%CV

REOTNO, A -KA7,)Et817

c"""-TAR19M OF V~ios T . SA, 1t~l 1 Q FO

PdEVA9KA2'232 )PME4T,~k TER'-O

ru l eN rdsrbmn -lmtd

Wcrivs~r1 PenslGoi

- . .~z mtt utew e &~mnaajj- - - - - -

U IC LAS S IFI E D

DOCUMENT CONTROL DATA.- R & 0Security classification of title. body of abstract a,.J .rsdr..n ann~jtvtin ic cr ir Vr r- - a r'.. z- fled

I ORIGINA TING ACTIViTY fCOf'POIte AUthor) ~A. P. ~ lOtSECWR.-V LSCC~f.

AERO 4ATERIALS DEPARTMENT I UNCLASSIFIEDNAVAL AIR DEVELOPMENT CENrRRIh cC.

WARMINSTE R, PA. 18974 NFE3 REPORT 7ITLE

A COXPARISON OF VARIOUS TEST METHODS FOR DETECTING HYOr-)GEN E.XBR1TTLEMENT

4. DESCRIPTIVC HOTES-Type of report andinciustsve calet;

PROGRESS REPO:!T ________________________________5 AU THORISI (Firut name, mniddi& initial. last namt.)

E. J. JANKOWSKY

6. REPORT DATE 1. O-AL N~O - MAGES I.. - .c ous.

8 June 1971 1 20 2- S9&. CONTRACT OR GRANT ftO.

INADC- MA- 7066b. PROJECT NO-

AIRTASK A32052031202B/ 1F51 541201I-. Work Unit OD :3-R REWPORT *.. An35 1,:-" ru=!ers ft&. ma- b,

d. JN10 DISTRIBUTION STATEmENT

APIROVED FOR PUBLIC RELEASE; DISTRIBUTION U L!IMI T ED.

It- SUPPLEMENTARY N4OTES soso. VI r

NAVAL AIR SYSTEM4S COMMIANDNO NE DEPARTMENT OF THE NAVY

.WASHINGTON, D. C. 2036013. ABSTRACT

Four hydrogen ernbrittlement test methods were evaluated using three paintI

strippers as the embrittling media. Resilts were compared with those obtainedwith notched C-rings, the method now prescribed in paint stripper specifications.

In general, all tlb.. methods give good results and good correlation. The mainI

differences were in ease of use.

FORM~~47 (PAGEl1)UN L S i I )DD ~ I OV 4f7-,147 r SUTI FI

UNCLASSIFIED

Security C1MsificationKEY WORDS LINK A LINKS 4N

"C p

KE:Y V.ORO$ -

ROI. T. ROLE V T ROLE ,T

Hydrogen Embrittlement Testing

High Strength Steels I

laint Removers

r7

I I

I 1

UI II I

tt

FORM ___

1473 _____ _____

5E'ltt lsjfcto

DEPARTMENT OF THE NAVYNAVAL AIR DEVELOPMENT CENTER

WARMINSTER. PA. 18974

Aero Materials Department

REPORT ,11. NADC-.I&-7066 8 June 1971

A CO-WARISON OF VARIOUS TEST MEFHODS FORDETECTING HYDROGEN EMBRITTLEME.'r

AIRTASK A32052031202B1 IF51541201WORK L'NIT-OD

Four hydrogen embrittlement test methods were evaluated using threepaint strippers as the embrittling media. Results were compared withthose obtained with notched C-rings, the method now prescribed inpaint stripper specifications. In general, all the methods give goodresults and good correlation. The main differences were in ease ofuse.

REPORTED BY:

Chemical Metallurgy Branch

AlT ROVED BY: .. /jS. J. KETCHAM, HeadChemical %etallurgy Branch

F. S. WILLIAMS, SuperintendentMetallurgical Division

Approved for public release; distribution unlimited.

NADC-Mi-7O66

S U M AR Y

For some years the laval Air Development Center has used sustainedload tests with notched C-rings or notched tensile bars to evaluatehydrogen embrittlement dama~ge to high strength steels. These specimensare sensitive and give reproducible results but are expensive to inachineanid require long testing times.

Several instrumental methods have been developed in the past severalyears. Two involve the use of probes, the third utilizes metal strips.A comparative study of three -~f these has been completed and resultr-correlated with those obtained from rotcbed C-rings. A notched boltwith a miniature calibrated proving ring was also included in the inves-tigation. l-aint strippers ;*Pre used as the -z.bri'-tline media largelybecause !4IL-R-81294A (reference (a)) now contains an embrittle-ment testusing notched C-rings.

This work was performed under work Unit F5i 541 201 OD as part of thecontinuing effort to find better means of me.as-ur4.ng hydrogen embrittlementresulting from pickling, electroplating and related proizessine.

S1NVIARY OF RESULTS

Results of the three instrumental methods and the two notchedspecimens showed excellent correlatfon. One of *he paint strirpersused is very embrittling, the second somewhat less so and the third isnot embrittling. All the methods were as effective in assessing thedegree of embrittlement of each paint stripper as was the norchej C-ring.

CONCLUSIONS

The three instrumental methods varied in complexity of use. TheFmethod utilizing strips was the simplest to use. The no-tched boltwith the calibrated proving ring, howevez, is taie simplest test toconduct and the bolts and rings are readily available, but at the presenttime there is only one source of supply.

REaMMM'IDATIONS

It is iecommended that consideration be given to preparation ofa Federal Test 'lethod Standard for llydrozen Embrittlement Testing whichwould include the notched C-ring and the other methods covered in thisreport.

VADC-M1A-7066

TABLE OF CONTENTS

Pase

SU- ARY . ........ . . . . . . .... iii

Introductior ..................... iiiSuinary of Results . ................ . 1i

Conclusions ...................... 1M

Recomendations . """

SEXER.XE4T-A PROCEI'RE . ...................

R 7%LTs. .......................... 2

DISCUSSION ....... a. .. . ............

REVEEEESS .......... .. ....... .....

TABLES

I -Test Results .... ............ 6

FIGURES

1 - tched C-Rings with Timers . ................ 72 - The Laurence Hydrogen Detcctibn Gaur.. .........3 - The DrIlube tWId Ven Evaluation Geug . - ...... 94 - The i--1 -Tester. . ...... . . .... . ....... 05 - Drilube ?qotched Bolt Loading Device .............. II

~ ~ - --

EX PER IMENTAL PROCEDURE

Three tuaint etripping solutions were used for this evaluation.The strippers vere classified very embrittling, m-derately crbriL--tMug, a&d non-embrittling on the basis of resulcs obtained with

- notched C-rings.

The test method used for notched C-rings consisted of cadmiumplating the outside surface oi the rings in a Cyanide bath Wiithout

I brghteers(50 sf, ui), bkS for 24 hours at 3749", stressing

to 75% o~f the breAking strength) and then dipping the closed hrif ofthe ring into the paint etrippqer. M.jers ion ti-- was approimtely

1 1 ninute. Striper vas a!loved to remain on the ring surtae.ihnz insing after dippiLng. Tin-e to failure vas measured with elbctr-ictic-ars attached to nidcroswitches as shown Inm Figure 1.

Drilube Notch Bolts were prt-pared and treated irn the sane mauneras the C-rings, except that -ore-plate cleaning included grit blasting

I with 120 grit aluzinum oxide and baking tim- after plating was I hourin the ove furnished with the eq1ip~ent. Time to failure vas deter-.1nedby visual mxaninat iou at intervals, or In some cases, b-, an aud ible

-"anV mwnen the bolts oroee.

Drilube bolts are supplied with a miniature proving ring cali-brated to apply a 75% WEIS stress to a bolt pssing through itsdiameter. The ring is couwrested in a hydraulic loading device withthe bolt in place. W~hen th~e ring has bsen compressed & predeternineddistance (furnished by the manufacturer), a smoll collar -is swagedonto the bolt and the load on the hydraulic Jack ia released, leaviingI the bolt In tension. Each bolt comes with its own calilbrated ring.Bolts are utde of 4~340 steel, heat treated to the 260 - 280 .KSIS~strength level. The notch -9s 60e with a 0.005! radius.

I -The IIEP Tester consists of titdng deviAces attached to bank-s cfbent strip holders. Five steel strips are used for each test. andthe time for three strips to break is recorded as the "'=edian" failure

-~ tine. Strips are ly? X TV- X .0221 thick, with double tabs stamped1 through the center to act as nortclieip TLhe stress applied is prede-

teramied by the distance between the restraining ends on. the tester.J The ends of the strips are simly squeezed by hand until .they cAn be

popped into the tester. The design of the ETP Tester does not permitthe testing of liquids ar slurries with the specirsns unmder stress.it was therefore necessary to alter the procedure used for testingC-rings.

HEP strips were plated with unbrightened cyanide cadmiumi for 2t miutes at 50 asf, baked 24 hours at 375F, Zir:ed in ttripper for6 hours, washed in running water followed by aceone, and i~ediatelyplaced under stiess in the timing device.

NADC-M-7066

The Lawrence Hydrogen Detection Gauge makes use of a steel walledtube, similar to a 6V6 tube, as a hydrogen probe. The quantity ofhydrogen passing through the wall of the tube during plating orprocensing is measured electronically and displayed on a hydrogenindex scale. The maximum value measured after the tube is placed inthe built-in oven is called the heat peak, H_, and the time for thehydrogen index to decrease to one half the heat peak value is calledlambda,A. The heat peak is related to the amount of hydrogen intro-duced into the tube and lambda is a measure of the irternal clean-upand outward diffusion of hydrogen.

Tubes with the entire ead section grit blasted i ediately priorto use were used for the measurement of embrittlement by paintstrippers. A blast cleaning unit especially designed for grithlastir lawrence tubes, 120 grit dry aluminum oxide, and a totalblasting time of 30 * 2 sec. were used for the tube cleaning. Iriorto blasting, the tubes had been baked out on a special bake-out rackto remove residual hydrogen. The cleaned tubes were inrx.ediately platedwith unbrizhtened cyanide cadmium at 50 amps/sq. ft. for 2 minutes andagain baked to remove hydrogen. The plated section of the tube wasimersed in stripper for 6 hours, taken out, washed in water andacetone, connected to the rrobe socket of the Gauge, and the beatpeak and lambda value measured.

A auch more detailed description of the procedure is given inreference (b).

The Brilube Hydrogen Gauge employs standard 616 tubes tjith steelwalls. Biast cleaning is similar to that used for the Lawrence Gauge,but is considered less critical and is not timed. Following gritblasting, each tube is calibrated in a solution containing 50 gms/liter IaCN and 50 gaslliter NaOH. Tubes were reblasted followingcalibration, plated in an unbrightened cyanide cadmium bath, baked

to remove hydrogen, and immersed in stripper for 6 hours. Thestripper was cleaned off with water and acetone, and the heat peakand lambda values measured from the recorded curve of hydrogenIndex versus time.

The ebrittling characteristics of paint strippers often changewith age as demonstrated in reference (b). Care was taken to runthe various tests simultaneously to avoid differences caused bychemical changes in the stripper.

Results of the various tests are given in Table I.

Cadmium plated and baked, notched C-rings are required to sustain 7

a load equal to 75% of their breaking strength for 20 hours after -

imersion in paint stripper, according to reference (a): Stripper A

2

NADC-MA- 7066

caused failure of four specimens in times ranging from 1.3 to 1.9hours. A fifth specimen broke in 34.4 hours. Eight specimens weretested in Stripper B with failure times rangit~g from 3.0 to 34.9hours for seven specimens and one specimen not failing in 200 hours.Four notched C-rings tested i: Stripper C, a non-em;rittling stripper,did not fail in 300 hojrs.

All other types of specimens were cadmium plated and baked priorto immersion in paint stripper for comparison with C-rinqs.

HEP Strips, heat treated to a Rnckwell C hardness of 51, brokein a median time of 13.1 hours after immersion in Rtrirper A, 35 jhours after immersion in Stripper B, and did not fail after 50C hoursin Stripper C. (Median time is taken as the time the third strip ina series of five breaks).

Standard HEP strips, which are heat treated to a lower strenethLhvel rnan those described above, were not sensitive enouah to showembrittlement in the paint strip2ers tested. Two specimens brokewithin the 200 hour testing time, bu: the third "median" strip,on which failure time is based, did not fail.

Fot:- Drilube Notch Bolts brokf in tines ranging from 1.3 to2.4 hours in Strippe- A. Six bolts, tested in Stripper B, brokein times between 7 and 100 hours. Exact times were not obtainedfor this group bec2use they broke during.E long weekend shutdownand no timing device is used with the bolts. The four bolts testedin St- pper C did not fail in 5(j hours.

Heat peak ( HP) values above I are generally considered beyondthe "safe" hydrogen level for high strength s.eels. Lambda (>k)values are less important for paint strippers because they are anindication of the permeability of the plated coatinz rather than ameasure of the amount of hydrogen introduced. Stripper A gave HTvalues of 212 and 207, Stripper B, 4.2 and 6.F, and Stripper C, 0.23and 0.23, in good agreement with notched C-rine results.

The Drilube I values should be lower than 1 on the 7 scale inorder to be considered safe. Much higher values were obtained forStripper A, slightly higher for Stripper B. Stripper C was notreported because the masking undercut badly during the tests. How-

ever, the values obtained, even with the oversized area exposed, wereconsiderably below the hydrogen limit for safety.

3

F

NADC-MA-7066

D I S C U S S I 0 N

Any of the methods tried appears to be satisfactory for thetesting of paint strippers, althousgh ench one has some undesirablefeatures.

The Lawrence Hydrogen Detection Gauge shown in Figure 2 givesreproducible results that are very useful for determining the extentof embrittlement induced by paint strippers or electroplating processes.

Its principal disadvantage appears to be the extraordinary care thatmust be taken in each step of probe preparation and measurement inorder to obtain reproducible results. This disadvantage still allowsthe Gauge to be a useful laboratary tf, l, but limits its usefulnessas a prcduction or shop type measuring device. A more comprehensiveinvestigation of this instrument was reported in reference (b).

The Drilube Hydrogen Evaluation Gauge, Figure 3 , a devicesimilar to the Lawrene Gauge, appears to be somewhat simpler inprobe preparation and measurement than the Lawrence Gauge. How-ever, at this time, the measurement of embrittlement by paint strippersis still in the experimental stage and definite procedures have notbeen worked out. Its most useful function appears to be the deter-mination of embrittlement from plating processes, and especially, theprediction of the efficacy of baking for relieving embrittlement.

The HE" Tester shown in Figure 4, is a potentially excellentmethod for determining embrittlement. Its chief fault: at present,'s that the standard strips are not sensitive enough to fail underslightly embrittling conditions. This fault is being corrected bythe manufacturer and the new, RC 51 hardness strips appear to bemuch more sensitive. The principle advantages to this method arethat it is extremely simple, very low in cost, and adaptable to

shop use.

The Drilube Notched Bolt shown in Figure 5 is considered tobe ready for use as it stands. Bolts may be obtained from themanufacturer in any desired quantity at a relatively low cost, andthe loading equipment is simple and irexpensive. The only fault thismethod has is that at present there is no timing device to measurefailure times,and failures must be determined by visual examinationat convenient intervals. It should not be difficult, however, tomodify the equipment to include timers.

A4

NIADC- MA- 7066

(a) Specification IWlL-R-81294A, Remover, Palnc, Epoxy System

(b) Naval Air Developmen~t Center Report Nn. NADC-M-7052, Evalu~ation~sof the Lawrence Hydrogen Detection Gauge for the Frediction ofHydrogen Embrittlement, Progress Report, 9 October 1970

TABLE I

0

0 E -,00c

$- 4

'0 .-4uJ

Lf'n

0 000 C4 C14Lo0000 r- r-

-4 E4~d

0 v v v C1 V t nvtrAA A AI\0 0

-4 -

co C41 N --4

"C _ ) C

4JD -OEA'-4Mt tl

.00 0

0 cl0

C~C)

-;0I cc %00 -3' 00S- , =I

1-

4 cc

cc-C

I. Ci. 0

-40-[~1 '3'..

V_ _ _ _ _ _ _ __ _ _ _ _ _ _ __ _ _ _ _ _ _ __ _ _ _ __1__ _ _ _ _ _ _ _

I1.1 -

F

I

F-I:

cn

C-,

C

I'-I *.1)~-

I

-- 4 1*

'I

rr -

U

Ii

C,

I-

-1 2 t!rc. I

A

IA'7 _________

___________________________________________________________________________________

* -.

- ~ -

__ - A -- ~--~~T -iiajJ~:J (I - El

4 - 41r

C

__ I4:Fr p1PITt Ti 1-

L [ ---

I

A Ifir p

1r

-_ N. -I.

JEV

ILP-

____i

I

_________________ ________

_____ __________

-w~= ~-==~

_____________ ft -~

_ __ I

_ _ I______ .4-

~

_ - I_____ 3

I

I

1;

e