. ~. a r S r 3 f J F F4!2%jJ%

Hydrogen in Electrodeposits: mm $C &W+

Of Decisive Importance, But Much Neglected By Prof. Dr. Christoph J. Raub, 1992 AESF Scientific Achievement Award Honoree

Because water is the base of most electrolytes the electroplater must handle, hydrogen's influence on the properties of various layers in the electrodeposition process is very important. How and why this is true will be discussed in this, an edited version of the 34th William Blum Lecture, presented by the author at the Opening Session of SUWFIND '93-Anaheim.

he e1ement"hydrogen"was first iden- T tified in 1766, by English scientist H. Cavendish. Two years later, he also dis- covered that this newly identified gas, together with oxygen, formed water. The name, in fact, comes from the Greek words for 'hate? and "forming.'"

Earlier, Cavendishs fellow countryman J. Priestley had observed that water changes its properties if it comes in con- tact with electrical current. In 1781, Laurent Lavoisier and Simon Laplace continued Priestley's work, corroborat- ing in 1782 with A. Volta, who had just invented his famouscell. Their investiga- tionscame to aprematureend, however, in 1794, when Lavoisier was murdered near the end of the French Revolution- anearlycaseof politics havinga negative impact on science!

piblishedin 1840.

30

Among the many names also con- nected with electrolysis of water, W. Cruikshaw, W. Nicholson and A. Carlisle-all early 19th-century English physicians-are worthy of mention. A Germanchemist, V. Gruner, andaschooi teacher, K.W. Bockmann, also became interested in thisstudy. it was Bockmann who coinedthe expressions"oxygen wire" (anode) and "hydrogen wire" (cathode), on the latter of which metals deposited.'

The first one to realize that electrode- posited metals may contain fairly high amountsofgases(i.e., hydrogen) wasT. Graham, a professor at London Univer- sity and, later, Director ofthe Royal Mint, co-founder of the Chemical Society, and author of one of theearliest textbooks on chemistry-a German translation of which had been published by F.J. Otto in 1640.2 (See Fig. 1.)

The inventor of electroforming, Rus- sian-German M.H. Jacobi, studied the workofGraham,andin 1872 recognized that electrodeposited iron contained "volatile"compounds.3 He had Observed, for instance, that electrolytic iron-de- pending on deposition conditions-has adensityof7.675g/cm3, which increases in annealing to 7.81 1, but is lower than the7.87ofcast, pureiron. Hefoundthat annealing reduced hardness and im- proved brittleness. So hegot his friend,

1992-93 AESF President Richard 0. Watson, CEF, presenb Ihe Scientific Achieve- men1 Award to Prof Or. Christoph H. Raub, during SUR/ FiW '93-Anaheim.

R.E. Lenz, to study these effects in greater detail.

Lenz, a professor from St. Petersburg, found that electrodeposited iron contains appreciable amounts of gases, such as nitrogen, carbon dioxide, and particularly hydrogen, making it hard and brittle.' In annealing, the gases disappeared and the material got softer. He also noted that the side of the deposit that faced the anode contained less hydrogen and was softer than the opposite side. Lenz pub- lished an article on his findings, and a year later, Jacobi published one about Lenz's results in the Annals ofthe Acad- emy of St. Petersburg. 5,6

There are some very interesting side notes on the ObSeNatiOnS that came out of St. Petersburg:

* By controlling "gas" concentration and maintaining proper deposition param- eters, the Treasury Printing Press in St. Petersburg was able to prepare long- lasting electrotypes of iron or iron com- posites, as well as objects of art in iron

tury-more than 100 years ago!

~

during the last quarter of the 19th cen- .~

An example of the advantageous effect of electroplated iron is that cast electro- types normally lasted for about 50,000 prints, while a combination of eiectro-

Plating and Surface Finishing

Call for Papers SUR/J?INm '94 International-Indianapolis Abstract Deadline: November 1

SUWN"

Topics for Consideration: * AESF Research * Finishing of Printed Wiring Boards * Management 9 Analytical Methods - Electroforming - Electroless * Brush Plating * Electronics * Aqueous and Non-Aqueous Cleaning

Vapor Deposition Processes Plating lor Bonding &Assembly

* Finishing with Special Propenies * Amorphous Films * Fabrication of Suoerconductor Films

- tight Metals

- Organic Finishing - Quality& Statistical Methods - Decorative Finishing - Alloy Deposition * Chemical Pretreatment * Research * Puise Plating * Micmelectronics MetallizationTechniques

Material Characterization * New Finishing Processes * Diamond Films * Technoloav An

Environmental Control

_ I - DI) P r e s s e j lor Derorat . C AFP ~310"s * Aro3 c O I o e F 1"s 0" A .m *.m - S.*aceTocnrooa, 13rEecYonc AD^ calom * Coseo S,jtcmsaio Rec,c "c - S.1ace Treaine-'i?r Corrcrcn Prnecton - XRF Ap3.ca:ons :\11111. 111;

* Conpmza ton tor S.laceFnsnn$ - S.rsce 1Jo.n. Teci-ooq, hlS ? 1 + ) 1 Slal s CB Process Conro n S.nic: F n sn n2 - A ,:ran P3i.e P an1 6 A rirane F n sn "g

Dc3ort Cna~zclerstcr

B,,\ Mh * P a t n g C ~ r ~ s l I ) St reEi iect jc l - hfa Ap3 c s ons l l ~ r r . > h u r ~ . PI\ 171115 * I..\\ 717 S4I -3 l t3h

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SUWFIY' '94 Technical Conference Abstract Forin

Yes' 1'0 i<e 10 S a m I a paper lo1 SLR F h' 94 lar,ng p a c e in lnoianapo s notana. J-ne 20-23 1994. (Seno lo Dr. -)Jan We1 A m p nc MS 21-01. Box 3608. na r r i so l rg , P A 17105,

Sess on - P a p c r T t e .~ Adlhor-P ease SI all a m o r s (dse a separale sneel I necessary, no ca le w h c h a- lhor UI PrCSCnl m e paper

. ..

-

~

Address Telephone F A X

~I have enclosed my biography. Abstract-Type your abstract in the space below or on a separate sheet, 75.100 words, single-spaced.

A~thoiOfterSlheaboveabstractingoodfaith. anditthepaperisaccepted,shallensuie that a proper technical text is Submitted by March 1, 1994, in time for publication in the Proceedings.

Author Cellifies that this IS an abstract of original and unpublished work and is not pending publication elsewhere. Author understands that the acceptance 01 the abstract wiii be based on itstechnical content and adherencetothe deadline for submission, and that it is the author's respansibiiity to ensure that the subject is suitable for presentation a1 an open meeting.

Author fullher agrees that the paper WIIi be personally author-presented at the

conference, and that heishe is not under any restriction by law, contract. or othenvise. from making the presentation at an open meeting.

Author grants Copyrightto the American Electroplaters and Surface Finishers Society, as well 8s permission lo reproduce the above abstract.

Author's Signature SUFIIFIN* is a regisleredtrademark ofthe American Electraplaten and Sulface Flnishen Soclaty, 1°C.

. *. formed copper base and hard iron top layer, heat treated, can hold up for as many as seven million impressions.

Whatbeganinthelate 1970scontinues today: The task of reducing the weight of automobiles to save energy, which increased interest in lighter, aluminum engines. To maintain excellent wear between, cylinders are cast in an AI-Si alloy and later etched to generate pro- truding silicon particles. Pistons are coated on top with a layer of hard iron, havingabout thesame hardness (more than 800 HV) as hard chromium? Hard iron, properly deposited with ad- justed '"gas and impurity" content, whether heat treated or not, might be- come an environmentally safe substi- tute for hard chromium-all the more because the former's properties can be patterned in a wide range, to specific applications. The Russian observation of hard iron is a prime example of how accurate ob- servation and explanation of an unex- pected effect, which looked purelyaca- demic in the beginning, can lead to a tremendous improvement on an exist- ing product, as well as to economic advantages. Iron plating brought the Treasuty Printing Press in St. Peters- burg such fame that large companies from outside Russia sent their copper electrotypes there because they could not get hard iron coatings of such qual- ity anywhere else!

Theoretical Background Hydrogen Deposition Water is the base of most electrolytesthe electroplater must handle, so hydrogen deposition is critical. Some basics will be discussed, before attention is shifted to codeposition of hydrogen and its influ- ence on the properties of various layers. The influence of hydrogen on the base material (/.e., on steel) will not beconsid- ered, because this is quite a problem in itself-in many casesarising in such pre- treatment steps as electrolytic cleaning.

In water, the hydrogen ion is present in the hydrated form, with one water mol- ecule bound to it, making it, actually, H,O'. For its discharge, it must be dehy- drated in a preliminary step. This takes place intheoutersectionofthe Helmholtz double layer and can occur in several steps. The migration of the hydrated ion to its discharge is determined by at least three processes: The transition of the ion into the outer diffusion range of the Helmholtz layer, where migration is de- termined to be the concentration gradi- ent, convection and migration in the elec-

September 1993

tric field. Field strength is still rather loyl and cannot remove the water molecuk from the ion, but it is strong enough tc generate an orientation of the water di, poles, according to Gerischer? The hy drogen ion then migrates through the diffusion range of the double layer to its iixedsectionclosesttothecathode where iield strength is relatively high ( lo7 Vlcm: and H,O molecules are torn away from :he H*ion.The hydrogenisthenadsorbec at the metallic cathode and discharged This is, indeed, a simplified explanation, )ut it describes the essential steps.

Ofthethree possible mechanisms thai nay determine movements of hydratec ons, convection can be virtually elimi. iated by working at zero gravity, in an Iuter-space experiment. This was done about 15 years ago, and clearly revealed he differences in the behavior of hydro- gen bubbles, especially in their growth ind adhesion to various cathodes (e.g., gold and platinum) (Fig. 2). io

The overall equation for the discharge )f hydrogenatthecathodecan be written:

2W t 2e- = H, for acidic electrolytes

)r

2H,O t 2e = 20H- t H, for alkaline solutions

n this case, hydrogen is deposited di- ectly from water.

.here are, however, quite a number of

.teps until hydrogen gas finally fori"

. Transport of the hydrogen ions into the double layer

I. Discharge and change from a solva- tion bond with water, to an adsorption bond with the cathode metal

H'solv. t e- = H ads,

T. Erdey-Gruz and M. Vollmer, Vollmer saction)

Here, for the first time, we encounter H, atoms adsorbed.

The H ads. atoms now will react at the cathodic surface with other H ads. or H+ ads. atoms and form H, ads.

H ads. t H ads. = H, ads. (Tafel) H ads. t H ads. t e- = H, ads. (Heyrovsky)

a) Desorption of the adsorbed H, from the cathode and transition intothe electrolyte

b) Removal of H, by diffusion and disso- lution on the electrolyte or as hydrogen bubbles

Details of steps a & b can be seen nicely in the above-mentioned zero-gravity ex- periment.Atfirst, hydrogen bubbles grow whilesticking tothecathodicsurfaceand pushingeach otheraround as they grow. Another pari of the hydrogen dissolves in a zone close to the cathode, while the rest of the electrolyte remains rather poor in dissolved hydrogen, because of the missing convection. Once bubbles reach acertainsize,theypusheachotheraway from the cathode, into the electrolyte, and after drifting through a hydrogen- gas-enriched zone, they get smaller again, because of the dissolution in the remaining electrolyte, which is low in hydrogen.'0

This adsorption of hydrogen on vari- >us metals is technically important in a Jubble chamber for studying hydrody- iamic flow patterns around profiles. In a :hamberfilled with electrolytes, asystem 3f cathodically polarized, suspended fine 4u or Pt wires is moved through the ?lectrolyte, leaving rows of tiny gas Iubbles in solution for a while. A three- limensional line pattern is formed, ,hrough which the profiles to be tested ire pulled. If the system is illuminated rom the side, the movement of thestrings I f bubbles around the profiles can be Matched. For this system to function Iroperly, a uniform and equally sized generation of the bubbles, along with an Ippropriate detachment from the cath- )de, is essential.

The problem of the ad-hydrogen at- )ms or molecules is a serious one, for iearly every electrochemical process- romfuelcellto battery, andfrom electro- iating of zinc and chromiumto the large- icale electrolytic decomposition of water or energy production or storage.

The reversible equilibrium potential of iydrogen deposition is only observed vith aspecial Pt -electrode, covered with 't -black. In nearly every other case, it is iemmed-or "polarized-by various ef- ects, summarized by the term "over-

31

_____

~

Eh

voltage." This hydrogen over-voltage depends on a number of factors, such as temperature, kind of cathodic metal, struc- ture and composition of the cathode's surface, hydrogen pressure and compo- sition of the electrolyte, etc.

In water, with step number one, the concentration polarization is negligible. Of the rate-determining steps, the final one is the slowest, and rate depends on the deposition conditions.

It should be mentioned that the over- voltage AE varies with current density" i," according to the TAFEL-equation

AE = a - log i

Simultaneous Discharge Of Hydrogen and Metals In two detailed papers, A. Knodler stud- ied the simultaneous discharge of hydro- gen with metals and the penetration of hydrogen through electrodeposits."' i2

In electroplating, codeposition of hy- drogen with metals is a very important effect, and it can be positive, as well as negative for the electroplater. The posi- tive aspect is the very effective stirring action by gas evolution at an electrode. N. Ibl did calculations showing that this gasstirring isthe mostactiveagitation we have in electroplating-most actively re- ducing thickness of the diffusion layer. At the same time, hydrogen bubbles may shield the cathode, preventing deposi- tion; or also becoming incorporated, cre- ating defects in the layers or even form- ing alloys, as with palladium. Hydrogen, too, may penetrate the electroplate, es- pecially when it is still porous and defec- tive, and react with the base.""2

When electroplating from aqueous electrolytes, always keep codeposition of hydrogen in mind, because it can be polarized strongly, but depolarized as well, by constituents of practical electrolytes.

32

From a theoretical point of view codeposition of hydrogen can be treatec like alloy deposition, considering meta hydrogen systems as alloys, too.

For alloy deposition, imagine thret possible cases, illustrated in Fig. 3.'

In case "a," the potential-partial cur rent density curves of both elements t( be discharged are set quite apart. A first, the most noble partner,"A," is dis charged till, at El, its limiting curren density is observed. The potential ther becomes less noble and the depositior of the less-noble partner '"6," starts- this, in many cases, being hydrogen The deposition of partner " A will con tinue the same-absolutely-but dimin ish, relative to the total amount. In c a s "b," the two curves are closer. Deposi tion of the less-noble metal begins a potentials that are somewhat close to gether, with nearly the same partia current densities. Depending on thc slope of the potential current densit! curves, relative deposition of both con stituents may differ at low and higt current densities-as is observed, ir reality. If both curvesintersectwitheact other, the two partial current densitie! are equal. Before and after reachin! the intersection-at higher and at lowe current densities-the constituents be ing most-strongly discharged ma! change. This is the often-observed cur rent density effect in alloy deposition influencing such parameters as curren efficiency and composition.

This discussion presumes "ideal con ditions," but reality is often different. 11 most cases, for example, current poten tial curves of separate metals modif! each other during alloy deposition. An other strong influence on the shape 0 current-potential curves is shown by ad ditives, temperature and chemica changes in the electrolyte caused b! ongoing electrolysis.

For hydrogen deposition, in general, concentration polarization is of little di- rect influence. In a real electrolyte, how- ever, because of concentration or pH changes, it can alter during electrolysis and hydrogen-in the same electrolyte- will be codeposited, more or less. It is even more critical, if we consider the conditions at the various spots of the cathode-especially on a microscopic scale, with the surface of the cathode geometrically, chemically and structur- ally different, to name a few factors of influence. Theremay besitesofhighand low over-potential for hydrogen dis- charge; there may be electron tunneling effects through naturally existing oxide layers or through adsorbed monomo- lecular or even atomic films, all altering deposition conditions for the partners- especially for hydrogen.

Copper from acidic baths and silver from cyanide electrolytes normally are deposited, at the usual current density, with 100 percent current efficiency. On the other hand, if the limiting current density is exceeded, hydrogen can be enclosed in the layers.

A typical case involves silver flash elec- trolytes? The potential current density curves are analogous to Fig. 3b, wherein '"8" represents hydrogen. The curve with " A being hydrogencannotoccurinaque- ous electrolytes, because there is no limiting current for hydrogen in aqueous electrolytes, and because there isalways ample water available for discharge. Fig- ure 3c, however, represents a very com- mon case. If '"A is the metal, and "B" the hydrogen, the current efficiency for the metal decreases with total current. If " A is hydrogenand"Bthemeta1, thecurrent efficiency for metal deposition increases with total current, as in the case of nickel.

E. Raub noted thedecisive importance of over-voltage in his book,' and cites the example of zinc deposition: In about one molar (300 g/L) solution of zinc-sulfate, the equilibrium potential is near -0.76 V. At pH = 3, hydrogen at -0.177 Vis much nobler, sn it should bedischarged prefer- entially. Under these conditions, how- ever, zincdeposits at almost 100 percent current efficiency, because hydrogen has atremendous over-voltage on zinc, which shifts its potential to less-noble values. On the other hand, zinc cannot be plated on platinum from such an electrolyte be- cause hydrogen over-voltage on plati- num is rather low. This over-voltage of hydrogen on certain metals is the expla- nation as to why some metals can be electroplated, even if they are less noble than hydrogen-far from their equilib- rium potentials. D.P. Smith states that

Plating and Surface Finishing

-

~

~

~

hydrogen bubbles

hydrogen over-voltage increases from Pd, Pt, Fe, Au, Ag, Ni, Cu, Ta, Sn to Pb and Hg. Metals with the highest hydro- gen over-voltage, such as lead and mer- cury, do not dissolve hydrogen at all.

The influence of hydrogen in practical electrodeposition is often underesti- mated." Hydrogen not only influences the deposition mechanism and proper- ties directly, by incorporation in the metal and altering mechanical properties, but it has an indirect influence, as well. Hydro- gen bubbles formed, for example, can cause a superb stirring action, thereby narrowing down the thickness of the dif- fusion layer and, by this, increasing par- tial current density of metal deposition- generally a positive effect-but this, too, may change pH of the cathodic film and facilitate hydrolysis of metal salts that may be adsorbed at the cathode and incorporated into the growing layer. This isseen clearly in transition-metal electro- lytes, such as those of Pd, Ru and Rh, and Ni. Here, it must be remembered that hydrolysis constants and products are dependent on the kind of anions present: They are different, for instance, for Ni- sulfate and -chloride.

Another case of indirect hydrogen evo- lution effects is presented by Fe impuri- ties in nickel plating baths, inwhich iron is onlysoluble in the divalent state. Oxygen changes it into Fe3+, and at fairly low pH values, ferrihydroxide precipitates, which often exists as a colloid. Because it is positively charged, this colloid is trans- ported to the cathode by cathaphoresis, catches hydrogen molecules or gas bubbles on its way, and the two are in- corporatedintoagrowing layer.Thismay explain, at least in part, whya high hydro- gen concentration in a deposit usually goes parallel with other effects (e.g., in- corporated constituents of the bath), gen- erating lattice distortions that also favor hydrogen inclusion. Colloidal phenom- ena may play a role, too, in the advan- tages of using Li or K compounds vs. Na salts. It is wise to quickly oxidize iron in nickel baths, in order to avoid formation of such colloidal iron hydroxide systems.

September 1993

Adsorbed hydrogen atoms are very reactive and they may reduce other adsorbed molecules or foreign matter included in the substrate. E. Raub dis- cusses the beneficial, though second- ary, effect of hydrogen evolution on zinc plating from cyanide Only after passing the limiting current density, do zincdepositscrystallize in afibrous struc- ture, whichis brightandabletobechemi- cally polished. The explanation is that, becauseof hydrogen evolution, thecath- ode layer gets a lower pH value, and zinc is discharged favorably via the Zn(OH)x complex.

The final case of hydrogen is pre- sented by post-deposition reactions of incorporated water molecules4r pro- tons-reacting with the metal matrix, form- ing H atoms, which recombine to gas- eous hydrogen molecules and exert enough pressure to blow the metal lattice apart. This is discussed further in the section "Hydrogen in Zinc," below.

Under certain circumstances, a similar reaction can be observed in anodized aluminum, where agglomerates of me- tallic aluminum may react with OH- or H+-after formation of the anodized layer-generating hydrogen bubbles.

Last, but not least, it must be remem- bered that hydrogen can form stable or metastable hydrides by electrolysis (e.g., NiH, CrH or Pd-H), which, when present in smaller amounts, determine proper- ties. Some, but unfortunately not all of which can be produced easily on a com- mercial scale in this manner.

There is only one other alternative method of putting hydrogen into such metallic structures, with similar effects, and that is forcing in hydrogen ions (or protons) by ion implantation. This is the reason the author likes to call the dis- cussed electrochemical hydrogen-load- ing procedure 'The poor man's ion im- plantation method." An electrolysis cell can be made with a spent car battery and a used soft drink cup-costing a few dollars. The cheapest ion-implantation unit will be a little more expensive-in the few-million-dollar range!

Nowwewill lookatafewtypicalcases, most coming from the work of collabora- tors at the Forschungsinstitut fur Edelmetalle und Metallchemie (FEM), but relying on past and ongoing work of other researchers the world over.

Hydrogen in Zinc Even metals with nearly zero solubility of hydrogen at equilibrium conditions may contain hydrogen, though in smaller amounts, occluded mechanically. It was once believed that this hydrogen was

€

0 1w 2w 3w 400 5w Time to end of deposilan, min

ig. 5-Stress in nickel layers during electrodepo- tionandcathodi~polanti~" in sulfuriccacidtO.1 'L thiourea, with and Without Pd-ions.

icorporated in molecular form only in lese metals,' but Y. Okinaka and H.K. taschil showed that for electroless cop- er, hydrogen molecules can form by iffusion of hydrogen atoms and there is o reason why this should be different ,ith other metals.'6 They also demon- trated that molecular hydrogen can re- lain for quite some time in the layer and ot diffuse at room temperature. The tory is different for metals with high iffusion rates for hydrogen, such as pal- ldium and other platinum metal group iements (PMGs). Depending on the plating conditions,

nc and especially bright zinc may con- iin additives that are adsorbed at the rowing cathode and are incorporated. Iften, these brighteners are a combina- m of organics-sometimes polyalco- AS-with a fairly high molecular weight I d otherproprietarychemicals. If, under ?rtain conditions, too much of these jditions is incorporated-particularly if ley have adsorbed water molecules- ater molecules get into the layer. This corporation may be selective, depend- g on substrate defects or inclusions i d on local current density variations, :c. During storage, especiallyatelevated mperatures, these water molecules- . zinc hydroxide-can react according ,the equations:

Zn t H,O =ZnO t 2 H or H,

Zn t Zn(OH), = 2 ZnO t 2H or H,

his hydrogen cannot escape by diffu- on through the lattice, but only by diffu- on along grain boundaries or along ?avily distorted areas. It forms from oms and molecules that become gas JbbleS at sites favored for such a re- "ination-inclusions, defects, etc., __ 'r example. After enough pressure has Jilt up in these bubbles, they can

concentrate near the substrate (if they cannot penetrate into it); accumulate near a final top coat (wet iacquer or powder coating, because

33

~

u!w ' a u l Ob2 002 091 021 08 Ob 0

0

05 -5 _. m m g.

001 .=

c 051 0

5 0

s

ooz

01 I I I 100 4w 700 1000

Hardness, mHV

Fig. 8-Wear of iron deposits in dependence or their hardness.

It is not quite clear what is more domi- nant on the mechanical properties of nickel-the hydrogen in whatever form, or the incorporated, defect-generating substances, or a combination of both. Thesame is true forinternalstress, which alters during out-diffusion of hydrogen (Fig. 5).'6-'8

In nickel alloys, composition and kind of alloying partner determine the hydro- gen uptake, too. This is typically shown by nickel-palladium and nickel-zinc al- loy~.'.'~ln the latter, hydrogenconcentra- tion increases, in proportiontotheamount of zinc in solid solution. In alloys richer in zinc, various nickel-zinc alloys form, and the hydrogen content drops very low, in the range of the hexagonal zinc-nickel phase-around 50 percent zinc. Bright nickel shows a much higher penetration than matte nickel.11,12

Hydrogen in Chromium In a paper on stress and stress relief in chromium, D. Gabe discussed the up- take and release of hydrogen in chro- mium and its relationship to the proper- ties of the layer.zD In 1963, he summa- rized: "There has hitherto been no direct experimental evidence of this hydride ..." (chromium hydride). He arrived at the conclusion that prompted this statement through discussions on the diffusion of hydrogen through b.c.c. chromium and h.c.p. nickel, promoting decomposition ofapossibleCrH. Healsofoundastrong effect of arsenic on tensile stress, about which he stated that "the influence of arsenic further supports this idea ..." (of decomposition of an h.c.p. CrH).

At almost the same time, however, A. Knodler, of FEM, was able to synthesize, quantitatively, X-ray clean h.c.p. CrH by electrolysis. He determined its properties and studied the transformation to b.c.c. Cr.21,22ThisCrH isstable in air; unstable, if powdered and in vacuum (Fig. 7). He supported Gabe's assumptions of the

September 1993

high mobility of hydrogen in CrH experi, mentally and found that the formation o the hydride is an exothermal process while the uptake of hydrogen by b.c.c chromium is endothermic. If h.c.p. Cr t transforms into b.c.c. chromium, a VOI. ume contraction is found. As earliei guessed, this may explain the crack for. mation in chromium layers.

Chromium produced by decomposi. tion of CrH is chemically very active anc oxidizes at normal roomconditions. Solid- state properties were also s t ~ d i e d . ~ ~ , ~ ~

Pieces of CrH can easily be set afire with a match and burn with the bluish flame characteristic of hydrogen.

It is now fairly certain that-directly )r indirectly-the hydrogen (or CrH) :ontent of hard chromium layers is im- lortant to their properties. Because Cr- ixy-hydrates are also incorporated to a iigh extent, it is again difficult to estab- ish theactua1"hydrogen effect."In hard :hromium, we do have hydrogen )resent in solid solution, in the form 01 I.C.C. Cr, as well asvarying amounts of :rH, which is highly dependent on the :urrent wave form.

By low-temperature electrolysis (at 30 "C), we were able to synthesize X- ay amorphous Cr-H alloys, highly hy- Irogen-supersaturated (with oxygen im- lurities, too), whichdecomposeat room emperature, via CrH and P-W-type :r-0 to b.c.c. chromium. The hydrogen eleased is quite similar in behavior to ~al lad ium.~~

An interesting test-well known to plat- !%-is to drop a hard chromium sample ito hot oil and observe the hydrogen levelopment, originating from diffusion Nf hydrogen and decomposition of CrH- iuessing theamountofhydrogen present nd using this for production control.

lydrogen in Manganese )espite the fact that hydrogen in man- anese can present problems in indus- .ial electrowinning of manganese be- ause of its ignition, few details are ublished about the mechanism and i e possible existence of an Mn- hy- ride. Immediately after deposition, iereisavaluegivenof200ppm hydro- en, which was mostly removed by eat treatment at 500 0C.26 Equilibrium olubilities were determined by Sieverts nd Moritz in 1938, with electrolytically ydrogen-charged manganese. They tated that, after annealing at 600 "C, bout 8 ppm is in solution, but during Doling to lower temperatures, some is ?absorbed (to 35 p ~ m ) . ~ ' There is evidence that there are tre-

iendous variations in the H- content.

0.51 2 3 4 5 Current densily, M d M

I ~~~

Fig. %Dependence Of Ni- and H-concentration or Current density for Pd-Nideposition (30°C. pH 8.51

depending on the electrolysis conditions, as well as on the determination method. T. Agladze developed a process to re- place hard chromium with "hard manga- nese.'' He can generate manganese lay- ers with a hardness up to 1000 kgicm2.

All this shows how little information on electrodeposits of the Mn-H system is published, because the work was aimed almost exclusively in the direction of in- dustrial electrowinning. This prompted a collaboration on manganese electrodepo- sition among thoseat FEMand colleagues from Eastern Europe.

Hydrogen in Iron It is most amazing to realize that even if the electrolytically prepared Fe-H alloys are the earliest examples for hydrogen codeposition, detailed information on them is of questionable value. As we have seen by work from Jacobi, Raub and Wullhorst, hydrogen content goes parallel with the inclusion of other con- stituents, showing the well-known pic- tureof increasing with rising latticedistor- tion, higher hardness, but superb wear resistance, as well (Fig. 8).

The FEM is again taking up its earlier research on electrolflic iron and iron alloys, largely because iron is one of the safest elements, environmentally speaking.

Hydrogen in Palladium The electroplating of palladium has at- tracted much attention, since it showed great potential in the electronic industry.

zodeposition with hydrogen is thoroughly investigated. In a series of nearly 20 3apers, to date, the FEM's team of H.D. Hedrich, M. Butz, F. Friedrich and D. Ualz has been studying the influence of iydrogen and electrolytically prepared 3alladium and palladiumalloylayers. Re- :ently M. Monev, from Sofia; M.E. 3aumgartner; M. Kittel and A. Juzikis 7ave joined this group in its efforts. Fur-

35

~

This explains why palladium and its .~ ~~

~

ther work was also done in the UK, the US., and in Eastern Europe by such distinguished researchers as Y. Okinaka and J. Abys, both with AT&T Bell Laboratories.

A summary of the work, covering "the greatyearsof palladium,"upto 1982was published by the authorthatsameyear.3°

Metallic palladium is one of the most '"reactive" metals with hydrogen, and is used in its pure form, or its alloys, exten- sively for hydrogen purification in diffu- sion and fuel cells. Even though the first palladium layers were prepared from chloride electrolytes as early as the late 1800s, little attention was paid to the codeposition of hydrogen.

The main features of electrolytically prepared palladium-hydrogen layers are explained here. Palladium can dissolve fairly large amounts of hydrogen as a solid solution-depending on thedeposi- tion conditions-in itsf.c.c. - Pd lattice. At normal conditions, the H:Pd weight ratio may go upto about0.03. At higher hydro- gen concentrations, the p-Pd-H phase, tetragonally distorted, forms at an H:Pd ratio of more than 0.57. Its lattice con- stant is higher by about 3.8 percent, as compared to the aPd-H lattice.

The uptake of hydrogen-as is known from other metals-generatesstress, but even morecomesfrom the formation and later disintegration of this P-phase, at room temperature. During storage, hy- drogen diffuses out from areas of high H:Pd ratio, and the decomposition of the p-phase and accompanying volume change cause stresses that are released by cracks in the layers, sometimes even days after deposition.

This '"hydrogen problem" was well known in the beginning of the electronic use and one effort to cope with it involved heat-treating it as deposited palladium, at 500 "C. This only shifted the problem from the user backtothe plater, however. It wassolved through the development of processes that produce practically hy- drogen-free deposits. It becomes a bit easier, too, because Pd-alloys, in gen- eral, show less tendency to hydrogen uptake than pure hydrogen (Fig. 9). To- day, the"bad names"of nickel andcobalt may force the industry to go back to the laboratory and devise new palladium al- loy system^.^'

Organic additives increase hydrogen concentration and have an influence on the texture of the layer (Fig. 10). The hydrogen-to-palladium ratio may reach 0.8 in the as-deposited state-a high supersaturation. If placed in a liquid at room temperature, the hydrogen bubbles evolvingfromthelayercan beseen. Low-

Voltage during 350 Pd deposition 400r-----

Alloy-free Pd (NHJ, CI, electrolvte, pH 8.5 . . .

+-. Pd (NHJCI, electrolyte, 0.005 m nicotinic acid

i2 200 pH 8.5

Voltage following r,- . Pddeposition d / I

2' n

7 7

I' i

I' I

0 2 4 6 8 10 1 2 0 10 2 0 3 0 40 50 60 Deposit thickness, pm Time after deposition, min

7. lo-Stress of Pd- layers from a chloride . electroiyte, with an addition of nicotinic acid during andaft6 wmition.

temperature-deposition experimentsfrom a highly acidic Pd-electrolyte (at about - 20°C) areveryintere~ting.2~ Underthese conditions, extremely distorted and hy- drogen-supersaturated powders are pro- duced. If these powders, as deposited, are placed in liquid nitrogen, they de- velop a dense stream of hydrogen bubbles, showing their supersaturation. Layers are X-ray amorphous, and they transform, under heating, to room tem- perature, to the p-phase, and later to the X-phase. In these layers, as estimated from dehydrogenation curves, hydrogen must be bound in various forms, first to the amorphous palladium, then to the p- phase, and finally, to the X-phase. If the purely physically absorbed hydrogen and the H to H, recombination are taken into consideration as well, there are at least fivestatesof hydrogen binding involved- some of them even seen in dehydroge- nation experiments.

This method of producing unstable amorphous Pd-H intermetallics may be considered an example only of the possi- bilitiesofelectrolyticsynthesis methods- not so much a means for generating a '"cold fusion material."

M. Monev and collaborators demon- strated interesting effects in nickel-rich palladium-nickel alloys and the influence of trace amounts of palladium and others on the stress of electrodeposited ni~kel.'"'~

The extensive research in palladium electrodeposition has yielded much in- sight into the ways of incorporation of hydrogen into electrodeposited metals. This is transferable, at least in part, to other systems, too. It was of decisive importance, in that the hydrogen prob-

lem and cracking of layers is no longer a dilemma in modern palladium plating.

Hydrogen in Gold ~~

Under equilibrium conditions, gold does not dissolve any hydrogen in its lattice. Electroplated hard gold layers, as they are used for electric connectors, are al- loys of gold with iron, cobalt and nickel. They usually contain between 0.3 and a few percent of these, depending on the metal and on the deposition parameters. In addition, alloys can possess organics, with up to more than 1000 ppm C; and hydrogen, which can go up to 1000 ppm.

Inclusion is enhanced ever further by the fact that cyanide helps the recombi- nation of H atomsto hydrogenmolecules, promoting diffusion.

What is too-often overlooked, how- ever, in many explanations-most of which refer to the Au-Co system-is that the amount of incorporated foreign mat- ter, including hydrogen, isnearlythesame for all three elements of the iron group- a fact, complicating chemical-electro- chemical explanations much more. It bears mentioning that Co is present in two states in gold: A supersaturated Au- Co solid solution; and an inorganic com- pound, details of which are unknown, of a K-Au-Co-CN composition, to which hydrogen may be chemically bound. If Co is present in the electrolyte as the Co (IN) ion, itsconcentration in the layer and its foreign-matter content drop to com- paratively low values.

In Au -Fe baths, the incorporation of K- Fe-cyanides and hydrogen is also evident.

This has not yet been proven, experi- mentally, for Ni, but it is most likely. For

Plating and Surface Finishing

.-

~

36

I .

._ E

8

0-0 0 1 10 100

Annealing time, Hr Fig. 1 1-Porous S t N c f m of odd- 0.5% Nidemxit, afferanoealhg for fwo hrat Fig. IZ-Vahtion of hydrogen conlent and ducfilify with annealing lime a1 150

"C (depositplated wifh 10 mb/L K,Ni(CN), at 70 "C).'B - . .

700 "C.

Au-Ni and Au-Fe layers, thealloying con- stituents are, at least in part, in solid solution, which is easier than with cobalt, because gold exhibitsa rather high-equi- librium solid solubility.

All three systems, in the form of elec- troplated alloys, exhibit similar properties of hardness and wear resistance, de- spite theirdifferences in equilibrium chem- istryand metallurgy. It iswell established, too, that hardness and wear behavior cannot be caused by the metals in solid solution, but must be the result of incor- porated foreign matter, including hydrogen.

Toacertainextent, thisisconfirmed by the fact that the so-called "addition-free hard gold (AFHG) of Y. Okinaka, does not contain metallic elements, but shows up to 1000 ppm C (from inclusions of AuCN) and hydrogen. The same is true for deposits from Au (ill) cyanide baths (which may or may not contain cobalt) that reach hardnesses of more than 200 HV on similar grounds.

These hardnesses must be connected with the incorporated foreign matter-be it inorganic cobalt compounds, some or- ganics (tentatively called a "polymer" by G. Munier), or hydrogen. Hydrogen can be removed by heat-treatment or elec- trolysis at higher temperatures-a fact thatiscriticaltolaserplating.0n theother hand, all the layers tend to blister andlor crack, or form voids during heating to highertemperatures-another indication of a build-up in gas pressure.

Forthe Au-Co-H problem, acontribut- ingfactormaybethatcertaincobaltcom- plexes are highly reductive and can even decompose water. But what about the corresponding Ni and Fe complexes? Furthermore, up to now, only complex Co-cyanides are known to adsorb H at- oms. What is identical, however, are the factsthat all deposits discussed are lower in density than would theoretically be

calculated (remember Jacobi and eiec- trodeposited iron); and all three systems have pores, voids, and are full of defects interfering with slip planes.

These pores were first observed in gold-cobalt by Okinaka and Nakahara, and later shown to exist in gold-nickel as well, by Reid, Saxer and Fiuhmann. Okinaka and Nakahara assumed, for the firsttime, thatthesevoids andporeswere full of highly compressed hydrogen gas, formed by recombined hydrogen atoms during electrodeposition-lateralso found to be true of electroless copper.

The extremely pressurized gas gener- ates a wide distortion area around a void, which results in stresses and high hard- ness. The incorporated organics add to such a defect area, and facilitate diffu- sion and recombination of hydrogen in such sites during electrolysis. For metals that do not dissolve hydrogen, such as gold, silverandcopper, theeffect isespe- cially dramatic, because H can only dis- appear rather slowly from those voids.

It is confounding that this model by Okinaka and Nakahara, first developed forgold, was rarelyusedforotherelectro- plated metals, forwhich it works every bit as well. This was confirmed for electro- less copper in 1986, through an experi- ment by Okinaka and Straschil.'E

Hydrogen in Electroless Copper In the early days of copper electrodepo- sition, Knodler ObSeNed a relatively high development of H and stated that below 5bofthickness, layers can be penetrated by hydrogen--especially if the copper is

This may be important be- cause it is known that not only electrolytic deposits may contain '"volatile com- pounds," but electroless layers may, too.

Y. Okinaka and H.K. Straschil studied this carefully for electroless copper, of special importance to electronic produc-

tion. The layers contained up to 200 wt ppm hydrogen, and 100 to 800 ppm carbon and other impurities (0, N, Na) (Fig. 12).IE They suggest that hydrogen gas bubbles are entrapped during electrocrystaliization, and hydrogen at- oms diffuse through the distorted thin copper layer around them, to recombine and build up pressure inside. This pres- sure creates large defect areas around voids that hinder the movement of slip dislocation planes during deformation, or hardens the layer. The addition of ar- senic-astabilizerforatomic hydrogen- made it possible to do a reversible re- charge of an annealed electroless cop- per sample, after it had been annealed "empty of hydrogen." Ductility after an- nealing was 6.6 percent; after recharg- ing, it came down to only 3.8 percent (as deposited, 2.1 percent)-a difference from ion implantation?

In experiments, Y. Okinaka and a col- laborator were able to find clusters of molecular hydrogen, employing a method that uses a calorimetric measurement to determine the slightest amount of heat released during the conversion of ortho- to para-hydrogen, at the temperature of liquid helium.

In the abstract of his paper on the study, Okinaka remarks, "Results show that the hydrogen exists in two different forms: Diffusible, which escapes readily on annealing; and residual, which is not removable by annealing. Diffusible hy- drogen impairs ductility ... the diffusible hydrogen consists essentially of molecu- lar hydrogen, existing in micro-voids at high pressure ... both the pressure effect of molecular hydrogen and the structural effectofvoidsinfluencing theoverall duc- tility. Ductility-promoting additives inhibit both the inclusion of diffusible hydrogen and the formation of voids. Residual hy- drogen, together with other elements,

37

~~

~

~

September 1993

constitutes other compounds (e.g.. EDTA or H,O) occluded from the plating solu- tion ,.. . However, when the incorporation of an additive causes an increase in residual hydrogen and carbon contents, the beneficial effect ofthe additivecan be offset by its influence on grain structure."

This is, in this author's opinion, the perfect explanation of the role of hydro- gen in metal deposits formed electrolyti- cally and by electroless methods. It shows, too, how to influence properties of layers by hydrogen incorporation only: Look for additives in small amounts that promote building-in of small hydrogen bubbles, promote diffusion of H (e.g., by adding traces of Pd) (See work by M. Monev, e l ai.), and/or by distortion of the host lattice by an additive concentration in as small a quantity possible. For electrodeposited copper, A. Knodler confirms the use of a matrix through which hydrogen can only diffuse if the metal is thin and highly distorted. It is well known from processes involvingsteel that such areas may actas "traps" for hydrogen.

In any case, the "poor man% ion im- p1antation"method wasalready known. It would be interesting to ObSeNe real hy- drogen ion-implanted copper for similari- ties-a suggestion to Y. Okinaka.

Acknowledgment I wantto thankthe International Branch of the AESF for nominating me for the Society's prestigious Scientific Achieve- ment Award, and the Awards Commit- teeforgrantingitand inviting metogive this, the William Blum Lecture for 1993.

In a wider sense, I am grateful to my parents for enabling me to study the fascinating field of chemistry, and to my teachersforshowing me how exciting it is. I have to thank my wife for giving up a profession she loved in order to move with me to California.

My gratitude to my admired teacher, and later friend, Professor Bernd T. Matthias, who taught methatscience, if properly applied, is fun; and my indus- try colleagues worldwide who showed me that you should test your ideas and results in production; all my collabora- tors at the FEM, from whatever country they may have come as permanent residents, semi-permanent visitors, or guests, and my friends worldwide, who make life much more colorful. And, es- pecially in connection with precious- metal plating, I thank Dr. Y . Okinaka.

Last but not least, i feel that this award is shared by all who in any way contributed to the change of electro- plating from an alchemistic adventure to a modern, economic and quality- ruled process; and by the organizations and people supportive of our work, through funding. 0

38

References 1. K.W. Bockmann, Annalen der Physik

Halle, 8, 137-162 (1801). 2. Dr. T. Grahams, Lehrbuch der Chemie,

edited by Dr. Fr. J. Otto, printed anc published by FreidrichVieweg undSohn, Braunschweig, 1840.

3. 0.1. Pavlova, Electrodeposition of Met. ais-A Historical Study, edited by S.A. Pogodin, lzdatelktvo Akademii Nauk SSSR Moskva, 1963; Israel Program foi Scientific Translation, Jerusalem, 1968.

4. M.H. Jacobi, Bull. de Academy de Sf. Petersbug, 13, 40 (1868).

5. M.H. Jacobi, Buil. de Academy de St. Petersburg, 14, 252 (1870).

6. R. Lenz, Bull. de /'Academy, 14, Col. 337 (1869); GornijZhurnal, 4, No. 10 (1887).

7. E. Raub and K. Muller, Fundamentals oi MetalDeposition, Elsevier Publishing Co., Amsterdam-London-New York. 1967.

8. A Review on the Work of Russians in fhe Fleldof Electrical Engineering from 1806 to 1900, or ref. 181 of (3), Spb. 1900.

Elektrochem., 57, 604 (1963). IO. W. Wopersnow and Ch. J. Raub, Z.

Flugwiss. Welfraumforschung, 2, 341, Raumfahtt, 1978.

11. A. Knodler, Metalloberflache, 40, 515

12. A. Knodler, Metalloberflache, 41, 29 (1 987).

13. D.P. Smith, HydrogeninMefals,TheUni- versity of Chicago Press, Chicago, IL, 1945.

14. E. Raub, Galvanotechnik, 60,348 (1 969). 15. H.W. Dettnerand J. Elze, Handbuch der

Galvanotechnlk, Band 1 Carl Hanser Verlag, Munchen (1963).

16. Y. Okinaka and H.K. Straschil. J. Elec- frochem SOC., 133, 2608 (1986).

17. M.Monev, M.E. Baumgartner,O.Loebich, Ch.J. Raub, Mefallobefflache, 45, 77

18. M. Monev, M.E. BaumgartnerandCh.J. Raub, J. Electrochem. SOC., 138, L16

19. M. Monev,M.E.Baumgartner,Ch.J. Raub,

20. D.R. Gabe and J.M. West, Trans. Insf.

21. A. Knodler, Metallobeffiache, 17, 162

9. H. Gerischer and H.E. Stambach, Z.

( 1 980).

(1991).

(1991).

Mefallobefflache, 46,269 (1992).

Met. Finish., 40, 197 (1963).

(1963). 22. /bid., 331. 23. H.R. Khan,A. Knodler, Ch.J. Raub, A.C.

Lawson, Mal. Res. Bullelin, 9, 1191 (1974).

?4. M. Hanson, H.R. Khan, A. Knodler and Ch.J. Raub, J. LessCommon Metals, 43, 49 (1975).

25. 0. L0ebich.T. MuramakiandCh.J. Raub, Proc. of fhe Symp. on Elecfrocrys- tallizafion, Hollywood, FL; Ed. R. Weil, Proc. Electrochem. Soc.. 81-86 (incl.) , , (1981)

26. M.C. Carasilla and R.W. Fowler. J. Electrochem SOC. 104, 352 (1957).

!7. A.SievettsandM.Moritz,Z.Phys.Chem., 180, 249 (1938).

28. T.R. Agladze, Tifiis, private communica- tion.

29. E. Raub and B. Wullhorst, Miff. For- schungsges. Blechverarbeifung, 14, 279 (1951).

30. Ch.J. Raub, Plat. Mef. Rev., 26, 158

31. Ongoingworkatthe FEMonelectrodepo- (1982).

sition of Pd-Fe layers.

About the Author Dr. Christoph J. Raub, an AESF member since 1974, is head of the Forschungs- institutfur Edelmetalle und Metallchemie, in Schwabisch Gmund, Germany-apost he has held since succeeding his father.

He passed his Abitur (a qualification exam for university entry) in 1951 ; spent a practical yearworking for W.C. Heraeus Hochvaakuum, in Hanau; and attended the University of Munster from 1952 to 196l.There, hestudiedchemistry phys- ics, archaeology and art history, receiv- ing his degree in chemistry in 1959; and his doctorate in 1961, magna cum laude, with his dissertation"Thermodiffus1on and Thermodynamics in Ag-Zn Alloys."

From 1961 to1965, Rauband hiswife, Hanna, lived in the U.S. Raub was a research physicist at the University of California, San Diego, La Jolla, working in collaboration with several renowned researchers at Bell Laboratories, Murray Hill, NJ, and Los Alamos, NM.

Anoffer in 1965,fromW.C.Heraeusto head its Metals Laboratory in Hanau, Germany, tookthe Raubs home. Among hisdutiesat Heraeus, Raubdid research, development and quality assurance work in industrial production of precious and refractory metals and alloys.

As his father prepared to retire from his role as head of the FEM, Raub was offeredthepostbytheBoard. From 1970 until today, he has worked to make the Institute an international center in its field, in whch his father (now 86 years young) is still interested.

Raub's accomplishments are impres- sive, amongthem the publication of more than 450 papers, articles and books re- lated to surface finishing; work as editor- in-chief of the Journal of the Less Com- mon Meta1s;the acquisition of numerous patents; the 1964/65discovery, with B.T. Matthias and A.R. Sweedler, ofthe super- conductivity of multicomponent oxides. ~

Inadditiontothe 1992AESFScientific Achievement Award, Raub hasgathered such honors as the International Pre- cious Metals Institute's Henry J. Albert Palladium Medal and Prize, the Turkish Electroplating Society's Document of Merit, and the German Jacobi Prize.

(For a more-detailed biography of this distinguished industry leader, see Plat- ing and Surface Finishing, September 1992, pp. 27, 28.)

~

Plating andsurface Finishing