Metal Lid Ing

8/8/2019 Metal Lid Ing

1/26

PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [Indian Institute of Technology]

On: 15 November 2010

Access details: Access Details: [subscription number 731845207]

Publisher Taylor & Francis

Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-

41 Mortimer Street, London W1T 3JH, UK

Materials and Manufacturing ProcessesPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t713597284

SURFACE MODIFICATION BY ELECTROCHEMICAL METALLIDINGM. Kendiga; S. Menezesa; S. Jeanjaqueta; D. Raleighaa Rockwell International Science Center, Thousand Oaks, CA

To cite this Article Kendig, M. , Menezes, S. , Jeanjaquet, S. and Raleigh, D.(1989) 'SURFACE MODIFICATION BYELECTROCHEMICAL METALLIDING', Materials and Manufacturing Processes, 4: 3, 385 409

To link to this Article: DOI: 10.1080/10426918908956300URL: http://dx.doi.org/10.1080/10426918908956300

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.
http://www.informaworld.com/smpp/title~content=t713597284http://dx.doi.org/10.1080/10426918908956300http://www.informaworld.com/terms-and-conditions-of-access.pdfhttp://www.informaworld.com/terms-and-conditions-of-access.pdfhttp://dx.doi.org/10.1080/10426918908956300http://www.informaworld.com/smpp/title~content=t713597284


2/26

MATERIALS & MANUFACTURING PROCESSES, 4(3), 385-409 (1989)

SURFACE MODIFICATION BYELECTROCHEMICAL METALLIDINGM. Kendig, S. Menezes, S. Jeanjaquet,

and D. RaleighRockwell International Science Center

1049 Camino Dos RiosThousand Oaks, CA 91360

ABSTRACTControlled electrolysis in molten salts can improve the surface properties of metalal loys for wear and corrosion resis tance. Surfaces of Al II 00 and steel were modified bypassage of cathodic current under potentiostatic and galvanostatic control in moltensalts containing Ce 3+, BF, ", TP+ ions. Hard, poorly adherent CeAI, and CeAl 3 layersformed on AI surfaces at potentials which were 200 to 400 mY posit ive to the reductionof Ceo The deposition occur red by a reduct ive prec ip itat ion from a KCII LiCI melt at560C containing I mole96 Ce 3+ . Boride and titanide layers were found on steel in aLiF/KF melt containing the respect ive ions. A borided layer on steel exhibited a factorof ten improvement in hardness and a titanided layer exhibi ted a 3096 increase in hard-ness.

INTRODUCTIONAdverse environmental effects, economic factors, and structural demands oftenplace seemingly incompatible requirements on materials used in environments of highcorrosivity or wear. High-performance coatings can meet these demands. Particularlyuseful are coatings that fall into the category of "surface modification" such as nitridingor carburizing. Surface modification is distinguished from conventional coatings in thatthe surface layer is formed by in-situ conversion of the workpiece surface to a new alloyor compound. Surface modification eliminates many problems, such as cracking, spallingand pinholing, t ha t are associated with physically distinct coating layer formation.The metalliding surface modification process,3"'. 'I"" developed in the 1960's by

N. Cook.> entails the electrodeposi tion of a metal from a molten halide bath in the formof a surface alloy with the substrate metal. The mechanism for this process involves thefollowing reactions:anode (metallide metal M(I)): M(I) M(On+melt + ne- (I )

cathode (workpiece metal M(2)):

Copyright 1989 by MarcelDekker, Inc.

ne" + M(l)n+melt M(l)surfaceof M(2)

385


3/26

386

M(J)(surface of M(2 (diffusion) M(l)-M(2) alloyFor the metalliding process to proceed, the side reaction:

M(2) M(2)m+melt + me-

KENDIG ET AL.(3)

must be substantially less favorable than reaction (2). This can be assured by:1. Solid solubility of M(l) in M(2)2. The relative s tabi li ty of the fluoride of MO) over that for M(2).3. MO) being more e lect roposi tive than M(2)

Criteria 2 and 3 have led to a metalliding series which orders metals with respect totheir abili ty to metallide metals below them in the ser ies. 3 '1 5 Y, Ce and the transuranicelements reside at th e top of the l is t, while the noble metals remain near the bottom.Even for cases where all of the above cri ter ia are met, slow rates for any of th esteps may adversely affect the viabili ty of the process. For example, a slow diffusionprocess brought about by the presence of substrate impur it ies or a slow rate of alloyingwill minimize the kinetics of reaction (3). On the other hand, less electropositive metalscan metallide more electropositive metals in cases where alloy forma tion is highlyfavored, provided the substrate is int roduced to the mel t under applied potential.Typically, metal liding has been carried out under constant current conditions.Accordingly, the potential at the substrate cathode will change with time as the alloycomposition gradually increases toward tha t of pure MO), at which point conventionalelectroplating commences . However, if the process were carried out under potentialcontrol, and in the absence of or correction for ohmic potential drop within the melt, agiven alloy composition could be maintained at the surface as the metallide metal diffuses into th e bulk. In this case, the activity, am' of the diffusing metal at the surfaceremains constant according to the equation:am =exp(-(nFE/RT) (.5)

where E is the difference be tween the potential of the substrate and the reversiblepotential of a nonpolarized reference electrode of the metallide metal. The current forthis reaction would follow the equation:

where Cm is the surface concentration of the diffusing metalliding element.Fluoride melts provide an exce llen t medium for metalliding as a re su lt of theirability to dissolve residual layers of the surface oxides on the substrate materials.J" 14'l 5As a structural material, Al finds increased application due to it s relatively lowdensity. However, it is a soft, easily scratched metal, and corrosion, wear and fatiguelimit i ts ult imate usefulness . A rapid, economical method for hardening the surface ofAl components against corrosion and wear without altering the fabricated surface geometry would benefit a number of technologies associated with vehicle construction whereadverse environmental effects demand high corrosion or wear resistance.It has been the objective of this work to provide the electrochemical characterization required to make metalliding a reliable well-understood and well-controlled processfor metal surface modification, and to explore the possibili ty of producing a hard, corrosion-resistant metallide surface on AI.


4/26

SURFACE MODIFICATION

Table IThermodynamic Data for Metallurgical ReactantsFree Energyof Formationof the Fluoride Electro- 25*Metallide 1000 = negativity eMetal 6GF (Pauling) (Voll)kcal/mol

Ce -115 I . l 2.48Y -110 1.2 2.37

K -101 0.8 2.95Sc - 98 J.3 2.08Hf - 93 1.3 1.70

Zr - 92 1.4 1.53

Ti - 90 1.5 1.63

Al - 89 1.5 1.66n.a. Not applicable* Mn- + ne- =M (aqueous)

SolubilityMinAI.05 wt%0.8 wt% ateutectictemp.Insoluble- .05 at . %0.182 at . %at 662C0.068 at . %at 640C0.14at.%665Cn,a,

387

Metalliding Al with a number of elements should be possible, despi te i ts high electropositive ranking and its tendency to form extremely stable oxides that persist to lowlevels of residual oxygen, which could inhibit the diffusion of metallide metals. Since thehalide of the metalliding metal must have greater stability in the melt than th e halide ofthe substrate, in addition to being more electropositive for metalliding to occur, themetals Ce, Sc, Hf, Zr and Ti (as shown in Table I) would be expected to form a metallidein fluoride, given their fluoride stabilities, electropositive nature and solubility in AI. K,of course, is completely insoluble in AI, and therefore would not metall ide despite it selectropositive nature and stable fluoride. Ce, the most electropositive of the metalsconsidered, was chosen as a first candidate for metalliding AI. It lies substantially aboveAI in the metalliding series,' and several obscure reports have been made of the successful ceriding of Al from a chloride melt. ' Experiments were conducted to evaluate theability of Ce to metallide AI from KCI/LiCI and KF/LiF melts.

EXPERIMENTALApparatus

Figure I shows a schematic for the apparatus used in the metalliding experiments.A quartz cell sea led with an elastomeric gasket formed the outer cell which was placedin a thermally controlled furnace. The cell contained an alumina crucible that held the


5/26

388 KENDIG ET AL.seas 300&2

E

2

DIGITAL THERMOMETER

IVAC?""LOW 0~ O L'i ] A ~ SEAL M E T E ~ ' - II 111 ' U ' ~ x . . > 'r-+VENT III fVACUUMi' t- ,

~ T / C I N '--ALUMINA WEl l

V .....-MOLTEN SALT INj J I ALUMINA OR Ni CRUCIBL1~ F U R N A C E( -' - ,M2M,M,M3AUX REF WORKING TEMPERATURE CONTROLLER

POTENTIOSTATEap E I

RECORDERf- --------0 x-IORx-yTRANSFER FUNCTIONANALYZER

tCOMPUTER VAX 11I7BO -FIG. 1 Schematic of the apparatus.

fused salt. A number of stainless steel baffles placed above the salt container minimizedheat loss from the cel l, thereby protecting the gaskets of the cel l cap and electrode fittings. The electrodes and thermocouple well passed through holes in th e baffles, andcould be lowered into the melt by loosening gasketed compression fittings and pushingthe respective electrodes of the thermocouple well down into the mel t. The cell atmosphere could be controlled from a 5 mTorr to 3 psi atmosphere of high purity 1 ppm0'2) Ar, as shown in Fig. 1. Vented gases were directed through an O2 meter, followed byan oil -seal for minimizing any backdiffusion of atmospher ic oxygen. A potentiostat wasused to control the potential of the specimens (M,) vs a reference electrode (M3) . Asecond electrode (M1) served as the auxiliary electrode and anode during the metallidingprocess. Additional apparatus in the form of a computer-eontrolled transfer functionanalyzer (TFA) or a sweep generator (not shown) could be used in conjunction with thepotentiostat for respectively collecting the electrochemical impedance spectrum or


6/26

SURFACE MODIFICATION 389

cyclic vol tammograms. A PAR model 373 served for most of the metalliding runs. Fortests using IR compensation, the Stonehart BCl200 potentiostat was used with IR compensa tion via posi tive feedback. For the impedance measurements, a PAR 173 with amodel 276 interface was used, as described elsewhere. 7Melt Preparation

KCl/LiCI Melt Containing Ceo 107 gm of Anderson Physics Laboratories (APL)LiC!/KCI eutectic, 2.95 gm of CeCI 3 and I gm of Ce chips (Alfa 99.9'(6) were weighedout and transferred to the crucible and sealed in the cell in a dry Ar atmosphere. The Ceelectrodes (Al fa 99.9'(6) were abraded in the dry Ar atmosphere and connected to stainless steel leads running through AI,O, rods, which were sealed at the outer end with ahard wax.The cell containing the electrodes and salts was transported to the furnace andslowly raised to 560C with no further purification.KF LiP Melts Containin Ce. 16.85 gm of LiF (Reagent grade Anachemia AC5469), 37.02 gm of KF Aldrich, 99.999'(6), and 2.07 gm CeCI, (corresponding to I mole%)were weighed out in a dry Ar atmosphere and transferred to the me lt crucible. Thecrucible was slowly heated 'to 550C under vacuum to remove res idual water and thenwas allowed to cool. Ce chips were added and the system was elevated to 610-620C forthe metal lid ing. The vacuum treatment was necessary to remove res idual water fromthe mel t. Above - 600C, a vacuum could not be applied to the sys tem since it causeddecomposition-of the melt, forming K vapor. The vapor condensed on the cold portionsof the cel l and was later confirmed to be K from a flame test. A second fluoride meltwas made up in the same manner, except that 1 m/o of CeF, (Aldrich 99.9'(6)was used inplace of CeCl 3 and the 99.9'16 Ce was obtained from Aesar (Seabrook, NH) andGoodfellow Metals (Cambridge, England).KF LiP Melt Containin Ti. A Ni crucible con ta ined 44.4 g (1.7 moles) of LiF and91.8 g 1.58 moles of KF. I mole% (0.033 moles) of TiF 3 was added to these salts. Thecrucible was fi lled and placed in the cell in a glovebox purged with Ar. The cell was thenplaced in the furnace and connected to the system shown in Fig. I and evacuated t o 60Cand 200 rnTorr, The temperature of the evacuated cell was gradually increased to 100Cto drive water from the salts. After being evacuated for 24 h, the sal ts were brought toa working temperature of 560-600C for the metal liding exper iments . Ti served as thereference (M,) and counter (M/) electrodes.KF/LiP Melt Containing B. In a Ni crucible in a glovebox purged with Ar, 1.8 moleLiF, 1.8 mole KF and 0.133 mole KBF4 were weighed. The cruc ible was placed in thecell (Fig. I) and sealed. The cell was attached to the system shown in Fig. 1, evacuatedand slowly heated to 467C, keeping the pressure under 100 mTorr. The heat wasremoved and the cell kept under vacuum for 16 h before bringing to temperature for themetalliding experiments. Boron chips held in an Au basket served as the counterelec

trode (MI ') , and a graphite thermocouple well was used as the pseudo-reference electrode(M I ) .Metallography and Specimen Analysis

Selected specimens were subjected to scanning electron microscopic analysis(SEM), including x-ray f luorescence analysis (EDAX). The loosely adher ing scale wasremoved from the surface by light abrasion and was mounted in the microscope for thisanalysis. In several instances, metal/ographic cross sections of the treated specimenswere mounted in epoxy and were polished to 0.03 ~ finish in kerosene, followed by anelectropolish in MeOH/H 2S0,. Optical micrographs and hardness measurements wereobtained for the cross sections.

For one case where a metallic scale was formed on the specimen, the scale wasstripped from the specimen and was mounted with double-sided adhesive on a glass slidefor x-ray diffraction.


7/26

390 KENDIG ET AL.SC8530055A

-- KCLIliCi--- KFILiF

10 5 .--- - -.-- - - r-- - --.--- - --,- - - ,.....- - --,

r. 5FIG. 2 Typical log current vs log time curves for specimens treated in KCI!LiCI at560C and 2.5 mV (I, Il, [[I); 16.5 mV (IV, V); in th e KF!LiF melt between 580615C at 2.5 mV (II); 16.5 mV (11l) and 35 mV (IV)compared with a calculated

curve.

Electrochemical AnalysisElectrochemical impedance measurements were performed on Al test specimens,using techniques described in Ref. 7 and references therein.

RESULTSFormation of Ce Coatings

Chronoam ometr in CC and F- Melts. Figure 2 compares currents 0) vs time (t )behavior of Al spec imens in KF LiF and KCI!LiCl melts at various applied potentialswith a calculated plot computed from Eq. (6), assuming a solid-state diffusion coefficientof 10- '0 cm 2 s-. and a surface activity of 0.9. The calculated Curve represents an upperlimit for the metalliding currents. However, the i- t curves in both Cl" and F- melts showhigher currents. The currents ar e typically much higher in the F melt and decreaseerratically with time. The nonlinear behavior of the current as a function of t- 1/ 2 suggests AI-Ce compound formation rather than Ce diffusion. The behavior in th e CIappears to be d iffusion- limi ted with t- 1/ 2 dependence. Previous investigators claimedceriding Al gave better. results with the Cr" compared to th e F- melt.-Impedance Analysis

Changes in the resistance of surface films and relative reaction rates at Al electrodes were monitored with electrochemical impedance spectra. 7 '8 ' " Impedance spec-


8/26


SC85300616r - - - r - - - r - - - r - - - r - - - . , - - - - - . , - - - - - . , - - - - - - - ,

UiEJ:.e.!:N

NClo...J

2 __ CHLORIDE-... ........ ....o FLUORIDE< ...................------6

LOG co, Iw in radlsl- 2 ' - _ ' - - _ ' - - _ ' - - _ - ' - - _ - ' - - _ ~ _ _'__----'-2

Cl 60Ql't l ,..- ....,.; / "-...J / \CHLORIDElZ 40cl: I \w I \VI I 'l: I \J:..... / \- 20 \ \ \ \0-2 6

LOG co, (w in rad/s)

FIG. 3 (a) Electrochemical impedance, and (b) phase angle spectra for Al 1100 inLiCI/KCI melt at .560C and LiF/KF melt at 610C.


9/26

392 KENDIG ET AL.SC37070

6r--.----r--r--r--,--r--,-----,

..E.g

.5NNCls

wwctewoW....ClZc:(wIIIc:(:z:n.

o 2LOG w (w in rad/s)

o 2LOG w (w in rad/s)

4

4

6

6

FIG. 4 Comparison of (a) impedance and (b) phase angle spectra for Pt (1), I \ l (2), Alafter voltammetric cycling 0) and Al after 15 h electrolysis a t 0.25 V vsECe+3/0 .

tr a for Ali i 00 in th e Cl" and F melts ar e shown in Fig. 3. The spectra allow evaluationof th e polarization res is tance, Rp ' and th e apparent solution resistance, Rs' ~ equalsth e low-frequency limit of th e impedance spectrum minus Rs obtained from th e High-Irequency limit. R relates inversely to th e rate of charge transfer at th e metal/electrol yt e i n t er f a ce , al?d Rs depends on th e combined resistance of th e solution and th e inhibiting films. The spectrum for th e Al in the chloride medium ca n be resolved into a nonzeroR of_82 n and an Rs of 0.35 n for th e - 2 cm 2 specimen. No R could be resolved forttPe F melt, since Rp is substantially less than th e Rs of 3 n as a result of a rapid chargetransfer reaction, pdssibly due to corrosion. The decrease in Rs for th e CC m el t c om -


10/26

SURFACE MODIFICATION 393SC370691.0,------r-----r----,r---,-----.,..-------,-------,......

...NN

....II!...0"':NN.,;

Olo.,;..N.,;

...2=>oow>Wu:

7000001 l : - : 0 = - - - - - - : ! = - - - - - - = ' = - - - - - - : " : : - - - - ~ = - - - - - - - - = ' : = - - - - - ~

fIG. 5 X-ray dif fraction for the surface layer formed from electrolysis at 0.25 V vsCe in Ce-LiCI/KCI eutectic at 560C.

pared to the F" melt of nearly an order of magnitude suggests a higher solution res istance for the higher melting f- salt. On the other hand, R for Al in the cr melt isbetween one and two orders of magni tude higher than for the ~ melt, suggesting a lowercorrosion rate, possibly due to the presence of residual surface oxides in AI. Thus, the F"melt can dissolve res idual surface oxides on substrates, but provides an ironically moreresistive medium.Impedance spectra obtained for Ai and Pt in the CeCiTLiCI/KCI eutectic aftervarious electrochemical treatments appea r in fig. 4. Curves I and 2 compare theresponse at Pt and Al electrodes, with roughly the same surface area and geometry relative to the Ce counter and reference electrodes, and thus the same Rs' Significantlyhigher Rs for Al suggests the presence of res is tive films on the surface. The low-frequency por tion of the impedance spectra results from interface charging. The large r isein impedance for Pt at frequencies


11/26

394

FIG. 6

KENDIG ET AL.

Photomicrographs of Pt after passage of cathodic current in the Ce containingLiCl{KCI melt (a): AI after passing cathodic current (b); AI after passingcathodic and anodic current (c); AI after mainly anodic treatment (d); AI3Ce-Al4Ce on Al (e).


12/26


AI

395

..,.

OL, .....

,.

X-Ray Diffraction of Cerided AI

FIG. 6 CONTINUED

Visual inspection of Al specimens polarized at 250 mVvs ECe/C +3 for up to 15 hdiscerned formation of a grey-black film which was not always adl ierenf to the substrate.Some of these films were mechanically removed and characterized with x-ray diffrac-tion. The diffraction pattern shown in Fig. 5 ident if ies the film as comprising mainlyCeAI . . with some CeAl 3 and a-Ceo


13/26

396 KENDIG ET AL .8C8&3004l1

(a)

FIG. 7 SEM of (a) crystaJline surface layer , (b) black underlayer on the metaJlided (c )Al substrate.

PhotomicrographsCross sec tions of specimens sub jected to different electrochemical treatments

were examined optically and also tested for hardness . Photomicrographs in Fig. 6 i llustrate the formation of different layers and the range of adhesion encountered for thevarious cerided Pt and Al samples . Micrographs (a) and (b) were obtained on Pt and Alsamples after cyclic polarization between Eoc and 0 V vs ECe/Ce + 3 (cathodic polariza-t ion only) as desc ribed earlier. Conversion of the Pt surface to Pt ceride (a) occursnearly ideal ly , generat ing a compact adherent layer , thus confirming the feasibility ofthe metalliding technique, at leas t for cases with a l arge differential in electronegativityas for Ce and Pt . The AI-Ce coating (b) appears more porous than the Pt ceride, particu-larly nearer to the substrate , and less adhesive. When the substrate is anodically andcathodically polarized, two distinct layers are formed and the adhesion deteriorates, asseen in micrograph (c). The gap between the layers was found to widen (micrograph (dfor specimens potentiostatically cycled in a melt were the anodic dissolution occursrapidly, that is where large anodic currents are superimposed on the reduction currents.Despi te the lack of adhesion encountered in metallides obtained below 600C, it was possible to obtain a coating (micrograph (e comparable to that for Pt (micrograph (a byimmersing Al in a melt comprising CeCI 3 -L iC1 /KC l and metallic Ce at 660C. The Alsurface appears to have melted, forming an eutectic with Ceo Except for a few voids, nogap is visible between the coa ting and the substrate. The diffusion of Ce a long grainboundaries into the substrate is clearly apparent from micrograph (e).SEM-EDAX Analysis

Film topography var ied from granular crystallites to continuous, smooth or roughfilms. As shown by the SEMs in Fig. 7, the overlayer generally appeared to be crystaJline(a), while the underlayer consisted of a black scale (b) over the substrate (c).


14/26


(bl

lei

FIG. 7 CONTINUED

397

Results of EDAX analyses for the various regions are shown in Fig. &. The crystal-line layer (a) is AI-rich and probably compr ises CeAI.. The gap (b) between the l ayer scontains large quantities of CI in addition to K, Ce and AI. Such a gap was consistentlyfound in specimens that had been anodically polarized, as shown in Fig. 6c and 6d. Com-position of the black underlayer (e) varied with the electrochemical treatments, but theintensity of the AI lines was generally lower than that of Ceo This is also apparent fromtrace (b). Spect rum (d) shows the AI substrate under these layers. Spectrum (e), obtainedon a Ce counterelectrode, shows incorporation of AI on the surface.


15/26

398

AI

SC37139(a)

Ce

KENDIG ET AL.

o XENERGY, KeV

10

FIG. 8 EDAX for regions a, b, c and d (gap).


16/26


KNOOP HARDNESS TEST8C8. .34774

399

, , . -460 HK10025 HK100

I I/ 50 IJrn.IIFIG. 9 Photomicrograph of a metal lided specimen, showing indentations for Knoophardness tests on the (I ) Ce layer, (2) CeAl rCeAl4 layer, and (J) Al substrate.

Although the EDAX analys is in Fig. 1\ is representative, the precise ratio of thevarious components, particularly Cl, varied for the specimens examined. For the ratio ofrelative intensities, AliCe was found to vary widely for different samples and for layerswithin a sample from 1.2 to 0.2, the higher ratio representing the crystalline outer layer,possibly CeAl, ; The less well-defined inner layer generally accompanied large gaps andmay comprise more than that one phase, as seen in Fig. 6. The lower Al content in thi slayer implies a faster diffusion coefficient for Al in Ce than Ce in AI, which suggestsvariation in the diffusion coefficients of Al in the various intermediary Al-ceride phases.Microhardness Indentatioo Tests

Knoop hardness tests performed on the metal lided layers and on the Al substrateare compared in Fig. 9. Hardness measurements on the CeAl,-CeAI 3 layer averaged460 HK 10 0 compared to 120 HK 10 0 for the electroplated and 25 HK 1 0 0 for the Al substrate. Thus, the Al-Ce layer is much harder than either of the component elements.Discussion of Results of Ceride Formation

Occurrence of several anodic and cathodic reactions between the potentials of Alcorrosion (Eoc) and Ce deqlJ' it ion (ECe/Ce+ 3) was evident from voltammetric experi-


17/26

400 KENDIG ET AL.

ments described elsewhere. Cathodic depositions of the ceride at 460, 350 and 20 mYvs ECe /Ce+3, respectively, imply interaction between Al and Ce+ to form reduced compounds, more anodically stable than Ceo The CeAl 4 phase appears to form on Al under avariety of potentiostatic condi tions. Minor changes in the experimental parameters donot affect the stoichiometry of the outer layer , but affect the morphology and adhesionof the coa t ing.Formation of compounds CeAl, CeAI" Ce,Al3 and CeAl.,; via reaction

is a three-electron process. Formation of Ce3Al involves nine electrons. The respectivefree energies, 6G, for the 460, 350 and 20 mY processes ar e -32, -24 and -14 kcal/molefor a three-electron reduction and -9.5, -7 and -4 kcal/mole for a nine-electron reduction. 6H'96 values for CeAl ; and Ce3Al have been reported to be -39 and -22 kcal/mole, respectively.' Thus, while the formation of Ce-Al intermetallics is thermodynamically possible, the formation of AI-rich compounds is more favorable by 10-20 kcal/molethan Ce-rich compounds.

Two possible mechanisms for the formation of the observed CeAI,-CeAl. layer are:I) reduction followed by solid-state diffusion of Ce into the Al surface as previously outlined in Eqs, (1)-(4); or 2) reductive precipitation of the CeAI3-CeAI. compounds fromthe mel t onto th e surface of the AI. Previously described experimental resultsI' suggestthe latter reductive precipitation mechanism.The reductive precipitation mechanism results in the overall reaction of Eq. (7) bythe following steps. First, Al is sufficiently active to dissolve anodically in the CC electrolyte by the reaction:

followed by th e forma tion of a Ce chloro-aluminate complex in the melt:3- 3+A1Cl6 (melt ) + Ce (melt); CeA1CI6(melt)

(8)

(9)

The CeAlCl 6 complex is then reduced in the highly reducing, electronically conductingmelt to form a CeAl4 precipitate from the mel t onto the metal surface as:- C - 3+15 e (melt) + 4 eAICI6(melt) + CeAl 4 + 24CI (melt ) + 3Ce (melt)

(10)This mechanism is consistent with the presence of observed predominantly CeAl4films. Reduction of a Ce chloro-aluminate complex from the mel t is also consistent withthe observed CI- inclusions found in the films. However, after passage of sufficientcathodic charge, a highly reduced intermediate, possibly a csan complex, builds up int he mel t and results in the inclusion of the more reduced CeAl1 compound in the resulting films. This intermediate tends to undergo rapid anodic OXidation, making the meltappear more active (Fig. 9).The formation of Ce-AI metallide via solid-state diffusion must be ruled out, sincethis reaction requires a high rate for the diffusion of Ce into the Al substrate (k 3 in Eq.(3. To our knowledge, no data exist in the literature for diffusion coefficients for AI


18/26


Table 2Summary of Titaniding ExperimentsT E(OC) Substrate mVvs Ti Conditions Results

564 AI -350 Zero Vvs Ti Corrosion- 5 mA/cm'565 Al -330 -50 mA/cm' Black layergalvanostatic560 AI -290 -5 mA/cm' Silver white depositgalvanostatic containing K565 Steel 160 -200 mV vs Ti Ti layer under a scale(- 50 mA/cm ,) EOAX showed Ti andsome K, no Fe744 Steel 40 oVvs Ti Silver deposit, EOAX(- 10mA/cm') showed Ti and Fe730 Steel 42 ov vs.ri Silver deposit, EOAXshowed uneven distri-bution of Fe and Ti

in Ce (OAI/Ce) or for Ce in AI (DCeJAJ. If DA1/Ce DCe/AI' the cerided specimenswould have a Ce-r ich surface and a LCeJ gradient cfecreasing exponentially according toEq. (6) into th e bulk. Alternatively, success ive phases of Ce3AI, CeAI, CeAl, CeAI 3,CeAl, and At would follow a Ce-rich surface. The formation of only one or two At-r ichsurface coatings, CeAl, and CeAI 3, and the large gap between the Al substrate and thecoating as shown in Fig. 6 suggest that 0 AI/Ce 0Ce/AI'The solid solubility of Ce in Al is reported as being


19/26

402

'"...c:.,...e.,>....,a;'"

o

Steel-200 mY vs Til iF/KF + 1 mole ll: TiF,

5Energy. eY

KENDIG ET AL.

10

FIG. 10 EDAX for a steel surface subjected to electrolysis at -200 mVvs Ti/Ti3+ electrode in Ti containing LiF/KF molten salt.

Formation of TitanidesPotentiostatic and galvanostatic electrolysis of Al 1100 and 1010 steel in LiF/KFcontaining 1 mole% Ti gave the results summarized in Table 2. Al was more active inthe fluoride melt, as evidenced by the negative open-circuit potential vs the Ti referenceelectrode. For most cases, the Al electrode corroded when polarized at potentials nearzero volts vs Ti, For cathodic galvanostatic electrolyses at -50 mA/cm 2 and -5 mA/cm 2 ,a b lack salt layer remained on the surface with no apparent deposi tion of Ti or K.On steel electrolyzed in the molten Ti -conta in ing bath, electrolysis under verycathodic condi tions (-200 mV vs Til resul ted in apparent plating of Ti without formationof a metallide. EDAX analysis of the resulting surface showed it to be primarily that ofTi (Fig. 10). However, electrolysis at 0 V vs Ti (current equal to 10 mA/cm 2 ) produced asilvery deposit composed of Ti and Fe shown by the EDAX spectra in Fig. l l a, SEMrevealed that the deposit also contained small crystallites composed of Ti and Fe (Figs.11band 12). Metallographic cross sections of th e specimen (Fig. 13) revealed a thin surface layer having a hardness (KN 50 ) of 170 24 as compared to a substrate hardness of120 10.

Formation of a Borided SurfaceThe electrochemical formation of borided steel has been reported etsewhere s andreproduced here. A galvanic current of 12 mA/cm 2 for 51 min produces a nominally

400 urn layer , shown in Fig. 14, having a Knoop hardness (l00 g load) between 1100-1300,which is ten t imes that of the substrate (KNIOO = 106).


20/26


=....,c.,....oS.,>......0;a: Ti

Steelo V YS Ti744 .cLiF/KF + I moleX TiF.

Fe

403

- - - - ~ - -o

'"'iiic..

.5.,>...

II :

TI\ I. FevL.-ILA5

Energy, eV

Steelo V vs Ti744 cliF/KF + 1 mole X Ticrystallite

TI

10

o 5Energy, eV

10

FIG. 11 EDAX for a steel surface subjected to electrolysis at 0 V vs Ti/Ti'+ electrode(10 mAlcm 2) in Ti containing LiF IKF molten salt at 744C (a); crystalliteformed under th e same conditions (b).


21/26

404 KENDIG ET AL.SC39149

STEELLiF/KF BATH CONTAINING 1 MOLE% TiF3744C+40 mV vs Ti

FIG. l2 SEM of Ti-Fe crystallite.

Using similar and higher currents, no boriding of AI was achieved. Figure l5 showsthe current vs vol tage behavior for the steel and AI substrates in the boron-containingmelt. The potential vs a carbon pseudo-reference electrode was swept at a rate oflO mV/s. The current densities ar e very high and of the same order of magnitude forboth materials. The zero current potential Eoc for steel is much more positive than thatfor the more active AI by about 1 V. A large change in the cathodic overvoltage appearsfor the steel electrode after the f irst sweep in the cathodic direction. For AI, an apparent increase in anodic current occurs after the cathodic scan, indicat ing a reversibleredissolution of a cathodically deposited material.


22/26


SC39148

\ .-J"\ ' \,I'j \\ \... (

\


23/26

406 KENDIG ET AL.

IC3UU

STEELLiF/KF MELT CONTAINING 3.5 MOLE% KBF4-12 mA/cm2 FOR 51 MINUTES886C

FIG. 14 Photomicrograph of a borided steel surface.

is relatively low, as evidenced by the lack of current hysteresis upon reoxidation (Fig.15). The reaction that depolarizes the steel interface is the metalliding reaction (Eq.(2. It can be concluded that Al in the boron-eontaining melt is too active from a corrosion standpoint to facilitate the progress of the metalliding reaction.SUMMARY OF RESULTS

This work has concentrated on ascertaining the feasibility of hardening AI and steelby the electrochemical surface modification process of metaIliding in a molten salt elec-


24/26

SURFACE MODIFICATION 407SC3I92:l0.5,--------,--------,----------,

0.1.01

liF/KF3.5 MOLE% KBF4

-2. 5 ' - - . 1 - . . . 1 - . ----J0.001

wooa:tiw....wZo'":c(U..>

(f l

o>

CURRENT, AFIG. 15 Current vs potential behavior for steel at 744C and Al at 560C in a LiF/KFmelt containing 3.5 mole% BF"-.

2.00..... I- -Al 1100E l iF/KF + 3.5 mole Z KBF4u 1.60 I- -E 560C.c I- -Cc 1.20 I- --N

0.80 I- -... ....N I- -ell 0.40 I- -0...l I- -

0 ,-2 0 2 4 6Log w (w in rad/s)

FIG. 16 Log impedance modulus vs l o frequency for a steel electrode in a LiF/KFmelt containing 3.5 mole% BF,


25/26

408 KENDIG ET AL.

trolyte. Intermetallic AI-Ce compounds of CeAI, and CeAI 3 were found to form at560C in LiCI/KCl electrolytes at 350-450 mV more positive than the potential for thereduction of Ceo The resulting film is 20-50 t imes harder than AI, but shows poor adhesion to the Al surface. The formation mechanism most likely entails the reductive precipitation of the material from the melt. Solid solution does not form under the conditions due to low solubility and diffusion of Ce into AI. A possible means for forming ahard Ce on AI might result from hot-dipping of Al into the 580C, 95% Ce-5% Al eutect ic . Fur ther investigation into this 'possibility is recommended. While a poorly adherentintermetallic film forms on AI, an adherent diffusion layer of Ce readi ly forms on Pt .

Titanlde compounds formed on steel from a LiF/KF melt I mole% in TiF 3 melt, butdid not form on Al because AI is much more active than Ti in the fluoride melt. Steelcoated with Ti compounds at 744C exhibited greatly reduced corrosion in stagnantroom-temperature 0.5 M NaCI compared to 4340 steel in this medium. The titanidedsteel surface was 30% harder than the substrate.The ease of formation and hardness of the borided steel r ecommends it as a v iableprocess for protecting critical structural surfaces from degradation due to wear. On the

other hand, Al in the boron-containing electrolyte corrodes too rapidly to allow successful boriding to take place.Metalliding of the less electropositive metals such as Fe and Pt can readily be performed electrochemically in molten salt electrolytes. It is d if fi cu lt to form metallideson the highly electropositive metals such as Al due to its activity in the molten salt.For all sys tems considered, the relatively high solution resistance of the moltensalts relative to current density makes accurate potentiostatic control of the metallidingprocess difficult. In performing controlled potential metalliding, the ohmic drop must beconsidered, as it is very significant.

ACKNOWLEDGEMENTThis work was supported by the Army Research Office under Contract No.DAAG29-84-C-0006. The authors wish to acknowledge the work of Darie Wong whohelped in setting up the molten salt cell.

REFERENCESJ. Arens, C. , Metals and Alloys 22, 749 (1945).2. Blitz, W. and Priper, H., Z. Anorg, AUg. Chemie ill, 13 (1924).3. Cook, N.C., Scientific American, 38 (1969).4. Cook, N. and Hayes, W.J. , Ger . Offen. 1,933,010; Chern. Abstr , 72, 74037A (1970).5. Danzig, l.Fc, Dempsey, R.M. and Conti, A.B., Corrosion 27, 55 (1971).6. Fied le r, H. and Sieraski , R., Met. Progr, 99 (2),101 (1970).7. Kendig, M., Allen, A. and Mansfeld, F., J.Electrochem. Soc. i l l (4),935 (1984).8. Kendig, M., Allen, A.T. and Mansfeld, F., "Application of Impedance Spectroscopy to'the Evaluation of Corrosion Protection by Impedance Spectroscopy, in Electrochemica l Techniques for Corrosion Engineering, R. Baboian, ed., NACE, 19869. Kendig, M. and Mansfeld, F., Corrosion 39, 466 (1983).


26/26


10. Mansfeld, F., "The Polarization Resistance Technique for Measuring Corrosion Cur-rents," Chapter 3, Advances in Corrosion Science and Technology," M. Fontana andR. Staehle, eds., Vol. 6, Plenum Press, NY (1970).

II. Mansfeld, F., Corrosion 37, 301 (1981).12. Menezes, S., Kendig, M., Jeanjaquet, S. and Raleigh, D., J. Electrochern, Soc. 134(J 2), 2997 (1988).13. Nowoty, H., Zertschr, Metallkunde 34, 22 (1942).14. Sethi, R.S., J. Appl. Electrochem. 2, 411 (J 979).15. Withers, J .E., Perry, J.E. and Fosnocht, B.A., Techn, of Metals Research, Vol. 7,R.F. Bunshah, ed., 1972.

Metal Lid Ing

Documents

Transcript of Metal Lid Ing