil ring if process temperatures become excessive. 4. Corrosion of nonasbestos containing headers in magnesium containing alloys is improved for excessive temperatures. Oil soaking and coking does not seem to deteriorate mechanical properties of nonasbestos materials but does increase heat absorption and heat losses which could effect thermal properties and cause undesirable chilling of molten metal on the header in lieu of solidification at the oil ring. The early Nichias L-16 and L-17 materials both had exhibited excessive oil saturation. The L-18 material showed decreased oil saturation but was still at higher levels than the XM and MM materials which were on a par with MMM-HG. Pick up of chlorine was associated with a loss of strength for all the calcium silicate bonded materials. MWV:PAN:8 22 ATC 0035049 6.0 EXAMINATION OF NEW AND USED HEADER MATERIALS The composition of the various candidates were described in the initial section of this report. The purpose of this work was to examine fracture surfaces of new and used materials to determine the relative morphology before and after use and to improve understanding of the basic properties of the materials. Header materials are composites with the hydrated calcium silicate binder phase reinforced with one or more fibrous materials. A minimum amount of fibers are added to improve toughness and decrease chipping and cracking. If excessive amounts of fiber are added a condition could exist where too high a concentration of fiber of uniformly large diameter and length can create a material of high compressibility and lowered bending strength because of inefficient bonding and excessive void formation. The Information in the Appendix provides detailed comments on scanning electron photomicrographs (Figures A-l through A-17) of new and used candidate header materials. The structure of MMM-HG Is composed of chopped bundles of asbestos fibers (Figure A-l) that on close examination develops a material structure that is largely controlled by these bundles. It is apparent that some of the alternate materials tested are designed to duplicate the properties of asbestos fiber reinforced Mifi-HG by adding a substitute fibrous material. However, the matrix (hydrated calcium silicate) controls the material structure in these alternate materials because comparable levels of long, large diameter substitute fibers which are used In lieu of asbestos would tend to increase compressibility of the board and also Increase oil permeability. The more open network of crystalline Xonotllte that make up the bond of L-17 and L-18 could also contribute to excessive permeability. When MMM-HG Is 4- MWV:PAN:8 23 ATC 0035050 installed or clamped on the mold, the fine bundles of asbestos fibers serve to block these open channels and therefore reduce oil permeability and tend to act as a more efficient "gasketing" material. This blocking tendency would increase as the header is compressed during clamping and the bundles are broken into minute fibers that more efficiently fill void spaces. Special attention is given to placing the "screen side" of MMM-HG against the oil ring because of the higher levels of finely divided fibers present in the surface layers of that surface are more suitable for sealing. In all cases the used header materials are altered in the immediate zone that was exposed to molten metal. This zone seems to be denser and less of the original binder morphology is obvious. Reinforcement is almost completely depleted in the exposed surface of the carbon fiber containing L-17 material. This would result in the decreased hot strain-to-failure values observed for the Nichias materials (with the exception of N-14). Other than the densification observed, no obvious cracking, reaction or melting seems to have taken place in the used materials examined. Carbon pick-up related discoloration occurs, but no marked build-up of deleterious material was obvious. A continued post mortem examination of used header specimens should continue as part of this program. 7.0 DESCRIPTION OF PLANT TRIAL RESULTS in 11 12 14 To date, Permatech Sigma, Metal Mover, N-12 , N-14 ' , L-17 ", and XM materials have been evaluated in plant trials. Table VIII reviews the results of these trials. Trials with Permatech Sigma, N-12, N-14 and L-17 were not successful. MWV:PAN:8 24 ATC 0035051 Trial results to date Indicate that the XM material has currently proven to show the most promise. The new L-18 material, with decreased oil absorption over L-16 and L-17, may also prove to be a viable candidate. (The L-17 trial was deemed unsuccessful because of excessive oil pick-up.) Although Metal Mover showed an acceptable trial performance, processing problems (white spots) prevented it from being reliably produced and from being commercially available. The same Manville Corporation process equipment is being used for making MMM-HG, MM and XM. 8.0 OVERALL SUMMARY AND RECOMMENDATIONS The requirements for an acceptable HOC header material are fairly complex because of the number of material and process variables Involved. The primary goals of this work were to: (1) understand how MMM-HG functions In the HOC process, (2) characterize candidate materials and develop a database, and (3) recommend substitute materials based on combined Information from design analysis, material properties, material post mortem analysis and field experience. The desirable properties of MMM-HG are: ,1 1. Available in suitably homogeneous (macro) board or shape form. If in board form, the material must be readily machinable into a header shape within required tolerances. Machined surfaces should be relatively MWV:PAN:8 25 V ATC 0035052 smooth to minimize friction (actual surface roughness should be characterized). The working surface is often lightly sanded with emery paper to obtain the required surface texture. 2. Nonwetted and limited corrosion by magnesium containing alloys up to temperatures of 1300F (~700C). 3. Suitably low density to minimize heat losses and heat transfer from molten metal (for HOC process conditions). 4. Low permeability, absorption, and reaction with lubricating oil. 5. Limited and reproducible compressibility and load relaxation perpendicular to the board plane at room temperature and at operating temperature (=1100F). 6. Low thermal expansion in the in-plane directions to minimize cracking when constrained in the HDC mold. Near zero or negative thermal expansion perpendicular to that plane. Predictable and uniform dilation behavior throughout the process temperature range. Limited use beyond 1300F (~700C). 7. Acceptable thermal shock resistance (combined low coefficient of thermal expansion, low elastic modulus, and high strength). MWV:PAN:8 26 ATC 0035053 8. Adequate strength and toughness, low Poisson's ratio to effectively maintain a stable overhang when compressed against the oil ring. Adequate resistance to chipping during service in the HDC process. A summary comparing the suitability of various candidate materials for HDC headers is given below: SUMMARY: Parameters Analysis of Various Aspects of Suitability for Various Candidate HDC Header Materials ____________________ ______ Legend o = unacceptable / - acceptable Materials * no comment or unknown Permatech MMM-Hq MM XM N-14 L-18 Sigma 1. Available in homogeneous board form. Must be machinable into or available in accurate header shape. / 2. Non wetted and limited corrosion by Mg-Al alloys up to 1300F. // / / 3. Suitably low density to minimize heat losses and heat transfer from molten metal (HDC process). // / / 0 4. Low permeability, absorption and reaction with lubrication oil (HDC process). / 0 (continued on next page) MWV:PAN:8 27 ATC 0035054 SUMMARY: Parameters Analysis of Various Aspects of Suitability for Various Candidate HOC Header Materials ______ Legend o = unacceptable / acceptable Materials * no comment or unknown F'ermatech MMM-Hg MM XM N-14 L-18 Siqma 5. Limited and reproducible compressibility and load relaxation normal to the board plane. / // o (L-17 (Incompressible) unacceptable) 6. Low in-plane thermal expansion to minimize cracking when constrained in the HOC mold. Near zero or negative expansion normal to that plane. Predictable and uniform dilation behavior up to 1300F. / / / o 7. Acceptable thermal shock resistance. / / o 8. Adequate strength and / toughness, low poissons ratio to maintain a stable overhang during service in the HOC process. // o o 9. Heat treated to minimize offgassing during the casting process. / // / / Physical properties that more accurately describe these parameters are listed in Table IX. These properties should not be considered as specifications, but as guidelines for further materials selections. Material consistency Is obviously an important Issue once a given product has been selected. Detailed discussion of this aspect Is beyond the scope of this report and should be given separate treatment. MWV:PAN:8 28 ATc 0035055 Acknowledqments The writer wishes to acknowledge participation by the following key personnel: Willie Lansdale and Bob Poland - Warrick Operations for excellent cooperation regarding all phases of this project and Mario Caprio Purchasing for initiating our investigations on this project. Jim Colpo, Janet Krahn, Bob Lytle - Material testing. Mike Kulak, Rich Ray, Pete Vranka - Instrumentation, design analysis and material testing. Ray Heilich and Steve Libby - Petrography. John Jacoby, Joseph Kopanda and Bob Milauskas - Background information. MWV:PAN:8 29 ATC 0035056 REFERENCES 1. W. Lansdale to M. W. Milner (Warrick Operations), "Re: Manville Meeting - 1984 February 08 - Batch #49 Header Grade Pressing," 1984 February 17. 2. M. J. Caprio (Pittsburgh Procurement), "Re: Status of Header Grade Molten Metal Marinite," 1984 February 14. 3. M. W. Vance (ATC-ACC) to F. W. Baker, "Re: Trip Report: Meeting with Products Concerning Quality of Ingot Casting,Header Board Materials for Warrick Plant - 1984 February 27," 1984 March 01. 4. M. W. Vance and R. J. Milauskas (ATC-ACC) to F. W. Baker, "Re: Field Trips to Alcoa's Warrick Plant - 1984 April 09-11," 1984 May 01. 5. M. W. Vance (ATC-ACC) to T. L. Francis, "Re:. Proposed Program to Evaluate Alternate Header Board Materials for Sheet Ingot Casting Warrick," 1984 June 12. 6. M. W. Vance (ATC-ACC) to F. W. Baker, "Re: 1985 Program Plan: Program to Evaluate Alternate Header Board Materials for Sheet Ingot Casting Warrick," 1985 January 28. 7. M. Kulak, M. W. Vance, P. A. Vranka (ATC), "Re: Trip Report Measurement of Temperatures and Strains In a Molten Metal Marinite Header in the HOC Ingot Casting Process - Warrick Operations," 1984 August 16. MWV:PAN:8 30 ATC 0035057 8. M. Kulak and M. W. Vance (ATC), "Re: Recent Design Analysis of HOC Headers During the Assembly and Installation Phase," 1984 December 19. 9. M. J. Caprio (Pittsburgh Procurement), "Re: Specifications for Header Grade Molten Metal Marinite," 1979 January 09. 10. W. A. Deer, R. A. Howie, J. Zussman, Rock Forming Minerals, Volume 2 Silicates, Longmans, Green and Company Publishers, 1967, pages 234-248. 11. G. L. Kalousek and J. E. Kopanda, "Binders of Auto-Claved Products," Journal of Materials, Vol. 3, No. 2, published by American Society for Testing and Materials, June 1968, p. 304-311. 12. G. L. Kalousek, F. V. Camarda, J. E. Kopanda, Z. T. Jugovlc, "Analysis of Asbestos - Cement Binders," Materials Research & Standards, Vol. 6, No. , published by American Society for Testing and Materials, April 1966, p. 169-179. 13. Y. Willingham (Manville Corp. Technical Center) to M. W. Vance, telecon 1985 June. 14. M. W. Vance (ATC-ACC) to F. W. Baker, "Re: Trip Report - Observation of Manufacturing Processes for Manvllle's Header Grade Molten Metal Marinite and Metal Mover Materials," 1984 May 08. /v MWV:PAN:8 31 ATC 0035058 15. W. Lansdale (Warrick Operations), "Re: Nichias Corporation/Pyrotek Meeting at Warrick on 1984 May 11," 1985 May 25. 16. R. M. Lytle, ATC Data Letter, "Re: Determination of Effect of Casting Lubricant Saturation and Coking on HOC Header Integrity," 1985 June 24. 17. D. L. Kinosz, ATC Report No. 8-68-13, "Coatings for Molten Metal Marinite," 1968 August 30. 18. J. K. Krahn, ATC Data Letter, "Re: Thermal Expansion of Ingot Casting Header and Roll Caster Tip Materials," 1985 February 05. 19. J. K. Krahn, ATC Data Letter, "Re: Thermal Expansion of Ingot Casting Header Materials," 1985 March 28. 20. J. K. Krahn, ATC Date Letter, "Re: Differential Thermal Analysis (DTA) of Header and Roll Casting Tip Materials," 1985 July 29. 21. R. M. Lytle, ATC Data Letter, "Re: Molten Metal Immersion Tests of Ingot Casting Header Materials," 1985 August. 22. M. Kulak, ATC Report, "The Prediction of Temperatures and Stresses in the Molten Metal Marinite Header During HOC Casting at Warrick," 1985 March 13. 23. W. Lansdale (Warrick Operations) and M. W. Vance (ATC), Te'lecon on 1985 August 12. MWV:PAN:8 32 ATC 0035059 24. W. Lansdale, Warrick Operations Memo, "Re: Summary of 1984 June 27 Meeting with Nichias and Pyrotek - Lumiboard," 1984 July 02. 25. W. Lansdale, Warrick Operations Memo, "Re: Lumiboard N-14 Trials," 1984 August 26. 26. W. Lansdale, Warrick Operations Memo, "Re: L-17 Header Trials Nonasbestos Material," 1985 February 28. 27. W. Lansdale to H. G. Reavis, Warrick Operations Letter, "Re: HDC Header Materials (Asbestos Replacement) - XM Material (Manville Corporation)," 1985 August 15. 28. Virginia L. Hammersmith (CAC-ATC) memo to Dale F. Ostergaard (PPEO-ATC), "Re: Properties of R-680," 1982 December 08. MWV:PAN:8 33 ATC 0035060 TABLE I- CURRENT HDC HEADER BOARD REQUIREMENTS*9 I. Dimensional tolerances: Both lateral and longitudinal lengths must meet requirements. Taper - .010 inch/ft (MAX.) Flatness - Bow - .0156 inch/ft (MAX.) Twist - .0156 inch/ft (MAX.) Thickness - 1.00 +_ .0312 inch Effects of dimensional tolerances: 1. Material loss (cracking) 2. Header overhang (_+ .0625 inch) 3. Uniform continuous clamping pressure II. Physical Properties and Requirements: Density - 40 _+ 2 lbs./ft Transverse Strength (modules of rupture) - 800 psi min Thermal Conductivity (BTU - in/hr-ft^-F) at Mean Temperature of 1000F (538C) - 0.85 avg. Screw holding strength (withdrawal load for type A, No. 8 sheet metal screw in a No. 29 drilled hole) - .50 inch penetration - 100 lb/screw min. .875 inch penetration - 250 lb/screw min. Dimensional Change after reheating at 1100F length, width and thickness - +_ 0.4 1 Max. Durometer Hardness Units die side - 55 min. screen side - 55-70 * Reference 9 MWV:PAN:9 (Table I continued on next page) ATC 0035061 TABLE I - Continued III. Finish Standards: Sanded Surface Requirements 1. Roller Marks - .005 inch max. 2. Ridges, groves, scratches or gouges - depth and width .0156 inch max. Foreign material - None IV. Identification: Identification of the screen side is desired due to its higher density V. Shelf Life: Moisture pick-up requirements currently being developed (hygroscopic) MWV:PAN:9 ATC 0035062 co oo o 4- o 0) S' 4oj o E m 3 0) 6 h4- fD X Q o ** f- o c o U. o o> o *9 o fO o u CSI to s u in a? o. <T5 0) UJ z o z ox<4- o 1/3 01 V L. U- a; Cl o 9 oo au. *--1 o i (/) s o. M a: UJ <CO oz o--x ou o3 >>44-> *- 3 </> o c^ aj m on c_ 40-1> fO 31 in CSJ CSJ CO OS CSJ o o CSJ o CO o coo CSJ o r"< o CSJ o vo VO CO *-4 oo 1 LD CO CO CO vo vo co CO CSJ CSJ co CO c o o * o o o CO t CO o os CSJ csj csj oo 1 in r-- VO OS CO in oo 00 CO CSJ in in --I r--H --4 ^4 *4 ^4 vo o i VO o r-* 1 on CO CO GO n in On CO o o in -^ "4 1 CO ^4 c- VO t r*- o CO 1 COr-. r-. CO CSJ in *"4 m oCSJ in m --< CSI o COc-. r- C-* r*- r>. *T in Cn. *4 CSJ 1 CVOO VO nr^ r-* vo CfvO Os n icnr co CO CO CO CO CO CpS4J Os o VO co m CV^SO4J . oVO VCOO in VO iCnO CvOo vCSoJ VCOSI ro^. CO i< z oo CO rion-i* in . o r^. r* csj o o CSI os in ^r ON CO Os *T vo ^T *t m co ^r vo ^r CO 4f iH CO CO CSJ -4 <SJf +| o c o L> X Q CSI o O CO 01 oo o: z i z zy z z 4-1 4 VO c-. 00 M p4 o o E 2E rO X u o> OJ i 1 1 1 X <0 0) *" CL z z z X z -J _J -1 X a. CL to CO 1/3 01 * XC uo c 4-> u in a: ------1/3 1 C lO <o in CL O OO U u II a> --j x u * "O 4- < o o *-- x4-> *9 C7> OJ 4- OJ #-- X CL O. H < X n CL C o * *r-. <--* 4-> **" u /) o> a n w1 </> x4-> in O CT U co 01 4L-> OCJ */> E (J C 0J -^t- om. c jQa> uO- 4-1 JC C 4- o -<oo a. o aaI;> xL it X f- -C II 1A CL 'W' 4-> *-- u 4- rO3*) 4O in 3 *3QFC 3> c T3 01 CO kn --i|x Cl x co| CSJ II oo ,, u- 4s CM m Ml U o *---s in 0J o c E fO a> jO u c in 1^3 4o c o r-- 4- oE 0j CJ 4- XI 0J T5 X u E3 c e i- CSI X *- n <o II < * : <in VO in n s CSJ fVNO. co ft a.1 x t! < > \ ATC 0035063 (4) References 9 and 15. TABLE III - Thermal Properties of Various Header Candidates Thermal Material Bulk Density Thermal Conductivity* Diffusivity' lbs/cu ft Btu-1n/ft2-F-hr ft2/hr 3 Ceramic Fiber Board 20-28 0 600F 0.47 0 1000F 0.66 0 1000l 0.140.10 Molten Metal Marinite3 40 0.82 0.85 0.09 Metal Mover3 50 0.76 0.84 0.07 XM3 3 Marinite C 4 N-14 5 Permatech Sigma 46 45 52 125 0.77 1.18 1.33 4.60 0.88 1.22 1.38 5.08 0.08 0.10 0.11 0.16 Notes: 1) Thermal conductivity - k 2) Thermal diffusivity - k/pCp p = density Cp = specific heat Sources: 3) Manville Corporation - Reference 13 4) Nichias Corporation - Reference 15 5) Alcoa Technical Products Division - Reference 28 ATC 0035064 TABLE IV - Average Thermal Expansion Coefficient (70-1292F) -in/in/F x 10" Manville M M M H Nichias M M II II Pabco ZIRCAR Alcoa TPD Material MMM-HG MMM-RG MM-old MM-new XM-4 N-14 L-16 L-17 L-18 KX 32 AL-45 Permatech Sigma (R-680) Parallel to Board Plane 3.49 2.70 2.78 2.38 4.52 3.81 4.29 5.56 2.50 2.22 2.70 6.03 2.90 Perpendicul, Board PI a -0.95 2.86 0.71 -3.57 1.51 -1.59 3.57 3.02 1.50 -2.06 -0.24 5.63 2.90 MWV: PAN: 9 ATC 0035065 TABLE V - Observed Reactions and % Weight Change for Header Materials Sample Code R443 R280 R418 R386 R448 R370 R372 Description MMM-HG2 MMM-RG N-14 L-16 L-18 MM(A4.3-1) XM-4 Observed Reaction Temp. - DTA Samples Heated to 1100C* Low Temp. Reaction 1 Hi-Temp. Reaction 2 Accumulative X Weight Loss 1 ' 3 Temp. C 100-** Temp. C 775+ 110C 2.15 800C 6.112 1100C 7.05 None 775+ - - 8.50 100- NONE 0.60 7.25 8.22 100- 725- 0.27 6.47 6.57 100- 725- - -- 100- 800+ 2.76 9.40 10.40 100- 800+ 1.04 6.55 7.77 * i(Temperature conversions - 100C = 212F, 725C = 1341F, 775C = 1427F, 800C = 1472F, 1100C = 2012F.) - ** + exothermic, - endothermic Notes: 1. All samples were held at each temperature for 1 hour. Weight losses are accumulative for each temperature. The same sample was used throughout the test. 2. MMM-HG fired at 800C showed noticeable cracks, deformation, and shrinkage. 3. After these samples were fired at 1100C, they were allowed to cool at room temperature for 28 hours. MWV:PAN:9 ATC 0035066 TABLE VI Load Relaxation for Candidate Header Materials Material MMM-HG MM-(A4.3-1) XM-42 XM-62 N-14 L-172 L-182 Applied Load (psi) 400 800 400 800 400 800 400 800 400 800 400 800 400 800 Percent Load Relaxation r Temperature - 'HTJ--------- RT 1100 1406 5.7 13.8 98.0, 7.0 33.0 82.03 4.9 7.2 28.0 31.5 100.0 87.03 4.8 70.0 5.8 21.0 81.0 4.8 28.1, 100.0 6.1 21.23 66.03 4.6 16.8 56.0, 6.6 19.8 39.03 3.0 12.0 4.7 16.0 26.0 5.2 19.3, 48.0, 5.6 13.03 24.03 Notes: 1. Temperature conversions - RT(70F) * 21C, 1100F * 593C, 1400F * 760C. 2. XM-4 and XM-6 are different batches of the same material. L-17 is the same chemistry as L-16 and L-18 but intermediate In density. (L-16: 48.0 lbs/cu ft, L-17: 46.0 lbs/cu ft, L-18: 43.7 lbs/cu ft) 3. Relaxation decreased as a result of crushing that occurred when increasing applied compressive load from 400 to 800 psi. MWV:PAN:9 '9 ATC 0035067 c oo 4-> f- p c 4-o> c O O C7> CX -r- IO -r- o M- E u 4J **- C * 40 C cc r-- 40 cr> c_ o 0)T3 U *- z *- *3 *^ c v. tO tO 4-> 40 * <- *- L 33C 3CO oj 4-> u> a> 40 *-- IO 4-> f- o 4-> O 40 4> 4-> * to 40 4-> QJ O >> o o CO bc L. to ^ * c % u >>-*c to a> t- > to * 40 *t CL 3 C t_ 3o oo O L- U .C Z U CO to O 40 o r>c CM z r0H0 t/> CM c fc*. ,,__ a; eo o Ol <o jk 00 o SO fo* [fcB Os ^CMH --c Li. to O O o O CO to O1 --4 4-> to tO 4-> 3 4-> V) -C CJJ CO r_ * 4-> to o c tO a> * c .po. H* 3 * 4J 40 C tO CM. c 4J 4J (O O C .*:*-- (/) L. 0P) g <0 oE to E L. O O *co 1 r- U c t- to O JZ tZ W D 3 O 1Z to * > CO Lul K J 00 0) < OI -- 40 -H lO c tJZ to 01 4-> 40 c. OJ 3o EoC c o 3 *4J QJ 40 4J O) (. 4-> 40 Of ^ 3: c o 1 * o C L (O o x: *Z ifl TJ > o CM f- .*-> 40 c. 0) 3 O E oC *c o *3 * 4J 40 4-> OS U P IQ O -^ *-- * eo 1 *t" u c c_ to O Z -fZ tO 3 - >> 0L_) C7) c 4- -> *" -CCO 0) *-- *" 4^ *-- -- -F- IO to 4- X a> *E 3 Q> 0) Q) 4J ox: 4-> <0 4J X C QJ -r- > c c. o OX x Z to 40 O CO CM SO CO 0*1 QJ 4J to t "p3 EOC *c o *3 * 4J ai a) <o 4^ cn tP <Q O QJ #* c 1 **- o c to 0^ Z to 3 4-> -cCn V-- to o c c o 3 * 4-> C to 4-> a> t4-> 40 c 1 *- u C u to o x: -Z to 3 00 o o^ --40 Z 1-- o z1 4- 40 f- z z: 1 CO ^r z <: Z KlOt zX Z 1-- 00 1 -J x: 4^ 3 It to L. to o C JZ ^ 4-> u o> t- c JZ 4-> r-- 33 LL 40 40 o X) X5 x: x: 4-> 4^ oo 4-1 4-> -- r-- -- 40 40 L L. 40 40 CL cl 33 40 40 EE 4-> 4- cc EE <- c. 33 t*oo t<oo II to 4-> o z ATC 0035068 TABLE VIII - HDC Header Trial Results - Warrick Operations Material Date Results Alcoa Permatech Sigma (R-680) 1977 Material sensitive to cracking as a result of stresses during installation and start-up. No successful start-up on Warrick's HDC caster resulting in ingot tearing. Manville Metal Mover 1982 Successful HDC runs. Header material was inhomogeneous with the presence of white spots. Manville withdrew the product to correct processing problem areas. Nichias/Pyrotek N-12 Excessive shrinkage and oil absorption. Ingot tearing attributed to header shrinkage. Nichias/Pyrotek N-14 1984 Excessive absorption of mold lubricant. Cracking and chipping of the header. Excessive chilling caused tears and cracks in the Ingot surface. Nichias/Pyrotek L-17 1985 Termination of 5182 cast (26 1/4 hrs) resulting from excessive gaps forming between header and oil ring leading to ingot tears. Excessive oil absorption. Manvi11e XM 1985 Eight trials (10 1/2, 25 1/4, 43 3/4, 19 1/4, 45 1/4, 43 1/4 hrs - 5182; 75 hrs 3004; 114 hrs 5182/5042, and 77 hrs 5182/5352). No casting stoppages attributed to unusual header performance. MWV:PAN:9 '4 ATC 0035069 TABLE IX - Properties for Header Grade Marinite Bulk Density - 40 +2 Ibs/cu ft Durometer - 55-70 {screen side) MOR/MOE - 700 psi/100,000 psi at R.T. to 1000F Compressive Strength - 1000 psi (initial crushing normal to the-board plane) Compressive Load Relaxation - 5-10% at R.T,, 10-20% at 1100F at 400 psi load (normal to board plane) Poisson's Ratio Thermal Conductivity Thermal Diffusivity Average Thermal Expansion Data (70-1292F) - 0.10 - 0.85 Btu-1n/ft2-F-hr at 1000F mean temperature (normal to board plane) 2 - 0.09 ft /hr at 1000F mean temperature - o c 3.49 x 10" in/in/F (parallel to board plane) -0.95 x 10" in/in/F (normal to board plane) Oil Absorption (AA Castor Oil) - 10-15 weight % (hot oil) Permeability - No data MWV:PAN:9 atp. OO35070 i FIGURE 1 - DETAIL OF CLAMPING ARRANGEMENT ATC 0035071 OIL "A " ADSORBED BY MATERIALS FIGURE 2. Amount o f AA S tandard L u b ric a tin g O i* P ic k -u p f o r V a rio u s HDC Header C andidates. ato nns*in79 ATC 0035073 IN HEADER (HETAL CONTACT TO OIL RING) o m Q. CXI txi n 0 1 h O 0 1 ro* -J h< O ua:i to < tr h CO CO s & > CO CO UI CO lU >M CO CO tu CE a ao CXI u Ui u. sUI a < CO CD ^4 XTi e to | n cIn xx i _i J xx zi < TT A<3 I 11 IT COMPRESSIVE STRAIN I N / I N . x 10' >V i/% ia/%> uo Oo=>' ATC 0035074 0 a a0 (O tn in CD i hO 0 1 o a UJ X o< (0 M_l o zM X h to CO 3 CxO UJ CO CO UJ X w CO U>J w CO CO UJ X X 3 a cu CJ Ui J sUJ too CD CO "F o (o x v n T Trt I r< XI I X I XJJXZ< TI A*.I CMIo. ZM \ z \, ''>8 ATC 0035075 FIGURE 5. Com pressive S tre s s Versus S tra ln -O il-S o a k e d . < z 1-4 cn o aint ozo u M UKZISUJ Z3 u z -0279 MANSVILLE MOLTEN METAL MARINITE HEADER GRADE FIGURE 6 . Thermal Change as a F u n c tio n o f Tem perature MMM-HG: P a r a lle l and Normal to th e Board P lane. o oott s oo s0 g oo UJ 9! oo N Oo o ATC 0035076 OJ 01 Ot- < a UJ CL ULt X UJ 1-- t--< z *-< UJ > 0 X xX h- H z-( < --1 < 0 x<X u0 t- m CO CO UJ CD 1 < U3 XXXzaX O) rx ro rv 0 IS ru ru ru ru ru tn x XX XXX CO COMPARISON OF LONGITUDINAL CUT HEADER MATERIALS FIGURE 7. Comparison o f Thermal Chanpe B e h a v io r f o r V arious Header M a te ria ls - P a ra lle l to the Board Plane. ru cn UJ > < rx 0) u ix 0 ATC 0035077 CJ Ui x LU cr UJ H- UJ t- hH z t--( c<r CD I > oz -<J HLU Z 71 m o ru cc ru cr H z H cr < z C9 cr *<T z1 CD CD O) o CM cr Mil! ru cr COMPARISON OF THRU THICKNESS CUT HEADER MATERIALS . TEMPERATURE C FIGURE 8 . Comparison o f Thermal Change B e h a v io r f o r V a rio u s Header M a te ria ls - Normal to th e Board P lane. ru m o u m< a c\j cr Y-- c<n u o CO tcor r- m cr ATC 0035078 COMPARISON OF LONGITUDINAL CUT HEADER MATERIALS FIGURE 9. Comparison o f Thermal Change B e h a vio r f o r V a rio u s Header M a te ria ls - P a r a lle l to th e Board Plane. o it m cn v 'T i 03 r*->< o in mv xi ixi i < < X _j _i x _i _j << to x Ui a o Ui a a UI a o XL ui - X uj h 2d ui t- a: < tr < I CJ a u o a o cj cj z i i i tr a: cc a: cr aX X X X > > > H H X X X X D. Q. N N I ATC 0035079 CD I I Z ir o m m ^| VIDI TS< < LU Q O O I < UJ Q O CJ I Z X Ui O Ol-- uo i cc I _l X UJ z z z z >- > >- M M z z z z o. a Q. N N :iim I _l X UJ h o cc COMPARISON OF THRU THICKNESS CUT HEADER MATERIALS TEMPERATURE C FIGURE 10. Comparison o f Thermal Change B e h a v io r f o r V a rio u s Header M a te ria ls - Normal to th e Board Plane. x x XUi H- Otr mo i-nt i _j << cc < CccJ i _j cc < cCcJ ATC 0035080 uZD 3 u in a cn LU z ID u z oX ATC 0035081 R44B PYROTEK L -1 8 TEMPERATURE C FIGURE 11. Thermal Change B e h a vio r f o r L-18 M a te ria l - P a r a lle l and Normal to th e Board P lane. d) cn <n a 0 OJ m in m 1 "3 I o00 aCoO o u_ tt Q < o a: Ua) 2 3 O < X < _J LU CC fv Oo X < _! LtUr CD CD -r-l XI ZI IDI 7I (-*D"1 N n. n^ X 2 2 | I TT I x X x _J _] < z O (D X ATC 0035082 TIME IN MIN. CD -- XI I CD CD m TT Xl X X I "V I X X _l C 2 I_ 1 Ii on c? X <. * RELAXATION UNDER LOAD HOOF 400PSI CD Ol to 01 in n in co i _i Z> -3 1 in o in a o o TT Ll O O TT Q< o _J CL UJ a 2 n 2 o K < X < UJ cr "T x-i X < UJ cr co CO 21 I ID 'T O N PI Tf 2 I | ti n . 2 2 2 1 l^l 2 X X _J _J < Z I 1 li i . 1 1 iiT TIME IN MIN. X CJ I < 2 CD ID ATC 0035084 I dQ - 3arUVH3dW31 ATC 0035085 FIGURE 15. Process temperature data recorded at Uarrick showing metal temperature O ac C3 =: g-J 3LCU O H- ^U I CO < LU HZ- h- =3 CD H- LJ LQU_ CI-- & H-- C/) LU O CO J-- = c gg I a: =3 l> ATC 0035086  (M i-W ii-a fi*IM (lie b-H FIPURL ]7 - STEADY STATE CIRCUMFERENTIAL STRESSES FUR "FIXED* BOUNDARY CONDITION ATC 0035067 -FIGURE 18 Zones of tensile stresses in clamping direction an HDCheader during and AFTER START-UP. ATC 0035088 0) time 65 sec. E) time 137 sec. APPENDIX SEM PHOTOMICROGRAPHS OF NEW AND USED HEADER MATERIALS MWV:PAN:9 ATC 0035089 Material MMM-HG (New) MMM-HG (UsedOverhang Area) MM (New, A4.3-1) MM (Used, A3.1-4) APPENDIX SEM Photomicrographs of New and Used Header Materials Mag Comments 20X Relatively high concentration of oriented fibrous bundles without an obvious continuous binder phase (Figure A-l). 524X Fluffy binder phase interdispersed with fibers (<5 ~um) (Figure A-l). 30X 400X Material has a more monolithic appearance. Orientation of fibrous bundles is still obvious (Figure A-2). Small fibers still retain original morphology. An amoslte pencil also retains its unique size (cross section 25 x 50 urn) and shape. It acts as a flaw in the overall matrix and could be a cause of premature failure in the highly stressed overhanging region (Figure A-2). 400X Fiber bundles retain original morphology. Binder phase seems to agglomerate around the individual fibers and bundles (Figure A-3). 1000X Highly magnified view of fibers and associated binder agglomerates (Figure A-3). 1000X Another example of agglomerated matrix around asbestos fibers (Figure A-4). 20X A more continuous matrix. Fiber reinforced concentrations are lower than MMM-HG (Figure A-5), 100X Fibers and relics of kraft paper (cellulose) are highly oriented. The fibers are relatively large in diameter (5-10 pm) compared to asbestos fibers and do not occur In bundles (Figure A-5). 1000X Voids are created In the more continuous binder phase by the kraft paper which retains a flat appearance about 60 x 10 pm in cross section (Figure A-6). 30X Fracture surface shows a porous matrix with a lower fiber content than MMM-HG. Fibers still retain their morphology (Figure A-7). MWV:PAN:9 A_1 AATTCC 00003355009900 L-17 (New) L-17 (Used) L-17 (Used Tip) L-17 (Used) Interior) L-18 (New) L-18 (Used) Appendix - continued 1000X 30 X 200X 400X 100X 20X 521X - Matrix is similar to N-14 (Figure A-14). Fiber concentration is decreased at the tip because of possible oxidation. Matrix is heavily sintered (Figure A-15). Matrix is heavily sintered at the tip. Carbon fibers are no longer visible (Figure A-15). Tiny fibers in the matrix are no longer visible (Figure A-16). Reinforcement is still present in the Interior. Sintered texture of matrix is predominant throughout the cross section of the header (Figure A-16). Carbon fiber reinforcement is similar to L-17. Fibers seem to be shorter and less concentrated than L-17. The matrix is continuous but more open than L-17 (Figure A-17). The matrix is composed of minute fibers similar in morphology to L-17 (Figure A-17) No trial use as yet. MWV:PAN:9 A`3 ATC 0035091 20X 524 X Figure A-l - SEM Photos of New MMM-HG Fracture Surfaces MWV:PAN:9 A-4 ATC 0035092 30X 400X Figure A-2 - SEM Photos of Used MMM-HG Fracture Surfaces MWV:PAN:9 A-5 ATC 0035093 R 'flATEi 400X 1000X Figure A-3 - SEM Photos of Used MMM-HG Fracture Surfaces HWV:PAN:9 A-6 ATC 0035094 /E 1000X Figure A-4 - SEM Photos of Used MMM-HG Fracture Surfaces MWV:PAN:9 A-7 ATC 0035095 Figure A-5 - SEM Photos of New MM (A4.3-1) Fracture Surfaces MWV:PAN:9 A-8 jk ATC 0035096 1000X Figure A-6 - SEM Photos of New MM (A4.3-1) Fracture Surfaces MWV:PAN:9 A-9 ATC 0035097 200X Figure A-7 - SEM Photos of Used MM (A3.1-4) Fracture Surfaces MWV:PAN:9 A-10 ATC 0035098 1000X Figure A-8 - SEH Photos of Used MM (A3.1-4) Fracture Surfaces MWV:PAN:9 A-ll ATC 0035099 Figure A-9 - SEM Photos of New XM-6 Fracture Surfaces MWV:PAN:9 A-12 ATC 0035100 100X Figure A-10 - SEM Photos of Used XM-4 Fracture Surfaces MWV:PAN:9 A-13 ATC 0035101 1000X Figure A-ll - SEM Photos of Used XM-4 Fracture Surfaces MWV:PAN:9 A-14 ATC 0035102 20X 1000X Figure A-12 SEM Photos of New N-14 Fracture Surfaces MWV:PAN:9 A-15 ATC 0035103 inoox Figure A-13 - SEM Photos of Used N-14 Fracture Surfaces MWV:PAN:9 A-16 ATC 0035104 Figure A-14 - SEM Photos of New L-17 Fracture Surfaces MWV:PAN:9 A-17 ATC 0035105 Figure A-15 - SEH Photos of Used L-17 Fracture Surfaces MWV:PAN:9 A-18 ATC 0035106 100X - Interior Figure A-16 - SEM Photos of Used L-17 Fracture Surfaces MWV:PAN:9 A-19 ATC 0035107 521X Figure A-17 SEM Photos of New 1-18 Fracture Surfaces HWV:PAN:9 A-20 ATC 0035108 FROM M. W. VANCE CERAMICS DIVISION ALCOA TECHNICAL CENTER T0 MEMORANDUM CONFIDENTIAL 1985 October 10 RE: DIVISION REPORT - "EVALUATION OF CANDIDATE MATERIALS FOR HOC INGOT CASTING HEADER APPLICATIONAT WARRICK OPERATIONS" This report sunmarizes a program intended to improve the understanding and data base for selection of new HOC header materials as substitutes for asbestos containing Molten Metal Marinite. To date, two new candidate calciun silicate bonded materials have been identified for future trials at Warrick Operations. Although some of the properties presented in this report may be used for header material quality control purposes, material consistency should be addressed as a separate item. It is intended that this information may also be used as background for selection of headers for other HDC and FDC casting operations, for headers for FDC casting Alithalite alloys, and for potential selection of materials for other applications involving molten metal contact such as: troughs, basin linings, floats and dams. Future ATC involvement will consist of performing limited testing, making recommendations for implementing substitute materials, and continued participation in the HDC Header Task Force at Warrick Operations. Further inquiries should be addressed to the writer at 8-221-2815, Alcoa Technical Center. M. W. VANCE TH:MB:11 0 ALCOA ATC 0035109 r>4su (ncv n.) cc: ID - D R. Bachowski - ATC-B F. W. Baker - ATC-B P. R. Bridenbaugh - ATC-C K. E. Buckovecky - ATC-C T. L. Francis - ATC-B G. E. Graddy, Jr. - ATC-B S. C. Jacobs - ATC-B J. E. Jacoby - ATC-B R. S. James - ATC-B S. C. Libby/D. A. Weirauch, Jr. - ATC-B R. J. Milauskas - ATC-B R. E. Miller - ATC-B R. L. Rolf/M. Kulak - ATC-D R. R. Sawtell - ATC-C P. C. Scheble - Badin A. Borges/D. Scott - Lafayette M. Scherbak - Massena J. D. Boyce - Pittsburgh-19 M. J. Caprio - Pittsburgh-19 C. J. Cox - Pittsburgh-3WPH N. W. Neilsen - Pittsburgh-18 E. L. Rooy/H. G. Reavis - Pittsburgh-23 A. J. Sartschev - Pittsburgh-3WPH M. A. Thomas - Tennessee P. D. Thomas/F. D. White/M. Vrablic/J. E. White - Tennessee W. C. Harvey - Vernon C. R. Sammy - Vancouver J. E. Gunn - Wenatchee W. Lansdale - Warrick M. W. Milner/R. K. Boyles/R. A. Poland/T. A. Perigo/S. S. Owen - Warrick J. Reynolds/S. T. Davis/W. B. Steverson/S. L. Shelby - Warrick E. F. Smith - Warrick Ceramics Division Files: 10.322, 10.331, 10.380 SI ATC 0035110 REINFORCEMENT OF ALUMINUM WITH WHISKERS AND FIBERS OF BORON NITRIDE, ASBESTOS, ALUMINA AND GRAPHITE_____________ By C. N. Cochran Phys. Chem. Div. Report No. 8-68-3 s,V, a March 29, 1968 Copy No. "8 ATC 0035112 REINFORCEMENT OF ALUMINUM WITH WHISKERS AND FIBERS OF BORON NITRIDE, ASBESTOS, ALUMINA AND GRAPHITE By C- N. Cochran Phys. Chem. Div. Report No. 8-68-3 , .\ ' ' March 29, 1968 Copy No. 8 ATC 0035113 Aluminum Company of America Alcoa Research Laboratories Physical Chemistry Division Report No. 8-68-3 REINFORCEMENT OF ALUMINUM WITH WHISKERS AND FIBERS OF BORON NITRIDE, ASBESTOS, ______________ALUMINA AND GRAPHITE Signed C-/1 _________ C. N. Cochran, chief Physical Chemistry Division Date - March 29, 1968 TABLE OF CONTENTS Page ABSTRACT ............................................................................................................................................ 1 INTRODUCTION....................................................................-.......................................................... 2 BORON NITRIDE............................................................................................................................... 2 ASBESTOS ............................................................................................................................................ 3 ALUMINA ............................................................................................................................................... 4 GRAPHITE................................................................................................................................. 5 SILICON CARBIDE........................................................................................................................ 7 ATC 0035115 A---B--S--T-R---A--C--TComposite tensile specimens of boron nitride paper in ^ Is / high purity aluminum, R414 Safibers in 2025 alloy, graphite yarn in 2025 alloy (aged to the T6 condition) and alumina whisker tape in 2219 alloy were prepared by a high pressure liquid metal infil tration technique. In all but the boron nitride reinforced sample, the composites were weaker than the unreinforced alloys. The 5600 psi gain in tensile strength due to the boron nitride is equivalent to a -ensile strength of 72 ksi for this material. The poor performance of the alumina whisker tape was unaccounted for. The R414 Safibers were probably weakened in an attempt to enlarge the crystals by firing to 1850C. The graphite yarn reacted partially with the aluminum to form aluminum carbide. Microcrystalline asbestos fibers were incorporated into prealloyed 2024 powder by hot pressing but did not provide the claimed improvement in elongation. ATC 0035116 INTRODUCTION 2. Hot pressing of aluminum powder and infiltration with liquid aluminum (described in Report 8-67-8) were used to incor porate whiskers and fibers of boron nitride, asbestos, alumina and graphite into aluminum. A stronger bomb of 310 stainless steel (Figure 1) was constructed to permit liquid metal infiltration at pressures up to 2500 psi at 1000C. The high pressure reduced the porosity of castings by permitting metal infiltration in channels between whiskers, as narrow as 0.07 microns. Infiltration was performed at 750C rather than at 1000C to reduce strain in the bomb.. The boron nitride mold (Figure 2) was elongated by 1/2" to give a tensile specimen with a 1-1/2" long reduced section area. The vatted seal between the aluminum and alumina tube was formed at temperatures of 1000 to 1100C in vacuum before infiltration. For the most part, alloys in the 2000 series were employed rather than high purity aluminum, but the castings were not heat treated in all cases. The results for the specimens prepared for the present report are given in Table I. Summaries of the reinforcing agents and the composite tensile specimens included in this study to date are listed in Tables II and III. BORON NITRIDE Boron nitride paper obtained from Mr. Lynn Ogden of the Carborundum Company was infiltrated with high purity aluminum at ATC 0035117 3. 230 psi. The yield, tensile, and elongation.of the specimen con taining 0.233 volume fraction boron nitride and 2.1% porosity were, respectively, 5,300 psi, 10,500 psi and 10%. Transverse and longitudinal sections of this sample in Figure 3 show that the orientation approximates a two-dimensional random array. Using 6,600 psi as the tensile of high purity aluminum cast in the same 3 GO. 500 - (1-0.233-. 021) 660(0 fashion, a tensile strength of 0.233 " 72 ksi was calculated for the boron nitride for this orientation. This compares to tensile strengths of 75-120 ksi claimed for this material in the literature. ^ The other constituents showing in the photomicrograph have not been identified but may be Al-B or A1 ~-!n compounds. ASBESTOS During a visit to these Laboratories in the summer of 1967, Dr. A. O. Battista of Food Machinery Corporation offered to supply "Avibest" microcrystalline asbestos fibers for incor poration in aluminum. These are not strong fibers, but an improvement in elongation of aluminum was claimed. It was necessary to employ hot pressing of aluminum powder-asbestos mixtures for incorporation because the asbestos contains water of1 (1) Donald G. Sturges, Steel S-l to S-6 (May 30, 1966). ATC 0035118 4. hydration which would be released above 660 C if molten aluminum were employed. Chemical treatment of the powder is required to permit raizing of the aluminum powder and asbestos. A quantity of pre alloyed 2024 powder supplied to Food Machinery Corporation was treated, mixed with the "Avibest" and returned to Alcoa. Mixtures of 10 and 25 wt. % "Avibest* with 2024 powder as well as treated 2024 powder without "Avibest" were pressed into tensile -specimens by Mr. J. H. Dudas. The specimens were sintered 45 minutes at 1100F in nitrogen, cooled to 920F and cold-water quenched. Samples of each material containing "Avibest" were aged to the T6 temper. The tensile strengths were only 400 psi for the specimens containing "Avibest, " and the elongation could not be evaluated. The treatment of the 2024 powder required for mixing with the "Avibest" evidently caused deterioration because the tensile strength of one compact from the treated powder without "Avibest" was only 1,400 psi, and the other broke before it could be tested. ALUMINA Samples of R414 polycrystalline alumina "Safibers" with smaller diameters (approximately 20 microns) than previously available were received from Mr. A. Pearson of the Alumina and Chemicals Division. One lot that had been reheated to 1850C to promote grain growth proved disappointing in giving a tensile ATC 0035119 5- strength of only 4,100 psi in a 2025 alloy composite. This is less than expected for the matrix alone. Tests will be made with a second lot which is the same material as in the first lot but without the grain growth treatment at 1850C. A quantity of alumina sapphire whiskers in a PIB rubber tape was obtained without charge from General Technologies, Inc. The high weight fraction (0.44) of whiskers in the tape raised hopes of producing a composite with a high volume fraction of reinforcing agent. However, the low density of the rubber matrix in the tape reduces the high weight fraction to a volume fraction of only about 0.15 in the tape. The rubber was burned away by firing in air at 1000C, and the remaining whiskers were used to reinforce 2219 alloy (0.124 volume fraction). The infiltration was not successful because tensile strength (10,400 psi) well below that of the matrix alone was obtained. There appears to be no advantage of the tape over the whisker mats previously purchased from Thermo-Kinetic Fibers, Inc., a subsidiary of General Technologies, Inc. GRAPHITE A free sample of Thornel-50 graphite yarn obtained from Union Carbide Corporation was incorporated into 2024 alloy, and the composite sample solution was heat treated and aged to the T6 temper. Metal grade alumina ore was piled over the mold beneath ATC 0035120 6. the aluminum to prevent reaction with the graphite during the 1000--1100C treatment to form the AI-AI2O3 seal prior to infil tration. Vaporization of magnesium during evacuation prior to infiltrating changed the matrix composition nearer to 2025 alloy. The yarn, consisting of 1,440 filaments, each 7 microns in diameter, permitted packing to a high volume fraction of 0.336. Polished transverse and longitudinal sections (20x, 100X, and 500X) in Figure 4 show that the graphite was pushed to one side of the section during infiltration, but good longitudinal orientation of the fibers was obtained. The grey areas surrounding the fibers and the needles in the matrix in the 500X section may be aluminum carbide. The modulus of 50 x 10^ pounds/inch^, tensile strength of 400,000 psi, and density of 1.55 g/cc make this material extremely attractive for reinforcing aluminum. However, the density of 2.643 g/cc for the sample corresponds to 8.6% volume contraction during infiltration which may indicate reaction to form aluminum carbide. If all of the graphite in this specimen reacted to form aluminum carbide, the theoretical contraction would be 16.1%. The tensile strength of the composite was 20,800 psi, which is below that for the matrix alone. The modulus of 24.3 x 106 psi is identical with the value calculated from the rule of mixtures for this composition. ATC 0035121 7. Powder diffraction analysis (Film B54371) showed about 4076 Al, 40% AI4C3/ and 20% graphite while chemical analyses (68-011810) showed 7.1% A14C3. In an attempt to reduce AI4C3 formation in Run 11418-44, the infiltration temperature was reduced from the 785C used in previous runs to 705C, and the time at temperature before the start of casting reduced from about 1 hour to 2 minutes. This reduced the carbide to about 25% (Film B54424), but the sample did not release from the mold and broke.SILICON CARBIDE Photomicrographs of longitudinal and transverse sections (500X) of #13 alloy containing 0.103 volume fraction Carborundum SiC whiskers and 0.9% porosity are shown in Figure 5. This specimen had been described in Report 8-67-8. The alignment is, predominantly a two-dimensional random array with a slight trend toward uniaxial alignment. With 22,100 psi for the tensile strength of unreinforced #13 alloy cast in a similar specimen, the tensile strength of the silicon carbide is ---------*---------------oTI03------------------------------ " 126,000 psr. ATC 0035122 oc <44 os O rs. CO -H ^ WH 0"HC N m <aCl IN 00 <D 7 O' go go go 7 0DI> 0cI0* 00 CO CO o ri! u. < cco W 5 ui 3 4J **- C CO 4J O Li JZ -H o O' *i *w 'fn n IN nPI IN o IcN- O < uc. Zo\ otc a 4Vc0J* < Cc" <UC0c*4l cs wo '6DU3-h-Ol4 -4 3 X *0 U*-4 Z0 *0H *>Hj< uo Qtct D *<l> C0 ocO' *4 U *Q1 o o o IN oH0I0 U007 0r>l0 o pin- IN C O (A IN O o orc-i O* oo o O cro- rC>*L, o m -*h , *D C O -4 *J ~ can OO ^Cg4do a-. Cl *4 l* IN O *J XQi oot 33 -i e uV*JI s3O fMCl4 m tn oo o IN oo4* *D g 3 o r7- i h h lw y 4? < * 1Jt0tfWlHOAoOGCA u cn tt -- oIN IN VXJl O >Uc1 o C <H Co xLo a O OC H LI < w7i oIN IN Li tU* Xm oc < O' H IN IN ~co4 *co 14 is? C o z *ogD ? m oIN IN ATC 0035123 43 P d> < c -H 0 c <n P H A! P CO a * o O O Ht 0 X o O O CN H0 Gi -H 0P H* H H 43 0) 0 0 G >2 4> 0 Eh Q rH C .C -P *H o CN > CN r- o rH O 0 -P CJ 0 0 OS o oO O S3 oo O O P oooO O H CO H 4) * 3P3 * in m _ to 3" i i o Ht 1 in l in l Q >M Q H _C C 0 ' a o oin o om o CN o o o o o rH 0 % S* CN CC Eh co 4) P 4J H P p G O 02 nJ P 4) in 0w e 0 (0as e H <P A! P 0 S V\ m 1 cn l cn i rH 1 (0o p H rH 'rl rj rH ' -rH t--i --( rH o cn p S3 Q J a cn I CO CN in m lo cn rH o o < uM >. to P >3 H O r r~ r" r- in 01SC a. CJ c\ cn in cn rH CO ID 0 IT m cn cn 2a cn rH CN H W Ei < Q 1 1 1 0rH1 O o O a: O in ion O in 1 %% Oo O CN O Eh a w (Q e <u rH cn l cn 1 cn l H -rH o o o <d --i to o o o Go o o O rH 1 oin cn V m cn r- co u 0 D E< CM (N CN t> V m n rH o O CO CO h P 0 H(0 W H m a O O O o < o rH aa f--3( vo 31 oOo cn cn cn lll 3 0 o o o o O rH o 1o O' CO 10 o m 2H cj os C1O 1 1 s0 H X o rH o rH o rH o a 2 H 00a -0p p2 to 0 p oin 02 a o * Eh o * P to 0 -rH P a: _G rH 0 to to 0 a H P p to 0 -P AS G to 0 H aI J rH t0o rH 33 0 to G <P O 0 33 OS <3 2 2? D (0 o id 5 Oh p 0 A3 -OH CO -rt G tn 2a rH 03 - rH 0G G 0) tr < 000 p P P0 H -rH rH CO 33 aa 3aa3 3aa3 ** rH a) 0 0 01 CO CO CO OS a a p 3 O wA 2 Eh CO Eh Eh p a c 3 P CJ A p 0a 00 CJ Eh P 3 T3 G 3 P 0 A P U0 G 0 P 0 PQ 0 o 0 X 0 Eh -P rH m0 0 -P 333 ^o >k <V 0 p C2aJ h2oH O P -rl 0 3P3 >H 00 CJ P rH Co 3a3 *H 0 GP P C5 ATC 0035124 SUMMARY OF COMPOSITE T E N S IL E SPECIMENS MADE TO DATE I N P H Y S IC A L CHEMISTRY D IV IS IO N L iq u id I n f ilt r a t io n Used u n le s s O therw ise N oted CW <QD. to) ns Po I--0)1 H mc oo HiON oo<N <n oo oo * CM H rH CD 03 0 3 C H a H4-1 3(0 4J c C fa o & p < o a) p &c CD rH cr> <N rH rH O CO fa H ua O o o o rH o CO o rH o oH >o cH f<aD (00] O o oo ho3 rH H* % - m mm O m 0} O4 co CO CM o CO rH rH CM O oo -GoH +o nus &4 + C tO4 4O-1 <>04 fa 4J tn CM H pnps rH < rH < rH < rH c CM 2 fa9 9 a,9 fa fa OCM 35 35 35 rH OrH' CCMM rCHO < fa9 35 a a) Ocm H>oi Tp3 CM -4 HaS foa fpa a ns EH c c o i4 -r4 Pa CM HX ns Eh 1 1a f-4 r4 r4 ns rn*4s +> 4J 4J G 4) & 5 TO o o O o H JQ P O. TO v> VO X oP m Po p O < & C H O p 404 C P rf p r4 i-4 O O 0o a np (0 & s fa A P -4 4) P H p H 4*44 nj 01 .C a .C a A a P 4AH4 ns 01 H* m P A ^444 ns 01 i4 P -C P A a n&s 01 -urH CO 6 p A a ac 0 a ns Po fa -r4 s Z a r U p *o o A Cn 0p V *t4 2 B 4J m Vi a fa a Ol nas CO (Q0i 01 H H* fa 2fa f2a f2a 4 CO Eh Eh Eh W r4 fa P01I W r4 H# fa PCO5 W Eh oG CD 3 P 0 A P O 3 P O AnPs U >0 *< -P n 2fa <n5 >O =< -QP O P) 2 TO fa <5 H < 4O-> <40) A4HJ fa > rH <5 fa* 35 p G Jq r-4 H fa ao po CD O0 ns X Eh O p cH 01 O c ns C<aDQ a fPa 4J o w O m i-4 c p O P5 4 C *0 P h <d A >H Opns 4-1 CA -oH a(0 GP 3> e> ATC 0035125 FIGURE la --PRESSURE AID TEMPERATURE 3013 FOR PREPARING REINFORCED 'ALUMIrUuM -TEI'SILE BARS BY LIQUID I3TAL INFILTRATION (310 Stainless Steel) ATC 0035126  FIGURE lb EIH5 CAP FOR 3 Cl 3 SECT;!' IF FIGURE la (310 Stainless Steel) ATC 0035127 ELONGATED SPLIT BORON NITRIDE HOLD FOR CASTING REINFORCED ALtTilNLN TENSILE BARS ATC 0035128 ' FUGURE 2b ELONGATED 3CRCi\ NITRIDE SLEEVE FOR HOLDING SPLIT MOLD OF FIGURE 2a TOGETHER IN CASTING REINFORCED ALUMINUM TENSILE BARS. ATC 0035129 yiOTRB 3 (30OX) sBcxncm or bxqb purity aldmzbdm (sample 12339-31) CCaSAXSXSS 0.233 VOtLOMB PRACTICE Or CABBOBOTBOM BOBOS BZSBZXX PAPER ATC 0035130 T tra rp *T 6 e *1 (SOS) ATC 0035131 &QUXBSUD tSCTZQBB OT 202$ AUDI (AHlOk) CQBWmmP 0.1M VCSAHS re*crxcn o* bsxcs crmibb squbi*-$o aumxrs vzum ATC 0035132 FSSKXRB 4o FGLzaasD aacnoBt or 2025 aixcbt (Amo.) CORTMBHW 0U6 VOUM FRACTION OT UHIOS CAB8ZS6 ATC 0035133 FIGURE 5 (50CX) Transverse 0.103 VOLUME FRACTION CARBORUNDUM SiC WHISKERS IN #13 ALLOY, 0.5 TO 3^ DIAMETER, 100 TO 750^ LONG, RANDOM TWO-DIMENSIONAL ARRAY, DERIVED TENSILE STRENGTH OF WHISKERS - 126,000 PSI Longitudinal atc 035134 ROM: C. N. COCHRAN PHYSICAL CHEMISTRY DIVISION ALCOA RESEARCH LABORATORIES NEW KENSINGTON TO: DR. P. T. STROUP ALCOA RESEARCH LABORATORIES NEW KENSINGTON March 29, 1968 V RE: REINFORCEMENT OF ALUMINUM WITH WHISKERS AND FIBERS OF BORON NITRIDE, ASBESTOS, ALUMINA AND GRAPHITE Physical chemistry Division Report No. 8-68-3 The enclosed copy of the captioned report covers the preparation and properties of composites of aluminum with whiskers or fibers of boron nitride, alumina, graphite, asbestos and silicon carbide. Evidence of some small degree of reinforcement was obtained only with the boron nitride. These results illustrate the problems of whisker or fiber damage and reaction with the matrix that can prevent reinforcement. Possibly, the reinforcing agents will have to be coated with other materials to circumvent these difficulties in future work. C. N. COCHRAN CNC:dls Enclosure cc: W. A. Dean, ARL, N.K. J. W. Newsome, ARL, N.K. H. Y. Hunsicker, ARL, N.K. L. K. Hudson, East St. Louis K. J. Brondyke, ATC, N.K. Library, ARL, N.K. ATC 0035135