DAVID W. TAYLORNAVAL SHIP RESEARCH AND DEVELOPMENT CENTER

IBetheda, Md. 20

AN EXPERIMENTAL INVESTIGATION OF

BABBITT METAL BONDING

by

0 Michael G. Vassilaros

4.J

4JP4.!•i •Approved for public release; distribution unlimited.

i 4) MATERIALS DEPARTMENT

"AnnapolisRESEARCH AND DEVELOPMENT REPORT

May 1976

Report 76-0037

• p..

The Novel •Uip Romearch and Development Center as a U. S. JNevy center lor l•bierstoryeffort directed at •hcievinI Improved see and aie vehicles. It was fonned in Maorh 1967 bymaerging the Dinvid Taytor Model Basin at Carderock, Maryland with the Marine KfninmetingLaboratory at Annapolsa. Maryland.

Novel Ship Research and Development C~eterBetheida, Md. 20034

MAJOR NSRDC ORGANIZATIONAL COMPONENTS

TEC"NICAL DIRECTO 1

REPORT ORIGINATOR -

OFVICER.IN.CMARGE OF FICER.IN-CHARGECARDEROCK ANNAPOLIS

05 04

i~~i F -•AW Wc,, ... . .n

SYSTEMSDEVELOPMENT•! L DEPARTMENT I

hr AVIATION AND)SHIP PERFORMANCE SURFACE EFFVCTS

DEPARTMENT15 EOPARTMENWI

I16

S U COMPU TATION

DEPARTMENT AND MATHEMATICS

P17 D'PARTMENT

S A S PROPULSION AND""I�P !TCS, AUXILIARY SYSTEMS

DEP RT EN 19 DEPARTMENT 2

MATERIALS *CENTRAL

/ DEPARTMEN't INSTRUMENTATION26 DEPARTMENT

29

.• ,.. .......... 1

// ... .,,}•

min

t ~

UNCLASSIFIED______Sgr.UIRITY CLASSIFICATION OF THIS PAGE (Wheni Oe!a Entered)

REPORT DOCUMENTATION PAGE _ FREA INSTRUCTIONS O

1. 91POPRt NUM39EP 2. 00VT ACCESSION NO. 3 ~IPI V'S CATALOG NUMBERq

76-ý0057/___4.TITLE (and SbtIe)fo FRAk 0PRO OEE

(7 ~I~;xrERIMENTA.L ;NVESTIGATIOt4 OF,,DABBITT eer vl~ ,3TAL 'ýONDING. - - -~PfM~ _ItPOT UM

S. CONTOR11CT ORG FIRANT NUMISRF"

Michael G./Vassilaros

6. Petr~OmMINCA ORGANIZATION NAME AND ADORES.~ 10. PAOGRAM ELEMENT. PROJtCT, TASK~

AREA Q WORK UNI% NMu9111David W. Taylor Naval Ship R&D Center ak Area ZF .120Annapolis, Maryland 21402.

__________________________ Qrk. .Unixt 281 4-17011. CONTROLLING OFFICE NAME AND ADDRESS 12

Davi W.Taylor Nava.l h RD Cente myia,'Bethesda, Maryland 2008U AG17ES ____

lTrMN_1TORING AGENCY NAME A ADORESS0If different from ConItoIInjifffIe) 15. 5!CURITY CLASS .of this report)

Jnclassified __

Se ECLS~FICATIONIOOWNOfRAONo

14. DISTRIBUTION STATEMENT

Apprvedforpublic re~lease; distribujI lg iit-

17, DISTRIBUTION STATEMENT (of the r tetred In Block 20, It ffforenf from A.,jeorf)

1S. SUFPiCMENTAiR'NOTES

1.KEY WORDS (Conti~nue on rev@ers aide It noc..esez and l~nifyne by blc-ubr

Babbitt metal Bond strengthBearings CastingTinning

ABSTRACT (Ci)ntinuo on reveres tid. It nece.7eary and Identify by block numrber)

ISeveral methods of casting babbitt metal to steel were evaluatedfor bond streng~th by use of tensile-type specimens. The methodsinvestigated were the use of a tin bath, torch/tin stick, and"ELectroplated nickel to insure proper babbitt-to-steel bond, aswell as no tinning. of the methods performed, only the tin-bathmethod consistently produced a bond stronger than the babbitt

____ ____ ____ ___ ____ ____ ___(over)

DO ~ 7 1473 EDITION O0P 1 NOV 65 12 083OLIETE UNCLAZS IF lEDS/N OO2-O4~ 601 ISECURITi, CLASSIVICAT50ON e)F TMIS PAGE (Whom~ Veto Knisfed)

UNCLASS IFIED-1.141T'Y CLA~SSFICA~rIOM Of TH~IS V40I(WPIWI Date Efiored)

(1+ 20. Abstract (cont)4 ,metal itself. The bond strengths developed in the other

methods tested were erratic and always less than that obtainedwith the tin-bath method. The torch/tin stick method gavean adequate compromise between bond strength and convenientapplication. It was found that the oxide formed on the tinnedsurface which precludes good bonding could be effectivelyremoved lty the use of flux during babbitt casting.

(Author)

ft

UNCLASS IFIEDSECURITY CLASSIFICATION OF THIS PAGE(no' n Dots Entorod)

ADMINISTRATIVE INFORMATION

This report was prepared under Task Area ZF 61512001,Element Number 62766 N on Investigation of Babbitt Metal Bonding,Work Unit 1-28]4-170. The investigation was sponsored by NAVMAT(MAT 03L4) as part of the IR/IED program of the Centec underthe direction of Dr. A Powell (Code 01).

SACKNOWLEDGMENT

The author wishes to express his grateful appreciation forthe assistance and guidance given by Mr. Robert F. Gustafson(Code 2r(84) throughout the project.

LIST OF ABBREVIATIONS

o C - degrees CelsuisF - degrees Fahrenheit

m - metermmn - millimeterMPa - Megapascalspsi - pounds per square inch

7'

I;.S76-0037 i

A'l

TABLE OF CONTENrS

ADMINISTRATIVE INFORMATION iACKNOWLEDGMENT iLIST OF ABBREVIATIONS iINTRODUCTION 1

Background 1MATERIALS 2DESCRIPTION OF EQUIPMENT 2METHOD OF TEST 2

Bonding Methods 2RESULTS AND DISCUSSION 5CONCLUS IONS 7RECOMMENDAT ION 7TECHNICAL REFERENCES 8LIST OF FIGURES

Figure 1 - Drawing; Stainless Steel Fixture and Mild

Steel Blanks Assembly of which Forms theMold and for Casting the Babbitt

Figure 2 - Curve; Results of Tensile Specimen TestsPrepared by Methods I and II

Figure 3 -Curve; Results of Tensile Specimen TestsPrepared by Methods III, IV, and V

Figure 4 - Microprobe Scan photograph cf the Babbitt-to-Steel Interface of a Tinned SpecimenShowing the Distribution of Copper

Figure 5 - Microprobe Scan photograph of the Babbitt-to-Steel Interface of an Untinned SpecimenShowing the Concentration of Copper at theInterface

INITIAL DISTRIBUTION

76-0037 iii

P=MnG ?AM BLAN~WTn

INTRODUCT ION

BACKGRO(UND

The Navy's extensive use of babbitt bearings in surface ship"machinery requires that Navy tenders and repair ships have thecapability cf repairing babbitt metal bearings on board. Thepractice of casting babbitt within the limited facilities of atender's foundry has evolved over years to a "traditional" pro-cedure passed on by crew chiefs. This method is simpler thanthe bearing manufacture's methods which are designed to optimizebearing strength, and require a large, well-equipped facility. Inpresent tender practice, the babbitt lining of a bearing is bondedto the steel or bronze bearing shell by means of a tin interfacelayer. Normally, the tin is applied to the hot bearing shell,after which the shell is set up in a mold and molten babbittpoured in.

The primary difference in the methods is the way in which thetin interface layer is applied. Manufacturers are able to use abath of molten tin in which the hot bearing shell is immersed;however, aboard ship the shell is heated by a flame torch, anda stick of tin is melted on the shell, much as solder is appliedto connections. The hot tin layer oxidizes rapidly, whilecooling, and the subsequent bond is often weak and brittle.

Many turbine qear, and lineshaft bearings are renewed annuallybecause of poor bonding. Furthermore, rebabbitting bearingsfrequently requires excessive man-hours of work because ot theunsatisfactory bonding and consequent rework.

The traditional method, as well as modifications, andalternatives to this method were investigated to develop andrecommend a simple, improved method of bonding babbitt bearinglinings to metallic bearing shells. The investigation intendedto characterize the important features of a simple babbitt-bonding procedure for possible use in a tender's foundry.Various babbitt-bonding procedures were used to make tensile

of these specimens was used to evaluate the relative efficiency

.7.1c several bonding methods, as compared to the standard method(o' using a tin bath for applying the tin interface layer.

means of improving the torch/tin stick method as well as

methods to eliminate tinning, and electroplating methods werestiai.ed.

'Superscripts refer to similarly numbered entries in the Techni-cal References at the end of the text.

76-0037

MATERIALS

The babbitt metal used was a tin-based babbitt alloy number2 (Navy designation). The alloy is commercially available undera variety of trade names all with a nominal composition of 7%-8% antimony, 3%-4% copper, and the remainder tine (88% minimum).Commercially pure tin metal was used in the tinning processes.The flux, when employed, was a standard solution of zincchloride made from zinc metal dissolved in concentratedhydrochloric acid. Low carbon steel bars (AISI IC15), 1/2 inchx 1/2 inch x 20 feet (12.7 m x 12.7 mm x 6.096 m),* were used .as the base metal to which the babbitt was bc.ided.

"DESCRIPTION OF EQUIPMENT

The babbitt was cast into a stainless steel fixture (figure1) which held the steel bars in a gapped position during casting,so that the steel bars and the fixture when assembled formed themold. A stainless steel crucible in a pot of molten lead heldthe tin bath. When torch tinning was performed, an oxy-acetylenetorch adjusted to give a reducing flame was used to heat theassembly, and the tin was applied with a tin stick, a long thinbar of tin metal. The flux solution was applied directly with aflux brush, i.e., no flux-cored sticks were used,

METHOD OF TEST

The strength of the babbitt-to-steel bond was measured usinga tensile specimen composed of the two 1/2-inch square (12.7 mm)steel bars whose ends were joined in a slug of cast babbittapproximately 3/4 inch (19.05 Nm) long and 1/2 inch (12.7 min)square. The specimens were pulled in a universal testing machinewhich recorded the maximum load to failure. The maximum stresswas calculated from the maximum load and the cross-sectionalarea of the fracture.

BONDING METHODS

Five different babbitt-to--steel bonding methods wereinvestigated, and these are described briefly in the following.

*A list of abbreviations used appears on page i.

"76-0037 2

AWITW..

Method I - Standard Method (Tin Bath)

This method simulated the recommended procedure for achievinga strong babbitt metal-to-steel bond, and is used in the manu-facture of new babbitt bearings. The steps pezforinedwere asfo illows:

* Sand the steel surface (smooth) to remove rust.

e Degrease the steel surface with acetone.

o 7lux the steel with a solution of ZnCl.

* Hold in a tin bath for 1 minute at 7000 F (5710 C).

e Insert tinned steel blanks into stainless steelfixture and preheat in furnace.

0 Cast liquid babbitt into mold cavity.

* Stir molten banbitt and cool mold with flowingwater.

Method II - Torch TinninQ

This method simulated present Navy tender foundry practices.The characteristic feature of this method was the use of aheating torch/tin stick instead of a tin bath. The steps were asfollows:

o Sand and dc.grease the steel bars with acetone.

a Heat the steel with an oxy-acetylene torch whileapplying ZnC! solution and tin from a tin stick.

* Insert tinned blanks into casting fixture.

o Preheat fixture assembly.

o Cast liquid babbitt into mold cavity.

* Stir molten babbitt and cool mold with flowingwater.

Method III-- No Tinning

This niethod simulated a bonding method which eliminatedthe tinning steps to find if flux alone might give adequatebonding. The steps are listed below:

e Sand and degrease the steel bars with acetone.

e Place cleaned steel. blank into casting fixture.

76-0037 3

* Preheat fixture assembly.

* Pour flux into fixture just prior to casting.

* Cast liquid babbitt into mold cavity.

* Stir molten babbitt and cool with flowing water.

Method IV- Electroplated Nickel

This method incorporated a substitute for tinning, whereina portable electroplating rig was used to electroplate nickelonto the steel, instead of tinning the surface with molten metal.The method was as follows:

0 Sand and degrease the steel bars with acetone.

* Electrcplate nickel with commercially availableplating kit. 3

e Insert plated steel blanks into the cast fixture.

* Preheat fixture assembly.

e Pour flux into mold just prior to casting.

e Cast liquid babbitt into mold cavity.

SStir molten babbitt and cool with water.

Method V - RebabbittingtThis method simulated the rebabbitting of previously cast

bearings to determine whether the tinnihg or nickel platingcould be reused by melting off the old babbitt and recastingnew babbitt.

'1 • Melt off babbitt from previously babbitted speci-mens.

* Insert the steel blanks into casting fixture.

• Preheat fixture assembly.

• Pour flux into mold just prior to casting.

* Cast liquid babbitt into mold cavity.

e Stir molten babbitt and cool with flowing water.

'76-0037

It has been found' that a tin-bath temperature of 660' F(3490 C) to 9300 F (4990 C) is necessary for good tinning ofsteel; accordingly, a temperature of 7000 F (5710 C) was chosen

as the tin-bath temperature used for method I.

The babbitt temperature in the ladle at pouring vas 7000 F(3710 C). Since the mass of tha mold was approximately 190 timesthat of the babbitt and the mold material's heat capacity wasover 2.5 times that of the babbitt, the cooling and solidifi-cation was controlled by the mold preheat temperature. Thetemperatures used for preheating were chosen to cover the rangeof recommended mold and babbitt temperatures, as found in theliterature.102,4 Preheating was don~e in an electric mufflefurnace, with the temperature measured by a type K thermocoupleon the mold surface. The temlierature of the mold assembly atthe time of removal from the furnace was recorced; the babbittwas cast in the mold within 10 seconds after the mold assemblywas removed from the furnace.

RESULTS AND DISCUSSION

The results of the tensile tests performed on the babbitt-bonded assemblies are given in figures 2 and 5, including boththe average and range of breaking strengths for each group offour specimens cast by the method and at the temperature indi-cated.

The results for method I, the standard method, using a moldpreheat temperature of 6200 F (3270 C) and 7000 F (3710 C) showthat the breaking strength varied greatly in the eight specimenstested. only one of the specimens, with a breaking stress of16,200 psi (112 MPa) failed in the babbitt rather than at thebabbitt steel bond. The other seven specimens failed at thebond with breaking strength as low as 2,600 psi (17.9 MPa).All of the seven specimens with bond failures had oxidescovering 30% to 90% of the fracture surface. Theoretically,the oxide layer on the tinned surface should float to the sur-

I ' face of the babbitt arid, therefore, not degrade the bondintegrity. Method I was modified to alleviate the problem ofthe tenacious oxide film. Pouring a small amov;nt of fluxsolution into the mold prior to casting the babbitt allowed theflux to float on the babbitt and remove the oxide as the levelof the babbitt rose in the mold. This method, IA, gave excellentresults, as seen in figure 2, where all of the 20 specimenstested failed in the babbitt. The average breaking strength ofall the specimens was 13,500 psi (93 MPa). Hence. the bond strengthmust have been higher than the breaking strength. Although thismethod gave good results, the size of the tin bath necessary tohold a large bearing shell makes it impractical for shioboard use.

76-0037 5

Method II was the result of a survey of procedures used onfour tenders (AD 19, AD 26, AD 38, and AS 36). The only surfaceconditioning used was acetone degreasing and sanding. The blankswere tinned with an oxy-acetylene torch adjusted to give areducing flame. Torch tinning dces not requ:ire a tin bath orany floor space for permanent equipment. The results for eightspecimens done by method II are shown in figure 2. The failuresalways occurred in the bond. The series performed at a mold of6700 F (354" C) contained some trapped oxide on all of thefractures, and the average breaking strength was 23% to 46%' lowerthan for method IA. To remove the trapped oxides during casting,method II was also modified to include the additional step ofpouring flux into the mold prior to casting. The results for the4modified method, IIA, are shown in figure 2. The averagebreaking strength of method IIA was higher than method II butlower than method IA. Of the 20 specimens tested, four failedin the babbitt, and the remainder had from 10% to 50% of theirfracture in the babbitt. The fracture surfaces indicated thatthe bond strength approximated the strength of the babbitt.

Method III was investigated as a possible method ofsimplification which eliminated the tinning step entirely.Surface preparation was limited to degreasing with acetone andsanding. In all the tests, flux was poured into the mold justprior to casting. Although the average bond strength in sometests of method III approach that of method IA, all of thespecimens prepared by method III failed at the bond. This mayhave been due to a brittle, copper-rich phase of babbitt formingon the steel surface when a tin layer is not present. Thisdifference in phase distribution may be noted in comparingfigures 4 and 5, which are microprobe scan photographs of thedistribution of copper at the interface between the steel andthe babbitt.

Method IIIA was performed to determine the effect of anP oil residue on the ste3el surface during preheating. Alter the

steel blanks were degreased and sanded, they were dipped inmachine oil and then placed in the stainless steel fixture.Further preparation was as for the other specimens prepazedby method III. The specimens were cast with the mold preheatedto 6500 F and the average breaking strength was reduced 16, fromthe average of the eight specimens prepared by method III at 6500 F.The range of breaking stresses was much greater with method IIIA.All four specimens failed at the bond, and three of the four hadtrapped 3il residue on the fracture surface.

76-0037 6

Method IV utili',ed a nickel-plating process to replace thetLnning process. A nickel surface was electrochemically appliedt-o the degreased arid sanded steel surface with a portable brushplating apparatus. 3 The results for the specimens prepared bythis method, figure 3, showed the procedure to be less effectivethan IA, IlA, and III. The breaking strengths were from 9% to[r' lower than 1A, and all of the 16 specimens tested failed atthe babbitt-to-nickel bond.

Method VA was used to investigate the rebabbitting ofspecimens prepared by method IA which were previously preparedand tested. The babbitt was melted off the steel blanks, andnew babbitt was cast to the same steel blanks with no othersurface preparation. Method VB was similar to VA, but the steelblanks had been prepared by method III. The reduction ofaverage breaking strength varied from 22% to 77% of the strengthof the original value. All of the 24 specimens rebabbitted andtested failed at the steel-babbitt interface.

CONCLUS IONS

In comparing the results of the five different methods ofcasting babbitt, several conclusions may be drawn:

0 The best method was the standard (tin-bath) method.

* The deleterious effect of oxides on the tinnedsurfaces can be eliminated through use of flux just prior tothe babbitt casting.

* The tinning of a steel surface is required for abond strength greater than that of the babbitt metal.

* Nickel plating is not a worthwhile alternative totinning. Although a clean metal surface without a tin layermay have a high breaking strength, the bond will usually beweaker than the babbitt itself.

RECOMMENDAT ION

* Aboard Navy tenders and repair ships where the torch tinningmust be used, it is recommended that procedures for these shipsbe specially documented and include the step of applying fluxjust prior to casting the babbitt, the procedures describedunder method IIA of this report.

7

7c -0037 7

TECHN ICAL REFERENCES

1 - Prytherch, J. C., "Measurement of the Adhesion of Tin-BaseBearing Metals to Various Backing Materials," Journal of th(Institute of Metals, Vol. 67, London (1942)

2- "Bearing Babbitting Procedures," NAVSHIPS 250-644-2, Bureauof Ships, Navy Dept, Washington, D. C. (May 1961)

3 - "The Dalic Process-An Advanced Engineered System of Selecti'.Brush Plating," Sifco Matachemical, Division of SifcoIndustries Inc., Independence, Ohio (1968)

4 -Bassett, H. N., Bearing Metals and Alloys, Butler and TannerLtd., London, U. K. (1937)

7 8

76007

*

ASBESTOS TWINE

~r 71/1N IL

MILD STEEL SI EELSTEEL

STAINLESS STEEL FIXTURE

Figure 1Stainless Steel Fixture and Mild Steel Blanks Assembly

of which Forms the Mold and forCasting the Babbitt

76-0037

MOI D TEMPERATURE BREAKING STRESS, KSIjMETHOD F ___ 2 4 6 101 14 94

"" "C?20 40 60 80 100I 620 327 I-----I 700 371

IA 580 304IA 600 316IA 630 3431A 660 349IA 700 371

.T. 560 293i.- --

i] 670 354

TT A 550 288 -- --

'IA 614 329[I'A 650 343 : • 'rlA 700 3 71 I--•"[IrA 750 399

L IA 9.50 510

...-4RANGE OF BREAKING STRESSES• AVERAGE Of FOUR TESTS

Figure 2Results of Tensile Specimen Tests Prepared by

Methods I and II

76-0037

MOLD TEMPERATURE BREAKNG STRESS, KS1METHOD 100 12 14 1

i'{, MPo

r' C 20 4G 60 80 100

III 480 2491M1 570 299131 640 338

TT 650 343131 650 343M 750 399

MliA 650 343 1-,

is[ 500 21260650 343 -- 1-4

S660 349Iz 770 410 i

YA 550 288.A 650 343

YA 650 343YA 700 371

YB 600 316""B 650 343 _ _....__

Figure

Results of Tensile Specimen Tests Prepared byMethods III, IV, and V

76-0037

S.. . . . . . . . . ,I, • : • • ' • • • . . • , , , :. . . . . .. .

Babbitt

...Interface

Figure 4Microprobe Scan Photograph of the Babbitt-to.-Steel Interface

of a Tinned Specimen Showing the Distribution of Copper(white dots), 50OX

-.a-Babbitt

' --Interface

Figure 5Microprobe Scan Photograph of the Babbitt-to-Steel interfaceof an Untinned Specimen Showing the Concentration of Copper

at the Interface (white dots), 500X

* v76-00357

INITIAL DISTRIBUTION

copies CENTER DISTRIBUTION

2 NAVMAT copies Code2 (MAT 03W 4)1

2 075 NAVSEA1

1 (SEA 03)127I(SEA 035) 1 28)1 (SEA. 03523) 15 28114)

2 (SEA 09G32) 1 2832)30 5211.4)2 52ý21)

5 NAVSEC 2 5222'2 SEC 6101D) 2 5231$

1 SC6140)1I SEc 6148D)

1 NAVSECPHILADIV

'12 DDC

*Adrsse

6 03,My17