AD-A245 286

College of Earth andMineral Sciences

PENNSTATE

V

ANNUAL REPORT

to

OFFICE OF NAVAL RESEARCH

Contract USN 00014-91-J-1189

January 1992

PREDICTION OF HYDROGEN ENTRY AND PERMEATIONIN METALS AND ALLOYS

H. W. Pickering

Department of Materials Science and EngineeringThe Pennsylvania State University

, .31 1992 i.t

92-0237311 I I/,[II III ' ll " ii

PENN STATECollege of Earth andMineral SciencesUndergraduate MajorsCeramic Science and Engineering, Fuel Science, Metals Science and Engineering, Polymer Science;Mineral Economics; Mining Engineering, Petroleum and Natural Gas Engineering;Earth Sciences, Geosciences; Geography; Meteorology.

Graduate Programs and Fields of ResearchCeramic Science and Engineering, Fuel Science, Metals Science and Engineering, Polymer Science-Mineral Economics; Mining Engineering, Mineral Processing, Petroleum and Natural GasEngineering;Geochemistry and Mineralogy, Geology, Geophysics; Geography; Meteorology.

Universitywide Interdisciplinary Graduate Programs Involving EMS Facultyand StudentsEarth Sciences, Ecology, Environmental Pollution Control Engineering, Mineral EngineeringManagement. Solid State Science.

Associate Degree ProgramsMetallurgical Engineering Technology (Shenango Valley Campus).

Interdisciplinary Research Groups Centered in the CollegeC. Drew Stahl Center for Advanced Oil Recovery, Center for Advanced Materials, Coal ResearchSection, Earth System Science Center, Mining and Mineral Resources Research Institute, OreDeposits Research Group.

Analytical and Characterization Laboratories (Mineral ConstitutionLaboratories)Services available include: classical chemical analysis of metals and silicate and carbonate rocks;X-ray diffraction and fluorescence; electron microscopy and diffraction; electron microprobeanalysis; atomic absorption analysis; spectrochemical analysis; surface analysis by secondary ionmass spectrometry (SIMS); and scanning electron microscopy (SEM).

The Pensylvania State L'niversit. in compliance with federal and state laws. is committed to the policy that all persons shall have equal

access to programts. admission, and employment without regard to race, religion, sex. national origin, handicap, age. or status as a disabled orVietnam-era veteran. Direct all affirmattve action inquiries to the Affirmative Action Officer. Suzanne Brooks. -Of Willard Building,

University Park. PA 16802; (814) 863-0471.U Ed. 87-1027Produced by the Penn State Department of Publications

Form ApprovedREPORT DOCUMENTATION PAGE oMB NO 0704-0188

Paiaic lfoimg b'.14141' fOr this cOle 'On Of Information It estimated 10 a-Vtge I hour Del, rlrwe'n ,f nciC ng TC 1th t Ime t1 v . g iPInftru lo1 n, wa'(0 '9; V, sl, datl s Cr(es.

t n o ~.l e-rmrigtat~ he data hq e, el o lfe-g hl r ig~( the collecliOn Of inform ation Send comments regard,mg this tiirden estimate o, env othor asoeci of %his

c O fc It omaiion. .nldt n gelmf.'d ng this oi,,,Oen to viahtvg0o HeaOdaMers, Stilocca. D.,0018111,fo t0,lonao p~~o.aa'p, 5JteO = 6selt-ai.Saute 1204. itng o1. VA 22 -02 . nd t0 the Ofloc Cl M nament an S..ge. PaprmO. Aedud'Ob ProjeC (070- IS$) W aehXngl. " DC 20 0 3

1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED

January 1992 Annual 10/190 to 9/30/91L TITLE AND SUBTITLE S. FUNDING NUMBERS

Predictions of Hydrogen Entry and Permeation inMetals and Alloys N00014-91-J-1189_ _ _ _ _ __TO_ _ _ _ _431 50986. AUTHOR(S)

Howard W. Pickering

7. P ERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) B. PERFORMING ORGANIZATION

The Pennsylvania State University REPORT NUMBER

Department of Materials SCience & Engineering326 Steidle BuildingUniversity Park, PA 16802

9. SPONSORINGIMONITORJNG AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING /MONITORING

Scientific Officer AGENCY REPORT NUMBER

Materials Division Code: 1131MOffice of Naval ResearchArlington, VA 22217-5000ATMN: A. John Sedriks11. SUPPLEMENTARY NOTES

12L. OISTIUBUTION/AVAILABIUTY STATEMENT 12b. DISTRIBUTION CODE

Approved for public release; distribution is unlimited.

13. ABSTRACT (Maximum 200 words)

This report summarizes results of the past year on our continuing in-situ experiments directed to theproblem of hydrogen entry and degradation of materials both for planar surfaces and for the morecomplicated recessed surface. For the planar surface the hydrogen permeation and scanning tunnelingmicroscopy (STM) techniques are being used, and for the recessed surface the study uses a combinedmicroscopy/electrochemical probe technique and a crevice geometry.

Further modeling of the hydrogen permeation technique has led to an experimental procedure that yieldstwo previously unattainable rate constants that are important for controlling hydrogen absorption into amaterial from an aqueous environment, the hydrogen absorption (entry) and desorption (exiting)constants. STM has been further developed for in-situ study of hydrogen bonding at the atomic scaleby successfully imaging hydrogen adsorption on the model Si(100)2X1 surface from a low pressurehydrogen gas phase. For the recessed surface hydrogen entry from an aqueous solution occurs for amuch wider range of conditions than previously believed. Chloride ion and acidification indirectlypromote hydrogen ion reduction and entry on the recessed surface through their direct effect onpromouing the JR-induced mechanism of crevicing.

14. SUBJECT TERMS 15. NUMBER OF PAGES

KEY WORDS: modeling of hydrogen electro permeation, diffusion ..

of hydrogen, hydrogen absorption, chloride ion and acidification 16. PRICE CODE

effects, hydrogen evolution in recesses.17. SECURITY CLASSIFICATION 11. SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACT

Of REPORT OF THIS PAGE OF ABSTRACT

UNCLASSIFIED UNCLASSIFIED UNCLASSIFIED

NSN 750-01-280-5500 Standard Forr 298 (Rev 2-89)

291 102dtv .S id19'

CONTENTS

REPORT DOCUMENTATION PAGE .....................................................

INTRODUCTION ............................................................................

SECTION 1: PROGRESS SUMMARY ...................................................

PUBLICATIONS .......................................................

SECTION 2: HYDROGEN PERMEATION MODELING .............................

SECTION 3: RECESSED SURFACES ...................................................

Role of Acidification ....................................................

Role of Chloride Ion ....................................................

Aooession For

NTIS GP.&IDIN T: 7I

jlii

..................................... ,A

-4-1 i i'i'-I--.-

INTRODUCTION

This report is divided into three selections. Section 1 is a summary of progress and a list of

project publications in the past year. Section 2 is a report (reprint) on how to obtain the important

hydrogen absorption (entry) and desorption* (exiting) rate constants from steady state hydrogen

permeation data. Section 3 consists of two reprints, one that relates to the role of acidification and

the other to the role of chloride ions in shifting the electrode potential inside a recess in the direction

favoring proton reduction and hydrogen entry, and which present a combination of in-situ

techniques: direct viewing of the crevice wall (through a transparent media that forms the other

wall of the crevice) and simultaneously using an adjustable potential measuring prbe inside the

crevice.

SECTION 1

PROGRESS SUMMARY

The problem of hydrogen-induced damage and failure of metallic systems starts with hydrogen

entry. The hydrogen permeation technique has been successfully used for three decades for determining

some of the important hydrogen/material parameters, e.g., the hydrogen diffusivity and hydrogen

concentration inside the input surface, but most of the parameters related to hydrogen entry were not

previously attainable from permeation data because of the incomplete nature of the quantitative treatment of

hydrogen permeation available in the literature. Our work of the past few years on the project corrected

this situation so now a more rigorous quantitative treatment of the permeation process is available

(described in last year's Annual Report and various publications). As a result several additional

parameters related to hydrogen entry became attainable from permeation data. These are the proton

discharge rate constant, transfer coefficient and exchange current density, the hydrogen recombination rate

constant, the hydrogen coverage, and the ratio of (but not the individual) hydrogen absorption and

desorption rate constants. To obtain the individual absorption and desorption constants some further

model development was required, and has recently been accomplished. This extension of the model

The desmxion or exiting rate constant is referred to as the adsorption rate constant in the reprint.

requires that the steady state permeation data are obtained as a function of membrane thickness. This

extension of the model is described in Section 2.

The scanning tunneling microscope provides the opportunity for in-situ imaging of the surface over

a wide range of scale including the atomic scale. Imaging adsorbates on surfaces is also possible. Initial

STM studies on the project began with the easy-to-study silicon low index single crystal surfaces using the

ultra high vacuum STM. Then, studies of hydrogen adsorption on these surfaces was started with the

initial successful STM observations at the atomic level reported last year for the Si 11 l(7x7) surface (ONR

Technical Report, February 1990). This initial success in imaging adsorbed hydrogen atoms is

encouraging for the successful development of the STM technique, and alternatively the atomic force

microscope (AFM) which we are now employing in our studies of adsorbed monolayers, for in-situ

imaging of hydrogen adsorption on metals at the atomic level.

For most cases of nonuniform corrosion, the electrode potential varies over the surface, e.g., in the

case of oxygen concentration cells, being more noble at cathodic sites of high oxidant availability and more

negative at anodic sites. The shift of the electrode potential in the negative direction can be quite large and

its magnitude is unknown in service applications since the anodic sites are in recesses (crevices, cracks,

etc.). It follows that the tendency for the occurrence of the hydrogen evolution reaction, and thus for

hydrogen entry into the metal structure, is greatest in the recesses.

Thus, it becomes important to know how and to what extent the various parameters e.g.,

acidification of the local cell environment, associated with recesses in particular during localized corrosion,

affect the shift of the local electrode potential in the negative direction. This question does not appear to

have been addressed in the literature but is not trivial as we now know from recent results on crevice

corrosion of iron. During crevicing ever more negative electrode potentials to a limiting value exist along

the crevice wall with increasing depth into the crevice, accompanied by the evolution of hydrogen gas

(ONR Technical Reports, February 1988, January 1988 and October 1987, and reprints in Section 3 of

this report). During the past year the roles of acidification, chloride ion, recess geometry, and other

factors, on promoting the potential shift into the hydrogen evolution potential region within recesses have

been under study. Both acidification and chloride ion have been found to increase the magnitude and

frequency of the potential shift within recesses in the negative direction. Thus, the tendency for hydrogen

evolution and entry into the metal within recesses is increased by both of these parameters along with the

tendency for localized corrosion. Results available to date are presented in Section 3.

PUBLICATIONS ON THE PROJECT

R. N. Iyer and H. W. Pickering, "Construction of Iso-Coverage Tafel Plots to Evaluate the True H. E. R.Transfer Coefficient", J. Electrochem. Soc., 137, 3512-3514 (1990).

K. Cho and H. W. Pickering, "Demonstration of Crevice Corrosion in Alkaline Solution WithoutAcidification", J. Electrochem. Soc., 137, 3313 (1990).

R. Iyer and H. W. Pickering, "Current Developments in Modeling and Characterizing ElectrochemicallyInfluenced Hydrogen Evolution and Entry in to Materials", pp. 195-209 in Hydrogen Effects on MaterialBehavior N. R. Moody and A. W. Thompson, ed., TMS, Warrendale, PA (1990).

R. N. Iyer and H. W. Pickering, "Mechanism and Kinetics of Electrochemical Hydrogen Entry andDegradation of Metallic Systems", Annual Review of Materials Science, Vol. 20, Annual Reviews, Inc.,Palo Alto, Calif., (1990).

A. Valdes and H. W. Pickering, "IR Drops and the Local Electrode Potential During Crevicing of Iron",pp. 393-401 in Advances in Localized Corrosion, H. S. Isaacs, U. Bertocci, J. Kruger and S.Smialowska, eds., NACE-9, National Association of Corrosion Engineers, Houston, Texas (1990).

H. W. Pickering, "A Critical Review of IR Drops and Electrode Potentials Within Pits, Crevices andCracks, ibid, pp. 77-84.

H. W. Pickering and T. Sakurai "Scanning Tunneling Microscopy and Its Applications in CorrosionScience", Sym. on Surface and Interface Characterization in Corrosion, NACE, Houston, Texas, PreprintNo. 81, CORROSION 91.

SECTION 2

HYDROGEN PERMEATION MODELING

Reprinted from JOURNAL OF TtHE ELE(CIO(IIHEMiCA. SO7IETYVol. 137, No. 11, November 1990

Printed in USA.Copyright 19O

Construction of Iso-Coverage Tafel Plots to Evaluate the HERTransfer Coefficient

Rajan N. Iyer and Howard W. Pickering*Department of Materials Science and Engineering, The Pennsylvania State University,

University Park, Pennsylvania 16802

This communication is a supplement to a recent paper R = gas constant, 8.314 J (g-mol K)', and(1) describing a new model of electrolytic hydrogen charg- T = temperature, K.ing and permeation in metals. This model shows how theforward and backward rate constants of the hydrogen Equation [1] can be rewritten asentry step, kab. and k.ds, respectively, can be obtained from ice,permeation data as a function of membrane thickness, and OR [2]that the hydrogen evolution reaction (h.e.r.) transfer coef- H 1 ,,ficient, a, should perhaps not be considered a constant forthe entire range of hydrogen coverage, 011, on metallic sur- Equation [2] can be termed the polarized adsorption iso-faces. In the present paper we show that if, in addition, ex- therm for a reaction sufficiently far from equilibrium soperiments are conducted for membranes of certain, rather that the back-reaction can be neglected (1). Differentiatingthan random, thicknesses, the a values at constant hydro- Eq. [2] with respect to -9 and simplifyinggen coverage can also be obtained.

For even small variations of a with OH, iso-coverage a do,1 i'e"- ( d In i,[values are needed in the newly developed (I-P-Z) model d ac + [3

(1, 2) in order to accurately determine the above-men-

tioned and other previously unobtainable parameters of Equation [3] shows that if the coverage, OR, can be fixedthe h.e.r. and hydrogen permeation processes. And it has while polarizing the electrode, i.e., if OH = constant, thenbeen previously pointed out that the value of a is impor- dOH/d~j = 0 andtant in determining the polarizability of the electrolytic re-duction of protons to form adsorbed hydrogen atoms on a 1 d In i, RT d log i [metallic electrode surface (3, 4). Especially when the sur- a(0H constant) = - - 4 - - [41face hydrogen coverage is significant (01 ; 0.1), corre- a d-9 2.303F dsponding to significant percentages (>2% for iron) (1) ofthe charging current (i,) permeating through the mem- where a (OH constant) is the h.e.r. transfer coefficient ob-brane, rather than evolving as hydrogen gas, it will not be tained at a particular OR coverage. Equation [41 is the famil-possible to evaluate a from the conventional Tafel plot, log iar Tafel equation, except that dOH/d-9 = 0 has been appliedi, vs. T, since these plots will not be straight lines. This per- as a condition to obtain an actual a value.haps suggests the influence of OR on a. Examples of non- The normal tendency for 6R to increase with increasing rjstraight Tafel plots are shown elsewhere (5-7). Since they (in the absence of permeation into the metal) can be con-occur for a wide range of hydrogen coverage, including trolled by decreasing the membrane thickness, L, therebyvery low OH, it now seems reasonable to consider, as we are increasing the steady-state permeation flux, j, (or equiva-doing in this note, that for some conditions a may vary sig- lently the permeation current, i.), in hydrogen permeationnificantly with OH. In the following analysis, a novel tech- experiments utilizing the Devanathan-Stackurski cell (8),nique for analysis of hydrogen permeation data as a func- i.e., the i. value can be manipulated by varying L. This istion of thickness is advanced to enable construction of possible because the steady-state hydrogen permeationTafel plots as iso-coverage lines from which a (at constant through the bulk metal can be described as a simple rate-OH) can be accurately evaluated, controlling diffusion process (8) and i,, 61, and L are re-

The Butler-Volmer equation for hydrogen discharge lated by (1)when the backward reaction can be neglected (1 > > RT/F)is t1, 3, 1) - [51

(I - O[1] FD, k"

where i, is the cathodic current density, where

i", D, = the bulk hydrogen diffusivity in the metal, andV = the thickness-dependent hydrogen absorption-

adsorption constant = kb,/Ik~d. + DI/L].i,, =exchange current density, Rearranging V gives0, equilibrium surface H coverage,0j, (polarized) surface H coverage, 1 1 D, 1

F - [61a -=38.4V 1k" k' k.b, L

RTSh.e.r. transfer coefficient, where k' = the (thickness-independent) hydrogen absorp-

H overvoltage = polarization, tion-adsorption constant under conditions of negligibleF Faraday constant, permeation = k,k,,.

It is now apparent from Eq. 15] that we can achieve a con-Electrochemical Society Active Member. stant 61, (termed the iso-coverage) condition by setting i. L

J. Electrochem. Soc., Vol. 137, No. 11, November 1990 © The Electrochemical Society, Inc. 3513

(k") constant. Then, using Eq. [51 and [61 one can prove wherethat

g = FD1 k'o 1 [Ila]At equilibrium of the h.e.r., i. is small but finite, corre-

where i... and i .... refer to i, values for membranes n and sponding to the equilibrium hydrogen coverage, 0., i.e.. atm of thicknesses L. and Lm. Equation [7] is applicable to T = 0, i. = i ; 011 = 0.; g = g". Then, Eq. [11] becomesany number of membranes, N. of different thicknesswhere m = 1,2,3.... ; n = 2,3,4.... N; and n > m. Equation 9 D,[7] describes the relationship between the iso-coverage i. L . - - [11b]

values for membranes n and m. However, we have to know V. k~d,

the DI/k,d, value or eliminate this unknown quantity from Whereas 6e is considered to be independent of L, i. de-Eq. [7] with other known quantities, as done below. pends on L. By plotting L vs. 1/io (where io. is obtained by

If the h.e.r. occurs by a discharge-recombination process extrapolating the i. vs. q plot tonl= 0), we can evaluate g" =and if the (chemical) recombination process is not rate- the slope, and -DI/kd. = intercept. This value of -D,/kd,limiting, it has been shown (1) that can be used in Eq. [7] to calculate iso-coverage i. values re-

i = Ki,2 K(i,. - ij)2 [8] cursively for various membranes.

It is to be noted that data need to be obtained on mem-where K is a function of L. Since ir is constant (for constant branes whose thicknesses span one or two orders of mag-oi) and independent of L, it follows from Eq. [7] and [8] that nitude. Depending on the hydrogen diffusivity, mem-

branes with thicknesses ranging from about 50 im toi,... = (K,/K) . [9] 5 mm can be used. Still, it may be that the intercept of L vs.

where K,. and Km refer to the K values for membrane thick- l/i° is undeterminable by extrapolation. In these casesnesses of L, and Lm, respectively. It may be noted that K is (with 0, > 0), one may instead use the slope. g". It can bea constant for each of the membranes. From Eq. [8] and [9], easily shown using Eq. [111 for iso-coverage of two mem-one obtains after simplification branes, m and n, that

i,. = i,, + ((K./K.) - 1) i.._ [10] 1 1 (Ln - L)7-- + -[Ille)

Equation [10] can thus be used recursively to find iso- i,., i ,r 9Mcoverage i,.. (and hence T),) values from polarization andpermeation measurements on membranes with varying where g. = g, (Eq. [Ila]) since 6H., = OH.n (iso-coverage).thickness. Since K is inversely proportional to L, it is clear Again, using Eq. [11] and [llb] for membrane m, one canthat K,/K. 6 1. This means that different iso-coverage i, show that(and hence -i) values, can be obtained from Eq. [10] byusing membranes of different thickness. But it is neces- gm g'sary to vary the thickness by one to two orders of magni- - [lid]tude depending on the magnitude of i.. Since i. will best i., i,.M

be obtained for a number of i, values, i, should always below enough to avoid introducing nonsaturable traps, and From Eq. [le] and [lId]should be applied in a sequence of decreasing magnitudesin order to have a stable trap density and to minimize film _ _ _ _

effects. The presence of a film will also have to be deter- i -i" [12mined as described elsewhere (9,10) in order to know if I+{(L L,)Li°1)lhydrogen is adsorbing on a film or on the metal surface. LI

If the assumptions in Eq. [8] (1) are not met, for exampleif the h.e.r. occurs by a more complex process such as dis- Once iso-coverage i. values are obtained on differentcharge followed by simultaneous chemical and electro- membranes (by recursively using Eq. [10] or alternativelychemical desorption, or if the recombination process is using Eq. [7] or [121 in conjunction with L vs. 1/i. and Eq.rat,-limiting, then Eq. [8] will be modified. An example of a [ 1 lb]. corresponding iso-coverage ij (i.e. -,.) and i, (i.e.. i,.,)modified relationship between i. and i. occurs for the case values are easily obtained. Thus dOH./d'T, = 0 (61., repre-of H 2S poisoning the h.e.r. and enhancing hydrogen entry senting iso-coverage) and Eq. [4] will hold asinto iron in acidic solutions (7, 9) for which the relationshipis In (Vt=i,) (ctifb/k") i. + In {b(Fk3 )0 5/k"}. Since this rela- -RT A log in, RT A log i . 1tionship contains the unknowns a and k", it cannot be used 0(H constant) - = [131to find Di/kd,. Instead, an alternative procedure has to be 2.303F A-% 2.303F (-n)

used as follows. The iso-coverage Tafel plots (log i, vs. ,.) will yield the aFrom Eq. [5] and [6], one can obtain values for the different hydrogen coverages, 8H. The advan-

[L D tage of using a (of constant) is in reducing the scatter in log.=Di vs. Y and i,eaa vs. i. plots for the I-P-Z model (1). and

FD k' k.,, thus yielding more accurate values of the various rate con-

stants, surface coverages, etc. In reactions where signifi-i.e. cant coverages as well as both chemical recombination

and electrochemical recombination rates for the h.e.r. areI D, prevalent, such as in poisoned electrolytes (7, 9), it may be-

FD, k' 9. "- = L + -k~d, come mandatory to use the iso-coverage Tafel plots, de-

rived here, to determine a.In conclusion, iso-coverage Tafel plots should be used to

since determine the most accurate a values, especially if cover-ages are significant. The data base needed for this con-

k' 1 struction is simply the measured cathodic overvoltage (,i),- cathodic current density (). and steady-state permeation

ka, kd, current (i.) for membranes with one to two orders (- 50 jImto 5 mm) of variation in thickness.

or Acknowledgments

D, Encouragement and financial support by A. John

g. - = L - [Ill Sedriks and the Office of Naval Research. Contract No.i kd., N00014-84K-0201, are gratefully acknowledged.

3514 J. Electrochem. Soc., Vol. 137, No. 11, November 1990 (:, The Electrochemical Society, Inc.

Manuscript submitted Jan. 22, 1990; revised manuscript 4. J. O'M. Bockris and A. K. N. Reddy, "Modern Electro-received June 15, 1990. chemistry," Vol. 2, pp. 918, 1007, Plenum Press, New

York (1970).The Pennsylvania State University assisted in meeting 5. C. Kato, H. J. Grabke, B. Egert, and G. Panzner, Cor-

the publication costs of this article. ros. Sci., 24, 591 (1984).6. E. G. Dafft, K. Bohnenkamp, and H. J. Engell, ibid., 19,

591 (1979).REFERENCES 7. R. N. Iyer, I. Takeuchi, M. Zamanzadeh, and H. W.

1. R. N. Iyer, H. W. Pickering, and M. Zamanzadeh, This Pickering, Corrosion, 46, 460 (1990).Journal, 136, 2463 (1989). 8. M. A. Devanathan and L. Stackurski, This Journal,

2. R. N. Iyer, H. W. Pickering, and M. Zamanzadeh. Scr. 111,619 (1964).Metall., 22,911 (1988). 9. R. N. Iyer and H. W. Pickering, Annu. Rev. Mater. Sci.,

3. E. Gileadi, K. Kirowa-Eisner, and J. Penciner, "Interfa- Vol. 20, p. 299, Annual Reviews Inc., Palo Alto, CAcial Electrochemistry," pp. 54, 109-111, Addison- (1990).Wesley, Reading, MA (1975). 10. R. N. Iyer and H. W. Pickering, In preparation.

SECTION 3

RECESSED SURFACES

ROLE OF ACIDIFICATION

ROLE OF CHLORIDE IONS

Reprinted friom JO'RN At. OF THE ELE(-TRUxIIEMICA. SMx IErYVoL. 137, No. 10, Oc0l,4r l. ")

Printed m U.S.A.C pyright t.)

Demonstration of Crevice Corrosion in Alkaline Solution WithoutAcidification

K. Cho* and H. W. Pickering**

Department of Materials Science and Engineering, The Pennsylvania State University, University Park.Pennsylvania 16802

Brown, et. al. (1) showed that significant changes in the strongly iron complexing ammoniacal species and itspH occurred inside cracks. Then Pickering and co- strong buffering action both preclude hydrolysis, so thatworkers (2-4) showed the same for electrode potential, no significant acidification can be expected.E, inside pits and, even more clearly, inside crevices.Although generally under recognized, the importance of EXPERIMENTALE in stabilizing local cells has gained in stature over theyears so that today E is considered as important as pH. Pure iron (Ferrovac E) and 1M NH4 OH -IMThe roles of pH and E in stabilizing localized corrosion NH4 NO 3 solution were used and the experimental set-upare shown in Fig. 1. Acidification within the crevice and sample arrangement are shown in Fig. 2. Thecauses the passivation potential, Epp, to move to the crevice consisted of one metallic side, four inertright, thereby extending the active region of the crevice (plexiglass) walls and one opening to the bulk solution.electrolyte towards the outer surface potential, E(x=o). The crevice width was larger in the upper section withAternatively, E(x) is a function of distance x into the dimensions 0.05 cm x 0.5 cm x 0.5cm and the lowercrevice, shifting in the less oxidizing direction or to the narrower section, 0.001 cm x 0.5 cm x 0.5 cm. Inleft in Fig. 1. Hence, at some distance into the crevice, preliminary experiments it was found that crevicing onlyE(x) is in the active region. In the limit, either the shift occurred in the narrower section where the condition JRin pH or in E can solely stabilize the local cell process, > IR* was met. Lacquer was used to eliminate creviceswhereas in general both E and pH participate in the between the iron and Teflon holder. Observations of thestabilization. A criterion for the occurrence of stable events inside the crevice and photographing them waslocal cell action is that IR > IR* (Fig. 1) since E(x) = possible with a stereomicroscope. Luggin-capillariesE(x=o) - IR, where IR* is determined by Epp which is were used to monitor E(x=o) in the passive region on thestrongly pH dependent and by E(x=o), andlR is outer surface and E(x) inside the crevice. Reproducibledetermined by the factors associated with charge results have been obtained using a Luggin measuringtransport in electrolytes (3,4). probe with a diameter 0.005 cm, to measure local

potentials in crevices of the same dimensions (3,6). AFollowing the recognition that large E(x) gradients silver oxide electrode was developed into a microprobe

contribute to the stability and mechanism of local cell for monitoring the pH inside the crevice. It wasprocesses (2), Valdes (3,4) measured large E(x) confirmed using buffer solutions of different pH that itsgradients within crevices in iron in acid solution (pH equilibrium potential was a linear function of pH. Thus,2.8) for which there is no tendency for hydrolysis and in both E(x) and pH could be measured as a function of thebuffered acid solution (pH 4.6) for which the hydrolysis depth x within the upper segment of the crevice but nottendency is suppressed by the buffer. In these in the lower segment because of its narrow opening.experiments for which the solution was already acid and The pH was also measured by extracting solution with aacidification per se was not a factor, he found that active syringe and along the crevice wall by applying p-Icrevicing only occurred when the IR>IR* condition was papers.met. Furthermore, in the absence of an active loop inthe polarization curve as is generally true for iron inalkaline solution, active crevicing did not occur unlessstrong acidification occurred. In this case, acidification With the outer top surface of the iron sampleis needed to form an active loop thereby making it anodically polarized in the passive region at +200 mV,possible to meet the IR > IR* condition. Conversely, if SCE, as illustrated in Fig. 1, the dissolution processesan active loop exists in an alkaline solution, crevicing observed on the iron crevice wall are shown in Fig. 3.

should occur even without acidification. The The horizontal line halfway down (x-=0.5 cm) the creviceiron/ammoniacal solution is perhaps unique in this regard wall is the boundary between the two segments of the

in that an active loop exists at pH 9 to 10 (Fig. 1). Also, crevice and the vertical white line down the middle is thefine Luggin-capillary used to measure the electrodepotential at the halfway mark, E(x=0.5 cm). The

*Electrochemical Society Student Member. outer(top) surface of the iron sample is in the passive**Electrochemical Society Active Member. region, as is the crevice wall down to the (lower)

3314 J. Electrochem. Soc., Vol. 137, No. 10, October 1990 © The Electrochemical Society, Inc.

horizontal line (x==0.7 cm) indicated by the arrow in the16 hour photograph of Fig. 3. Below the arrow(brighter region) iron is dissolving or crevicing. In this Microscope with cameraregion a dark-green corrosion product and an PROBE IREincreasingly aggressive dissolution were observed. The |latter was also indicated by an increasing current flowingout of the crevice with time, as shown in Fig. 4. The METAL WEtime dependent electrode potential, E(x--0.5 cm), at theopening of the narrower crevice is also shown in Fig. 4.At 18 hr, E(x--0.5 cm) is almost 600 mV less oxidizingthan E(x=o) at the outer passivated surface, and it can be 11 ...............

easily inferred that at a greater depth into the (narrower)crevice (specifically below the arrow in Fig. 3), E(50.7cm) < Epp, so that active iron dissolution or crevicing Fig. 2. Experimental set-up.occurs. m this experiment a 1:1 correspondencebetween active crevicing and a sharp E(x) gradient, sothat IR > IR*, was always observed. The measured pHof the crevice solution was always the same within theexperimental error (approx. + 0.5 pH unit) as the bulk(pH = 9.7) solution.

REERNE

1. B.F. Brown, C.T. Fujii and E.P. Dahlberg, ThisJournal, 116, 218 (1969).

2. H. W. Pickering and R. P. Frankenthal, ThisJournal,, 1.12, 1297, 1304 (1972).

3. A. Valdes and H. W. Pickering, in Advances inLocalized Corrosion (ed. H. Isaacs, et. al.) NACE,Houston, in press; H. W. Pickering, ibid.

4. H. W. Pickering, Corrosion Sci., 22, 325 (1989)5. J. W. Lee, K. Osseo-Asare and H. W. Pickering,

This Journal, 132, 550 (1985).6. B. G. Ateya and H. W. Pickering, This Journal, 6 hr 16 hr

M2, 1018 (1975).

ACKNOWLEDGMENTS Fig. 3. Photographs of iron crevice wall. E(x=o) =This work .as sponsored by the Office of Naval Research, Contract +200 mV, SCE (top surface).

No. N00014-84K-0201. We gratefully acknowledge helpful comments regard-ing the Iron/ammoniacal system by Y. T. Kho.

Manuscript received April 26, 1990.The Pennsylvania State University assisted in meeting the publica-

tion costs of this article.

100 300 100E(x=O) .

162 -l EX 0 b --o- o e,,.100

'IQ 1B-00 10"*1T 2Ex ---- PotentialE 10 R* .100 - Current

10 00301

.400

-0.8 -0.4 0 04 0.8 0 5 10 i5 20

E(SCE),V Time at E (x=O)

Fig. 1. Anodic polarization curve of iron in deaerated Fig. 4. E(x=0.5 cm) and current flowing out of creviceammoniacal solution, pH 9.7, 0.1 mV s-1 (5). as a function of time.

ll'-pIrilnt- f.|'l ... ok R. .( ''IE I crwH Ai. SI .I I , l iil n

Przited ii U.S.A.C:opYright 19,91

The Role of Chloride Ions in the IR > IR* Criterion forCrevice Corrosion in Iron

K. Cho* and H. W. Pickering**

Department of Materials Science and Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802

ABSTRACT

Although chloride ions have long been known to promote crevice corrosion in metals, little detailed understanding is avail-able In this paper, measurements of the potential, pH. and current inside the crevice, and simultaneous viewing of the crevice wallthrough a transparent portion of the crevice, have provided new results, and understanding of the role of chloride in promotingcrevice corrosion. The results also are a proof of the potential shift theory of crevice corrosion.

Acidification (of neutral or alkaline solutions) and a buildup amine complex ion. It was found that crevicing occurs in thisof the so-called aggresive ions, e.g., Cl ions. occur inside system when the electrode potential at the crevicing site shiftscrevices of iron. and both promote the crevicing process (1-3). into the active region, and that the pH holds constant in theRecently, it has been shown that i.) relatively large IR drops, crevice at the pH 9.8 bulk value (6). In these experiments on thesufficient to shift the electrode potential within the crevice into IR drop criterion for crevicing and on the role of acidificationthe active region, are necessary for crevicing to occur in iron (4-7), there were no chlorides present in the solutions, so it is(4-7): and ii.) acidification promotes crevicing by enlarging the clear that chloride anions are not necessary for crevicing toactive loop (the active/passive potential shifts to more noble occur in iron over the acid-to-alkaline pH range. Nevertheless,values), and is sometimes necessary to create an active loop, it is known that Cl ion promotes the crevicing process. Thebut is not in the case of acid and some alkaline solutions for purpose of the present paper is to investigate the role of chlo-which an active loop is already present (4-7). An example sys- ride in the crevicing process using the same experimental pro-tem in which crevicing occurs in the absence of acidification is cedure that has led to the above-mentioned identification ofiron in the alkaline ammoniacal solution. for which an active the IR drop mechanism of crevicing and improved understand-loop exists in the alkaline region because of the stable iron ing of the role of acidification in the crevicing process

* Electrochemical Society Student Member Thus, crevice corrosion has been found to occur when WfElectrochemical Society Active Member .A. °. where A(!) is the voltage drop in the electrolyte between

J. Electrochem. Soc., Vol. 138, No. 10, October 1991 t The Electrochemical Society, Inc. L57

~- XXLim XL Xp 0: : : : : : : : : : : : : : : : : : : : : : : : : :. ... ..... ...., , -,..I

crevcng . . - OtrPsurface

Fig. 1. Schematic of the po-tential shift mechanism ofcrevice corrosion.

Chloride-free

Epass Epi Surface

1POTENTIAL (+)

the cathodic site at the outer surface (where oxygen is plenti- elsewhere inside the crevice, in order to monitor the local elec-ful) and the crevicing site, and F is defined in Fig. 1 as the trode potential inside the crevice as a function of position anddifference between the cathodic site electrode potential. time. The pH was measured by extracting solution with a sy-ESufac,,, and the active/passive potential of the crevice electro- ringe connected to a fine glass capillary and by applying pHlyte,. . The measurable voltage drop, ',, is the product of papers along the crevice wall after disassembling. All of thesethe metal dissolution current flowing out of the crevice (in the procedures have been tested and described (4-8).-x direction) and the resistance of the electrolyte path to cur- eresultrentflow, i.e., ei = -IR, and IR > is equivalent toi 1' > the ResuiIn practice, the reconvolution of the IR product is not trivial, During crevicing, the visual changes on the iron crevice wallsince both the current flowing out of the crevice and the resis- were the same as reported previously (4-7), except pitting oc-tivity of the current path vary with distance x into the crevice. curred on the outer surface and part way into the crevice to aThe latter is especially evasive because H2 gas bubbles regu- certain distance along a horizontal front at xp, as shown inlarly form deep inside the crevice where E(, is in the hydrogen Fig. 3. Figure 3b and other magnified photographs of the crev-evolution potential range, as indicated in Fig. 1. In some re- ice wall were taken continually during the experiment. The si-gions of the crevice, wall, XL - X - XL,, in Fig. 1, the condition multaneously measured E(x) value at the pitted/unpitted inter-A~b >AV is met, meaning that E(x) is in the active region, face, xp in Fig. 3, corresponds to En,, (Fig. 1). The presence ofE..- E - EL-,,. Here E(x) is related to Ad by the equation, E(x) = the pitting front on the crevice wall (Fig. 3b) is another inde-Esrce A. Acidification shifts Epa,. to significantly more pendent confirmation of the large potential distribution alongnoble potentials inside the crevice for most neutral and alka- the crevice wall during crevicing and its correspondence toline bulk solutions. As a result, A( * is decreased and AM) may the anodic polarization curve of the dissolving metal with itsbe increased because larger metal dissolution rates can then active and passive regions (4-7). The measured limiting poten-occur in the active region. Thus, the condition A(i - Ad,* is metmore easily within the crevice as acidification occurs. In thispaper we address the question: does Cl ion similarly decrease -SCE - PtA(." by shifting Epa,,, in the noble direction, or does it increase Camera with macro-lensesI WEAMi (without an accompanying A(V" shift), and if so how? P I1_

Experimental PROBEAn iron,'Plexiglas crevice (Fig. 2) described elsewhere (4, 5) I , ' .

was anodically polarized on the outer surface at t600 mV vs.saturated calomel electrode (SCE) (passive region) in a buf- L ------------------------------------------------ --fered (pH 4.6) solution containing chloride. This system is '4 E ............................... .. . . .known to contain an active loop and to hold a constant pH X ................(4.6 t 0.5) inside the crevice during the crevicing process I(4, 5). The crevice opening was 0.5 mm x 5 mm and the depth ;RLwas 10 mm. Lacquer was used to eliminate crevices between Athe iron walls and the Teflon holder. The lacquer edges, them- ................ -...... . - ....selves, underwent some inconsequential crevicing, and prefer-ential pitting when the solution contained chloride anions.In situ observation of the crevicing action on the iron wall was . .._........................ .......... ........accomplished through the transparent Plexiglas using a ::- ::-.photographic camera with attached macro-lenses. A three- TEFLdirectional micrometer stage enabled precise positioning ofthe Luggin measuring probe at the opening of th- crevice and Fig. 2. Experimental set-up.

L58 J. Electrochem. Soc., Vol. 138. No. 10, October 1991 The Electrochemical Society, Inc.

Unpirted Pass..e A I'Itcd

Crevicing "

Region X X[--' + , ., , [p tc

Lightly 0 CvEtched 0 C

ED 00 0

Fig. 3. Schematic (left) and photograph (right) of the iron crevice wall taken through the Plexiglas during anodic polarizationof the outer surface at +600 mV vs. SCE in 0.5M CH 3COOH - 0.6M NaC 2H30 2 0.03M NaCI.

tial (EL,-) (5) in the crevice was approximately 550 mV vs. can amount to a rather significant increase in the current flow-SCE (Fig. 4), being about 35 mV more negative than the equi- ing out of the crevice, since the peak active current increasedlibrium potential for hydrogen evolution, consistent with the a few mA cm 2 and, hence, so also would the metal dissolutionobserved continuous formation of H2 gas bubbles in the rate at the crevice site (xL x • XL-. Other contributions to adeeper regions of the crevice, as reported before (4-8). The po- larger current flowing out of the crevice are the observed pit-tential gradient in the crevice, Atli. is thus as large as 1150 mV. ting on the crevice wall at x - x, (Fig. 3) and a higher passiveH2 gas bubbles were also observed coming continually out of current (schematically shown in the polarization curve inthe pits and adhering on the surface, confirming this sighting Fig. 1) located at x .- XL on the crevice wall The higher currentin a prior investigation (8) in which the gas was identified as causes a larger A(I) per unit length into the crevice, therebyhydrogen, a strong proof in itself that large potential shifts also yielding the .(f, - Ad,* condition at a smaller x, value. Thus, theexist within pits in iron, shift in XL to a smaller value is mainly due to a larger _16, with a

Crevicing more readily occurred in the presence of Cl ions, smaller contribution coming from the decrease in -'. Theas shown by the smaller XL value: XL = 1.3 and 0.7 mm for the measured constant pH values in the crevice showpd that thechloride-free and chloride-containing solutions, respectively, buffering action was effective in the crevice and that the pHFigure 4 shows the increase in current in the presence of chlo- was not a variable in the above experiments. In brief, chlorideride. Chloride anions were found to have a small effect on the anions promote the crevicing process by increasing the cur-potential of the activepassive transition: Ep... shifted 30 to rent flowing out of the crevice and. hence, increasing .16, and40 mV in the noble direction when the electrolyte contained to a lesser extent by decreasing Ad)'. Thus, based on the0.01M chloride. This resulted in a modest decrease in .&* and .l6 At* criterion, one can predict that crevice corrosion willan increase in the size of the active loop. The latter however, occur in shallower crevices or in crevices with a larger gap be-

tween the crevice walls when chloride anions are added to thesolution.

00 .. .00 Acknowledgments

E. j, at outer surface Financial support by the Office of Naval Research, Contract600 - - No. N00014-91-J-1189, is gratefully acknowledged. Konrad

Weil, Janusz Flis, and Yuan Xu provided constructive sug--a-- Potentiajl (chloride gestions.

4 00 Manuscript submitted May 8. 1991. revised manuscript re-.,. ' Current(chloride) U ceived July 22, 1991.

200 - - - Currenttchlonde-free) E Pennsy;vania State University assisted in meeting the publi-.. 10...., o cation of this article.

." 0 , ... REFERENCES

1. A Pourbaix, Corrosion. 27, 449 (1971).-00 L) 2 G(Sandoz. C. T. Fujii, and B. F. Brown, Corros. Sci., 10, 829

c. -00 **'~(1970).

-3 M G Fontana. Corrosion Engineering. 3rd Ed., pp 53-55." .400 , McGraw-Hill (1986).2. 4. A Valdes and H W Pickering, in "Advances in Localized

Corrosion, H S Isaacs, U. Bertocci, J. Kruger, and S.-600 L Smialowska, Editors, pp. 393-401. NACE, Houston (1990):

0 10 20 30 40 50 H. W. Pickering ibid. pp 77-84.5. H W Pickering. Corros Sci . 29, 325 (1989). Pickering, Cor-

rosion. 42, 125 (1986)Time(min.) 6 K Cho and H W Pickering. This Journal. 137, 3313 (1990)

Fig. 4. Current flowing out of the crevice (includes outer 7 H S. Kim. Y T Kho. H W Pickering. and K Osseo-Asare.surface contribution) and electrode potential at the bottom of ibid., 138, 1599 (1991)the crevice, E(, 0.,,, as a function of time for E , 0 8 H. W. Pickering and R. P Frankenthal, ibid., 119, 1297t600 mV vs. SCE. (1972)

r/.BASIC DISTR!?UTION LIST

Technical Reports and Publications Feb 1990

Organization Co'es Oroani:ation Copies

Defense Documentation Center Naval Air Propulsion CenterCameron Station Trenton, NJ 08628Alexandria, VA 22314 12 ATTN: Library

Office of Naval Research Naval Civil Engineering LaboratoryDept. of the Navy Port Huereme, CA 94043800 N. Quincy Street ATTN: Materials Div. IArlington, VA 22217 3ATTN: Code 1131

Naval Research Laboratory Naval Electronics LaboratoryWashington, DC 20375 San Diego, CA 92152ATTN: Codes 6000 1 ATTN. Electronic Materials

6300 1 Sciences Division 12627 1

Naval Air Development Center CommanderCode 606 David Taylor ResearchWarminster, PA 18974 1 CenterATTN: Dr. J. DeLuccia Bethesda, MD 20084

Commanding Officer Naval Underwater System Ctr.Naval Surface Warfare Center Newport, RI 02840Silver Spring, MO 20903-5000 ATTN: LibraryATTN: Library 1

Code R33 1

Naval Ocean Systems Center Naval Weapons CenterSan Diego, CA 92152-5000 China Lake, CA 93555ATTN: Library 1 ATTN: Library

Naval Postgraduate School NASAMonterey, CA 93940 Lewis Research CenterATTN: Mechanical Engineering 21000 Brookpark Road

Department Cleveland, OH 44135ATTN: Library

Naval Air Systems Commriand National Institute of StandardsWashington, DC 20360 and TechnologyATTN: Code 310A 1 Gaithersturg, MD 20899

Code 53048 1 ATTN: Metallurgy Division 1Code 931A 1 Ceramics Division I

Fracture & DeformationDivision

Officelof Nava'l ResearchResident RepresentativeOhio State University Research Center1960 Kenny RdColumbus, OH 43210-1063

Naval Facilities Engineering Defense Metals & CeramicsCommand Information Center

Alexandria, VA 22331 Battelle Memorial Inst.

ATTN: Code 03 1 505 King AvenueColumbus, OH 43201 1

Commancan: of the Marine Corps I Oak Ridge National Laboratory

Scientific Advisor Metals and Ceramics Div.Washington, DC 20380 P.O. Box XATTN: Code AX Oak Ridge, TN 37380 1

Oak Ridge, TN 37380 1

Army Research Office Los Alamos Scientific Lab.

P.O. Box 12211 P.O. Box 1663Research Triangle Park, NC 27709 Los Alamos, NM 87544

ATTN: Metallurgy & Ceramics ATTN: Report LibrarianProgram 1

Army Materials Technology Laboratory Argonne National Laboratory

Watertown, MA 02172-0001 Metallurgy DivisionATTN: Research Program Office 1 P.O. Box 229

Lemont, IL 60439 1

Air Force Office of Scientific Brookhaven National Laboratory

Research Technical Information Division

Building 410 Upton, Long Island

Bolling Air Force Base New York 11973Washington, DC 20332 ATTN: Research Library 1

ATTN: Electronics & MaterialsScience Directorate I

NASA Headquarters Lawrence Berkeley Lab.

Washington, DC 20546 1 Cyclotron Rd

ATTN: Code RM 1 Berkeley, CA 94720ATTN: Library 1

David Taylor Research CtrAnnapolis, MD 21402-5067

ATTN: Code 281 1Code 2813 1Code 0115 1

RE/1131/88/75

4315 (036)

Supplemental Distribution List Feb 1990

Profs. G.H. Meier and F.S.Pettit Dr. G. D. DavisDept. of Metallurgical and Martin Marietta LaboratoriesMaterials Eng. 1450 South Rolling Rd.University of Pittsburgh Baltimore, MD 21227-3898Pittsburgh. PA 15261

Prof. K.K. Birnbaum SProf. P.J. MoranDept. of Metallurgy & Mining Eng. Dept. of Materials Science & Eng.University of Illinois The Johns Hopkins UniversityUrbana, 111 61801 Baltimore, MD 21218

Prof. H.W. Pickering r Prof. J. KrugerDept. of Materials Science and Eng. Dept. of Materials Science & Eng.The Pennsylvania State University The 'ohns Hopkins UniversityUniversity Park, PA 16802 Baltimore, MD 21218

Prof. O.J. Duquette > Dr. B.G. PoundDept. of Metallurgical Eng. SRI InternationalRensselaer Polytechnic Inst. 333 Ravenswood Ave.Troy, NY 12181 Menlo Park, CA 94025

Prof. 0. Tomanek Prof. C.R. ClaytonMichigan State University Department of Materials ScienceDept. of Physics and Astronomy & EngineeringEast Lansing, MI 48824-1116 State University of New York

Stony BrookDr. M. W. Kendig Long Island, NY 11794Rockwell International Science Center1049 Camino Dos Rios .=Dr. J. W. OldfieldP.O. Box 1085 Cortest Laboratories LtdThousand Oaks, CA 91360 23 Shepherd Street

Sheffield, S3 78A, EnglandProf. R. A. Rapp

Dept. of Metallurgical Eng. Prof. Boris D. CahanThe Ohio State University Dept. of Chemistry116 West 19th Avenue Case Western Reserve Univ.Columbus, OH 43210-1179 Cleveland, Ohio 44106

-Dr. R. W. Orisko ,Prof. G. SimkovlchCode L-52 Dept. of Materials Science & Eng.Naval Civil Engineering Laboratory The Pennsylvania State UniversityPort Hueneme, CA 93043-5003 University Park, PA 16802

Dr. R.D. Granata Prof. M.E. OrazemZettlemoyer Center for Surface Studies Dept. of Chemical EngineeringSinclair Laboratory, Bid. No. 7 University of Florida-Lehigh University Gainesville, FL 32611Bethlehem, PA 18015

Dr. P. S. Pao Prof. J. O'M. BockrisCode 6303 Dept. of ChemistryNaval Research Laboratory Texas A & M UniversityWashington, D.C. 20375 College Station, TX 77843

Dr. N. S. Bornstein Dr. V. S. AgarwalaUnited Technologies Research Center Code 6062East Hartford, CT 06108 Naval Air Development Center

Warminster, PA 18974-5000Prof. R. M. LatanisionMassachusetts Institute of Technology Prof. Harovel G. WheatRoom 8.202 Dept. of Mechanical EngineeringCambridge, MA 02139 The University of Texas

ETC 11 5.160.Dr. R. E. Ricker Austin, TX 78712-1063National Institute of Standards andTechnology

Metallurgy Division -. Prof. S. C. DexterBldg. 223, Room 8-266 College of Marine StudiesGaithersburg, MD 20899 University of Delaware

700 Pllottown Rd.Dr. F. B. Mansfeld Lewes, DE 19958Dept. of Materials ScienceUniversity of Southern CaliforniaUniversity ParkLos Angeles, CA 90089

Dr. W. R. BitlerDept. of Materials Sci. and Eng.115 Steidle BuildingThe Pennsylvania State UniversityUniversity Park, PA 16802

Dr. S. SmialowskaDept. of Metallurgical EngineeringThe Ohio State University116 West 19th AvenueColumbus, OH 43210-1179

Dr. R. V. Sara-Jnion Carbide CorporationUCAR Carbon Company Inc.Parma Technical Center12900 Snow RoadParma, Ohio 44130

Prof. G.R. St. PierreDept. of Metallurgical Eng.The Ohio State University116 West 19th AvenueColumbus, Oh 43210-1179

Dr. E. McCaffertyCode 6322Naval Research LaboratoryWashington, D. C. 20375

2