Electrolyte Solution and Process For High Speed Gold Plating

Background of Invention

This invention relates to the electrodeposition of gold from an aqueous soluble gold cyanide plating bath. More particularly, it relates to obtaining bright gold deposits at high temperatures and high plating rates, e.g. without degra- dation of the quality of the deposit.

In U.S. Patent 2,905,601, an electrolytic bath for plating gold or a gold alloy is disclosed which contains a cyanide of gold, a base metal salt such as of cobalt, nickel, indium, etc. , and citric acid plus sodium citrate or acetic acid plus sodium acetate. The combination of, e.g. citric acid and its salt is clearly intended to act as a buffer to maintain the bath within a pH range of about 3-5. The use of other weak acids such as lactic, formic, etc., is mentioned, but there is no demonstration of a bath containing formic acid. A current density range of 1-100 ASF (amperes per square foot) with only 10 ASF being demonstrated, and a temperature range of 50 ^β to 120°F. with 70"F. being preferred, are disclosed. U.S. Patent 3,104,212 differs from the above in that the base metal salt is omitted. U.S. Patent 3,672,969 discloses a gold plating bath which contains an organophosphorus chelating compound, typically a phosphorus acid, e.g. amino-tri (methylphosphonic acid) or l-hydroxyethylidene-l,l-diphosphonic acid. As an improvement, a water soluble citrate is included in the bath. However, there is no mention of formic acid.

The production of gold-copper-antimony alloys is dis¬ cussed in U.S. Patents 3,380,814 and 3,380,898. A complexing

agent such as ethylenediaminetetraacetic acid (EDTA) is employed in the bath and a weak acid and salt thereof to provide a pH of 4.5 to 6.0, exemplified by KH2PO4 (a partially neutralized acid salt), partially neutralized citric acid, tartaric acid or acetic acid.

The use of nickel or cobalt chelates as brightener/ hardeners is taught in U.S. patents 3,149,057 and '058. The use of aliphatic acids of 2 to 8 carbon atoms such as acetic, citric, tartaric, etc., when properly neutralized to act as buffers to maintain a pH between 3 and 5, is described. U.S. Patent No. 4,186,064, discloses phosphate salts and citric acid salts as the conducting and buffering agents, and cobalt or nickel chelates of an organophosphorus compound such as nitrilotri (methylene phosphonic acid) . U.S. Patent 4,253,920 discloses a gold plating bath which includes potassium dihydrogen phosphate, a Cu or Ni hardefter/brightener and, as chelating agent, 1-hydroxyethyli- dene-1, 1-diphosphonic acid. No weak organic acids are present. In U.S. Patent 4,197,172, the chelating agent is nitrilotris (methylene) triphosphonic acid (sold as Dequest 2000). U.S.

4,396,471 states that virtually any conductive acid or salt may be used as electrolyte and the composition of the electrolyte is not critical, mentioning weak organic acids such as malic, formic, and especially citric. Potassium citrate plus citric acid to buffer the bath, is recommended.

Commercially, parts to be plated can be plated on a continuous ^' basis on reel-to-reel selective plating machines, see "Continuous Reel-to-Reel Plating for' the Electronics Industry" by Jean ochet et al, an AES Electronics Lecture. Such machines are very expensive and perform all the plating steps on a con¬ tinuous basis, including cleaning, activation, undercoating, and

final plating of the parts by processing the parts, in successive steps, through the complete plating cycle. Basically, their pro¬ cessing speed is only limited by the deposition speed, i.e. the ability of the plating baths to produce acceptable deposits of required thicknesses rapidly. It can be seen as a matter of economics that high deposition rates are highly desirable, since the higher the production is, the lower the unit cost becomes. The introduction of continuous selective high speed plating required gold solutions capable of plating at much higher speed and current densities. At first, when low gold prices pre¬ vailed, this was not met simply by increasing the gold concentra¬ tion of the bath, because as a general rule higher gold concen¬ trations permit higher efficiency, current densities and plating rates. That is, in the typical gold bath of U.S. Patent 2,905,601 with 8 grams per liter of gold, this was increased to 32 grams per liter and even higher to obtain higher current den¬ sity and plating speed. However, with the advent of greatly increased gold prices, this became impractical. For economic reasons (lower inventory, lower drag out, etc.) gold contents should be kept as low as possible. Consequently, other routes were sought to obtaining high speed gold plating baths with lower gold concentrations and high acceptable current densities.

Formulations were proposed making use of so-called cur¬ rent extenders. Typically, such current extenders increase the bath's ability to plate at high current densities without the deposit being burnt. A burnt deposit is spongy and black. It will be understood that higher current densities mean higher rates of deposition, since theoretically one ampere will deposit a definite amount of metal in one second. As illustrative, in U.S. Patent 3,929,595, the current extender is a heterocylic azohydrocarbon sulfonic acid or salt

thereof. In U.S. patent 4,436,595, glycolic acid with a salt thereof is used as current extender. However, the addition of heterocyclic azohydrocarbon sulfonic acids or salts thereof or of glycolic acid and its salts, to gold plating solutions, reduces significantly the current efficiency, expressed as mg/ ampere-minute, to very small values, rendering the buildup of the thick bright deposits difficult or impossible in high speed applications in which thick deposits have to be built up in a very short time, termed "retention time". That is, the low cur- rent efficiency works oppositely to the effect of high current density. The low efficiency of these baths at high current densities could be overcome by increasing the temperature from the usual maximum range of 120-130*F. to 150*F. However, when that is done, the resulting deposits become dull or even burnt, hence unacceptable. Thus, such current extenders, although improvements for certain applications, are of limited interest or impractical for some high speed applications. As stated in U.S. 4,436,595 at column 3, lines 25-29, the lower the temperature, the brighter the deposit, but the slower the plating speed, and vice versa; and as a compromise between brightness and plating speed, an operating temperature of 130°F. is preferred. In fact, in practice, very few if any known acid gold plating baths give bright deposits at 150 ^*> F., whereas, as will be seen in the ensuing description, the reverse is true for the baths of the present invention.

Furthermore, in many instances, the deposits produced by some high speed electrolytes still fall short of expectation for the following reasons:

High current density plating in the order of 500 to 1000 ASF at the cathode results in similar, and in some cases because of very small anode areas, in even higher anodic current

densities. Such high anodic current densities are highly unde¬ sirable because of anodic oxidation phenomena.

In most cases, the cobalt and/or nickel brighteners/ hardeners usually present in the valency of 2 are oxidized to the higher valency of 3 and/or even changed to the highly undesirable inactive potassium cobalticyanide KβCCo CNjg] or similar hydroxy complexes of the same family. The gold is also, in some cases, partially or even fully oxidized to the higher valency of 3, hence considerably reducing the efficiency and the rate of plat- ing. Also, oxygen is often absorbed by the electrolyte and decreases efficiency and worsens metal distribution, as discussed in U.S. Patent 3,669,852 recommending several methods to remove oxygen from gold plating baths.

In U.S. 3,475,290 an alkyl or alkylene quanidine com- pound is used in the bath and a large quantity of reducing agent such as formic acid is used to prevent its decomposition.

U.S. patent 3,904,493 discloses gold sulfite plating baths containing organophosphorus compounds such as phosphonic acids. A brightening agent such as nickel may be included in the baths. The addition of mineral or organic acids, bases or buf¬ fers to control pH, within a range of 5 to 11, is mentioned but the choice is not critical. Current densities useful for the baths are rather limited, e.g., of the order of about lA/dm- ² .

Other disclosures of general interest in this area are: U.S. Patent No. 3,893,896'

U.S. Patent No. 4,075,065; U.S. Patent No. 4,076,598; U.K. Patent Application No. 2,039,532A; "Selective Plating Equipment - What Are the Options?," by Douglas R. Stewart, AES Symposium on Economic Use of and Sub¬ stitution for Precious Metals in the Electronics Industry, Sep-

-6-

tember 16-17, 1980; "Multi-Lateral Thicknesses in Individually Plated Stripes", by Brian C Dowling, Second AES Symposium, October 5-6, 1982; "Super Selective Reel to Reel Plating", by Peter Meuldijk, Second AES Symposium, October 5-6, 1982; and Products Finishing, pp. 21-22, 24-25, January, 1941. SUMMARY OF THE INVENTION

It has now been found that in the electroplating of gold from an aqueous bath containing a soluble gold cyanide, e.g. an alkali metal gold cyanide, formic acid in a critical concen- tration and an organophosphorus chelating agent, in particular a phosphonic acid, act in a synergistic manner to allow high plat¬ ing speeds. The preferred phosphonic acids are 1-hydroxyethyli- dene-1, 1-diphosphonic acid, sold under the tradename of Dequest 2010 and aminotri(methylene phosphonic acid) sold under the trade name Dequest 2000, both available from the Monsanto Company. It is believed that in this composition the formic acid acts as a current extender, permitting high current densities and also high temperatures to be used thereby achieving high plating speeds. For convenience, the amounts of formic acid are given in milli- liters. The formic acid should be present in an amount of at least 20 ml/liter to about 150 ml/liter, preferably above 40 ml/ liter to about 90 ml /liter, more preferably above 40 ml/liter to about 50 ml/liter, based on the standard purified or C.P. grade containing approximately 90 weight % of formic acid. Concentra- tion of the organophosphorus compound in the range of 50-150 ml/1 have given good results. The electrolyte or conductivity salt may, in some cases, be a mixture of an alkali metal monophosphate and a phosphonic acid or mixed phosphonic acids.

Cobalt or nickel which may be introduced as their salts or chelates, e.g. as the sulfate, may be used as brightener/ hardeners. The cobalt or nickel concentration may be in the

range of 350 to 600 mg/liter, preferably about 500 mg/liter.

The pH is also critical and it has been found that when cobalt is present the pH should be in the range of 4.0 to 4.2, and that when nickel is present the pH should be in the range of 3.8 to 3.9.

The gold concentration may range up to 30g/liter, pre¬ ferably may be in the range of 8 to 20 g/liter, e.g. 10 to 20 g/liter, but for some plating techniques may be lower, e.g. from 2 to 4 g/liter. As will be seen in the following description, the addi¬ tion of formic acid and a chelating organophosphorus compound to the bath gives unexpected results, i.e., produces bright gold deposits at high deposition speed.

The plating may be accomplished by any of the commer- cial means available such as barrel, rack and strip plating equipment and high speed continuos selective plating equipment. The products are useful for industrial purposes, especially for making electrical connections, e.g. as connectors. Depending on the type of equipment used, plating may be carried out at tem- peratures in the range of 90 ^β to 160*F. and at current densitites from about 0.5 to an excess of 1000 ASF. The process yields deposits having a cobalt or nickel content of .15 to .2% and a hardness in the range of 130-200 Knoop. BRIEF DESCRIPTION OF THE DRAWINGS The Figure is a graph showing the effect of gold con¬ centration on plating speed at 100 ^β F. DETAILED DESCRIPTION

The invention will be described with reference to the ensuing tests and Examples, which are intended to be illustrated but not limitative.

Extensive testing was carried out in order to provide a stable solution capable of plating at a higher rate and higher current densities than the currently commercially available solutions, without the problems discussed above. ; The formic acid used throughout the testing was the standard purified or C.P. grade containing approximately 90 weight percent of formic acid. The weight of 1 liter (or 1000 ml) of this grade is 1248 grams (or 1 ml = 1.248 grams). Since 90% of the weight is formic acid, it follows that 1 ml. contains 1.1232 grams of formic acid. Other amounts can readily be calcu¬ lated, e.g., 100 ml. contains 112.32 grams of formic acid; 50 ml contains 56.16 grams formic acid, etc. Other grades of formic acid can also be used and in such case equivalent amounts to those disclosed herein can be calculated, for example, at half the concentration of the C.P. grade, twice the number of milli- liters of formic acid would be used.

The formula used in these tests for cobalt hardened gold is given by Formula A below unless otherwise indicated.

FORMULA A 1 liter

Potassium Monophosphate 70g Dequest 2010 50 ml

Formic Acid 50 ml

Potassium hydroxide to pH 4.0 Cobalt-metal (as sulfate) 500 mg

Au-metal (as PGC) 20 g

PGC is an abbreviation for 68% potassium gold cyanide. The cobalt may be any suitable soluble compound such as the sulfate or the complex of a suitable, compatible chelating compound or that of the organophosphorus compounds used in the formula.

The testing method used basically employs a 1 liter beaker with platinum coated anodes, a thermostatically controlled heater, a means to provide mild agitation and a suitable recti¬ fier in which are plated copper wires of about 1mm in diameter and 320mm in length turned around a wood cylinder of 2mm in dia¬ meter. These have the advantage over panels of giving a better idea of the bath overall plating abilities. All the wires plated in all the tests have a minimum gold thickness of 50 to 100 microinches. Test 1

Conditions: Plating Temperature: 150 ^β F. - gold concen¬ tration 20 g/1 pH 4.0.

Current Density 40 ASF 80 ASF 120 ASF 150 ASF Efficiency 54 mg* 32 mg 43 mg 28 mg Appearance Sem- bright bright bright bright (*) mg or milligrams per ampere-minute

The bath of test 1, employed in a controlled depth cell such as described in "Continuous Reel-to-Reel Plating for the Electronics Industry" ibid, gave an excellent deposit at 120 ASF.

EXAMPLE I The same formulation was used in a high agitation cell like the one described in U.S. Patent 4,431,500, with a gold con¬ centration of 15g/l. An excellent, about 54 microinch thick, bright gold deposit was obtained at a current density of 980 ASF, a temperature of 150"F., and a line speed of 25 feet/ minute. The retention time was 3.5 seconds and the efficiency was 39.3 mg per amp. min. , which gives a plating speed of 15.48 microinches per second of retention time. It should be noted that a high agitation cell like the one described in U.S. Patent 4,431,500 allows current densitites that are much higher, i.e. 6 to 10

-10-

times higher - depending on the cell configuration - than obtained in a beaker (with thin wires), or in a controlled depth cell. For such high agitation, high speed plating, the tempera¬ ture may suitably be in the range of 100° to 150 ^β F. TEST 2

Conditions: Plating temperature: 120*F. - gold concentration 20 g/1 pH 4.0.

Current density 80 ASF 120 ASF 150 ASF Efficiency 32 mg 26 mg 24 mg Appearance semi- bright bright dull TEST 3 Condition: Plating temperature: 150 ^β F. but gold concentration increased to 30 g/l: Current Density 40 ASF 80 ASF 120 ASF 150 ASF 200 ASF 300 ASF Efficiency 53.75 mg 59.5 mg 61.8 mg 62.6 mg 58.8 mg 49.1 mg Appearance dull bright bright bright bright hazy

Such a bath is capable of producing a bright gold deposit at higher current densities, higher temperatures and higher plating rates than that of the prior art.

The above tests show that the deposits' brightness is better at higher temperature (150 ^β F.) than at the lower tempera¬ ture of 120"F. That is completely unexpected since - as reported in U.S. Patent 4,436,595 at column 3, lines 25-29, as well as in other authoritative technical works - brightness usually decreases as the temperature increases. In practice very few acid gold baths, if any, are bright at 150 ^β F. As mentioned above, the reverse is true for the baths of the present inven¬ tion, which is totally unexpected. Further testing shows that another unexpected phenom¬ enon takes place. Tests 4, 5 and 6, set forth below, show that

at 150*F., deposit brightness decreases at lower current densi¬ ties.

TEST 4 Conditions: Plating temperature: 150"F. - gold concentration 20g/l pH 4.1.

Current density 40 ASF 80 ASF 120 ASF 150 ASF 200 ASF 250 ASF Efficiency 51.75 mg 57.1 mg 58.7 mg 56 mg 47.3 mg 42.2 mg Appearance hazy bright bright bright hazy hazy However, as shown in Tests 5 and 6, when the gold con- centration is reduced to"4 grams or 10 grams per liter, the deposits become brighter at 40 ASF.

TEST 5 Conditions: Plating temperature: 150 ^β F. - gold concentration 4g/l, pH 4.1. Current density 40 ASF 80 ASF 120 ASF 150 ASF Efficiency 30.75 mg 20.3 mg 15.75 mg 13.7 mg

Appearance bright bright dull dull

TEST 6 Conditions: Plating Temperature: 150 ^β F. - gold concentration lOg/l, pH 4.1.

Current density 40 ASF 80 ASF 120 ASF 150 ASF Efficiency 56.25 mg 43.6 mg 38.1 mg 36.4 mg

Appearance bright bright bright bright

Thus, when gold concentration is decreased, deposits at lower current densities become brighter. However, other tests show that at 4 g/1, the deposits are not bright at higher current densities in the order of 120 ASF.

When a bath is prepared with formic acid neutralized to pH 4.0 ^' with potassium hydroxide and cobalt is introduced as the sulfate, as soon as the potassium gold cyanide is introduced into the bath, one can observe the immediate formation of an insoluble

-12-

pinkish - white compound which renders the plating bath unuseable. However, when the cobalt was introduced in the form of a complex cobalt salt of Dequest 2010 (1,1-hydroxyethylidene-1,1-dipho- sphonic acid) a more stable bath resulted. The formulation used in Test 7 was Formula B below.

FORMULA B

1 liter

Formic acid 150 ml

Potassium hydroxide 130 grams

Cobalt-metal (as complex) 500 milligrams

Au-metal (as PGC) 10 grams pH 4.0

TEST 7 Plating temperature: 150 ^β F. The object of the test was to find out the limits of the bath at high current densities. Thus, the current density was increased until burning of the deposit took place.

The results were as shown below:

Current Densities Efficiency Appearance

ASF mg/amp.min.

40 28.9 Bright

80 32.6

120 30.8

150 30.5

200 24.85

300 19.9

400 16.35

500 13.3

600 11.2 Hazy

700 10.2 Burnt

-13-

The above results are somewhat unexpected since the deposits remain bright up to above 500 ASF. However, the color of the deposit was somewhat whiteish, suggesting a high cobalt percentage in the deposits which is considered undesirable in deposits to meet certain specifications.

It appears that an excess amount, i.e., an amount of free chelating agent such as Dequest 2010 (over that contained in a cobalt chelate) is necessary to stabilize the bath and assure the proper concentration of cobalt in the deposit. A new bath was prepared with an excess of Dequest 2010 in the formulation of the bath used in Test 7 as another attempt to obtain a yellower color. The following concentrations were used, as Formula C below, in Test 8.

FORMULA C 1 liter

Dequest 2010 50ml Formic acid 50ml

Potassium hydroxide 68 grams pH 4.0 Cobalt (as cobalt sulfate) 500 mgs.

Au (as PGC) 10 grams

TEST 8 Plating temperature: 150*F. Current Density 40 ASF 80 ASF 120 ASF 150 ASF 200 ASF Efficiency 52.9 mg 45.6 mg .40.0 mg 34.4 mg 30.0 mg Appearance bright bright bright bright burnt Results: Color is now a rich yellow and the cobalt content sig¬ nificantly reduced to within the range of .1 to .2%, which is perfectly acceptable. The bath gives results similar and com¬ parable to the bath used in Test 1. This demonstrates that a

•14-

free chelating agent of the organophosphorus type is required to work in cooperation with the formic acid and inhibit the cobalt deposition.

A bath was prepared similar to that used in Test 1, but without excess Dequest 2010, as shown in Formula D below, and was used in Test 9.

FORMULA D

1 liter

Potassium phosphate monobastic 70 grams

Formic Acid 50 ml

Cobalt (as complex of Dequest

2010) 500 mgs pH 4.0

Au (as PGC) 10 grams

TEST 9

Plating Temperature: 150 ^β F. Current Density 40 ASF 80 ASF 120 ASF 150 ASF Efficiency 49.6 mg 55.6 mg 43.8 mg 37.0 mg Appearance bright burnt burnt burnt Results: As the efficiencies obtained with the baths of Tests 1 and 8 are comparable, the aspect of the resulting deposits in Test 9 is far from being as good and is unacceptable at 80ASF. From the above tests, it is quite clear that a synergistic effect takes place in the baths of tests 1 and 8, which produce superior bright deposits at 150"F. The synergistic effect appears to be primarily between the Dequest 2010 and the formic acid.

A number of tests were run to determine whether the concentration of formic acid is critical. A bath without formic acid but similar to the bath used in Test 1, was prepared. The formulation is given as Formula E below.

FORMULA E

1 liter Monopotassium phosphate 100 grams Dequest 2010 50 ml Potassium hydroxide to adjust pH to 4.0

Cobalt metal 500 mgs

Au-metal (as PGC) 10 grams The thin copper wires of the type used in Test 1 were plated in the above solution and the results are set forth below. The plating temperature was 150"*F.

TEST 10 Current Density 80 ASF 120 ASF 150 ASF Efficiency 47.0 mg 29.0 mg 25.0 mg Appearance dull dull dull

In the above tests none of the deposits obtained were bright at 150"F.

To the bath used in Test 10, formic acid was added to obtain an increasing concentration in order to run tests with the following concentrations of formic acid: 5, 10, 20, 30, and 40 ml/liter. For all the tests, the pH was adjusted to 4.0 with K0H, and the gold concentration to 10 g/1. The plating tempera¬ ture was 150*F. The results are given below: Test 11 (5 ml of formic acid): Current Density 80 ASF 120 ASF 150 ASF Efficiency 45.0 mg 30.0 mg 27.0 mg

Appearance dull dull dull

Test 12 (10 ml of formic acid):

Current Density 40 ASF 80 ASF 120 ASF 150 ASF Efficiency 52 mg 45.0 mg 33.0 mg 27.0 mg

Appearance semi- semi- bright bright dull dull

-16-

Test 13 (20 ml of formic acid): Current Density 40 ASF 80 ASF 120 ASF 150 ASF Efficiency 54 mg 45.0 mg 37.0 mg 30.0 mg Appearance semi- semi- semi- bright bright bright dull Test 14 (30 ml of formic acid) : Current Density 40 ASF 80 ASF 120 ASF 150 ASF Efficiency 56.5 mg 45.0 mg 34.0 mg 30.0 mg Appearance semi- semi- semi- bright bright bright dull

Test 15 (40 ml of formic acid):

Current Density 40 ASF 80 ASF 120 ASF 150 ASF

Efficiency 41.0 mg 32.0 mg 27.0 mg

Appearance semi- semi- bright bright bright dull Although concentrations as low as 20 ml/1 begin to show an improvement at lower curent densities, all of the tests 10-15 show that the minimum effective concentration of formic acid to assure acceptable, consistent, high build, bright deposits at 150 ^β F. over 40 ASF is above 40 ml/1. Preferably, the concentra¬ tion should be about 50 ml/1 as shown by Test 1.

To the bath used in Test 10, formic acid was added to obtain a concentration above the level considered optimum of 50 ml per liter. No adverse effect was observed other than a slight decrease in efficiency as can be noted below. The gold concen¬ tration was 10 grams per liter and the plating temperature was 150 ^β F.

Test 16 (Formic acid 75 ml) : Current Density 40 ASF 80 ASF 120 ASF 150 ASF Efficiency 54 mg 43 mg 37 mg 30 mg

Appearance dull bright bright bright

-17-

Test 17 (Formic acid 150 ml):

Current Density 40 ASF 80 ASF 120 ASF 150 ASF

Efficiency 54 mg 42 mg 32 mg 26 mg

Appearance dull bright bright bright

Surprisingly, addition of other weak acids does not achieve similar results. For instance, a bath was prepared by replacing formic acid with citric acid, as shown in Formula F below, and was used in Test 16.

FORMULA F

1 liter

Monopotassium phosphate 50 grams

Dequest 2010 50 ml

Citric Acid 50 grams

Potassium hydroxide to pH 4.0

Cobalt-metal 500 gs

Au-metal (as PGC) 10 grams

TEST 18

Plating Temperature: 150*F. Current Density 80 ASF 120 ASF 150 ASF Efficiency 50.0 mg 36.0 mg 32.0 mg Appearance semi- bright burnt burnt

The results obtained are far from being comparable with those of Test 1. Even at 120 ^β F., the deposit was unacceptable at 80 ASF since the deposit on the wire was unevenly bright and burnt in the high current density areas. Increasing or decreasing the concentration of citric acid between 10 to 100 grams per liter did not show any significant change or improvement.

Nickel may be substituted for cobalt in similar formu¬ lations, however, the preferred pH for more consistent color is 3.8 to 3.9 instead of 4.0 to 4.1 for cobalt. The formulation

-18-

used in Test 17 was Formula G below:

FORMULA G

1 Liter

Dequest 2000 150 ml Formic Acid 50 ml

Potassium hydroxide to pH 3.8 (about 110 grams) Nickel-metal (as sulfate) 500 mgs Au-metal (as PGC) 10 grams

TEST 19 Plating Temperature: 150" F.

Current Density 80 ASF 120 ASF 150 ASF Efficiency 37 mg 29 mg 24 mg Appearance bright bright semi-bright

Although the efficiency is significantly less than for the cobalt solution, such formulation has proved to be eminently suitable for high speed applications (in controlled depth cells as well as high agitation cells) in which nickel-hardened gold is a requirement. It should be noted that nickel-hardened gold deposits are specified in some higher temperature applications since nickel-hardened gold does not discolor as readily as cobalt-hardened gold.

The nickel content of the deposits was found to be in the range of .2 to .3% depending on the conditions of deposition. Tests have also shown that the addition of alkali phos- phates, with the exception of ammonium phosphate, is not desir¬ able, as they have a tendency to render the bath unstable result¬ ing in precipitation of the gold and the nickel in the form of one or more unknown compounds. Furthermore, Dequest 2000 is preferred over Dequest 2010 and Dequest 2041 in the above formula- tion for nickel.

-19-

Other unexpected results occur when the bath used in Test 1 is used in conventional barrel plating or in related equipment like the Vibrobot. ^®

For optimum results, the bath of Test 1 was modified in order to optimize distribution of the gold deposit. The cobalt content is suitably kept in range of 350 to 415/mg/liter, prefer¬ ably at about 380 rag/liter. The following formulation, desig¬ nated Formula H, was used in Example II with the gold concentra¬ tion at 4 g/liter. FORMULA H

1 liter Monopotassium Phosphate 75 grams Dequest 2010 50 ml

Formic acid 50 ml Cobalt-metal (as sulfate) 380 mgs

Au-metal (as PGC) 2 to 4 grams pH 4.0

A plating temperature in the range of 90* to 110 ^β F. was selected mainly because it gave a color identical to that of the high speed formulation of Test 1. Higher temperatures may be used.

As can be expected, increasing the plating temperature increases the efficiency and the plating speed; however, at the relatively low current densities used in this type of plating, it decreases the brightness.

A plating temperature of 100 ^β F. appears to be the best all around compromise for uniform color and efficiency.

It should be noted that the standard barrel gold baths of the prior art are limited by their allowable maximum plating current density. When that current density is reached, the resulting deposit becomes burnt and hence unacceptable. That is

-20-

not the case with the bath of the present invention. It has been found that it is virtually impossible to burn the deposit in such a bath, in a barrel or related equipment, at voltages below 10. Even such voltages are impractical since the limiting factor, usually, is the voltage that can be handled by the rectifier and the platinum coated anodes, reported to be at not above 6 to 7 volts. Above that voltage, the platinum coating is slowly stripped, which renders the anode unsuitable for plating, so that the anode has to be replaced. Tests show that the cobalt-hardened gold bath of the present invention does not behave like the standard barrel gold plating bath of the prior art and is a quick building bath.

One of the plating requirements on a typical connector was 55 microinches of deposit. Such connectors are plated in bulk in the Vibrobot ^® by loads of several thousand parts having a total area between 60 and 90 square feet. As shown in the fol¬ lowing, the load was plated in 33 minutes, whereas the very same part plated with a conventional barrel formulation of the prior art required a plating time of 112 minutes. EXAMPLE II

At a concentration of 4g/l of gold, tests show that plating speeds between 1 and 1.75 microinches per minute are achieved in a Vibrobot. The part plated is a small typical con¬ nector with an area of .183 in ² which requires a minimum thick- ness of 50 microinches. A mean thickness of 56 microinches is obtained in 33 minutes at an average current density of 1.5 ASF with a standard deviation of less than 2. Details of some runs using the same part are presented below.

All parts were bright, uniform in color (without any color variation from lot to lot in spite of different current densitites) and passed solderability specifications.

It should be noted that the main variable is voltage which was not increased above 11 volts for the reasons discussed above. The area plated varied between 62 and 80 sq. ft.

TABLE

BASKET TOTAL TOTAL AVERAGE AMP. THICKNESS EFFICI

DIAM AREA AMPS VOLTAGE ASF MIN. MEAN S.D. MICRO-INCH/MINUTE (Mg/AMP

500mm 76.6 40 5.5 .523 4900 59.68 1.33 .4871 34

500mm 64.6 40 5.5 .62 3400 56.42 1.04 .6650 34

400mm 62.2 40 7.0 .64 3000 56.52 1.76 .7326 36

500mm 64.6 66 8.0 1.02 3550 60.2 2.0 1.1214 33

500mm 64.6 80 9.0 1.238 2965 52.98 1.57 1.429 38

500mm 79.9 98 10.0 1.226 4150 54.82 2.42 1.3 34

500mm 75.3 112 11.0 1.4864 3700 55.88 1.987 1.69 35

The following observations were made.

(1) Distribution is exceptionally good and the stan¬ dard deviation (S.D. in above table) remains around or below 2.0.

(2) Throwing power is very good and superior to that of the prior art.

(3) Color is consistent and uniform from lot to lot regardless of the current density used. Furthermore, the color is identical with the color obtained with the high speed plating at 150 ^β F. in Test 1. (4) Solderability is exceptionally good and consis¬ tent.

(5) Cobalt in the deposit remains below .3% at all voltages.

(6) Little or no consideration need be given to cur- rent density since the bath does not burn within the parameters given above.

As a general rule, higher gold concentrations allow higher efficiency, current densitites and plating rates. How¬ ever, for economical reasons (lower inventory, lower drag out, etc.), gold contents are kept as low as possible, i.e. around 4.1 g/1 and as low as 2 g/1.

Optimum gold concentration depends on the application and should be adjusted accordingly. The graph of the figure shows the effect of gold concentrations on efficiency and plating speed. Plating rate in miocroinches per minute is ^' plated against current density for gold concentrations respectively of 4 grams per liter and 6 grams per liter, at 100 ^β F. It will be seen that higher gold concentration increases plating rate.

It will be apparent that the invention is capable of numerous variations without departing from the scope of the invention and without sacrificing its chief advantages.