r 66 969 THE EFFECT OF BRKING AND STRESS ON THE HYDROGEN CONTENT 1/1- OF CADNIUN PLRTED..(U) NRVRL RIR DEVELOPHENT CENTERHRRM[NSTER PR RIRCRRFT RHNO CREW S.. D R BERMRN DEC 85

LCLASSIFIED NADC-8635-60 F/G 11/6UNLSIAINIE PIRCAF ADCRIIIIAIEIN EC8EIIIIIIIIIIIIl

9 - 1 4 .0 2 .

1A"

1.0 34 *2 1.8

111111-1.

MICROCOPY RESOLUTION TEST CHART

NAIOA OUP~EA.U OF STAN0400S - - A

REPORT NO. NAOC-86035-60P

00 0

THE EFFECT OF BAKING AND STRESS ON THEHYDROGEN CONTENT OF CADMIUM PLATEDL

HIGH STRENGTH STEELS

0. A. BermanAircraft and Crew Systems Technology Directorate

NAVAL AIR DEVELOPMENT CENTERWarmitnster, Pentnsylvania 18974-5000

DTICDECEMBER 1985 ~L G EqIMAY 8 1986

PROGRESS REPORT

Approved for Public Release; Distribution Unlimited

1.4.- ~Prepared for . .'4

NAVAL AIR SYSTEMS COMMANDDepartment of the NavyWashington, DC 20361

865 8

NADC-86035-60

I",.e

NOl ICES

REPORT NUMBERING SYSTEM - The numbering of technical project reports issued by the NavalAir Development Center is arranged for specific identification purposes. Each number consists ofthe Center adronym, the calendar year in which the number was assigned, the sequence number ofthe report within the specific calendar year, and the official 2-digit correspondence code of theCommand Office or the Functional Directorate responsible for the report. For example: ReportNo. NADC-78015-20 indicates the fifteenth Center report for the year 1978, and prepared by the i,Systems Directorate. The numerical codes are as follows: .-

CODE OFFICE OR DIRECTORATE

00 Commander, Naval Air Development Center01 Technical Director, Naval Air Development Center02 Comptroller10 Directorate Command Projects20 Systems Directorate30 Sensors & Avionics Technology Directorate40 Communication & Navigation Technology Directorate50 Software Computer Directorate y ,60 Aircraft & Crew Systems Technology Directorate70 Planning Assessment Resources80 Engineering Support Group

PRODUCT ENDORSEMENT - The discussion or instructions concerning commercial productsherein do not constitute an endorsement by the Government nor do they convey or imply the ' "license or right to use such products.

APPROVED BY. G L' DATE :,-/f'"1CPT,-1 SC, L .NAVY

° • %°

* '

Unclassified NADC-86035-60 A -A '(7tC ,R!Tv C..ASS,F.CAT1ONj OF --. s 51MI r

REPORT DOCUMENTATION PAGEa REPORT SEC' RI7Y CLASSIFCA7;ON 1b RESTRiCTivE MARKINGS

UnclassifiedZa SEC'.,RITY C.ASS.CATiON AUTHORITY 3 DISTRI8UTION,'AViAABILITY OF REPORT

'b )ECASS FiCATION DOWNGRA~GS1EU. Approved for Public Release;Distribution is Unlimited

4 PERFORMiNG ORGANIZA7rlON REPORT NUMBER($) 5 MONITORING ORGANIZATION REPORT NUMBER(S)

NADC-86035-60 N/A6a %A.VE OF PE.RFORMING ORGANIZATION 6o OFFiCE SYMBOL 7a NAME OF MONITORING ORGANIZATION

Naval Air Development Center N/Ak ADDRESS O~ty, State. and ZIP Code) 7b ADDRESS (C, State, and ZIP Code)

Warm inster, PA 18974 N/A3a NAME Or ; %DNG SPONSORNG 8b OFCE SYMBOL 9 PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER

OPGAAI;ZA- ON (If applicable)* Naval Air Systems Command

3 c ADDRESS (City. State. and ZIP Cooe) 'O SOURCE OF FUNDING NUMBERS

PROGRAM PROJECT ITASK WVORK UNIT

Washington, DC 20361 ELEMENT NO NO NO ~ ACCESSION NO

'!LE (Include Security Classification)

* The Effect of Baking and Stress on the Hydrogen Content of Cadmium Plated High Strength Steels

* 2PERSONAL A- I40R(S)D.A. Berman

3a -*DE 0: IEDORT 13b - ME COvERED 14 DATE OF REPORT (Year. Month Dy 15PG OUNT* Progress :Rom. TO__ ___ 1985 December

6 S .PPLEVENTARv NOTA-ON

7 COSA71 CODES 18 SUBjECT TERMS (Cont inue on reverse of necessary and identify by block number)

GROP SB-GOUP Cadmium Plating Hydrogen E mbrittlement

High Strength Steels Hydrogen Determination9AjS7RAC' 'Continue or reverse of necessary and idenrifye by block numnber)

* Variations in the diffusible hydrogen concentration of cadmium plated AISI 4340 pnd 300M steels due tobaking and stress are reported. Long baking times of greater than 100 hours at 190',C are needed to remove

* all the hydrogen from bright cadmium plated specimens, and around 40 hours for dull cadmium platedones. Very short baking not only does not remove hydrogen, but actually pumps it in. Baking for 96 hoursis insufficient to remove hydrogen embrittlement from sharp, double notched 300M tensile specimens havinga notched tensile strength of 1800 MPa. It is shown that there is no difference in the hydrogen concentra-

* tion whether baking immediately follows plating or there is a delay. Also, the hydrogen concentration doesnot change with time for cadmium plated specimens standing for one month at room temperature. Stressing,on the other hand, causes an increase in the hydrogen, presumably by allowing it to enter from the plate due

* to lattice deformation.

*20 D5PS'4BuT ON AVAILABILITY OF ABSTRACT 21 ABSTRACT SEC,-RiTY CLASSiFiCATION

%2 A%'- 0 RESPONSBLE '%DivIDUAL 2,b TELEPHONE (include Area Code) 22c OFFCE SYMBOL WD.A. Berman

* DO FORM 1473, 34 VAR 83 APR eto lADce .sec urtii exntaistec SEC_.IITV CLASSFICATION OF -HIS JAGEAvll tcV Coors a~e Loso'ete

NADC-86035-60

TABLE OF CONTENTS

Page

List of Tables. .. .............................. iiList of Figures.. ...................... .. I.. .........Abstract .. .......... ......................... IIntroduction .. ........ ......................... IExperimental .. ......... ........................ 2Results and Discussion .. ......... .................. 3Conclusions. .......... ........................ 6Acknowledgements .. ......... ..................... 7References .. .......... ........................ 7

DTIC VTf

~ELECTE~MAY 8 1986

NADC-86035-60 '

p 4.

-. LIST OF TABLES b

Table Page

I 12-Run Plackett-Burman Experimental Design ...... ............. 8* 2 Variables and Limits for Plackett-Burman Experiments .... ........ 8

3 Hydrogen Concentrations and Sample Calculations for Plackett-Burman Screening Experiments ............. ............. 9

4 Factor Effects and Minimum Significant Factor Effects, [MIN), forthe Plackett-Burman Experiments ......... .................. 9

5 Two-variable, Two-level Experimental Design and Results ....... . 106 Effects of Baking and Stress on Failure Time of 300M Steel

(Preliminary Experiment) ............... .............. 107 Effects of Baking and Stress on Failure Time of 300M Steel .... .... 118 Effects of Baking and Stress on Hydrogen Concentration of 300M

Steel ............. ............................... .. . 119 The Effect of Short Baking Times on the Hydrogen Concentration of

Cadmium Plated 300M Steel ....... ..................... ... 1210 The Effect of Very Short Baking Times on the Hydrogen Concentration

of Cadmium Plated 300M Steel ....... .................... ... 12

LIST OF FIGURES

Fi gnrePage

I Efficiency of baking treatment for removal of hydrogen from brightand dull cadmium plated AISI 4340 steel .... .............. ... 13

2

C0RR0S10PAPER NUMBER

The International Corrosion Forum Devoted Exclusively toThe Protecton and Performance oi Materals192 March 25-29, 1985 1 Sheraton Hotel

Hynes Auditorium I Boston, Massachusetts

THE EFFECT OF BAKING AND STRESS ON THE HYDROGEN CONTENT OF CADMIUM

PLATED HIGH STRENGTH STEELS

D. A. BermanNaval Air Development CenterWarminster, PA 18974-5000

ABSTRACT INTRODUCTION

Variations in the diffusible Electrodeposited cadmium is widelyhydrogen concentration of cadmium plated used as a sacrificial coating to pro-AISI 4340 and 300M steels due to baking tect low alloy, high strength steelsand stress are reported. Long baking from corrosion, especially in marinetimes of greater than 100 hours at environments.' It is generally recog-190 C are needed to remove all the hydro- nized, however, that hydrogen producedgen from bright cadmium plated speci- during this plating can diffuse to areasmens, and around 40 hours for dull of high stress causing embrittlementcadmium plated ones. Very short baking and possibly leading to catastrophicnot only does not remove hydrogen, but failure of a structural part, such asactually pumps it in. Baking for 96 an aircraft landing gear.2 Becausehours is insufficient to remove hydrogen of this, plated parts are baked toembrittlement from sharp, double notched lower the hydrogen concentration and100M tensile specimens having a notched thus reduce the possibility of failure.tensile strength of 1800 MPa. It isshown that there is no difference in the Specifications for baking arenydrogen concentration whether baking often ill-defined, sometimes callingimmediately follows plating or there is for as little as 1-5 hours at a widea delay. Also, the hydrogen concentra- range of temperatures or stating thattion does not change with time for the time and temperature should becadmium plated specimens standing for determined Por the material and in-one month at room temperature. Stress- tended use.3 Thus, in the aircrafting, on tne other hand, causes an industry, cadmium plated parts areincrease in the hydrogen, presumably baked for 24 hours at 375 F (90 C) forby allowing it to enter from the plate the hignest strength steels. This has4ue to lattice deformation. been shown to be insufficient to elim-

inate embrittlement from a part,

%,.

NADC-86035-60

especially when it is covered with a of a material to hydrogen embrittlement. .bright cadmium plate electrodeposited It is therefore important that methodsfrom conventional cyanide baths.4 ,5 are found to correlate the diffusible

hydrogen content and susceptibility toCommon practice and specifications embrittlement, especially of cadmium

are usually to "bake as soon as pos- plated high strength steels. Hence,

sible after plating."3 This is done the present work to study the effectsto prevent cracking due to residual of baking and stress on the hydrogenstresses. Also, it is often assumed concentration and delayed failure wasthat early baking is necessary undertaken.because the hydrogen is more easily R._removed at this stage. Baking immedi- In recent years, statisticalately may be good housekeeping as it methods have been used, not only inassures continuity of the procedure so analyzing experimental data but, asthat the baking is not omitted; how- importantly, in designing the experi-ever, whether or not there is any ment so as to get the most informaton.benefit from a hydrogen viewpoint has from the least number of specimens.'

9

not been established. Due to the complexity of the plating/baking procedures, statistical design

lecause of the difficulties with concepts were therefore used nere.hydrogen embrittlement when usingbright cadmium, low embrittlement EXPERIMENTALplating baths are often used. Theseproduce duller, more porous plates Materialswhich lose hydrogen more easily. Thedull cadmium is not as attractive or The specimens used in the screen-:rotective as the bright, however. ing experiments were 38 mm square,Therefore, the latter is still pre- 1.0 mm thick pieces of AISI 4340 steelferred for many uses. heat treated to 260-280 ksi (1800-1900

MPa) strength level. The remainder ofTraditiona.ly, mechanical testing the studies were conducted with 300M

has been the only way to determine the steel heat treated to 280-3CC k5ieffectiveness of the baking procedure. (1900-2100 MPa) strength level. TheSince the development of the barnacle double notched 300M tensile specimenselectrode method,-D however, it is were 6 in. long and 3.063 in. thicknow possible to measure diffusible (150 and 1.60 mm), and had a geometri-temorittling) hydrogen and thus to cal stress concentration factor, k.,monitor baking efficiencies. The of 5.6. The notched tensile strengthoarnacle electrode technique is based (NTS) was 260 ksi (1800 MPa). Theon the electrochemical hydrogen per- 300M coupons were 75 x 35 mm, eithermeation method. :f hydrogen is in a 2.0 or 6.0 mm thick. All specimens:art which is made the anode in an were fin'shed by surface grinding.electrochemical cell, the hydrogen isoxidized as It diffuses to the surface. ProceduresThe oxidation current can then be re-iated to the hydrogen content of the The bright cadmium electroplatepart. Correlations with mechanical was applied from a cyanide bath, con-testing are, of :ourse, still neces- taning brighteners, at a currentsary to leter.ine safe levels of hydro- density of 25 A/ft2 '25 mA/cm2 ); thezen for the material and use. dull from a cyanide bath, containing

carbonate, at a :urrent density ofStressing procuces a :attice ex- 6C A/ft- 60 mA/cm-'. A!- plates were

zansicn wn:c causes hycrcgen to move 3.3005 in. ,15 um) tnick, and alla.zng the stress gradient and. oyer a scecimens were oaKed at 37= 7 !'..:ertzd f ime, t3o :uild u o ign The cadmium was removed for nydrogenLevels. Secause of this time depend- measurements ising ammonium nitrace -l

once. measurement :f -he -ime to solution, follcwed ny :ignt acradingfai-ure n a sustained zons:ant) loac with Scotznorte.test -s ised :c cred 4 " susceot*-t::tv

2'2

NADC-86035-60

All hydrogen measurements were determined by comparing the hydrogenmade using the barnacle electrode concentration for the specimens inmethod. Details of the equipment and which that variable was at one levelprocedures are given elsewhere.5 The to those at the other. This was done "P.30-minute extraction current densities by adding the Y-values for all thewere used to give the relative diffu- pluses in the column, subtracting tiesible hydrogen concentrations. Measure- Y's of the minuses, and dividing byments on stressed specimens were made six (the number of pluses or minuses)as soon as possible after failure, at to get the factor effect for theleast 25 mm away from the fracture variable. This factor effect was thensurface. compared to the minimum significant

factor effect, (MINI, to determineThe tensile specimens were loaded its statistical significance.

in constant load stress rupture ma-chines for the time to failure The [MINI was calculated usingdeterminations, the unassigned columns. First, the

unassigned factor effect, UFE, forIn all experiments, randomness was each column was calculated in the

used where necessary, such as in the same way as above. This was thenplating bath, oven and loading frames used to determine the significantto ensure statistical reliability, factor effect, SFE, which is an esti-

mate of the experimental error,Statistical Experimental Design

Screening experiments were run to qdetermine the effect of five variables i2

on the diffusible hydrogen concentra- SFE UF)

tion of both bright and dull cadmium izlplated 4340 steel coupons. The 12-runPlackett-Burman experimental design8 ,1 0 qis shown in Table 1. Each column has

the limits for a particular variable where q is the number of degrees of

given, the plus or minus standing for freedom (the number of unassigned

high/low, #1/#2, etc. Column Y is for -m T

the response, the measured hydrogenconcentration. The variables and their [MIN] t'SE (2)limits are given in Table 2. As theumnscoicefassigning variableS to col- where t was obtained from the t-distri-

;arlables were assigned to the first bution table using q=6.

five columns of Table 1. This auto- To determine the importance of thematically defined which of the two variables, their factor effects werelevels (e.g., baking for 8 or 48 hours) then compared to the [MINI. If thewas to be used for each specimen. absolute value of the factor effect isThus, for column X1 (Table 1), trials larger than the [MIN), then thespecimens) 1, 2, 4, 5, 6 and 10 are variable is statistically significant.8

ai a pluses wnich means plating batchDne Taole 2) for the bright cadmium RESULTS AND DISCUSSIONexperiment and batch two for the dull.Likewise, specimens 2, 3, 4, 8, 10 and Screening ExperimentsiI were baked for 48 and 20 nours fortne orignt and dull, respectively. The hydrogen concentrations for7he remaining six columns XA- XT, the 12 specimens, as measured 1y rhi(,live a Indcm Jistributinn 3f pluses narnacle electrode method, are !voriind inuses wnicn were used -o deter- in Table 3 for both the brignt 3nt Iull-nine the experimental errozr. cadmium exoeriments; also given is

column (3 from Table 1 and a sampieStatistical Analysis :alculation for the factor effect f~r

:aking. The average difference in the7he effect cf each variable was hydrogen concentrations for the

3

............................................................................. . .-..

- %;

NADC-86035-60

specimens bared for the short and long long times at very high humiditiestimes was calculated by taking the could alter these conclusions.aosolute difference of the sums anddividing by six. Thus, for the bright Because of the concern voiced bycadmium experiment, the sum for speci- many plating people that a delaymens 2, 3, 4, 3, 10 and 11 is 3.26; for between plating and baking could bespecimens 1, 5, 6, 7, 9 and 12 is 5.36. important, it was decided to run a

. The difference is 2.10; and the factor check experiment with just the twoeffect is 0.35. This factor effect was variables: the baking time and thethen compared to the minimum signifi- delay before baking. Bright cadmiumcant factor effect, [MINI, to determine plated specimens were baked for 3 andits importance. The latter is given in 72 hours, with dealys before baking of7 able 4 for both the 90 and 95% confi- . and 24 hours. Duplicate specimensdence levels. The factor effects for were run for the four combinations ofthe five variables are also given in high and low values as shown in

S-able . One can see that for the Table 5. Also given in Table 5 areorignt cadmium, only the baking time the diffusible hydrogen concentrationsis significant, whereas with t"he dull as determined by the barnacle electrode:admium, tne batch is also significant method. As can be seen, the measure-starred values. The reason for this ments on the specimens oaked for :he

-z3 probably that small changes in con- short time are in excellent agreement,

,entrat on of the bath ingredients, as are those taked for the long time,especialy carbonate, can have a sig- regardless of the delay time; i.e.,

* niflcant effect on the nature of the trials 1 and 2 agree with each other,iecosi: from a dull cadmium bath. In as do 3 and 4. It ras thus been shown,onis experiment, the two batches were unequivocally, that there is no effectPlated one week apart. Tn the case on tne hydrogen concentration wnetterof the Drignt cadmium, the batch could or not the baking is done "as soon as '"oecome a significant factor for more possible" after'plating.extreme ohan.ges :n toe plating condi-:crs, :.n a1. subsequent tests in Effect of Baking4nicn two or more plating oatcnes wereised, trne blating was always done on As we have seen, baking (and to-ne sane lay oy the same technician, some extent, the plating batch) is theAso, scecimens were distributed so as only variable which was found to affect0c eliminate a possible batch effect. the final diffusible hydrogen :cncen-

tration. The effect of baking time on.he factor effects for the other the removal of hydrogen from cadmium

variaoles were shown to be statisti- plated 4340 steel is illustrated inta.-7 ncn-significant, Tamle 4. Thus, Figure J. The curve for the Jullt:aKing 4.mediately, or a day after cadmium plated specimens is the best-:atini will not affect the hydrogen fit for the 12 points from the five-:oncentratlon. 'This does not, of variable screening experiment; thatcourse, zreclude the possibility that for the bright cadmium plated is fromnaterials under stress might be damaged the two-variable exoeriment. For theif takinz is delayed.) The delay in latter, data from an intereCiateneasurtna tne hydrogen also was found taking time of 37.: hours not snown int. te icn-signzfcant jp to 690 hours Tacle 5 was also used. ExtracclatIcnsacr~x~mateT7 .one month) at rocm of the cur,:es show tnat zcmalete taKe-.zemeriture and Tcderate numidities. out would ce oDtained at times 4e-!-Tis -eans that e!Zyed failure testing tevond any normal taking :rocecures.,s c:ffuzle z'drcTen measurements can -t wcul ta.e over " c hours for scec--. _ an hvig e 'rrzgt :acmium ::atzg, "*c race worn -:r..f-ece tat tne vdrc- mens nav:ng .nee_. :orcen:rsci:n will nc: :anze witn and even 4C -curs !:r tne :ull. "-e

a:-. "ia eans -nat :nterlaccra- "no n.'drogen" level is -ne neasuremen':r oer testing, inol;:-i'ng :Ca:ned -ion olacec ar.n szrizcec T

:e .aved rTen measuremens :an ce scecimens whrcn -were al'owec t ezas:cne n:tr: need : ocrc:na-e tre in a .esz4:azor. -ne verv lw concen-"1 foe :r Eeasure-ents :xoreme-" -ratlcnS 3f rvdro:en neasurec fzr o.e

44[2

. . . . . .. . ."

_,-'J -_'~ . _ -,t"J J'Z 4 i i L_- X ," - --.-.... .-...-. _-.. . . . .- ".. . .-. ''....... ."."- -"--- - - ... - ' -. --

NADC-86035-60 V

longer baking times are much too small ment are given in Table 7. Again theto be calculated quantitatively, but trends are apparent. Thus, specimensare well under 0.1 ppm. 5 baked for 72 hours and stressed at 75%

NTS failed in shorter average time thanEffect of Stress those in the 96 hr/75% condition, and

the former in shorter time than those

Next, constant load experiments stressed at 25% NTS. Comparingwere run with 300M steel tensile speci- Tables 6 and 7, it can be seen that themens. Because the screening experi- largest spreads in times to failure arement showed that the hydrogen with the lowest stresses (25% NTS).concentration does not change over a This can be explained as follows. Whenmonth, it was thought that the hydro- stressing, hydrogen moves from the bulkgen could be measured on the failed of the specimen to regions of high 7specimens, and that this would give stress, due to lattice deformations.the same value as if it had been When loading at low stress levels,measured before loading. In early however, these lattice deformations aretrials, however, this did not seem to slight. Therefore, small variationsbe the case. There was some indica- in hydrogen concentration, or metallur-tion that the hydrogen concentration gical and mechanical factors are prob-may have been affected by the stress- ably very important in affecting theing. Even though it is well known life of the specimen. Contrast thisthat stressing produces diffusion of with the effects for the 96 hour bakehydrogen to regions of high stress, it and 90% load. Here the lattice de-is not apparent that the bulk hydrogen formation is great, thus causing theconcentration is affected to any hydrogen to build up to damaging levelsmeasurable extent upon stressing. in relatively short times, even thoughTherefore, the effect of stress on the the starting hydrogen concentration washydrogen concentration was investigated low after the long baking time.by measuring the hydrogen in control(unstressed) specimens and comparing To determine the effect of stressit to that in stressed and failed ones. on the hydrogen, the concentration of

diffusible hydrogen was determined onA preliminary two-variable, two- the stressed specimens after failure,

level, constant load experiment was as well as on the unstressed ones.conducted to determine appropriate The results are given in Table 8.baking times and loads for the 3COM Looking at the hydrogen concentrationssteel. Duplicate specimens were run for the control specimens, Table 8, itfor each bake/load condition. The can be seen that over 100 hours ofresults of the time to failure tests baking at 190 C is needed to removeare given in Table 6. The failure all the diffusible hydrogen. This istimes were generally very short, with in agreement with the screening experi-very little discrimination in times for ments using 4340 steel. The very lowtne different bake/load conditions. level of hydrogen after baking forThus all specimens failed within seven 96 hours is still embrittling, however,hours except for the 48 hr!25% condi- as seen in Table 7. It should be notedtion. The expected trend is obvious, that the background current densityhowever: that is, the shortest times level for the screening experiment isto failure are at the lower left different from that in the stressingsnort bake/'high load) and increase for experiment. This value must be deter-Longer taking time or lower load. mined for each set of specimens, as it

depends on both the material and theue to the short failure times in surface finish.

-he ::rel,_m;nar'y experiment. Longer "

.:aKZ.., -es ,iere selected for the Zomoartsons of the nydrogen con-,nain ex:eriment 4n order to reduce :entratlons for the stressed ana un-the nydrcen levels and extend tne stressed specimens ,tne Latter :omoined"imes to failure. Tripii:ate speci- according to oaki<ng time without retardners were run for each oake/load to load). Table 3, snow that stressing:cnc:tIon. The results *Lf tnis exoeri- toes, indeed, affect the nydrogen

5' "

NADC-860 35-60 L

concentration. For all baking times, 2 mm and 6 mm thick specimens, respec-the stressed specimens had higher tively, significant differences at thehydrogen levels than the controls. 95% confidence level were found. Thus,

Further, a two-factor computer analysis it has been shown that very shortof variance was run using the Statis- baking (1-5 hours) does not remove anytical ?ackage for the Social Sciences hydrogen and may, in fact, cause an¢SPSS).I1 It was found that there was increase.a significant difference in the hydro- sdwe rngen concentration for the stressed CONCLUSIONS ..:.

versus the unstressed specimens at the~95% confidence level. Several experiments were run to 2

study the effects of baking and stressOne would normally expect a loss of on the diffusible hydrogen concentra-

hydrogen through the fracture surface tion of cadmium plated nigh strengthduring cracking, a process that often steels.takes several hours to complete. There-fore the increase in hydrogen can only Screening experiments were per-be explained by the addition from a formed to determine the statisticalsource, such as from the cadmium olate significance of various parameters on 1(or possibly from traps in the steel). the hydrogen content of bright andStressing causes a lattice expansion dull cadmium plated, and baked,in both tne steel and the cadmium. 4340 steel. It was shown that theThe hydrogen in the cadmium could then baking time was essentially the onlymove into the steel, and along the significant variable influencing thestress and hydrogen concentration final hydrogen concentration as deter-gradients to the notch area, giving mined by the barnacle electrode method.noth the buildup to cause failure, and However, variations between platingan increase in the bulk hydrogen con- batches were found to be significantzentrat4on. for the dull cadmium plated material.

Sffect of Short Baking A delay of 24 hours in bakingafter plating had no effect on the

* Since even long baking times (up final hydrogen concentration. Also,to 16 nours) were shown to be insuf- the hydrcgen concentration was notfcient to remove hydrogen, it was changed if the plated specimens werea cec.ded to investigate the effect of held for a month at relative humiditiesshort oaking as specified, to see if up to 50% after baking.:here was any benefit at all. Rec-tangular coupons (76 x 38 mm) of 300M It was found that the hydrogensteel were plated with bright cadmium, measurements decreased exponentiallyand baked for different times. Table 9 with the baking time, and that at leastgives nydrogen concentrations for 6 mm 100 hours (at 190 C) were needed totnick specimens baked for times up to bake out all the detectable hydrogen24 hours. it can be seen that there is from 4340 and 300M steel specimensno removal of hydrogen in these times. plated with 0.015 mm of bright cad-:n fact, there is some indication that mium, whereas 40 hours were requirednydrogen may even !e increased during with dull cadmium.the first few hours. To check thisout, a second experiment was run using it also was shown that baking up

noth 2 and A m thick coupons. The to 24 hours had no effect on removing

n'crogen :concentratlons are given in hydrogen from oright cadmium platedTcle " t~: is clear that baking for material, and that hydrogen actual-7:nly an nour or two causes an increase increased during the first few nours'n -Iffu'sile nydrogen. tne time 3f caking.:ecencing cn the tnlckness of the

'secinen. s t-test :aiculaticns " :n constant load exDeriments. L:

c toinare :ne nydrogen concentration was founc :nat embrittlement still-,r [re ;roaKed zoectnens against those existed after 16 hours of caKin; it

* acCO Iir :ne and two nours. for :ne _'0 C fir orignt zacnmum nlatec 2>.

L 6::

. . . . . . .. . .

NADC-86035-60

steel having a NTS of 1800 MPa and a 6. D. A. Berman, W. Beck and

sharp notch, and that the hydrogen J. J. De Luccia, Hydrogen incontent of the steel increased after Metals, I. M. Bernstein and

stressing. It is proposed that hydro- A. W. Thompson, Eds., Americangen enters the base metal from the Society for Meals, pp. 595-607,plate during stressing. Therefore, the 1974possibility of hydrogen embrittlementof very nigh strength steels plated 7. J. Vokl, Ber. Bunsenges. Phys.with bright cadmium is always present, Chem., Vol. 76, p. 797 (1972).

given the right combination of stressand time. S. Strategy of Experimentation,

E. I. duPont de Nemours andACKNOWLEDGEMENTS Company, Wilmington, DE, 1975

The author wishes to acknowledge 9. J. W. Dini and H. R. Johnson,the assistance of L. Biggs and Plating and Surface Finishing,P. Sabatini in the experimental work; Feb, p. 52 (1981)M. Russo for assistance with thecomputer program as well as helpful 10. R. L. Plackett and J. P. Burman,discussions; and to V. S. Agarwala and Biometrika, Vol. 33, p. 305 (1946) il

J. J. De Luccia for stimulatingdiscussions. li. N. H. Nie, C. H. Hull,

J. G. Jenkins, K. Steinbrenner andREFERENCES 0. H. Bent, Statistical Package

for the Social Sciences, 2nd

F. A. Lowenheim, Guide to the Ed., McGraw-Hill, 1975.Selection and Use of Electroplated

and Related Finishes, ASTM STP 785, 12. J. C. R. Li, Statistical InferenceAmerican Society for Testing and "o1. I, Edwards Bros., Inc.,Materials, Philadelphia, PA, Ann Arbor, MI, p. 146, 1964p. 46, 1982

2. W. Beck, E. J. Jankowsky andP. Fischer, "Hydrogen StressCracking of High Strength Steels,".Naval Air Development CenterReport, NADC-MA-7140, December,1971.

3. Annual Book of ASTM Standards,Part 9, B242, Standard Practice for?reparation of High Carbon Steelfir Electroplating, AmericanSociety for Testing and Materials,Philadelphia, PA, 1980

. 3. F. Brown, Effect of Baking on-eiayed Fracture of Electroplated"!itra High Strength Steel, NavalResearcn Laboratory Report,MTRL- ,89, October 1956.

.e uccia and D. A. Berman,

Electrochemical Corrosion Testing,AS74 ST? 727, F. Mansfeld and

3ertocci. Eds., American Societyf:r Testing and Materials,? i:adei:nia, PA4 pp. 256-273, 1981

7

NADC-86035-60

TABLE 1 - 12-Run Plackett-Burman Experimental Design

Trial X1 X2 X3 X4 X5 X 6 X7 XXa X 9 X10 X11 Y

2 1. 6 7 8 9 0 l

3 - 4 4 4 - - 4

4-.

* - 4 - - - -"

3 4 4-0 - 4 461 4". 4" -

72

U - - 4 - 4 4 - 4 4 4

:)4_.frence -:

TABLE 2 Variables and Limits for Plackett-urman Experiments

Limits

Column Variable Bright Cadmium Dull Cadmium.-

* 2 - -- - - - . - - - 4 .

:'-°

X, 1ating batch 2 1 1

X, : Time between platingand baking, 1 24 2

X( Baking time, .n 48 9 20

X '. --me between bakingand measuring, n 2 i@0 2 590

*X*K_ elative humidity % C 25 0 0

• :o.2n zorresponas 'o -able* . over :rieri:e: Ii, average ambient: ,3, :ver saturated Ja-r.,%.

a i<'"

• .-

.- . ...;-.. ",v ,,.,... ... .'......'..v .,; "- .''v .. -.v .'." ',.'v v ." "--.,-- ' " "v -,."'""""" " , -""" -""" "8:

NADC-86035-60

TABLE 3 - Hydrogen Concentrations and Sample Calculations for Plackett-BurmanScreening Experiments

Hydrogen Concentration, uA/cm2

Trial Bake Bright Cadmium Dull Cadmium

1 1.15 0.682 .34 .273 + ~ .74 .304 .54 .50

5.97 .566 .97 .657 .78 .428 .33 .349 .58 .36

10 .60 .36ii - .71 .3012 .91 .43

Sum 3.26 2.07* Sum- * 5.36 3.10

Difference 2.10 1.03Factor Effect (1 6) 0.35 0.17

* See column X_, Table 1. 2* * Background livel, 0.22 UA/cm

Absolute value.

-. - Factor Effects and Minimum Significant Factor Effects, [MINI,for the Plackett-Burman Experiments

Variable Factor EffectBright Cadmium Dull Cadmium

X Batch 0.087 0.145

X2 Delay before baking .127 .038

X"I Baking time .350 .172

X, Delay after baking .040 .065

X_ Humidity .077 .005

.FE' experimental error) .119 .048

-MINI 95% confidence level, t=2.45 .291 .118,MIN] -0% confidence level. t:i.94 .230 .093

Significant factor effects

9

S - . .... .... ....•...-..... ... .. . . . . . . . . . .. .. . .. .. ",. ...- °- ... . . . * -..*. ':-- *- ... . A- - - ", - "i-

J' ' ' * " ", • XE

NADC-86035-60

TABLE 5 - Two variable, Two-level Experimental Design and Results

Trial Baking Time, h Delay Time, h Hydrogen Concentration, uA/cm2

Avg

3 0.881.26 1.07

2 3 24 1.081.02 1.05

72 .26.31 .28

4 72 24 .28.31 .30

2A* Background level, 0.22 PA/cm2

TAEL - Effects of Baking and Stress on Failure Time of 3OOM Steel(Preliminary Experiment)

Times to Failure, hLoad Baking Times, h -,

24 -'8

-5 4.7 41.55.7 177.0 L

-~ 0.0..3.

*14 otzhed tensile strength. 1800 MPa.

'9%

10

?.

.. .. . .. . .. . .. .. . .. . .. . ... . . .. f

-% 7-M x N -W- -. -Lin Vim

NADC-86 03 5-60

TABLE 7 _ Effects of Baking and Stress on Failire Time of 300M Steel

Times to Failure, hLoad Baking Times, hfoNTS 48 72 96

25 16104>500

50 4814

75 0 13

5 3551 46

90 303289 "

* Notched tensile strength, 1800 MPa.

TABLE 8 - Effects of Baking and Stress on Hydrogen Concentration of 300M Steel 4

Hydrogen Concentration,* uA/cm"

Baking Time, h Stress No. Specimens Avg. Std. Dev.

0 No 4 0.60 0.076

48 No 4 .45 .086Yes 3 .50 .118

72 No 4 .42 .07]Yes 5 .52 .070

'?6 No 4 .8 .03-Yes 6 .44 .256

* Background level, 0.35 uA/cm "

'i

NADC-86035-60

I.TABLE 9 - The Effect of Short Baking Times on the Hydrogen Concentration of

Cadmium Plated 300M Steel I,

Baking Time, h Hydrogen Concentration, UA/cm2

Avg. Std. Dev.

0 0.38 0.0262 .42 .0634 *43 .025

3 .42 .03314 .44 .04924 .w3 .073

* Three 6 mm thick specimens each.

-ABLE 10 - The Effect of Very Short Baking Times on the Hydrogen Concentrationof Cadmium Plated 300M Steel

Baking Time, 1. Hydrogen Concentration, uA/cm2

mm Thick i mm Thick

Avg. 2td. Dev. Avg. !td. Dev.

0.44 0.023 2.43 24 3

1 .54 .082 . .36.50 .041 .52 .029

4• "5 .015 .52 .039

7our specimens each.

12

S. . . . .. . ... .. -- ... .

s .... ...........

NADC-86035-60

NE

0.8

S0.6 DULL

0.4

0.2 -- - - - - ----- ".. ... N O ro OQIENl ------. ....... . . .. .-

l.0

* I. I I I I I I I !

i tSAK O HStO

FIGURE 1 Efficiency of baking treat- ." merit for removal of hydrogen from

bright and dull cadmium plated AISI". 4340 steel. Plating thickness, 15 pm.

Baked at 190 C."r

1.

0.2 --------------------NONYDUGE-------- --------- 3

I. I 1111 1 I a 1111

4340 stel..Plaing thckness,15 ,'m

,'. '. .'. '., ,, - .. • .-..-..- ,-.......'."...-..'....,-..-...,,...-.-.,-.-.-..,....-.....-,.....,...-.,..-.-....,,., ,.,,.,..,.,....

r M IRL P61 V6 W. 1% 4L

NADC-86035-60

DISTRIBUTION LISTReport No. NADC-86035-60

No. of Copies

Commander.................................6Naval Air Systems CommandWashington, DC 20361

(2 f or 00D4) (2 f or retention)(1 for 03D)(1 for 5304)(1 for 5304D)(1 for 5300D2)

Office of Naval Technology.............. . . .. . .. .. .. .. ..*800 N. Quincy St.

Arlington, VA 22217(Code 0725)

Office of Naval Research................. . . ... ... .. .. .. ..800 N. Quincy St.

* Arlington, VA 22217(Code 431)

* Dr. A.J. Sedriks

* Commander............. .................... 2Naval Air Rework Facility

Naval Air Station* Alameda, CA 94501

(Code 340)(Code 341)Mr. T. Moore

CommanderNaval Air Rework FacilityNaval Air StationJacksonville, FL 32212

(Code 340)

Commander . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Naval Air Rework Facility*Naval Air Station

Norfolk, VA 23511(Code 340)

Commander .............................Naval Air Rework FacilityNaval Air StationNorth IslandSan Diego, CA 92135

(Code 340)

14

NADC-86035-60

No. of Copies

Commander.1Naval Air Rework Facility*Naval Air StationPensacola, FL 32508

(Code 340)

Commander........................ .. . . ...... . .. .. .. .. .. ..... 1Naval Air Rework FacilityMarine Corps Air Station

* Cherry Point, NC 28533(Code 340)

Commander........................ .. . . ...... . .. .. .. .. .. ..... 1Naval Air ForceU.S. Atlantic FleetNorfolk, VA 23511

(Code 5281)

Commander........................ .. . . ...... . .. .. .. .. .. ......* Naval Air Force

U.S. Pacific FleetNaval Air StationNorth IslandSan Diego, CA 92135

(Code 7412)

* ~Naval Sea Systems Command.................. . . .... .. .. .. ..... 1Washington, DC 20362

(1 for Mr. I. Kaplan)

David Taylor Naval Ship Research and Development Center ........... 3Annapolis, MD 21402

(Code 281)(Code 2813)(Code 2814)

Naval Research Laboratory ....... ................... 3Washington, DC 20375

(Code 631)(Code 6312)(Code 630)

Naval Surface Weapons Center............. . ... . .. .. .. .. ..... 1White Oak Laboratory

* ~Silver Spring, 11D 20910 ..(Code CR-32)

Naval Surface Weapons Center............. . ... . .. .. .. .. ..... 1Dahlgren, VA 22448

(Code G-52)

NADC-86035- 60

No. of Copies

Naval Weapons Center ....... ........................ . .. .IChina Lake, CA

(Code 3854)

Naval Air Propulsion Center ........... ....................Trenton, NJ 08628

(Code PE-72)

Naval Underwater Systems Center .... .................... ....New London, CT 06320

(Code 4493) _1.

Air Force Office of Scientific Research ... ...............Bolling APBWashington, DC 20332

(I for Dr. A. Rosenstein)

Wright-Patterson AFB ....... ........................ .....OH 45433

AFWAL /MLSAMr. B. Cohen

Air Force Logistics Center ...... ..................... ... .Warner-Robbins AFBWarner, GA 31098

MMEMCMr. W. Thompson

U.S. Army Research Office ...... ..................... ...-.P.O. Box 12211Research Triangle Park, NC 27709

Dr. R. Reeber

U.S. Army Materials Technology Laboratory .. ..............-...Watertown, MA 02172

DRXMR-EMMr. M. Levy

U.S. Army R&D Command ............ . . . . . . . .. 1Dover, NJ 07801

DRSMC-LCU-SSMr. W. Field

U.S. Army Mobility Equipment R&D Command ... ................Fort Belvoir, VA 22060

DRDME-VCMr. D.A. Emeric

U.S. Army AMCCOM ......... .......................... ... 1

Rock Island, IL 61299SMCAR-ESMMr. J. Menke

16

PA"

NADC-86035-60

No. of Copies

National Bureau of Standards ..... .................. . . 2Washington, DC 20234

Metallurgy Div. Dr. C.G. InterranteRm B254, Bldg 223 Dr. E.N. Pugh

Defense Advanced Research Projects Agency ............6th Floor, 1400 Wilson Blvd. I.Arlington, VA 22209

Dr. P. Parrish

Chief, Materials and Processes .................Grumman AerospaceBethpage, LI, NY 11714

Chief, Materials and Processes . ..-.Vought CorporationP.O. Box 5907Dallas, TX 75222

Chief, Materials and Processes ....Rockwell International4300 E. Fifth Ave.Columbus, OH 43216

Chief, Materials and Processes . .Boeing AerospaceP.O. Box 3707Seattle, WA 98124

Chief, Materials and Processes ..Lockheed Aircraft Corporation2555 North Hollywood WayBurbank, CA 91503

Chief, Materials and Processes . . . . . . .Cleveland Pneumatic Corporation3781 E. 77th St.Cleveland, OH 44105

Department of Metallurgical Engineering and Materials Science . .

Carnegie-Mellon UniversityPittsburgh, PA 15213

Dr. I.M. Bernstein

Rockwell International Science Center ..............

1049 Camino Dos Rios

Thousand Oaks, CA 91360Dr. M. Kendig

Department of Materials Science and Engineering .Johns Hopkins UniversityBaltimore, MD 21218

Dr. J. Kruger

1 17

NADC-86035-60

No. of Copies

Materials Engineering Department ..... .................. ...Virginia Polytechnic InstituteBlacksburg, VA 24061 .W

Dr. M.R. Louthan, Jr.

SRI International ....... ........................ . .. 1333 Ravenswood Ave.Menlo Park, CA 94025

Dr. D.D. Macdonald

Rockwell International Science Center ... ............... ."."1049 Camino Dos RiosThousand Oaks, CA 91360

Dr. F. Mansfeld

Department of Metallurgy and Materials Science ............. ......Case Western Reserve UniversityCleveland, OH 44106

Dr. J. Payer

METTEK ........... .............................. ......-1805 E. Carnegie Ave.Santa Ana, CA 92705

Dr. L. Raymond

Department of Materials Science and Engineering . ...........Thornton HallUniversity of VirginiaCharlottesville, VA 22901

Dr. G. Stoner" L-,

Homer Research Laboratories...................Bethlehem Steel CorporationBethlehem, PA 18016

Dr. H. Townsend

Chemical Engineering Division . .... . . . . . . . . . . ..Argonne National LaboratoryArgonne, IL 60439

Dr. R. Varma

Sinclair Laboratory 7.....................Lehigh UniversityBethlehem, PA 18015

Dr. R. Wei

Department of Metallurgical Engineering ,.... . . . . . . ...... 1The Ohio State UniversityColumbus, OH 43212

Dr. B. Wilde

18

4. .- .**... . .

-77 7- C-7 -- -- ,.-

NADC-860 35-60

No. of Copies

Defense Technical Information Center .. .................. 12Cameron Station, Bldg 5Alexandria, VA 22314

NAVAIRDEVCEN. .... ........................... 128(3 for 8131)(125 for D. Berman 6062)

19

db7

- -6o,86

y~.' ~ ~ - - . -- - . - -~---------

-