\ mm \H H1

  

\

Iooo
ova—x

 

J—INA

 

1% ma or mums ON
~- ‘ »m my. WITH 595cm.

 
 

 

This is to certify that the
thesis entitled

Tfl EFFECTS OF IMPURITIES ON ELECTRODEPOSITED NICKEL

WITH SPECIAL REFERENCE TO CALCIUM

presented by

Ralph Albert Bacon

has been accepted towards fulfillment

of the requirements for

degree in

Doctor of Philosophy; ./'

I" ,1 I
/ _
I .- /

Major professor

  

Date May 172 1957

THE EFFECTS OF IMPURITIES ON ELECTRODEPOSITED

NIGEL WITH SPECIAL REFERENCE TO CALCIUM

By
RALPH A? BACON

A THESIS

Submitted to the School of Advanced Graduate Studies of
Michigan State University of.Agriculture and Applied
Science in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Metallurgical Engineering

1957

[,1-3 is 57’

I? a? . 1' A
I , ' 2'"
tr” / 2/ f/ g.--”

ABSTRACT

Four different types of buffered nickel sulfate plating solutions
were investigated with respect to the quality of deposit produced
‘when contaminated'with various metallic impurities. The physical
properties investigated were: appearance, adhesion, ductility, hard-
ness, throwing power, and corrosion resistance. Calcium when present
as an impurity in an amount less than the solubility limit had
negligible effect on the above properties. Methods studied fbr the
removal of calcium.from.these solutions included high pH precipitation,
electrolytic removal, high temperature precipitation, and a precipi-
tation with hydrofluoric acid. The roughness produced by copper
impurity was investigated by metallographic techniques, chemical
analysis, current density variations,'base metal changes and

polarization considerations.

ii

 

TABLE OF CONTENTS

INTRODUCTION...................................................
EXPERIMENTAL...................................................
EFFECTS OF CALCIUM AS AN IMPURITY IN NICKEL PLATING SOLUTIONS..
REMOVAL OF CALCIUM FROM NICKEL PLATING SOLUTIONS...............
DEPOSIT ROUGHNESS AS A RESULT OF METALLIC IMPURITIES...........
CONCLUSIONS............................................ ...... ..

WEIJCESOCOOOOOOOOOOOO ..... 0.0...0.0000000000000000.000......

iii

Page

11

26

28

63

 

INTRODUCTION

The metal nickel possesses a fortunate combination of physical
and chemical properties which.have established it as one of the most
important metallic coatings used for protective and decorative purposes.
It exhibits relatively high strength, ductility and hardness while at
the same time displaying a high resistance to corrosion. The favorable
properties of nickel together with the deve10pment of coating techniques
have lead to a.marked increase in the use of nickel coatings in recent
years. This is particularly true of the electroplating industry which
experienced wide expansion following the development of "bright nickel"
plating. A further advantage is the close control and.wide range of
thickness that can be obtained.

The principal limitation of nickel as a coating metal lies in its
electropositive relationship toward iron in most environments. Deposit
discontinuities are highly undesirable in the case of cathodic coatings
of this type where exposed base metal becomes anodic and dissolves away.
In fact the base metal which is exposed through discontinuities in a
cathodic coating can deteriorate at a much faster rate than if the
coating were not present at all. This is because of the localized
electrochemical cell which is set up in the presence of corrosive
electrolytes. Therefore, cathodic coatings must be as continuous as
possible to obtain a maximum in corrosion protection.

The complexity of present day nickel plating baths as well as the

variety of uses to which electrodeposited nickel is applied has prompted

 

extensive research concerning the variations in the physical and
mechanical properties of the nickel brought about by changes in bath
composition, temperature, pH, current density, and level of impurities.
The detrimental effects of certain metallic impurities has long been
known; however, previous to this research project no systematic study
of the effects of individual impurities as well as of combinations of
these impurities has been undertaken. Impurities find their way into
nickel plating baths in a number of ways of which "drag in", impure
nickel salts and a contaminated water supply are the chief contributors.
Sponsored by a fellowship grant from the American Electroplaters
Society a series of metallic impurities was investigated with respect
to their individual affect upon the appearance, adhesion, ductility,
hardness, throwing power, and corrosion resistance of the electrode-
posited nickel. These properties are particularly affected when the
impurity either co-deposits or is occluded in some manner in the
nickel deposit. The results of these investigations were reported
merely on a relative basis, that is, the deposits from solutions of
controlled impurity were compared with deposits obtained from a highly
purified bath and any variations in a certain property were noted.
Some very surprising phenomena in regard to the behavior of different
metals electrodepositing simultaneously were revealed in these studies.
It would normally be expected that two metals as similar to each other
as nickel and copper being completely soluble in all proportions,
having the same crystal habit, nearly the same lattice constants,
and the same valence, would be deposited together with little or no

interferences. Such was not the case, however, as the nickel deposits

 

containing trace amounts of copper were rough,'black, and completely
unacceptable from.a corrosion resistance standpoint. On the other
hand lead, an element far removed from nickel in the periodic listing
of the elements and whose lattice constants vary considerably from
those of nickel, provides a general brightening effect on both the
'Watt's type and bright nickel deposits. Also zinc, which crystallizes
in the hexagonal lattice system, When added to the nickel baths as

an impurity produced a slightly brighter deposit whose corrosion
resistance showed a general increase with increasing concentrations

of zinc. Other elements investigated such as aluminum, iron, manganese,
and chromium although differing from.nickel to varying degrees
produced no significant changes in the nickel deposit. Since no
attempt was made in any of the previous reports on impurity studies

to explain these rather unexpected results it was undertaken to search
for further evidences of these peculiarities and thereby to be able

to clarify and explain the actual mechanisms involved. It was

decided to carry out a thorough investigation of the element calcium
which is known to be present in nickel plating baths and is known

to have adverse affects on the electrodeposit under certain conditions.
The principal source of calcium.in a nickel plating bath is the use

of hard water in the make-up of the bath.

In spite of the common occurence of calcium as an impurity in
plating solutions, very little work has been performed in investigating
its effects. A comprehensive review of the literature has revealed
that the only case of an intentional addition of calcium to a nickel

plating bath dates back to 1916. In that year B. E. Miller (1)

advocated the use of calcium chloride as a conducting salt. It was
subsequently shown that the action of the calcium chloride was to
precipitate calcium sulfate, thereby replacing the sulfate ions with
chloride ions. To accomplish this same and more directly Robins (2)
advised the use of nickel chloride. No further reference to calcium
appeared until 1939 when Meyer (3) pointed out that through the use of
hard water rough deposits resulted due to the suspended calcium sulfate.
About this same period Hogaboom (h) also reported on the detrimental
effects of hard water on the nickel deposit. In l9hl Diggin (5) reported
the results of his investigations on the effect of various water
impurities on nickel deposits. He found that calcium itself appeared
to have little effect, but the formation of calcium sulfate caused
bright nickel to plate dull and spotted. These conclusions were
corroborated by further work by Diggin (6) and also by Kushner (7) who
studied impurity build up from drag-in of rinse water. Since that

time little further has appeared on this subject.

Any possible existing correlation between the surface roughness
caused by calcium and that caused by copper was investigated with
special emphasis being placed on the mechanism of formation of the
irregularities produced by copper. The report on the effects of
copper impurity on electrodeposited nickel (8) indicates that amounts
of copper in excess of 50 mg/l produce a deposit which tends to be
rough and dark. Also shown is the drastic drop in corrosion resistance
with the addition of trace amounts of copper. As far back as 1911
(9) a definite darkening of the electrodeposit was observed when

copper was added in small amounts to the nickel solution. Some time

 

later investigations carried out at the National Bureau of Standards
(10) (11) again revealed the detrimental effect of copper which produced
a dark, spongy deposit. These results were later confirmed by
Eckelmann(12), Johnson (13), and Francis-Carter (1h). Up to the present
time the reason for this type of deposit has not been explained. The
exploratory procedures utilized in this investigation for the purpose
of discovering the actual physical phenomenon responsible for this
undesirable electrodeposit include: a cross-sectioning of the spongy
deposit which made it possible to take photomicrographs of the
structure; spectographic analysis of the spongy nodules and of the
foil; current density tests utilizing the Hull cell; voltage versus
amperage characteristics at various copper levels; and buffing and
plating experiments.

This work was supported by the American Electroplaters' Society
by means of a fellowship grant, and the author is deeply indebted to
this society for this necessary financial assistance. Also the
technical and experimental assistance received from the project
committee; Mr. C. Nixon, Mr. L. Borchert, and Mr. L. Morse is greatly
appreciated. Finally my most sincere thanks to Dr. A. J. Smith for

his invaluable guidance and encouragement throughout this work.

 

 

EXPERIMENTAL

In this study nickel deposits were prepared from four different
types of plating bath (Table I). The Watts'-type nickel plating
solutions (at pH 2.2 and 5.2) were employed to produce dull nickel
deposits. The bright nickel deposits were obtained from a nickel-
cobalt alloy type bath (pH 3.75) and from an organic type solution
(pH 3.2) containing nickel benzene-disulfonate and triaminotoly-
diphenylmethane. Very small amounts of an antipitting agent in the
form of sodium lauryl sulfate were added to all but the nickel-cobalt
bath. The methods of preparation and purification of these solutions
are outlined in detail in an article resulting from early work on this
project (15). It was first attempted to detennine the approximate
solubility of calcium at normal plating temperatures. Calcium was
added as calcium carbonate in considerable excess. After equilibrium
had been reached at the proper pH, and settling was completed, the
clear solutions were withdrawn and analyzed. The following appears

to be the limit of calcium solubility in the respective solutions.

Watts' 2.2 pH 1000 + ppm
Watts' 5.2 pH 700 ppm
Organic 3.2 pH 1000 + ppm
Nickel-Cobalt 3.75 pH 1000 + ppm

It was determined (16) that on an average, commercial plating
baths contained a maximum calcium content of 700-750 ppm, while the
average concentration was approximately 390 ppm. With these figures

as a guide, and since the consensus of opinion was that the calcium

 

ion was relatively harmless it was decided to limit the experiments
to three levels of calcium concentration; a pure bath (0 ppm Ca), an
average calcium level (390 ppm Ca), and a level corresponding to the
upper limit feund in normal plating solutions (700 ppm Ca). The
calcium was added in the fOrm of calcium carbonate so as not to alter
the sulfate-chloride ratio of the individual baths. Following these
additions the pH was adjusted back to the normal operating range with
sulfuric acid. Panels of four different thicknesses were plated from
each bath and at each level of calcium concentration. The actual
plating process, including the cleaning of the panels,has been described
in detail in earlier publications from this project (17) (18) (19).
The thicknesses were chosen with regard to the ultimate tests to be run.

The evaluation of the deposits obtained was accomplished by a
quantitative comparison of the deposits from solutions containing
varying amounts of calcium with panels produced from pure solutions.
Deviations were reported as percent change from the properties of the
pure deposit, since in this case trends were the objective instead of
absolute data.

The analytical method originally employed for the determination
of calcium was the colorimetric procedure outlined by Serfass (20) in
which the nickel is separated from the calcium in the form of a nickel
hydroxide precipitate. The calcium is then precipitated as the oxalate
which is filtered and then dissolved in sulfuric acid. The color is
subsequently developed by the addition of a standard potassium perman-
ganate solution, the degree of color being proportional to the amount

of calcium oxalate present. The calorimeter readings are then referred

TABLE I

NICKEL PLATING SOLUTIONS

 

 

 

Watts' Organic Nickel-
Type Cobalt

Nickel Sulfate 21.0 g/l 262 g/l 21.0 g/l
Nickel Chloride ES 60 hS
Boric Acid 30 3h 30
Nickel Formate LS
Cobalt Sulfate 15
Nickel Benzene Disulfonate 7.5
Triaminotolyldiphenylmethane 0.1L ml/l
Temperature °F 55 60 60
pH (electrometric) 2.2 and 5.2 3.2 3.75

Current density
(amperes per square foot) hO ho ho

 

to a standard curve which reveals the calcium content. This procedure
which involved several filtrations and the evaporation of a considerable
amount of liquid proved to be very tedious and time consuming. Conse-
quently it was endeavored to accomplish these determinations by a much
more direct and equally reliable means; namely, a direct spectrographic
determination of the calcium in a measured quantity of plating bath.

In setting up this spectrographic procedure the usual spectro-
graphic carbons were not used for two reasons. First, it was
determined that even the most highly purified type of carbons still
contained a certain measureable amount of calcium and secondly, the
bath aliquots when placed on the carbons tended to soak in nonuniformly,
thus giving erratic results. Pure copper electrodes were selected
and proved to be satisfactory. A simple machining operation on a
lathe cleaned the surface and readied the electrodes for use. The
procedure consisted of sampling the bath and by means of a micro
pipette placing .025 ml. on the flat copper electrodes. By placing
the electrodes with sample under an infra-red lamp the solution was
dried in a matter of minutes.

The means of excitation chosen was the a.c. spark rather than the
d.c. arc, due to its more uniform nature in energizing the entire
surface of the electrode and also the fact that the excitation energy
derived from the spark is more than adequate to excite the calcium
and nickel atoms in question. The internal standard method was
employed in order to correct for any variations in the time of exposure,
plate emulsion characteristics, or developing conditions. It was

decided to use a weak nickel line for this purpose rather than adding

 

 

10

an additional constituent with excitation characteristics similar to
calcium. An example of this would be strontium; however, in this case
an insoluble strontium sulfate precipitate would form rendering it
useless. The exposure time was thirty seconds, a period which allowed
the sample to be completely burned, but at the same time maintained a
negligible level of background in the spectrum.

Standards were made up by taking five portions of a pure nickel
plating bath and introducing calcium in amounts increasing from twenty
to one thousand mg./l. The standard solutions were run along with
every series of unknowns and from them a standard curve was plotted.

The unknown microphotometer readings were then referred to this curve

 

for the final analysis. The analysis of samples by both the color-
imetric and the spectrographic procedures established that the spectro-
graphic method resulted in an equal or greater degree of precision as
well as being much more rapid. The instruments used for these
investigations were a Hilger large littrow type spectrograph and a

Jarrell-Ash Recording Microphotometer.

EFFECTS OF CALCIUM AS AN IMPURITY IN NICKEL PLATING SOLUTIONS

Appearance:

For the evaluation of appearance with increasing amounts of
calcium in the bath, panels of three different thicknesses were plated;
namely, .003, .0010, and .0015 inches. The steel panels themselves
were bent so that a vertical and a horizontal surface was available
for plating. The vertical section had an overall current density
range of hO amperes per square foot. The horizontal section posseSSed
low, medium, and high current density areas ranging in value from 10
to approximately 60 amperes per square foot. A cathode of this type
permitted an additional evaluation, that is the effect of current
density on the appearance.

The actual evaluation was kept as uniform as possible by maintain-
ing the same relative position of the diffuse light source, the
illuminated panel and the line of sight of the observer. The gray
deposits were compared with the Eastman Gray Scale, while the bright
deposits were classified as ranging from dull to mirror bright. In
each case any change in appearance was compared to the deposit produced
from a pure solution.

Both the Watts' pH 2.2 and 5.2 baths Were judged to have an
Eastman Gray Scale reading of 1-2 and there was no change in scale
reading with increasing calcium content. The roughness of the panel
lips (horizontal section) increased slightly with increasing calcium

content. The increased calcium levels produced no appreciable change

 

\

 

 

 

 

 

12

in grain size or treeing effect at the panel edges. No change was

noted in the appearance of the organic or nickel-cobalt bright deposits
with increasing calcium.content with the exception of a slight increase
in roughness at the higher calcium levels. These panels were all rated

as mirrorAbright.

Adhesion:

In a general sense the adhesion of electrodeposited coatings to
their base metal depends primarily upon their intimate contact.
Consequently the presence of oxides, greases, or other extraneous
material will naturally reduce the adhesion. Special care was exer-
cised in cleaning the steel panels in order to maintain as nearly as
possible a uniform, impurity free surface. The cleaning procedure
is outlined in an earlier report from this project (15).

To test the affect of calcium on the adhesion of the nickel plated
foil to the base metal a bend test was selected, since in this case the
plated specimens were easily bent and destruction of the panels was not
a consideration. The test consisted of bending the lip of the cathode,
upper side outwards, 180 degrees around a 3/16 inch.mandrel in the middle
and across the short dimension (21). For this test only the lips from
the 0.001 inch series of panels were used. The bend was examined under
a microscope at 20 diameters for any evidence of cracking or flaking as
well as any correlation between this property and the thickness of the
nickel deposit as revealed from the front to the back of the lip surface.
In this study, at all concentrations used, the calcium did not discernibly

affect the adhesion of any deposit to the base metal.

13

Ductility:

Deposits of .001 inch thickness were made from the four baths at
the designated calcium concentration levels. These deposits were
made upon a passivated nickel surface so that they could be stripped
in the ferm of a nickel foil. The ductility test which was then
carried out on this foil consisted of repeatedly bending, creasing
and then flattening out the deposit across a predesignated area with
the fingers until a break was noted. The number of times that the foil
can be creased and opened before fracture is a measure of ductility.
The fact that the number of bends possible with these nickel foils is
small places a definite limitation on the reliability of the data. To
help minimize any human error each foil was sectioned and given to
several individuals to test. The results of the tests are compiled
in Table II and are reported as percent change from the ductility
value of the pure deposits. The Watts' 2.2 bath showed no appreciable
change in ductility over the entire range of calcium.impurity. The
“watts' 5.2 bath while exhibiting the identical degree of ductility as
the'Watts' 2.2 bath at a zero calcium level showed a gradual decrease
in ductility with increasing calcium content.

The foils from the pure organic bath exhibited a much lower degree
of ductility than.the foils from either of the pure Watts' baths, but
the addition of calcium to the bath had little effect on the ductility.
The foils from the N100 bath were by far the least ductile of all the
experimental foils. In this case the ductility was so low that the
test was rendered inaccurate; however, there appeared to be no significant

change in the ductility with increased amounts of calcium.

TABLE II
EFFECT OF CALCIUM ON THE DUCTILITY

OF ELECTRODEPOS ITED NICKEL

 

Percent Change

Calcium concentration Watts' Watts' Organic Nickel-Cobalt

 

0 0 0 0 0
350 0 -37 -16 50
700 9 -b 5. 7 -12 SO

 

15

Hardness:

Electrodeposited nickel may be obtained in hardnesses over a range
of about 500 Brinell. These differences in hardness depend upon com-
position and structure and are Obtained by adjustment of'bath composition
and plating conditions.

To obtain a true indication of hardness, consideration.must be given
to the thickness of the nickel deposit to insure against the effect of a
comparatively soft base metal. In order to eliminate this error as much
as possible a relatively thick nickel deposit (.002 in.) was plated from
each of the baths. In this way the influence of the base metal was
diminished and at the same time the deposit thickness was not so large
as to alter the properties. Several instruments have been developed
for measuring hardness of relatively thin deposits by virtue of their
ability to employ light loads and measure accurately the degree of
impression. These instruments measure the length of indentation or
depth of penetration of a diamond tip under various degrees of pressure.
The hardness measurements for this project were carried out by the
National Bureau of Standards with a Tukon hardness tester employing
Knoop type of diamond indentor.

The results which are tabulated in Table III reveal only slight
changes in hardness with increasing calcium in the baths. Although
the‘Watts' 2.2 bath shows a slight increase and the Organic bath a slight
decrease in hardness the changes are not significant and no definite
trend is established. As expected the bright deposits show a much
higher hardness than the Natts' dull nickel, with the N100 deposits

possessing the highest hardness.

16

TABLE III

EFFECT OF CALCIUM ON HARDNESS

OF ELECTRODEPOSITED NICKEL

 

Percent Change

 

 

Calcium concentration ‘Watts' ‘Watts' Organic Nickel-Cdbalt
mg/l pH 2.2 pH 5.2 pH 3.2 pH 3.75
0 0 0 0 0
350 1.6 5.5 -5.1 -6.1
700 10.1 -ll.7 -7.6 3.0

 

l7

Throwing Power:

It is usually desirable that the deposited metal be distributed
uniformly over the surface to be covered even though this surface may
be very irregular. In order to accomplish this purpose, solutions of
rather complex composition are often employed that are particularly
effective in depositing the metal in recessed areas, thus exhibiting
what is known as high "throwing-power".

Throwing power was evaluated by a procedure outlined in an earlier
paper from this project (15). The method consists of sectioning the
horizontal lip of the bent cathodes which were plated to a.thickness of
.002 in and after clamping, polishing, and etching, the thickness of the
electrodeposited nickel relative to the pure deposit was measured by
means of a Bausch and Lomb research metallographic microscope. As
described in another project paper, the throwing power measurements
were performed on'both the top and bottom of the lip to minimize the
variation in the results due to a non~horizontal cathode lip during
the plating (22). The distribution of electrodeposited metal depends
on the distribution of current and the respective cathode efficiencies
at the prevailing current densities. Since no gassing was noted in
the concentration ranges studied as long as plating conditions were
controlled, the efficiency was assumed to be unchanged. Therefore,
any change in the deposit thickness was assumed to be due to a change
in the throwing power of the solution.

The results of this test which are summarized in Table IV indicate
an overall reduction in throwing power with increasing calcium content.

However, the measurement itself involves an error of approximately MO

18

TABLE IV
EFFECT OF CALCIUM ON THE THRONING PCMER

OF ELECTRODEPOSITED NICKEL

 

Percent Change

 

Calcium concentration Watts' Watts' Organic Nickel-Cobalt

 

mg/l pH 2.2 pH 5.2 pH 3.2 PH 3.75
o o o o o
350 -005 "loo “703 ”1.308

700 -1005 -500 ‘803 -203

 

19

per cent and so the variations are not considered significant and no
clear cut trend is established.
Corrosion Resistance:

Undoubtedly the primary reason for the electrodeposition of one
metal on another, other than decorative purposes, is to eliminate or
at least hinder the destruction or deterioration of the metal by direct
chemical or electrochemical reaction with its environment. To accomp-
lish this protection of steel with electrodeposited nickel the coating
must be continuous and completely cover the basis metal. In order to
evaluate the protective qualities of nickel.deposited in the presence
of increasing amounts of calcium it was necessary to employ some form
of accelerated corrosion test that would duplicate as nearly as possible
normal service corrosion. Employed was the most widely used test at
the present time, the salt spray (fog) test, the conditions of which
are outlined in the A.S.T.M. Tentative Method of Salt Spray (fog) Testing
Bll7-h9T. This test which was designed as a comprehensive accelerated
corrosion test has been the subject of a considerable amount of con-
troversy and is now undergoing intensive efforts at standardization.
Because of this almost universal skepticism.concerning the results of
the salt spray test it was deemed necessary to substantiate the results
by employing other corrosion tests on similar test panels. The acetic
acid modification of the salt spray which incorporates a lower operating
temperature and a lower salt concentration was employed. Also used
was a recently developed electrochemical test developed by Pierce and
Pinner (23) in which the nickel coating is the anode and a copper wire

the cathode in a cell which utilizes sodium chloride as the electrolyte.

2O

Triplicate panels of 0.0010 inch thickness were prepared from each
solution to be tested. These were placed in the salt spray (fog)
cabinet and allowed to remain until breakdown was judged to be complete.
Although there was an obvious difference between the corrosion resist-
ance of the bright and dull deposits, the bright deposits lastingtwice
as long as the dull deposits, there was no indication as to the effect
of increasing calcium content on corrosion resistance of the nickel.
The period of time required for failure of the panels varied greatly
within each level of calcium concentration, thus making it impossible
to establish any kind of trend. The type of corrosion encountered
nearly always involved three or four spots which gradually grew until
the panel had failed. These spots were very likely strongly anodic
and were protecting the remaining nickel surface. In a few instances
the panels developed more of an overall corrosion pattern with a large
number of rust spots appearing. This type of corrosion would tend to
indicate a more porous nickel structure. In cases where the panels
lasted for 120 hours the corrosion seemed to come to a standstill and
further cabinet time would have been futile. The water insoluble gray
film mentioned in previous papers (22) was evident and perhaps was the
protective agent preventing further corrosion.

Through the cooperation of the General Motors Research Laboratory
it was possible to test a series of panels in their acetic acid salt
spray chamber. For these tests a series of twelve panels of 0.0010
inch thickness were plated from each of the test solutions. The tests
were carried out under the direction of Mr. C. F. Nixon according to

the specifications outlined in A.S.T.H. B287-5hT which essentially call

21

for a 5 percent sodium chloride solution, pH 3.h with acetic acid, at
35°C. The panels were exposed for eight 2h-hour periods and following
each period the panels were rinsed with cold running water and inspected.
A rating number was assigned to each panel at each interval and the
findings were reported in terms of Performance Index numbers.

The rating scheme was as follows:

Amount of Failure Rating
None ------ no failure . 5
Slight --~— 1 to 10% of significant surface affected h
Moderate -- 10 to 30% of significant surface affected 3
Severe ---— 30 to 70% of significant surface affected 2
'Total ----- over 70% of significant surface affected 1
The Performance Index Number was based on these ratings for each of

the eight inspections, as shown in an example below.

Hours 2b L8 72 96 120 lhh 168 192

Rating 5 h b h 3 3 2 l

The sum of the ratings (26) is multiplied by an arbitrary constant 2.5.
The product is by definition the Performance Index Number. The possible
values range from a minimum of 20 to a maximum of 100. Number 20 repre-
sents a part which has totally failed in 2b hours and number 100, a part
which shows no evidence of corrosion at the conclusion of 192 hours of
testing.

The results are tabulated in Table VI and each value represents
the average of twelve Performance Index Numbers. Here the results

of the Salt Spray Test are substantiated in that there is revealed no

 

22

TABLE V
EFFECT OF CALCIUM ON THE CORROSION RESISTANCE
OF ELECTRODEPOSITED NICKEL AS DETERMINED

BY THE SALT SPRAY (FOG) TEST

 

Average Number of Hours to Produce Failure

 

Calcium concentration Watts ' Watts ' Organic Nickel-Cobalt
l
0 b1 55 103 113 r! A
350 72 148 79 103 A‘

700 L1 55 79 113

 

23

evidence of either an increase or decrease in corrosion resistance with
increasing calcium in the plating bath. Again the NiCo bright deposit
exhibited the greatest corrosion resistance.

As an additional check on the validity of these corrosion resistance
results a third corrosion test was utilized, namely, the electrochemical
method of Pierce and Pinner (23). This approach to the problem of find-
ing a positive means for evaluating electrodeposits with respect to their
corrosion resistance was based on some original work by Shome and Evans
(2h). These investigators developed a cell consisting of a copper
electrode, a thinly nickel plated steel electrode and a sodium chloride
electrolyte. The potential generated by the Shome and Evans cell was
found to be approximately 0.2 volts. In the Pierce and Pinner corrosion
cell which is described in detail in an earlier paper (25), the potential
was increased to 0.3 volts by the addition of a battery powered potential
source in order to increase the sensitivity when applied to thicker
deposits.

Electrochemical corrosion tests were carried out on duplicate panels
of 0.0003, 0.0010, and 0.0015 inch nickel thickness frOm each of the four
plating baths and at each calcium concentration. The thin (0.0003 inch)
deposit was tested for two hours, the 0.0010 inch deposit for three
hours, and the heavy (0.0015 inch) deposit for four hours.

In Table VII the figures representing the number of corrosion
spots are the average of both panels. The results here proved to be
very erratic and no significant change in corrosion resistance was

indicated.

 

EFFECT OF CALCIUM ON THE CORRCASION RESISTANCE

TABLE VI

OF ELECTRODEPOSITED NICKEL AS DETERMINED

IN THE ACETIC ACID SALT SPRAY (FOG)

2b

 

 

Calcium Performance Index
Concrelrglpiation Average Range Egggggn
Watts' 2.2 pH 0 57.7 117.542.; 6.9
350 56.3 50.0-62.5 h.5
700 56.7 50.0-62.5 3.9
Watts' 5.2 pH 0 55.8 h7.5—65.0 5.6
350 57.5 has-65.0 6.5
700 55.1: h2.5-67.5 7.8
Organic 3.2 pH 0 53.8 15.0—65.0 5.5
350 52.3 has-70.0 8.8
700 55.6 h2.5-62.5 5.7
NiCo 3.75 pH 0 59.2 50.0-65.0 um
350 61.7 52.5-72.5 5.9
700 63.1 55.0-72.0 14.9

 

25

TABLE VII
EFFECT OF CALCIUM ON THE ELECTROCHEMICAL CORROSION

RESISTANCE OF ELECTRODEPOSITED NICKEL

 

 

Calcium Average Number of Corrosion Spots ~
Concentration (2 hours) (3 hours) ()1 hours) '
mg/l 0.0003 in. 0.0010 in. 0.0015 in.
Watts' 2.2 pH 0 7.5 7 3
350 In 7.5 6
700 8.5 11.5 11
Watts' 5.2 pH 0 6.5 11.5 3
' 350 12 8.5 13
700 10.5 19 7.5
Organic 3.2 pH 0 9 18 5.5
350 5 13 5
700 2.5 16 13
NiCo 3.75 pH 0 1|.5 6.5 12
350 2 9 9

700 h 5.5 3.5

 

26

REMOVAL OF CALCIUM FROM NICKEL SOLUTIONS

The removal of calcium was attempted by several methods, including
low current density electrolysis, high pH precipitation, high tempera-
ture removal, and a chemical procedure involving treatment with
hydrofluoric acid. The procedures for both the electrolytic and high
pH removal were described in a previous publication by Ewing, Rominski
and King (15).

Standard operating conditions of pH, temperature, and rate of
agitation were maintained in the study of the electrolytic removal
of calcium. Also care was taken to use nickel plated cathodes since
it had been shown in an.earlier investigation that at low current
densities the steel cathode was not covered quickly enough to prevent
the dissolution of iron and subsequent contamination of the solution.
The current densities investigated were 2, 5, 10, and hO amperes per
square foot for each of the four plating baths containing 600 mg/l
of calcium. The solutions were sampled periodically and.analyzed
for the calcium content by means of the spectrographic method outlined
earlier. The results indicated no removal of calcium at either the
low current densities or at the operating level of to amperes per
square foot. An additional check on these results was possible through
a spectrographic analysis of the dummy cathodes used in the electrolysis.
These electrodes revealed no calcium being deposited, thus agreeing
with the analysis of the plating baths before and after electrolysis.

The removal of calcium by high pH precipitation was attempted by

the addition of nickel carbonate to the baths containing 600 mg/l of

27

calcium until the pH reached the 5.6-6.0 range. The baths were held
at this pH for 72 hours with samples being taken periodically for
spectrographic analysis. The results of these analyses indicated that
no calcium was being removed by this treatment.

The effect of increasing temperature on the calcium level was
investigated by sampling a Watts' plating bath containing 600 mg/l of
calcium ion at temperatures ranging from 25 to 70 degrees centigrade.
The bath was allowed to reach equilibrium at each temperature level
before sampling. The results of spectrographic analysis revealed
that this range of temperature had no effect on the concentration of
calcium at this concentration level.

An unpublished procedure for calcium removal involving a precipi-
tation of calcium fluoride with hydrofluoric acid was investigated.
The method consisted of heating the contaminate nickel solution to a
temperature of 170°F and in another container heating a dilute solution
of hydrofluoric acid to the same temperature. The nickel solution was
then transferred slowly and with stirring into the acid container. In
theory calcium and magnesium fluorides should form at this point and
the precipitate could be filtered off. Several concentrations of
hydrofluoric acid were employed in these experiments but in no case
did this technique result in the removal of calcium from the plating

solution.

 

28

DEPOSIT ROUGHNESS AS A RESULT OF METALLIC IMPURITIES

Reviewing the entire study concerning the effects of metallic
impurities on electrodeposited nickel which included investigations
into the effects of copper, iron, zinc, chromium (+3 and +6), lead,
aluminum, manganese, and calcium, it is observed that one of the most
if not the most incompatible metal is copper. Copper when present in
a nickel plating solution in an amount usually considered as a trace
ruins the appearance of the nickel by producing a deposit which is

rough and black (25). Also when calcium is present to an extent

 

exceeding 500 mg/l a certain roughness is evident. Any correlation
between these two rough deposits could most certainly aid in the
explanation of the mode of formation of rough, undesirable deposits.
A definite loss in ductility of the electrodeposited nickel is also
detected when copper ions are present in the plating solution. In
addition the detrimental effects of copper are made manifest in the
commercially important physical property referred to as corrosion
resistance. The nickel deposits showed a definite loss of corrosion
resistance, a loss which increased with increasing copper content.

The question arises as to the reason underlying these detrimental
effects of copper. Why should two metals which are completely soluble
in one another in all proportions behave in such an illogical manner
when they are both forced to plate out together from an electroplating
bath? Examination of the crystalline habits of each metal reveals

that they both exist with a face-centered cubic lattice. The ao

29

values for copper and nickel as reported in the A.S.M. Metals Handbook
are 3.6080 and 3.5167 respectively and the closest approach of atoms is
2.551 for copper and 2.h86 for nickel. The very close similarity of
these two metals in regard to their crystalline structure almost
certainly precludes the possibility of a lattipe disturbance of any
consequence.

Considering the decrease in corrosion resistance of electrodeposited
nickel brought about by copper ions in the plating solution it may be
well to first discuss the corrosion resistant properties of a pure nickel
deposit. It is well known that electrodeposited nickel is limited in
its corrosion protection properties by pores and imperfections in the
'plate due to several factors some of which are surface defects in the
base metal, solids carried over in the plating bath, and the deposition
of hydrogen gas along with the nickel. Thickness is the most important
factor.in controlling the porosity of nickel coatings. The number of
pores per square decimeter below a deposit thickness of .0005 inches
increases at an extremely rapid rate. The primary cause of this initial
gross porosity in a nickel deposit is undoubtedly related to the
hydrogen bubble formation which commences with the deposition of nickel.
If these bubbles are not dislodged by the action of surface tension,
agitation, or other disturbing factors, they will tend to grow at
least as rapidly as the deposit and will prevent a closing up of the
cavity. The result of this type of growth will be a pit which is
visible at the surface. Figures I and II illustrate this type of fault
in electrodeposited nickel.

With these facts in mind we may now consider the effects that

 

copper may possibly have upon this mechanism. The first approach
to the problem was a study of the cross-section of these deposits
to determine the nature and mode of fonnation of the rough, spongy
structure. Eighty milligrams per liter of copper in the form of
CuSOh-SHZO was added to a nickel-cobalt bath which was known to be
plating bright. Panels were plated from this impure bath under normal
operating conditions with the periodic addition of copper solution in
order to maintain the impurity level at 80 mg/l (25). These panels were
sectioned perpendicular to their surface and mounted in bakelite molds
to allow polishing and etching of the nickel structure. The etching
was carried out in two steps (26) the first being for the purpose
of removing the coldaworked surface. For this a mixture of one
volume concentrated nitric acid, two volumes concentrated hydrochloric
acid and two volumes glycerine was used. To reveal the grain structure,
inclusions, and voids the surface was subsequently etched with a mixture
of equal volumes of concentrated nitric acid and glacial acetic acid.
This roughness which develops as a result of copper impurity
in the bath is apparently the result of a very unusual deposition
phenomenon. Upon close examination this roughened surface is revealed
to be a collection of individual nodules growing out perpendicular to
the deposit surface. These nodules are seen to originate deep in the
nickel deposit near the base metal if not actually at the base metal.
As the deposition continues these nodules appear to grow at an ever
increasing rate and the indication is that they become self propogating
and branch out to an extent far exceeding any build up attributable

to ordinary deposition. Illustrations of this unusual development are

 

 

31

FIGURE I

 

A cross-sectioned nickel electrodeposit showing
a cavity originating at the base metal and continuing
through the nickel deposit.

Magnification: 800x
Etchant: 1:1 conc. nitric acid and glacial
acetic acid.

 

32

FIGURE II

 

 

 

 

 

A cavity in the nickel deposit which may or may
not have originated at the base metal but which has con-
tinued through the nickel and has resulted in the formation
of a nodule type imperfection.

Magnification: 800x
Etchant: 1:1 conc. nitric acid and glacial
acetic acid.

33

FIGURE III

 

 

 

The crane-section of a typical nodular sanctum
which illustrates the annual growth phenomem. weight
is the typical bright nickel striated structure and how it
has been forced awards into the mule base.

Magnification: 000x
Etchant: 1:1 cons. nitric acid and glacial
acetic acid.

3b

shown in Figures III through VI. Figure III reveals this unusual
growth phenomenon and it is seen that the size of the nodule is far
out of proportion to the deposited layer. The position of the true
origin of this particular nodule is not really known. It is very
possible that this particular plane of polish does not reveal a
deeper imperfection running slightly back into the nickel. Likewise
it is equally possible that the original growth pattern has already
been polished away. The slight depression noted in the deposit
surface directly under the nodule is believed to be due to a simple
shielding action, that is, the deposit is shielded from.the anode
and a fresh supply of nickel ions. The characteristic bright nickel
lamellar structure is seen to bend upwards into the nodule and is
even noted well up inside the nodule. Figure IV represents the same
growth area as in Figure III but at a higher magnification. It can
be seen that as the nickel deposit is layed down an inclusion or
obstruction of some sort has caused the normal banded structure to be
bent upwards. This slight distortion is carried on through successive
layers until at approximately three quarters the deposit thickness
additional obstructions are encountered and the banded structure is
forced to assume nearly a vertical direction. From this point on the
growth of the nodule continues at a rate far greater than normal
deposition. Figures V and VI are illustrations of similar nodule
formation.

Figure VII was obtained by polishing a rough panel Specimen in
a direction parallel to the surface, thus in this picture we are

looking directly down upon growing nodules which have been cross-sectioned

35

FIGURE IV

 

at a higher magnification. Clearly visible is the effect
of an occluded mass in forcing the electrodep0sit to bend
A upward and form a nodule.

The same nodule growth as shown in Figure III but

A Magnification: 1700x
A Etchant: 1:1 conc. nitric acid and glacial
1 acetic acid.

 

36

FIGURE V

 

A cross-sectioned nodule in which the central core
and the resulting vertical buildup of deposit are evident.

Magnificationé 800x
Etchant: 1:1 cone. nitric acid and glacial
acetic acid.

 

37

FIGURE VI

 

A cross-sectioned nodule which has developed from
several inclusion sites.

Magnification: 800x
Etchant: 1:1 conc. nitric acid and glacial
acetic acid.

V 38

FIGURE VII

 

A cross-section parallel to the deposit surface
revealing the striated bright nickel structure extending
up into the nodules.

Magnification: l700x
Etchant: 1:1 conc. nitric acid and glacial
acetic acid.

 

39

perpendicular to their principal growth direction. These polished
nodules in the unetched condition revealed no holes or faults; however,
the nodule sites were readily distinguishable by a difference in
metallic luster. A very slight etch, one second in the 93% concen-
trated nitric acid - 50% glacial acetic acid revealed within the
nodule areas voids or regions containing a material other than nickel
immediately below the polished surface. These so called voids
correspond to the dark areas shown in Figures III through VI.

An explanation of these nodules was sought by considering several
different possibilities for their mode of formation. At the outset
the phenomenon was believed to be a base metal problem. That is,
imperfections in the steel panel whether pits, cracks, protrusions,
impurities, or voids were given prime consideration as the possible
sites of primary disturbance which eventually resulted in nodule
formation. It was further postulated that these abnormalities were
continued through the electrodeposited nickel and by a mechanism
involving the copper ion the observed nodules developed. In an
effort to substantiate or refute this theory by experimentation
several investigations were undertaken in the laboratory. The first
was designed to reduce or eliminate the effects of any base metal
imperfections by initially depositing a smooth, bright nickel layer
on a properly cleaned steel panel. Over this base layer then could
be deposited an additional layer from a bath containing copper impurity
and the degree of nodule formation noted. The first of these panels
was made by plating a .0005 inch nickel deposit from a purified

bright nickel bath and then without interrupting the plating, copper

b0

sulfate solution was added to the bath to bring the copper impurity
level up to 80 P.P.M. Following this addition another .0005 inch

of nickel was deposited. Inspection of the final deposit revealed

it to be rough and black with considerable nodule formation. Therefore
although the steel was first coated with a smooth bright nickel
deposit, the nodule formation resulting from copper ions in the bath
was apparently unaffected. Repeated trials in which the deposit
thickness and copper ion concentration was varied produced results of

a similar nature. This would tend to contradict the theory that the
nodules were originating at base metal imperfections.

Another investigation consisted of altering the base metal surface
by buffing the steel panels on a cotton buffing wheel, the purpose
being to smooth out the surface as much as possible and present to
the nickel ions a comparatively level surface upon which to deposit.
The technique used was to buff one half the panel surface and then
note any variations between the buffed and unbuffed areas after
plating. Repeated trials failed to reveal any discernible difference
in regard to nodule formation between the areas which were buffed and
those which were not. An interesting phenomenon which became evident
in this investigation was that the position of the greatest concentra-
tion of nodules was dependent upon the direction of propeller agitation;
that is, the nodules tended to concentrate along one side of the panel
or the other depending upon the direction of bath rotation. It was
invariably found that the side of the panel upon which the solution
initially impinged as it traveled from the anode end of the rectangular

tank, was the side possessing the majority of nodules. When the

 

bl

direction of propeller rotation was reversed the area of nodule concen-
tration was likewise shifted to the opposite side of the panel.. Other
investigators have also noticed this relationship between solution
movement and position of faulty deposit (10). This phenomenon of the
dependence of the position of nodule concentration upon solution
movement further refutes the base metal theory and strongly suggests
that the underlying cause of this type of undesirable deposit lies in
the solution and not with the base meta.

Further evidence to refute the theory of base metal originating
nodules was obtained by polishing the surface of a rough, black
deposit in a direction nearly parallel to the panel surface. Polishing
in this manner afforded the opportunity of viewing the deposit over
its entire range of thickness. The resulting microstructure after a
suitable etch revealed a relatively pore free initial deposit continuing
to a thickness of between .0002 and .0003 inches.

Although it cannot be said that nodules are not originating at
deposit thicknesses of this order of magnitude or less, certainly
there is little evidence of their formation when viewed at a magnifi-
cation of 800x. Figure VIII is a photomicrograph of a portion of this
tapered section at this magnification. It shows the bare steel and
the gradual build up of nickel deposit to a thickness of approximately
two tenthousandths of an inch. The granular nature of the nickel
structure renders the detection of imperfections rather difficult; how-
ever, the uniformity of the surface would indicate the absence of any
nodules at this stage of the deposition. Figures IX, X and XI are

photomicrographs of this same tapered surface but at positions of

h2

increasing deposit thickness. Figure IX represents a thickness
ranging between .0003 and .0005 inch. At this point in the deposition
nodules are very definitely forming and are very much in evidence at
this magnification. The increasing number of nodule sites with
increasing deposit thickness is very noticeable. In Figure X the

size and number of the sites have increased considerably and the
deposit, if the plating were ceased at this point would begin to take
on a dark appearance. The deposit thickness in this photomicrograph
increases from approximately'.0006 to .0008 inch. In the final
photomicrograph of the series, Figure XI, the deposit thickness has
reached .0010 inch and the surface is almost completely covered with
this abnormal form of deposit. The amount of metal removed by
polishing from the particular portion of the surface in the lower

half of the photograph was very slight, therefore the structure
represents the basis of the rough, black deposits produced from copper
contaminated nickel solutions. The conclusions that may be drawn from
this series of photographs are that nodule formation does not necessarily
originate at the base metal interface, and secondly that by means of
some ever increasing reaction the frequency of nodule formation
increases with deposit thickness. These photomicrographs, coupled
with the results of the previous experiments already mentioned,
indicate quite definitely that base metal imperfections have very little
to do with the formation of these detrimental nodules. The origin of
these unusual structural growths appear to be located at random
throughout the deposit.

At this point in the investigation it became evident that an

1:3

FIGURE VIII

   

An electrodeposit polished nearly parallel to the
deposit surface showing the steel base metal and the gradual
build up of nickel to .0002 inch. Nodules are not apparent
at this magnification. '

Magnification: 800x
Etchant: 1:1 conc. nitric acid and glacial
acetic acid.

FIGURE IX

 

Electrodeposited nickel polished nearly parallel
to the surface. The deposit thickness ranges from .0003
to .0005 inch with nodule formation evident at the thicker
end.

Magnification: 800x
Etchant: 1:1 cone. nitric acid and glacial
acetic acid.

lab

 

 

145

FIGURE X

s'. .
‘ .‘ift. '
l k
‘fi'm, >
',',,.,.a‘)vy'4_

'. ’ \,

 

Electrodeposit polished nearly parallel to the
surface. The deposit thickness ranges from .0006 to
.0008 inch from top to bottom.

Magnification: BOOX
Etchant: 1:1 conc. nitric acid and glacial
acetic acid.

116

FIGURE XI

 

Electrodeposit polished nearly parallel to the
surface. The deposit thickness increasing from .0008 to
.0010 inch from top to bottom.

Magnification: 800K
Etchant: 1:1 conc. nitric acid and glacial
acetic acid.

117

understanding of the true mechanism involved in nodule formation
would necessitate the evaluation of a number of different aspects

of the plating operation. These factors would include first of all
the chemical analysis of the nodular structure compared with the
analysis of the smooth, bright nickel deposit in order to determine
any possible compositiondifferences. Other influencing factors which
warranted investigation were the effects of current density changes,
polarization effects, and cathode film relationships. Such plating
parameters as solution temperature, ionic concentration, and pH

were held constant during these investigations; however, preliminary

I:
1.
. O

 

tests had previously indicated that moderate fluctuations of these
factors had very little if any effect on the so-called nodule formation.
Composition differences between the smooth, bright deposit and
the rough black areas could be very informative regarding the
mechanism of formation of the rough deposits. Therefore it was
undertaken to prepare a series of these black, rough deposits over a
passivated film layer so that the deposit could be stripped and
divided according to quality of deposit. It will be remembered here
that the black areas were not uniform over the deposit surface but
tended to concentrate on one side or the other depending on the
direction of solution movement. This then afforded the opportunity
of obtaining from a single deposit a normal smooth deposit area and
an area of black nodules for analysis. The nodules were scraped from
the surface so as to isolate them from the normal type of deposit
and thus afford a means of analyzing them separately. The strip

deposits were prepared in a manner much the same as the deposits used

 

b8

for ductility measurements. A .0003 inch, bright deposit was obtained
from a purified bath and this surface was passivated by a fifteen
second alkaline cleaning operation followed by a fifteen second acid
dip in 20% hydrochloric acid. The panel was then placed in a nickel
bath containing 80 mg/l copper ions and a deposit .0010 inch in thick-
ness was produced which was readily removable from the passivated
surface. The black nodules were scraped from the surface to separate
them from the smooth deposit areas. TheSe two sample types, the
nodules and the smooth bright nickel, were then analyzed by the very
sensitive colorimetric method described by Serfass and Levine (27).
According to this procedure, after the dissolution of a weighed

amount of metallic sample, the copper is isolated from the other
constituents by precipitation with 2 - Mercaptobenzothiazole and
extraction of the precipitate with amyl acetate. The color is
developed by addition of dibutyl amine and carbon disulfide. After
the solution has been diluted to 25 ml. with ethyl alcohol its
transmittancy is measured and this value compared to a standard curve
to determine the milligrams of copper present. The results of these
determinations are summarized in Table VIII.

The values presented in Table VIII represent relative amounts of
copper in the normal smooth foil and in the nodules. These milligram
values are not significant in themselves but only the comparison
between the two. Considering the lowest foil value and the highest
nodule figure, there is revealed a moderate increase in copper content
in the nodules. This figure, however, only points out the possible

concentration extremes. A more reasonable comparison would be

 

119

TABLE VIII

RELATIVE AMOUNTS OF COPPER IN SMOOTH FOIL AND IN NODULES

 

ml: Mg Copper Average SMgCu)
Smooth Foil .OOSh

" " .00h9 .00h8

" " .0037

" " .0053
Nodules .0058

“ .0058 .0059

n .0060

 

between the two averages, and in this case the increase in copper
content in the nodules amounts to about 18%. The indication here is
that there is a slight concentrating effect on the copper ions in the
nodules. This effect, however, is not pronounced and it would be
erroneous to draw any definite conclusions on this basis along. If,
however, other experimental data point to a probable concentration of
copper in the nodules these results would offer definite support.
Considering now the various factors affecting plating operations
it is noted that the character of the deposit obtained from a bright

nickel bath depends upon a number of factors namely current density,

 

temperature, pH, concentration of brighteners, concentration of
additives, impurities, and others. Fortunately the variables affect-
ing deposition are not all independent and the effect of nearly all
the above factors will vary with the current density. Because of this
dependence upon current density, more information as to the extent
and kind of difficulty encountered can be obtained if deposits are
made at a number of current densities. A convenient means of achieving
this with a single plate is afforded by a device known as the Hull
Cell. This is a plating cell which has the cathode inclined with
respect to the anode with a resultant variation of current density
with distance along the plate. With this means of evaluation then it
was decided to determine the effect of current density upon nodule
formation from a bright nickel plating solution containing copper ions
as an impurity. The Hull Cell was cleaned and arranged in a water

bath so that the normal operating temperature could be maintained.

The impure plating bath was prepared in standard plating tanks and

51

agitated until completely homogeneous. From this solution was with-
drawn the cell samples and the plating was carried out using steel
panels which had been cleaned in the usual manner. The results of
these tests although not producing any clear cut evidence of either
high or low current density requirements for nodule formation did
tend to indicate that at higher current densities the nodule growth
was favored. This was evidenced by the fact that the concentration of
nodules was slightly higher at the higher current density end of the
panel. This fact is borne out in plating practice also by the build

up of nodules along the edge of the panel and in the position of

 

maximum nodule concentration on the flat panel.

At this point it seemed that the most logical electrolytic
phenomenon to consider was the very broad field of polarization.
Polarization which may be defined as the change in potential experi-
enced by an electrode immersed in an electrolyte under the impetus of
an applied current, is a phenomenon attendant upon all electroplating
operations and has been the subject of widespread exploratory research.
Even today research continues in an effort to verify and explain
certain phenomenon associated with polarization. It has been fairly
well established at the present time that polarization: 1. Increases
with current density. 2. Increases with the duration of electrolysis .
for a constant current density. 3. Decreases with an increase in
temperature. h. Decreases with agitation. S. Depends upon the nature
of the electrode surface. 6. Varies with the material of the electrode.

No single existing theory of polarization is in agreement with

1
all six of these generalizations; however, the theory of concentration I
l

52

polarization is in harmony with the first four of these statements (28).
This form of polarization which has been well established results in
a thin layer of solution adjacent to the electrode, the concentration
of which differs from that of the bulk of the electrolyte. This
layer may be called the diffusion layer and its apparent thickness
depends on those factors which influence the rate of diffusion. In
an effort to determine the effect of copper ions on this polarization
or overvoltage factor, measurements were taken of voltage versus
current with increasing amounts of copper in solution. 'These values
were obtained using existing laboratory equipment which unfortunately
proved to be inadequate for the requirements and the test was abandoned.
When a solution contains two or more kinds of positive ions such
as we have here following the copper impurity additions, then on
electrolysis the deposit may contain one or both of the metals (or
hydrogen) depending on the circumstances. Since all aqueous solutions
contain the positive ion hydrogen in addition to any metallic ions
which may be present, a consideration of the behavior of hydrogen is
in order. ‘When a hydrogen ion is discharged at the cathode the
resulting atom: 1. May remain as such. 2. May unite with another
to form a molecule. or 3. May unite with the electrode material to
form a "hydride", solid or gaseous, which may be either a definite
compound or an absorption complex. The compensating removal of the
gas may occur by: l. Diffusion of the gas into the electrode.
2. Diffusion into the electrolyte. 3. Chemical action (if oxygen
or an oxidizing agent is present). or h. Bubble formation. This

latter process by which hydrogen escapes as a gas is the most important

 

 

 

53

means of hydrogen release and is a major factor in overvoltage. It

is well known that this process does occur and that the solution pH
rises in the vicinity of a cathode operating at less than 100%

current efficiency. This rise in pH has been demonstrated by several
investigators (29) (30) (31). The consequences of this increase of

pH in the vicinity of the cathode are of extreme importance since

some metals, including nickel, form insoluble hydroxides and basic
salts at pH values of about 7. Most bright nickel plating solutions,
including the one used in this investigation, are operated at a much
lower pH than this, namely in the 3-h. Therefore even though consider-

ing the pH rise within the cathode film the possibility of such hydroxide

 

or salt formation of the cathode surface is rather remote in as much
as the differences encountered here would be very unlikely to boost the
pH to the value of 7. However, if by some solution adjustment these
hydroxides were able to fonn at a lower pH, their presence could be a
very real consideration. In the event of their formation these particles
of hydroxide or basic salt would probably be extremely small in size
(of colloidal dimensions), and if these sub—microscopic particles
carried positive charges, they would be transported cataphoretically
to the cathode where their charges would be neutralized. Simultaneous
with this discharge of colloidal particles, metal would continue to be
deposited by the usual process and could conceivably form around these
hydroxides and entrap them.

This sort of nickel solution alteration by which nickel hydrates*
are permitted to form at lower pH levels has received some attention

in Russia. In experiments there it has been demonstrated that a I

*Hydroxides or basic salts.

I.

5h

number of factors tend to effect a change in the pH level of beginning
hydroxide formation (32). It was shown that such things as the concen-
tration of nickel sulfate, NaCl additions, boric acid concentrations,
and temperature all have a definite effect on the pH at which nickel
hydrates begin to form. The pH at the instant a hydrate begins to form
was investigated by plotting potentiometric titration curves. The
circuit in which the measurements were made included a glass electrode
and two saturated calomel electrodes which were connected to the test
solution or to the solution filling the glass electrode by means of an

electrolytic switch. All measurements were made at a temperature of

 

93 I 2°C with a RAPS type potentiometer, using a Sensitive mirror
galvanometer. The instant at which the hydrate began to form was
likewise checked by the scattering of light in a Tyndall Gone. The
light scattering was manifested at the same point as the sharp bend
on the titration curves, though in ordinary light the solutions
appeared to be still completely transparent, turning cloudy only after
appreciable amounts of alkali were added.

The solution used in this investigation was that which is
employed for the electrolytic refining of nickel and which consists
of nickel sulfate, Sodium sulfate, boric acid, and sodium chloride.
This solution does not duplicate our electroplating solutions but the
similarity is close enough so that direct comparisons may be made.
Summarizing the particular results of this investigation which pertain
to the problem at hand we find that the pH at which nickel hydrate
begins to form is lowered somewhat by increasing the nickel ion

concentration. Boric acid shows a marked effect in lowering this

55

pH value, while sodium chloride additions and temperature increases
both have a slight effect in this same direction. Translating these
results into values which would pertain to the bright nickel bath
used here, the possibility of hydroxides forming at the cathode layer
now becomes more of a consideration since the pH differences are
reduced. According to these data it may be estimated that it is
possible for the pH of beginning nickel hydrate fbrmation in bright
nickel plating solutions to approach a value as low as 5.5.

The above results are very revealing and offer a great deal of
information which may be useful in formulating a possible explanation
for certain poor deposits; however, the detrimental effects we are
concerned with are due to the addition of copper ions to the nickel
plating bath. ‘What then is the effect of these copper ions on the
electrolyte since all indications so far point to a solution adjustment
of some type? In a second article from this same Russian laboratory,
the effect of several metallic impurities in a nickel solution upon
the pH of beginning hydroxide fonmation was studied.(33). The
experimental techniques employed were the same as described above
and included in the list of impurities studied was the copper ion.

To explain the behavior of copper in the nickel electrolyte different
amounts of copper, in the form of CuSOb-SHZO, were introduced into

a nickel electrolyte of the composition NiSOhoé H20 - h0.6g/liter,
Nazsoh - bOg/liter, NaCl - Sg/liter, and H3BO3 of two concentrations
0 and 20g/1iter. The first break in the curve obtained by potentio-
metric titration with a glass electrode and the appearance of a

diffused glow in the Tyndall Cone were regarded as the beginning

S6

of copper hydroxide formation since it has been demonstrated that
copper hydroxide will precipitate before the formation of nickel
hydrates (3h). The pH values at the beginning of copper hydrate
formation in relation to the concentration of the copper were deter-
mined. It was found that with a copper content of 0.00hg/liter the
titration curve had the same appearance as the titration curve of
nickel sulfate, and the pH value at the beginning of hydrate formation
agreed with that for nickel. However, when the copper content of the
solution was increased to 0.05 and.0.l9g/liter, the pH value at the
beginning of hydrate formation was noticeably lowered. The presence
of boric acid likewise tended to lower this pH level as was the case
with nickel.

Considering now the significance of these investigations in
relation to the results already obtained from.the chemical analysis
of the nodules and the smooth deposit, the current density study,
the various plating studies involving buffing and plating, multilayer
deposits, and solution agitation changes, and finally the nodule
cross-section studies, it is possible to formulate certain conclusions
regarding nodule formation. First of all the plating experiments and
the dependence of nodule position on solution direction furnished
strong evidence that nodule formation is not a problem originating
with the base metal. The structure of the nodules revealed a central
core of a dark etching material surrounded by what gave every indication
of being electrodeposited nickel. Even the striated bright nickel
structure was in evidence in this outer nodule material. Chemical

analysis of the nodules and smooth foil revealed more copper in the

S7

nodules than in the smooth bright deposit. This would indicate that
the dark core material contained more copper than the surrounding
structure.

Having eliminated the possibility of a base metal dependence it
became apparent that consideration should be given to the changes
occurring in the plating solution following the addition of copper
as an impurity. The pH of the solution was maintained at 3.75,
consequently it was difficult to postulate the formation of copper
or nickel hydrates. However, a consideration of polarization effects
at the cathode surface indicated the possibility of a significant
alteration of pH within a very thin envelope surrounding the cathode.
This increase in pH coupled with the information that boric acid and
an increased copper concentration both tend to lower the pH at which
copper or nickel hydrates will form, points very strongly toward the
presence of copper or nickel hydrates within this cathode film. If
these hydrates are ferming in the cathode film than the preferred
location of the nodules which depends upon the solution agitation
direction could be the result of a simple mechanical pressure on the
film layer which would tend to move these hydroxides or hydrates
closer to the cathode surface. In the rectangular tank used in these
studies the solution travels from the anode to the cathode along one
side and then moving across the surface to the other side and continu-
ing on back to the anode. If hydroxides or basic salts are forming
adjacent to the cathode they probably are produced in a colloidal
condition and carry a positive charge. Therefore it is reasonable

to assume that they would be electrodeposited on the cathode together

 

58

with the normal metal. This build up of hydrate surrounded by nickel
metal proceeds at a rate far exceeding the normal deposition of metal
alone. Assuming that these hydrates do fornlthere are several possible
reasons for this rapid growth behavior. First of all it seems logical
that the polarization at these particular sites of compound deposit has
been lowered resulting in localized areas which are more receptive to
further deposition. Very possibly this more receptive nature is
limited to further basic salt deposition so that the build up of a
spur results. As the normal depositing metal encounters such an area
it is forced upwards and the ultimate result is the cored type nodule
which.we observe.

A second possible explanation for this apparent uncontrolled
growth centers around the fact that from the surface of a particle
of deposited hydroxide hydrogen may possibly be liberated as a gas
more easily than from the pure metal. This localized increase in
hydrogen evolution would in turn result in an increased concentration
of hydroxyl ions which means a more basic pH value and thus further
hydrate or basic salt formation.

Thirdly, and equally as significant a possibility relates to
the current density increase which is associated with sharp edges or
projections. Thus nodules once they have started will experience a
greater current density and consequently more rapid cathodic reactions.
The Hull Cell tests and actual plating operations have indicated
that nodule fonnation is somewhat favored at the higher current density
areas.

A comparison of these nodules with the roughness produced when

 

59,

calcium is present in the nickel solution in excess of the saturation
limit, reveals that the two are entirely unrelated. The typical
treed structure observed when copper is the impurity is not found in
the calcium produced roughness. Further the roughness produced by
calcium occurs mainly on the horizontal lip of the panels indicating
more of a settling mechanism than any polarization phenomenon. Beside
other differences such as size, shape, and location of the nodules or
projections it is usually the case that the calcium produced roughness
can be alleviated by a filtration and pH adjustment treatment; such is
not the case with copper in solution. The logical explanation for
calcium produced roughness is the occlusion and entrapment of precipitated

salts, probably calcium sulfate.

 

CONCLUSIONS

The problem of determining the mechanism involved in nodule
formation when copper is present in a nickel plating solution was
pursued by investigating all the possible factors which relate to
this anomaly. Crystallography and atomic considerations of copper
and nickel eliminated the possibility of a lattice disturbance
phenomenon. Photomicrographs of the cross section of these growths
revealed a central core extending throughout the nodule surrounded
by an apparent nickel sheath. The possibility that these nodules
originated at imperfections in the base metal was disproven by several
experiments. The first consisted of depositing a nickel film from
a purified nickel solution and then adding copper to the solution as
the deposition continued. Other experiments consisted of buffing
one half the panel prior to deposition; taper sectioning the rough,
black deposit; and finally altering the direction of solution agitation.
In each case the results repudiated a base metal dependence for the
nodules.

Polarization studies and the likely fonnation of hydrates
resulted in a logical explanation for nodule growth. This expanation
was reinforced by a chemical analysis of the nodules as well as the
smooth nickel foil which revealed a slightly higher copper content in
the nodules. Additional support was obtained by current density
considerations which involved the use of the Hull Cell and indicated

that nodule formation was increased at higher current densities.

61

Thus by the elimination of all other possible factors which could
produce these nodules all the evidence points to the fact that the
roughness in electrodeposited nickel caused by copper as an impurity
results from a complex solution adjustment at the cathode surface
resulting in a pH precipitation and localized polarization differences
with subsequent deposition.

A comparison of the roughness in electrodeposited nickel due to
calcium as an impurity and that due to copper reveals that the two
are not the same. The calcium roughness results from the occlusion
and entrapment of precipitated salts, probably calcium sulfate.

The results of the investigation concerning the effects of
calcium in four different nickel plating solutions on the physical
properties of electrodeposited nickel may be summarized as follows:

Appearance - Both the Watts' pH 2.2 and 5.2 panels were
judged to have an Eastman Gray Scale reading of l-2, and there was
no change in scale reading with increasing calcium content. There
was no change in the mirror bright rating given to the organic and
nickel-cobalt bright panels with increasing calcium. On panels from
all four solutions there was a slight increase in roughness at the
higher calcium levels.

Adhesion - The adhesion of the nickel deposits from all
four plating solutions was unaffected by increasing calcium concentration.

Ductility - The watts' 2.2 deposits showed no appreciable
change in ductility over the entire range of calcium impurity. The
Watts' 5.2 deposits revealed a gradual decrease in ductility with

increasing calcium content. The foils from both the organic and

 

62

nickel-cobalt solutions exhibited a much lower degree of ductility
than the foils from either of the Watts' solutions, however, the
addition of calcium to the bath had little effect on the ductility.
Hardness - As expected the bright deposits revealed a
much higher hardness than the Watts' dull nickel, however, the slight
variations in hardness with increasing calcium content were not
considered significant.
Throwing Power - The results indicate an overall reduction
in throwing power with increasing calcium content. The variations,
however, are not considered significant and no clear cut trend is

established.

 

Corrosion Resistance - Evaluation of corrosion resistance
was accomplished by three different methods thereby verifying the
result that calcium does not affect the corrosion resistance of any
of the four types of deposit.
Removal of Calcium - Electrolyte removal, high pH
precipitation, increased temperature, and precipitation with
hydrofluoric acid all proved to be ineffectual methods for the 1

removal of calcium from the four nickel plating solutions.

CD

(2)
(3)
(h)
(5)
(6)

(7)
(8)

(9)
(10)

(ll)

(12)
(13)
(1h)

(15)
(16)
(17)
(18)
( 19)

63

REFERENCES
Miller, B. E., Monthly Review of the American Electroplater's
Society, 2, No. 3, 5-6 (1916).
Robins, G. s., ibid., g, No. b, 11-12 (1916).
Meyer, w. R., Metal Industries, (N.Y.), 23, 117-121 (1939).
Hogaboom, G. B., ibid, 21, 165—167 (1939).
Diggin, M. 3., Metal Finishing, 22, 13-15 (19b1).

Diggin, M. 8., Monthly Review of the American Electroplater's
Society, 22, 513 (l9h6)-

Kushner, J. B., ibid, 22, 751-770 (l9h2).

Rominski, R. J., "The Effects of Trace Amounts of Copper Upon
Some Physical Properties of Electrodeposited Nickel." Thesis
for the M.S. Degree, Michigan State College (l9h9).

Anonymous, Brass World, 1, bS-Sé (1911).

Thompson, M. R., and Thomas, C. T., Transactions of the American
Electrochemistry Society, 52, 79 (1922).

Haring, H. E., Monthly Review of the American Electroplater's
Society, 11, h-13 (192k).

Eckelmann, L. E., ibid, 2;, 18-35 (193h).
Johnson, L. W., Metal Industries (London), £9, 281—286 (1937).

Francis-Carter, 0., Proceedings of the American Electroplater's

Ewing, D. T., 33.31., Plating, 29, #11, 1137 (19h9).
Private Communication.

Ewing, D. T.,gt.gi., Plating, 32, #9, 1033 (1952).
Ewing, D. T., $5.31., Plating, £9, #12, 1391 (1953).

EWing, Do To, fioflo, Plating, 9-1;, #11, 1307 (19511).

 

(20)

(21)

(22)

(23)

(2b)

(250

(26)
(27)

(28)
(29)

(30)

(31)

(32)

(33)
(3h)

6h

Serfass, E. J., 22.31., Plating, 26, 103h (l9b8).

Ewing, D. T., et.a1., American Electroplater's Society Report,
Serial No. 23_(l9§3).

Dow, H. 0., "The Effects of Iron and Aluminum on Electrodeposited
Nickel." Thesis for the Ph.D. Degree, Michigan State University
(1955).

Pierce, W} J., and Pinner, W. L., Proceedings of the American
Electroplater's Society, El, 176 (l95h).

Evans, U. R., and Shome, S. C., Journal of the Electrodepositor's
Technical Society, 26, 137 (1950).

Rowe, R. J., "The Effect of Manganese on Electrodeposited Nickel."
Thesis for the Ph.D. Degree, Michigan State University (1956).

Brenner, A., 33.21., Plating, 22, 865 (1952).

Serfass, E. J., and Levine, W. 8., Monthly Review of the American
Electroplater's Society, SA, L151), (19117)-

Bandes, H., Metal Finishing, HQ, 516 (l9h6).

Henricks, J. A., Transactions of the Electrochemistry Society,
§2, 123 (l9h2).

Brenner, A., Proceedings of the American Electroplater's Society,

Graham, A. K., and Read, H. J., Transactions of Electrochemistry
Society, lg, 279 (l9h0).

Rotinyan, A. L.,.and Zeldes, V. Ya., Journal of Applied Chemistry,
(U.S.S.R.), 22, 717 (1950).

Rotinyan, A. L” and Zeldes, V. Ya., Ibid, g2, 936 (1950).

Britton, H. T. 5., Hydrogen Ions, 2nd Edition (1932).

 

 

M USE ONLY _
R0 MM USE 3ND!

=.~.(.':.-,~—.>,v'-n 'iu -.
1,1-.1 . M

 

 

 

 

 

 

“"'CIT1'FI)ITILTI)M\1 flfiflflfflxfiiflflflflflfilfl'“