ELECTROLESS GOLD PLATING WITH USE OF (EXTERNAL) SOLID NICKEL CATALYST
N. V. MANDICH, CEF HBM Electrochemical & Eng. Co.
2800 Bernice Road Lansing, IL., 60438
ABSTRACT
The object of this study is to find out if "Hypo Gold" ffl.G.1 formulation is a true gold electroless
process and if It can be used to plate gold flash over the selectively plated electronic contacts. The purpose
of gold flash to plate the entire surface of the gold plated contact with thin layer of soft gold for the
purpose of SoIderatMty. Thickness of this gold flash should be from 5-8 microinches. It is found that
addition of solid nickel when barrel plating with H. G. formulatlon, converted this system from immersion
type to true auto cataiytic process.
ELECTROLESS PROCESSES: DEFINITIONS
Because electroless deposition does not involve the passage of externally
applied current to the system, some confusion over the use of the term electroless
has resulted. Electroless deposition has been used synonymously with chemical
deposition which can result from the following processes:
(1) DisPlacement reactions. Depending on its position in the electrochemical
series, a metal higher up in the series may be covered (plated) with the metal lower
down in the series. A well known example is the coverage of iron with copper in an
acidified copper sulphate solution. Two reactions, one anodic and the other cathodic,
take place simultaneously at the surface of the iron.
e svf/Fim 1992, Atlanta, GA
823
AESF Annual Technical Coneerenee SUWF~'N@ 'SDZ
- June 3CSE-Z5,1Cr-221
.Atlanta, CeorgCa .
The American Electroplaters and Surface Finishers Society, Inc. (AESF) is an international, individual- membership, professional, technical and educational society for the advancement of electroplating and surface finishing. AESF fosters this advancement through a broad research program and comprehensive educational programs, which benefit its members and all persons involved in this widely diversified industry, as well as govemment agencies and the general public. AESF dissemi- nates technical and practical information through its monthly joumal, Plating and Surface Finishing, and through reports and other publications, meetings, symposia and conferences. Membership in AESF is open to all surface finishing professionals as well as to those who provide services, supplies, equipment, and support to the industry.
According to the guidelines established by AESF's Meetings and Symposia Committee, all authors of papers to be presented at SUWFIN@ have been requested to avoid commercialism of any kind, which includes references to company names (except in the title page of the paper), proprietary processes or equipment.
Statements of fact or opinion in these papers are those of the contributors, and the AESF assumes no responsibility for them.
All acknowledgments and references in the papers are the responsibility of the authors.
Published by the American Electroplaters and Surface Finishers Society, Inc. 12644 Research Parkway 0 Orlando, FL 32826-3298 Teiep"re: 407L%I -6441 : Fa: "2
Copyright 1992 by American Electroplaters and Surface Finishers Society, Inc. All rights resewed. Printed in the United States of h-~er ica. This publication may not be reproduced. s t p s m 'tted in whole or mrt, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise without the prior written Dermission of AFSF, 12644 Research
-anao. FL 328263 298.
Printed by AESF Press
SURIFIN'is a registemd trademark of the American Electroplaters and Surface Finishers Society, Inc.
; .
824
Fe- Fe2 + 2e (Anodic) E" = -0.44V
Cu2 +2e =) Cu (Cathodic) E" = 0.337V
<1>
<2>
Once the iron electrode is fully covered with the thin layer of copper, the process
comes effectively to a halt, and no further thickening takes place. The deposit is
normally thin (< 1 .O /I m); and i ts adhesion is not satisfactory.
(2) Galvanic dissolution reactions. In these, the workpiece (M) is coupled to
a less noble metal M,, and the assembly immersed in the plating solution containing
ions (M,) of a more noble metal M,. In this case the less noble metal M, goes into
solution (anodically) and metal M, deposits onto the work piece.
M,- M,"' +ne
M2"+ +ne 3 M,
The metallization in this case can continue for as long as dissolution of the
sacrificial anode (M,) is possible. There are commercially available gold plating
solutions which are claimed to produce thicknesses in excess of 2.5yin, the preferred
substrates being silver, copper, brass, nickel, tin or mild steel coupled to a zinc or
aluminium wire. The donation and acceptance of electrons is an integral part of the
above processes - as it is indeed with all aqueous plating processes.
To avoid confusion, we should examine the term electroless deposition. Since
electron donation during deposition is essential, we have to rely on methods other
than substrate displacement, galvanic dissolution of the sacrificial anode or supplY of
current via an external power unit (electrodeposition), to achieve this. A chemical
2
"reducing agent" (Red) is the electron donor in all truly electroless plating processes
and the process is catalyzed by the deposited metal.
<5> Red,, OX*o,'" + "= CMaWic sUff.U
where Ox is the oxidation product and Red is reducing agent.
Thus, once the substrate surface becomes covered with the plated deposit
continuation of the process relies on the latter to catalyze further deposition. The
term "auto-catalytic" is therefore used to describe this type of plating. So far, this
term has not suffered from the same confusion as the term "electroless". Before a
gold plating process can be described as autocatalytic one important requirement is
needed, that is, the system in question must be capable of depositing gold on a gold
substrate. In the following discussion the terms electroless and autocatalytic will be
used interchangeably.
SURVEY OF KNOWN PROCESSES
A large number of "electroless" gold plating bath formulations exist in the
literature, both in the form of technical papers and patents. Several of these
formulations have been reviewed by Okinaka'. Ali and Christie'have tested some of
these baths, based on hypopho~phite~", hydrazine4i8, thiourea7 and the results were
disappointing. No gold plating was achieved on a gold substrate, and it was
concluded that these baths were not auto-catalytic. In the case of the thiourea bath'
they were unable to deposit gold onto a gold substrate despite the claim of Okinaka'
that it was possible to do so from this solution.
Although the Brookshire hypophosphite bath' was not tested by them, it is
3
,
interesting to note that Okinaka' found that the plating of gold on nickel occurred
whether hypophosphite was present or absent (though no plating rate was indicated).
This is in spite of Brookshire's claim that hypophosphite was a necessary ingredient,
and that decreasing of its content reduced the plating rate. What is more surprising
in Brookshire's work is his claim that the plating rate was completely independent of
the pH of the solution. An electroless gold bath which has been widely tried and
reported upon in the recent literature is Okinaka's borohydride bath'. Plating was
reported to take place on copper, nickel, cobalt, iron, palladium, platinum, Kovar,
Permalloy and steel. It was also reported to occur on gold and thus demonstrating
that the process was truly autocatalytic. In their later work, Okinaka et allo concluded
that in their borohydride bath, Ni (11) decreased the plating rate and increased bath
instability. Gany and Mahapatre," in an excellent review, stated that no formulation
has been developed that provides a reliable, stable, fast-building process for
widespread application.
Okinaka bath, although truly autocatalytic, has several limitations on their own,
such as stability and intolerance toward contaminants and inability to plate on several
desirable substrates. Several i n v e s t i g a t o r ~ ' ~ ~ ' ~ improved Okinaka formulation but
limitations are still present.
C. 0. lacovangelo and K.P. Zarwich" worked with hydrazine based bath which
used hydrazine as source of electrons, but like one truly auto catalytic bath, it would
not plate over gold. However it did plate over nickel and they used the term
"substrate catalyzed". In their subsequent work" they used hydrazine and
dimethylamine borane as combo reducing agent and they obtained truly autocatalytic
process capable of plating on a wide variety of substrates including Au, Ni and Pd.
4 826
EXPERIMENTAL PART
A. EXPERIMENTAL DATA:
In order to evaluate the HG gold solution, we used selectively gold plated
contacts and (1" x 1 ") copper coupons . The intent was to plate the contacts all over
with electroless gold rather than using the standard barrel electroplating process.
Base Metal: Cd-Copper (about 2-4% Cadmium)
Gold Thickness on Contact Pad: 25pin minimum; Surface Area/contact:.l 25in2
1. Cleanina Stem
For the preparation of the surface area for electroless gold plating, the
perchlorethylene vapor was used to remove the remaining ink from the contacts prior
to plating, followed by commercial alkaline cleaner and acid salts immersion with
rinses in between and after.
I I . Electroless Formulation (H.G.)
Potassium Gold Cyanide (67% gold)
Ammonium Chloride
Sodium Citrate
Sodium Hypophosphite
PH
Operating Temperature
2 g/l
75 g/l
50 g/l
10 g/l
7.0 - 7.5
190-21 0°F
5 027
B. EXPERIMENTAL PROCEDURE
1. Electroless Gold Plating
A number of parts were plated with barrel gold using the above HG formulation
and thicknesses of gold deposits were measured to be about 1 microinch, within the
experimental error. This indicates that HG gold has more characteristics of an
immersion gold solution, rather than catalytic.
I I . Electroless Gold + Nickel "Catalvst"
An approach to use of the "catalyst" to boost the thickness of gold deposition
was explored with pure nickel wire (.022" dia.). Results proved to be encouraging.
It is interesting to note that nickel ions must be present in the hypophosphite - reduced electroless comer solution in order to get continuous copper deposition. It
was found that nickel deposit can catalyze the oxidation of hypophosphite". As far
as we know, our work is the first application of nickel ions in electroless gold
deposition using hypophosphite as a reducing agent.
Next, five sets of experiments were performed in order to investigate some
of the basic electroless plating variables and, at the same time, to determine optimum
conditions for plating of electronic contacts:
Amount of electroless gold solution vs. number of contacts;
deposition vs. plating time, operating temperature, amount of Ni-catalyst, and
electroless gold deposition over electroplated gold or copper.
Electroless gold
a28
1. Amount of Electroless Gold Solution vs Number of Contacts
The contacts were plated always using the same amounts of HG solution and
6
nickel "catalyst". Plating time (t) was set at 40 minutes, enough time to plate out
practically all gold from these solutions. Number of contact varied to the extreme of
1 :6 ratio.
Six different batches of contacts were plated under the same conditions:
HG solution volume (V): 40 ml; temperature (T) 190-201°F; plating time: (t) 40 min;
nickel "catalyst" - wire: (8 in/1000 contacts).
30 contacts from each group was measured for thickness by Betascope. Thickness
variation with number of contacts is presented in Table 1.
Experiment Batch Size Average Thickness Number No. of contacts) (p in.)
1 300 3
2 250 4
3 200 7
4 150 10
5 100 13
6 50 25
TABLE #1 Thickness Distribution vs number of contacts.
Actual thickness corresponds with theoretical calculations of thickness; e.g.
experiment No. 3:
1 liter of HG solution = 1.34 gms of gold/l = .0536 9/40 ml
40 ml of HG solution = .0536 gms of gold
200 contacts plated = 25 in2 of surface area
Thickness = .0536 = .0536 = .000007 in = 7 p i n 298x25 7450
7 829
,
830
Fig. 1 Represents data from Table 1.
2. Gold Thickness vs. Platinq l7me ?
v = 100 ml; 500 contacts; "catalyst": Ni-wire, 5 in long
Samples of 8 contacts each, were plated for 5,10,15,20 and 35 min. Gold
thickness was measured via Betascope and presented in Table 2 and Fig. 2.
Plating Time (Min) Ave. Thickness (p inch)
5 3.2 10 4.5 15 5.2 20 5.7 35 7.0
TABLE 2 Ave. Thickness vs. Plating Time
Practically all gold from the solution was deposited in 35 minutes and less than 2%
of gold remained in plating solution.
3. ODeratinq Termerature of the Solution vs. Gold Thickness
When Ni catalyst (in the form of 2 in long wire) is used with HG
electroless gold plating solution, plating starts at approximately 170°F. However, a t
this temperature, deposition rate is slow. Therefore two higher temperatures were
used: 180°F (Table 3) and 195-201°F (Table 4). As expected, chemical reaction of
deposition is faster a t higher operating temperatures due to higher mobility of ions and
increased free energy.
The original HG electroless formulation requires 198-201°F. as operating
temperature. It is obvious that Ni catalyst initiates and accelerates plating reaction
and is recommended for extended operating temperature of 190-201°F. After both
8
plating experiments all the solutions were tested for remaining gold content (v = 40
ml). Results are presented in Fig.3 which is derived from Tables 3 and 4
Plating Time(Min1 Residual Gold in Sol'n (G/L)
- 1.34 0
5
15
25
35
.925
.483
.275
.048
1.34 0 5 .570
0
15
20
25
30
35
.350
180
.I25
.066
.025
.oo I
TABLE 4 Gold Depletion vs Time, at 195-2oI"F.
9 831
From the results it follows that when recommended operating conditions are
kept, plating time should be 35 minutes. If, for some reason, this time is extended,
the solution will start attacking gold deposit, worsening the color and the grain
structure. When plating time is less than 30 minutes a significant amount of gold will
remain in solution.
4. Influence o f Nickel Catalyst on Thickness
Plating rate depends on the amount of nickel used as catalyst in regards to
number of the contacts or, in other words, to the total surface area exposed.
The following experiments were performed in order to find out the optimum
ratio: surface area of the contacts vs. amount of nickel (wire or shots) used
expressed as Ni surface area.
Five samples of contacts were plated in HG solution under the same operating
and with varying Ni-catalyst surface area.
Data: 250 contacts; V = 40 ml, T = 195-2Ol0F, t = 35 min
No. of Ni-Wire/250 contacts Ni Wire1250 contacts Residual Gold Experim. (linear in .) (in') in solution (G/L)
1.3
1.5
1.7
2.0
2.2
.OS98
. l o36
.1174
.1381
.1519 -
.5495
.2610
.1343
.0013
.0008
TABLE No.5 Nickel Catalyst vs Gold Content
10 832
Samples from experiments 1 to 4 were acceptable, therefore optimum ratio
sample vs. catalyst is 31.25 id.1381 in. = 225:l under given conditions, (Fig.4).
5. Electroless Gold Deposition 0 ver Electroplated Gold
This set of experiments was done to establishwhether HG gold with Ni-catalyst
would plate over electroplated gold.
250 contacts previously selectively gold plated were used. Thirty contacts
from this group were marked with a diamond pen for identification purpose. Selective
gold thickness was tested prior to electroless gold plating.
Data: v = 40 ml, catalyst 2 in wire; T = 195-201" F; t = 35 min
After plating, marked contacts were separated and the composite gold deposit
thickness (electroplated + electroless gold) was measured on the same spots.
Average thickness after HG plating: 31 pin and 26 pin before plating
Electroless gold thickness is given by the difference: 31-26 = 5 pin.
It is obvious from Table No. 6 that HP gold with Ni-catalyst plated satisfactorily
over previously electroplated gold as well over the copper.
11 a33
Electroplated Gold (pin)
Electroplated + Electroless Gold (pin)
Electroless Gold (pin)
56
51
57
67
67
66
66
62
66
76
76
76
10
1 1
9
9
9
10
36
44'
41
45
55
53
9
1 1
12
Ave: 10p in
TABLE No.6 Electroless plating over gold underplate
12 834
CONCLUSIONS
A. HG gold plating solution as is has mostly immersion characteristics. In most cases
thickness of this deposit was in a 2 microinch range.
B. If nickel wire is used as a "catalyst" or "activator," higher deposit rates resulted
when plating over the copper portions of electrical contacts.
C. Our experimental results demonstrated that the final gold flash over previously
gold plated contacts can be applied using the HG electroless gold. Five to ten
microinches thick gold flash will plate over the entire surface of contact, including
selectively plated gold pad and rest of contact area made of copper when using Ni
catalyst.
D. An enclosed plastic barrel needs to be designed for these particular plating
applications. Electroless solution, Ni catalyst and contacts have to be placed inside
of this barrel and the barrel must rotate slowly, 2-3 RPM.
Notes: Purity of HG gold deposit was not tested; "We suspect that there will be
some codeposition of nickel along with gold;" Some gold was deposited over nickel
wire. In most cases this deposit was not uniform and will "brake-off" to allow further
activation effect; Some nickel metal is dissolved in the solution but we don't have
analytical data at present.
It may be of interest to mention two electroless solutions patented long before
Brenner. U.S. Patent 1,207,218 (1918)!) states the following formulas:
13 835
Ni-citrate 10 g/l
Ammonia 10 g/l
AI kaline Hypophosphite 10 g/l
gold cyanide 10 g/l
Na-phosphite 10 g/l
Temp 212°F
It is never too early to invent! Is it?
REFERENCES 1. Y. Okinaka, in "Gold Plating Technology", edited by
E. H. Reid and W. Goldie, Electrochemical Publications, Ayr, Scotland, 1973
2. H.O. Ali, R.A. Christie, Gold Bull. .l7,(4),118,( 1984).
3. S. D. Swan, E.L. Gostin, Met-Finish., =,(4),52,1961.
4. E. L. Gostin, S. D. Swan, U.S.Patent 3,032,436 (1962).
5. T. Ezawa, H. Ito, Jpn-Patent 40,1081 (1965).
6. B. M. Luce, U.S.Patent 3.300,328 (1967).
7. T. Oda, K. Hayashi, U.S.Patent 3,506,462 (1970).
8. R. R. Brookshire, U.S.Patent 2,976,181 (1961).
9. Y. Okinaka, Plating, =,(9),914,(1970)
10. Y. Okinaka, at al, J. Electrochem. SOC. =,(1),56,(1974).
1 1. G.M. Gany, S. Mahapatra, J.Sci.1nd.Res. ,46,(4),154,( 1987).
12. M. El Shazley, K.B. Baker, U.S.Patent 4,337,091 (1982)
13. M. El Shazley, A.A. Halechv, U.K. Patent GB 212,144A
14. M. Matsuoka at al, J.Met.Finish.Soc. of Japan, 38,(2),19,(1987).
15. M. Matsuoka at al, P1at.Surf.Finish. 5,102,(1988).
16. C.D. lacovangelo, K.P. Zarnoch, J.€lectrochem.Soc., 138,(4),983,( 1991 1.
17. Ibid, p.976.
18. A. Hungjord, K. E. Chen, J.€lectrochem.Soc.,1,72,(1989).
14 836
28
27 20 2s 24 23 22 21
20 13 10 17 16 1s 94
13 12 11 10
9 0 7 0 S
4 3 2 1
150 200 250 0 so loo NO. OF CONTACTS
Raure 1. " M A R OF CONTACTS PLATED VS. THICKNESS
t 1.50
1.34
1.25
(w
1.00
.90
.EO
.70
.60
.50
.40
3 0
.20
.10
O\
30 IMW 15 20 25 6 10 0
Figure 3. PLATING TIME VS. RESIDUAL GOLD
I
I G N
838