Chemical and thermodynamic studies of dental gold alloys with special reference to homogenization, electrochemical corrosion and cluster formation

AKADEMISK AVHANDLING

SOM MED VEDERBÖRLIGT TILLSTÅND AV

ODONTOLOGISKA FAKULTETEN VID UMEÅ UNIVERSITET

FÖR VINNANDE AV ODONTOLOGIE DOKTORSEXAMEN

OFFENTLIGEN FÖRSVARAS I SJUKSKÖTERSKESKOLANS

AULA, LASARETTET, UMEÅ,

LÖRDAGEN DEN 6 MARS 1976 KL. 09.15 f.m.

AV

MAUD BERGMANleg.tandläkare

UMEÄ 1976

UMEÅ UNIVERSITY ODONTOLOGICAL DISSERTATIONSABSTRACTS

No. 6From the Department of Dental Technology and the Department of Inorganic

Chemistry, University of Umeå, Sweden

Chemical and thermodynamic studies of dental gold alloys with special reference to homogenization, electrochemical corrosion and cluster formation

by

MAUD BERGMAN

Umeå 1976

This thesis is based on the following publications, which will be referred

to by their Roman numerals.

I Heat treatment of soldered joints in dental casting gold alloys. An

electron microprobe analysis. In collaboration with Åke Björnham.

Acta Odont. Scand. 1972, 30:415-439.

II Dissolution rate of cadmium from dental gold solder alloys. In

collaboration with Olle Ginstrup. Acta Odont. Scand. 1975, 33:199-210.

Ill Structure and properties of dental casting gold alloys

I. Determination of ordered structures in solid solutions of gold,

silver and copper by interpretation of variations in the unit cell

length. In collaboration with Lars Holmlund and Nils Ingri. Acta

Chem. Scand. 1972, 26: 2817-2831.

IV Structure and properties of dental casting gold alloys

II. Emf-measurements using stabilized ZrO^ as solid electrolyte

as a means of studying cluster formation in gold-copper alloys.

In collaboration with Erik Rosén. Acta Odont Scand. 1975. Preprint.

Centraltryckeriets Offsetavd Bröderna Larsson Umeå -76

CONTENTS PAGE

INTRODUCTION 4

THE AIM OF THE INVESTIGATIONS 7

PART ONE (I, II) Background 8

MATERIALS 9

METHODS 11

RESULTS 16

DISCUSSION 26

CONCLUSIONS OF PRACTICAL IMPORTANCE 32

PART TWO (III, IV) Background 34

MATERIALS 35

METHODS 36

RESULTS AND DISCUSSION 39

CONCLUSIONS IN SUMMARY 53

GENERAL SUMMARY 57

ACKNOWLEDGEMENTS 59

REFERENCES 60

4

CHEMICAL AND THERMODYNAMIC STUDIES OF DENTAL GOLD ALLOYS WITH SPECIAL

REFERENCE TO HOMOGENIZATION, ELECTROCHEMICAL CORROSION AND CLUSTER FORMATION

INTRODUCTION

Gold alloys have a wide application in dentistry mainly because of their

high corrosion resistance in the oral environment together with suitable

mechanical properties. The ADA specification No. 5 covers dental casting

alloys based on gold and provides a range of alloys for various kinds of

dental restoration. Generally the alloys with simple high-gold-content

compositions are relatively weak and soft and their use is therefore

limited to restorations which are exposed to a load of low magnitude.

When the restoration has to carry heavy stress concentrations alloys with

greatly improved mechanical properties are needed. To meet this requirement

the simple gold-si 1ver-copper alloys are modified by the addition of metals

from the platinum group, especially platinum and palladium. Most commercial

dental gold alloys contain in addition small amounts of zinc which act as a

scavenger by combining with any oxide present. The presence of zinc also

reduces the fusion temperature of the alloy thus increasing the ease with

which the alloy can be cast. Indium is sometimes used in low concentration

as an alloy component because it is a less volatile scavenging element.

Indium also contributes to casting fluidity and a more uniform grain size.

Iridium in very small amounts is sometimes added as a grain refining agent

(Nielsen & Tuccillo 1966, Phillips 1973, Craig et al 1975).

Gold solder alloys are used for assembling a bridge and, sometimes, for

building up a crown locally to ensure a correct relationship with adjacent

teeth. These alloys have compositions which are similar to those of the

casting gold alloys. This is necessary to guarantee colour, mechanical pro­

perties and corrosion resistance comparable to those of the alloys being

5

5fC

joined. However, the fusion temperature of a solder must be 50-70 K below

that of the casting alloy to be soldered, in order to allow a sufficient safety

margin and to make the alloy flow more easily during soldering. Thus the

fusion temperatures of the solder alloys are reduced by an increased amount

of zinc or by the addition of tin and cadmium for example. Phosphorous may

sometimes be added to gold solder alloys as a deoxidizer. If a "white"

coloured solder is required, nickel can replace copper in the alloys (Taylor

& Teamer 1949, Phillips 1973, Asgar et al 1975).

For special techniques, for example when porcelain is fused to a dental

bridge the requirements of both the casting gold alloy and the solder alloy

generally implies that the compositions of the alloys are further modified.

The use of dental gold restorations with different requirements regarding

their physical properties thus makes it necessary to have a differentiated

assortment of alloys. The responsible manufacturers of dental gold alloys

make a creditable effort to develop and improve their alloy materials as well

as laboratory methods. However, less responsible manufacturers seem to use

mainly "trial and error" methods in producing some of their gold alloys

which, among other things, results in the availability of an unnecessarily

large number of dental gold alloys. When judging the corrosion resistance of

dental metallic restorative materials it is desirable to consider the entire

oral environment. Thus an alloy might be corrosion resistant per se when ex­

posed to different kinds of tests but when used in the oral cavity together

with other gold alloys of very different compositions a situation which pro­

motes corrosion may appear due to synergetic effects. From this point of

view it is undesirable to have too many gold alloys for use within the same

field of application.

*In the present thesis the temperatures are given in the SI base unitkelvin. t/°C = T/K - 273.15.

6

The main requirements for dental gold alloys can be summarized as follows:

First, the alloys should resist tarnish and corrosion in the oral environ­

ment and second, they should have physical properties which meet the require­

ments of the various types of dental restoration.

The first demands that the noble metal content of the alloy must be kept at

a level which will ensure tarnish and corrosion resistance. It also implies

that the melting of the alloy as well as the further laboratory working

moments must be carried out in such a way that any promotion of eventual

corrosion is avoided. The second requirement demands the production of alloys

with for example different melting temperatures and different mechanical pro­

perties. In general the tensile strength and proportional limit of these

alloys increase with hardness, while the elongation decreases. An increased

hardening of the temperable gold alloys is attained by suitable treatment,

usually involving heating and cooling.

The heat treatment of the ready-made dental restoration in routine dental

laboratory techniques often involves a homogenizing heat treatment followed

by a hardening heat treatment. The achievement of the maximum increase in

the hardness of an alloy depends, among other things, on the preceeding heat

treatment and quenching temperature. Thus it is undesirable either to use a

scheme which limits the quenching temperature after casting to a standard

value or to use any standard procedure at the subsequent heat treatments.

Such measures tend to restrict unnecessarily the possibility of securing an

optimum result as concerns the dental gold restorations (Wise 1964, Phillips

1973), especially as many manufacturers provide individual laboratory in­

structions for their alloys.

7

THE AIM OF THE INVESTIGATIONS

The composition of a dental gold alloy together with the different solid

state reactions which occur during heat treatments are thus factors of basic

interest for the evaluation of the technical quality and durability of the

final product, the dental gold restoration. It was the general purpose of

the present work to elucidate some of these factors. The more specific aims

were as follows:

1. To investigate whether, in gold-soldered assemblies, any changes in the

distribution of available alloy components occur in the region of the

soldered joint after various kinds of conventional heat treatment (I).

2. To determine in vitro the dissolution rate of cadmium from commercial

dental gold solder alloys of those types which are commonly used in

soldered bridge constructions (II).

3. To indicate the composition and amount of clusters in solid solutions of

gold, silver and copper through interpretation of variations in the unit

cell length (III).

4. To determine the activity of copper in alloys of copper and gold to obtain

additional data for testing the previously (in Paper III) proposed cluster

model (IV).

Papers I and II are dealt with in part one of the present work while Papers

III and IV are dealt with in part two.

8

PART ONE (Papers I and II)

Background. Dental gold castings have been found to be far from homogeneous

(Björn & Hedegård 1965, Söremark et al 1966, Eick & Hegdahl 1968) and this

lack of homogeneity in the gold restorations has been seen as one of the

contributery factors in the electrochemical process of oral corrosion (Spreng

1940, 1963, Hedegård 1958). Not only does such corrosion give rise to un­

desirable effects on the dental restoration itself, but it may also be a

factor in the production of pathologic changes in biological tissue. Metallic

elements from dental gold restorative materials have been shown to migrate

into hard and soft oral tissue (Söremark et al 1962, Bergenholtz et al 1965).

A number of symptoms and effects attributed to galvanic corrosion in the

oral cavity have been reported, e.g. Lain et al 1940, Schriever & Diamond

1952, Rheinwald 1953, Hedegård 1958, Frykholm & Hedegård 1968. A few cases

of hypersensitivity reaction to gold alloys are described in the literature

(Charpy 1952, Tornei! 1962, Frykholm et al 1969, Eigart & Higdon 1971).

According to Eick & Hedegård (1968) and Hedegård (1974) high temperature

heat treatment of a dental gold cast has a tendency to level out concentra­

tion gradients due to intracrystal 1 ine segregation. Hedegård (1949, 1958,

1974) therefore recommends that any cast or cast and soldered dental gold

restoration should be exposed to so called homogenizing heat treatment

before being finally installed in the oral cavity. This statement is perhaps

valid as concerns single casts but it cannot be taken for granted that the

proceeding should be extended to include soldered constructions.

Among those alloy elements which can be dissolved from dental gold alloys

by corrosion, cadmium has received a great deal of attention during the last

few years (Kawahara et al 1968, Kawahara et al 1975). This alloy element is

frequently used in dental gold-solder alloys where it contributes to a lower­

ing of the fusion temperature (Phillips 1973). Studies in progress

9

(Bergman et al 1976) show that cadmium from test pieces of a commonly used

dental gold solder alloy, subcutaneously implanted in rats, is released and

accumulated principally in the kidney and in the liver.

MATERIALS

-In Papers I and II only commercially available gold alloys were used.

In the paper dealing with heat treatment of soldered joints, (I), the three

casting gold alloys studied were of the types which are commonly used for

fixed crown and bridge restorations and for details of removable partial

prosthesis. The alloy A corresponded to type III and the alloy C to type IV

according to ADA specification No. 5,1973, while the alloy B lies between

type III and IV according to the same specification. The solder alloy used in

this study was of a type commonly used for the soldering of hard gold dental

assemblies. The chemical compositions of the four alloys are given in Table I.

Table I. Chemical composition in weight per cent of the alloys investigated. Code: 1 = specimen which was melted and quenched specially for the chemical analysis, 2 = specimen from one of the original parent samples of each alloy (marked "g"), 3 = alloy formulas provided by the manufacturer.

Alloy Code Au Ag Cu Pt Pd Zn In Ir

A 1 74.7 10.3 11.0 0.46 3.5 0.57 <0.1 0.012 74.0 10.4 11.0 0.46 3.5 0.51 <0.1 0.013 74.0 10.5 11.0 0.50 3.5 0.50

B 1 71.0 10.0 12.9 0.01 6.2 0.58 <0.1 <0.012 70.7 10.0 12.9 <0.01 6.3 0.71 <0.1 <0.013 70.0 10.5 13.0 6.0 0.50

C 1 66.9 10.6 12.3 9.6 0.1 0.52 <0.1 0.092 66.7 11.3 12.2 8.2 0.1 0.21 <0.1 0.073 68.0 12.0 12.5 7.0 0.50

Solder 1 70.0 5.2 15.2 2.87 <0.1 0.83 -5 0.053 71.0 5.5 15.0 2.75 0.75 -5

10

For each of the three casting alloys four parent samples were prepared as

described under "sample preparation" in Paper I. Each of these samples was

divided into twelve pieces of approximately the same size. The pieces were

each given a number and ten pieces were randomly grouped in pairs (marked

a, b, c, d, f) while the remaining two pieces (marked e and g) were used to

represent the "as cast" condition. The paired pieces were then soldered

together two and two. Various heat treatments followed by bench cooling for

one minute and then quenching in water were carried out. The marked specimens

originating from each parent sample were treated as follows:

a: no further heat treatment after soldering

b: heat treatment for 3 hours at 973 K after soldering

c: heat treatment for 1 hour at 973 K after soldering

d: heat treatment for 1 hour at 973 K after soldering, and then hardening

heat treatment from 723 K to 523 K for 30 min followed by quenching

e, g: representing the "as cast" condition (extra samples)

f: heat treatment for 1 hour at 1073 K before soldering and for 1 hour at

973 K after soldering.

The heat treatments were carried out in an inert atmosphere of argon except

the hardening of the d-marked samples which took place in the atmospheric

milieu of a common porcelain furnace. Four series of the a-d specimens and

two series of the f specimen were produced from each alloy giving in total

54 samples which were mounted in methyl methacrylate resin and then polished

using a standard metallographic procedure. The e and g specimens were intended

for reserve use and were not included in the electron microprobe investigation.

In the paper dealing with dissolution of cadmium from solder alloys, (II),

11

six different dental gold solders of the types commonly used for the solder­

ing of fixed gold restorations were studied. For comparison the dissolution

rates of copper and zinc were determined. The compositions of the alloys as

concerns the elements studied are given in Tables III-V. The analyses were

carried out by Analytica AB, Sollentuna, Stockholm. "Analysis 1" was made

using the roentgen fluorescence method on alloy samples with the same marking

as the electrode specimens. Before this analysis was performed the material

was melted. "Analysis 2" was made using atomic absorption spectrophotometry

on the residues of the electrodes after the corrosion experiments. In this

case the samples were directly dissolved in acid. The alloy material in band

form was coiled and mounted in a glass tube. The exposed surface of the alloy ?was about 2-4 cm .

The electrolyte used was physiologic saline solution. By addition of potassium

hydrogen phtalate the pH of the solution was buffered to 4.0. This electrolyte

solution was marked "A". In one test series the pH of the electrolyte was

further adjusted to 6.5 with sodium hydroxide solution; this solution was

marked "B". The gas phase used throughout these experiments consisted of an

oxygen-nitrogen mixture of known composition (0.209 ± 0.005 % 0^). Thus in

the experiments the partial pressure of the oxygen was 0.21 kPa as compared

to 21 kPa in normal air. As oxygen interferes with the metal dissolution and

it is practically impossible to remove all oxygen this composition was chosen

to obtain well-defined experimental conditions. All the experiments were

carried out at room temperature (298 K).

METHODS

Firstly the methods used in the electron microprobe study (I) will be

described and then the methods used in the potentiostatic study (II).

12

(I). The electron probe microanalysis of the 54 specimens was carried out at

the Swedish Institute for Metal Research (Stockholm) using a Cameca MS 46

microprobe analyzer. This and other electron probes as well as their fields

of application have been described by Kiessling (1960) and Brown & Thresh (1970).

Before the development of the electron probe microanalyzer there was no satis­

factory way of determining the distribution of the various constituents of a

material on a micron scale. However, the importance of the electron probe is

not its sensitivity but rather its ability to analyze non-destructively very

small volumes of material.

The element distribution across the gold-solder junction was determined by

moving the specimen progressively under the electron beam. The line analysis

was started in the cast alloy at a distance of 250 pm from the gold-solder

junction and proceeded across this junction into the middle of the soldered

joint. If the joint was very narrow the line analysis proceeded further into

the cast alloy on the other side of the soldered joint. In this way a line

of about 350-500 pm was scanned. Some of the heat treated specimens were

subjected to point analysis of the special demarcated domains, "microphases",

which were observed precisely at the gold-solder junction. Microphotographs

of each of the alloy specimens were taken.

Statistical treatment. The concentration curves were divided into two segments

by the gold-solder junction. Each curve segment has been regarded as a

realization of a stationary process. For each type of treatment, periodicity

i.e. regular segregation of elements, and some other properties of the process,

are given by estimating its mean level and covariance function. Corresponding

properties of the interactions between any two curve segments of the same alloy

are given in a similiar way. By comparing these results some conclusions about

the effects of heat treatments could be drawn, though the sample sizes were

small.

13

(II). The electrochemical cell is shown in Fig. 1. The working electrode, WE,

consisted of the sample. The current direction was anodic (metals were dissolved)

Me(s) -* Me2+ + 2e" (I)

The counter electrode, CE, consisted of a spiral of platinum wire mounted in

a glass tube with a plug of glass filter at the bottom. The electrolysis

current passed through the solution and the glass filter and evolved hydrogen

gas at the counter electrode:

2H+ + 2e~ - H2(g) (II)

The reference electrode, RE, was a commercial saturated calomel electrode

(Metrohm AG, CH - 9100 Herisau). Several reference electrodes were compared.

They agreed within a few millivolts.

The potentiostat (Fig. 2) sensed the potential difference between the working

electrode and the reference electrode. This potential difference was compared

with a preset voltage from a ten-turn potentiometer, E2. The potentiostat sent

an appropriate electrolysis current through the counter and working electrode

system to bring the potential difference El close to the preset voltage E2.

The tolerance was a fraction of a millivolt. The common value of El and E2

is denoted by E in the following. The electrolysis current was integrated by

means of an electronic integrator. The integrator reading was taken from a

digital voltmeter and from a recorder. The electronic components were adjusted

to give a direct reading for the amount of electricity, Q, passing the working

electrode. The accuracy of the adjustment was better than 0.1 %. The experi­

mental set-up was similar to that described by Johansson (1965). When an

appropriate amount of electricity had reacted (usually 30 ymol e") the experi­

ment was terminated and the electrolyte solution was analysed as regards

cadmium, copper and zinc by atomic absorption spectrophotometry (Department

14

Fig. 1. The measuring cell. Legend: CE = counter electrode;WE = working electrode (the spe­cimen); RE = reference electrode; Gas = gas inlet tube.

CELLPOTENT IOSTAT

INTEGRATOR

Fig. 2. Principal schematic diagram of the cell, potentio- stat and integrator. Legend: CE, WE and RE are the electrodes as in Fig. 1 ; El = potential differ­ence between RE and WE; E2 = adjustable control voltage; 163 = operational amplifier, Analog devices 163; R1 = precision re­sistor, 10.68ft; R2 = precision switchable resistors, 200 kft,2 Mft or 20 Mft; C = polystyrene precision condenser, 5.0 yF;SP656 = chopper stabilized operational amplifier, Phil brick SP656; REC = strip chart record­er, HEATH JR-18 M; DVM = digital voitmeter.

rm

15

of Analytical Chemistry, University of Umeå). The quantities measured were:9

The electrode area (mm ), E versus SCE (mV), Q (ymol e ), time (min), con­

centrations of the three metals in the solution (ppm), see Table I in Paper II.

Calculations. The calculated quantities were: firstly the current efficiencies

for dissolution of each of the three metals and then logarithms of dissolution"Merates. The current efficiency (2x-—) is the fraction of the amount of elec-

e~tricity used for the dissolution of one of the metals.

Electrochemically the most interesting quantity is the logarithm of the-1 -2dissolution rate, LDR, where "dissolution rate" has the unit: mol s m .

In the investigation the measurements were centered around LDR = -6 which-6 -I -2corresponds to a dissolution rate of 10 mol s m or a dissolution rate

-1 -2of 3.5 mg Cd year mm . The logarithmic scale was chosen because LDR is a

linear function of the applied potential E, see Fig. 3. According to electrode

kinetic theory (Vetter 1967) the following relationship, called Tafels equa­

tion, may be expected

n = a + b log i (1)

where n is the overtension, i is the electrolysis current and a and b are

constants. In the present case Tafels equation may be formulated:

LDR = Cl + C2 (E - 800 mV) (2)

where Cl and C2 are constants, referring to the measuring series and 800 mV

is a potential value chosen arbitrarily in the center of the measuring range.

A least squares method was used to fit the measuring points of each measuring

series to equation (2). The results are given in Tables III-V. In Table III,

the LDR for cadmium has been extrapolated to E = 550 mV.

16

O O

E/mV

Fig. 3. The logarithm of the dissolution rate (LDR) for Cd, as a function of the applied potential (E). The least squares straight line describing the measuring points of the series with specimen codes 2, 4, 5, 6 will be extra­polated to a potential range (500-600 mV) which can be found in the oral cavity. °

RESULTS

(I). The chemical compositions of the alloys as revealed by X-ray fluorescence

analysis and atomic absorption spectrophotometry agreed well with the alloy

formulas provided by the manufacturer (Table I). No inclusions due to extraneous

factors in the laboratory techniques were found in the analysis of the samples.

The demarcated domains, "microphases", which appeared on the microphotographs

17

of the gold-solder junctions (see Fig. 4 in Paper I) of all the homogeni­

zing heat treated samples of the alloy C and some of the heat treated samples

of the alloy B (Bid, B2b, B3b and c, B4d) were further investigated by point

analysis using the microprobe analyzer. These domains had compositions (Table II)

which varied a great deal from those of the original alloys (Table I).

Table II. Composition in weight per cent of the demarcated domains emerging in the region of the gold-solder junction after heat treatment.

Speci­men

Time forheat treatment(hours)

Domain localization Weight per centAu Ag Cu Pt Pd Zn In

C4b 3 Junction 45 2.0 15.6 28.5 1.2 5.7

Nearest to the junction in the cast alloy (two different domains)

48.761.6

4.612.2

21.8 14.9

23.04.2

1.00.2 0.3

Nearest to the junction in the solder alloy 64.0 6.7 17.5 4.2 0.5 3.6

Cl d 1 Junction 54.3 3.2 18.2 21.0 1.2 1.5

Nearest to the junction in the cast alloy (two different domains)

60.171.0

7.211.5

15.912.7

14.82.9

0.70.2

Nearest to the junction in the solder alloy(two different domains)

64.571.0

6.78.6

15.915.0

8.42.6

0.70.4

3.41.9

B3c 1 Junction 69.4 7.8 14.5 1.7

Nearest to the junction in the cast alloy 70.9 10.1 13.0 0.3

Nearest to the junction in the solder alloy 69.4 7.3 15.2 1.4 2.4 0.5 2.8

18

As regards the line scanning the distribution of each alloy element was

recorded on graphs. The distribution pattern of the elements was similar on

both sides of the soldered joint in those specimens where the line scan had

proceeded across the whole soldered joint into the cast alloy on the other

side. However, the main purpose was not to estimate the alloy composition at

a given point but merely to study changes in the element distribution in the

region of the gold-soldered joint. Since these changes are most pronounced

very close to the junction concentration values from only the shorter parts

(c. 70 ym) of the scanned line on each side of the junction were taken into

consideration. Thus a moving average of three terms for each of 10 sampling

points in the cast alloy and 7-10 sampling points (depending on the width of

the joint) in the solder alloy was determined. The modified concentration

points are plotted in Figs. 4-23. Each figure shows the obtained effects

from heat treatments for an element and an alloy.

As can be seen from the curves (Figs. 4-23) indium which is an alloy element

in the solder alloy only, already during the soldering procedure had diffused

into each of the three casting alloys. A reverse process occured when during

the soldering palladium diffused from the casting alloys A and B into the

non-palladium solder alloy. The hardening heat treatment carried out after

the homogenizing heat treatment on one series (d) of each parent alloy sample

did not seem to have any noticeable effect on the distribution of the alloy

constituents.

A comparison of the effect of alloy composition on the results of the various

heat treatments, see Figs. 4-23, indicates that the composition of casting

alloy and solder alloy was best levelled out in the alloy A. Heat treatment

for three hours (b) in general produced a more even distribution of alloy

19

components than heat treatment for one hour (c, d, f). However, the various

heat treatments did not produce concentration levelling-out effects to any

great extent as regards the alloys B and C. In the alloy C especially heat

treatment for three hours (b) did not level out concentration gradients but

had the opposite effect. These results were to some extent confirmed by the

microphotographs of the specimen surfaces which revealed the existence of

demarcated domains appearing at certain gold-solder junctions, and were further

confirmed by the results of the point analysis of these domains (Table II).

Figs. 4-23. Variation in distribution of available alloy components in soldered assemblies of the casting gold alloys A (Figs. 4-10), B (Figs. 11-17) and C (Figs. 18-23). Each curve is a mean value curve representing four (a, b, c, d) or two (f) average curves. Abscissa: equispaced (7 x 10"^m) sampled points. (Gold-solder junction between points 10 and 11). Ordinate: concentration in weight per cent. ■ = a, • = b, * = c, ▲ = d, — = f.

20

100X10

170X10]

21

22

(II). The sum of the current efficiencies for the dissolution of the three

metals cadmium, copper and zinc from the gold solder alloys studied was

considerably less than unity (with two exceptions which might be due to some

error). This fact indicates that other metals might have been dissolving

simultaneously.

The material "JS 730" was used in five measuring series of which those having

the specimen codes 2, 4, 5 and 6 are merely reproductions using different

specimen samples. As these four measurement series produced no significantly

'different results they were all treated together to yield a common linear LDR-E

relationship for each metal(Tables III-V). In contrast to all the other measuring

series, the fifth one with "JS 730" (specimen code 1) was carried out in the

electrolyte solution B with pH = 6.5. A comparison of the first two lines in

Tables III-V shows that the slopes are unaffected by pH, but the LDR values at

E = 800 mV are lower in solution B (pH = 6.5). The constancy of the slopes

indicates that the mechanisms of the electrode reactions are unaffected by pH.

The change in the LDR values indicates that the dissolution is slower when the

solution is less acid, possibly as a result of oxide film formation. As the pH

in the region of a soldered joint hardly ever goes down to 4.0 the results

obtained in solution A can be regarded as the worst case results.

When comparing the LDR at E = 800 mV with the metal element content it was

found that a low metal content was roughly associated with a low dissolution

rate (see Tables III-V). However, the variations are quite large and are

probably due to metallurgical variables - a point discussed later. Concerning

cadmium, the logarithm of the dissolution rate (LDR) as a function of the

appVied potential (E) is shown in Fig. 3. The straight line obtained from the

least squares calculations and describing the measuring points from the series

with the specimen codes 2, 4, 5, 6 was extrapolated to a potential range

(500-600 mV) which can be found in the oral cavity (e^g. Maschinski 1970).

rO t/)CJ >

c •1— CD -O E • roo CD O CD CD Q 00 o 1— r— o LO LO E

O O -P O-p f— C rO LD oo r— r— d r— r— OJ r— OCO o II a) LO CDCD “O O *o3 CO Cd Q. II *cr >> Q 4- rO CD cr» UD LO N O 00 ir> 4-

-C —1 c P LU CJ Oc CL o a: -P r— 1— o 03 00 o o

1 o Q X O ro r— r— 1 1 PII 1 _i CD -P > 1 i CD

CO _a>> co • Eo C LO to VO 03 o CO o ZJ

c o CD C cd i--- CO i— O OJ OJ CO co o O

ro -p CD •r— _l o o o O O d O CD-p 3 P ---- - -p JZ

<u 3 1— rö ^---V rO -p.c o >+-> o CO -a E > CO o

00 CD CD o -p4- CD -p O ~aO ~a rö O CD

o a 0Ü ~a 1 r-v. ^3- CO «Ö- ^3- LO LO 3CO X p OJ > -oCD o "O 1 rO o E3 P c C “O t/)

II -a LU £Z«3 o S_^ rO CO OJ 1— 1— OJ CO CO> z. JC CO OJ -P CD

CL CD •r— o OO c_> o o o o O o o i—CO E CD

•p— • 3 C -p + roCO T3 •1— rO -P>3 CD ~o P ro

“O o o CDrö "O CO CD 3 dlC rO JC II -prO JC 1— cr > > > > > > >

CD +-> cd E g E E E E E cCD -P 1 CD o-P rO 2 1 _l O o O O O O oO O o O o O O o c£= rO LO 1--- CD CO co 00 00 CO 00 00 CDCD 4-> CO C >-o -C LO CD CD •1— 1 1 1 1 1 1 1 ■i—

CL 1— CDCO O P 4- LU LU LU LU LU LU LUCD £= -P CD -p 10P CD CO rö sz r— C3 CD "O CD 1 1 1 1 1 1 1 ocr> O CD CD 1— •1— > > > > > > > •1—

•r- P -P C rö E E E E E E E +->4— "O to P CO CO CO co CO CO CO rO

>5 CJ P -P 1 i 1 1 1 3CD _c: O CO o o o o o o o cr

JC ~a to CD+-> E

=3ro rO

CDCOCD X X X X X X X ,__

•!— □= E P CD o *3- LO 03 03 00 CD-P CO Q- rö OJ 4-e CO CD 3 roCD ro . « -C cr + + + + + + + 1—CJ -P TD •P to

O CD OJ r^. o LO o o CDp CL “O O -pCD -o -p to C LO r"v LO LO LO LO -PCL s; rO rO O i i i 1 1 1 1

CD CD •r— E-P CD C -P ii ii ii II II II II O-C o ■P rö pcd rö p CD 3 Cd cd cd cd cd Cd Cd 4-

o CD _C cr Q Q Q Q Q Q QCD rö 4- 1— LU _l _l _l _J _l _l T3S -C -P CD CD

-p _c: P >Q. 1

c: 5 « 3 c: CDE c CO i— o O3 c CD -p O •i— CD

i— o CD £Z oo -p ca <C <c < C <c < PO •1— O rOo -p P -P CD

3 T3 CO OJ O 10CD i— >> C O_c O -C O to OJ _c+-> CO C_) "O* •1— 1— -p

E . c_> to 1c CD C5 ' CD >3 co 1— E

1—1 C P -p LO LO 00 OJ 03 CO O■1— CO rO c rO p1— CO CD C OJ 03 o 03 OJ OJ LO 4-

• ro ro C\J a < r— i— r— 1— •"O CO •P C_> to 1—o O p CD rö

Ql "O CD 3 T-p rO C i— CL to I— Po O Z rö •!— rö CD

4— rö -P to > -PCD i— ,— -C X r»> rO

to O O C_> CD CD 03 r>. CO CD E-P rO tor— O o L P cd C LO LO O O 03 i— CO CD r-13 •r— 1 3 C i— -C roCO to -C CD +-> C<D >) -p . « i— •i—P -C •1— ^—» -P rö OJ i— 4- CD

CL CD OJ O -r-i— •P rO O o O p P PrO II £C rO P co CO o P o P C OC O P z: CD c-. 1— OJ O r— o o

•i— <c •1— E CD O -p 3 -p •r- CDLl_ +-> C E E 03 < o CD o ■P -E

c o O O rö oo oo CD ■Z. p CD p rö -PO O C '-C ■“3 Q c a_ Q Q_

i—i •1— o -p CO > Ci—i -p CO 13 1— c CD -r-!—i 3 i— rO CD LO *o

CD O > E -oCD o C CO P LO >, CD

oo •1— CO CD u CD E L0-Q 1— •1— -P CD “O ^3- < 3rO 1 rö -a C O- O 1— »» O 03 CO OJ r— j.

1— 1 CO OO O OJ 1— 1— 1— *

Tabl

e IV. Final

resu

lts fo

r Cu.

For det

ails

see tex

t to Ta

ble H

I.

24

CO£Z cé CM o r— oo o '— CM CM y y CM no

+■>__i

O O O o o O O^— (Ö>- • i—E > no

CD Oo ~a «—o00 T3 X i LO no no no CD CD

c s- CM >1 (Ö O E

o T3LU c

• f— (Ö no i— CM CM CM no noCM 4->

+-> O CO c_> o o o O O o O

fö +

3 ,—o

crII

CL)Cd > > > > > > >Q E E E E E E E

1— _JO o o O o o O

CL) <D o o o o o o oC 00 00 00 00 00 00 00

4- •i—•— 1 1 1 1 1 1

fö4-> LU LU LU LU LU LU LU

1— _C v_X v__- •—^ 1—•" ^ " "—"CD <— «— •— y y y y

fO > > > > > > >s- E E E E E E E

+-> no no no no no no noCO

OiO O O o o o

CO i— i— «— «— ■— 1— 1—CD X X X X X X X

<0 o 00 *0- LO o o 00=3 CM 1— 1— 1— 1— CM 1crCO + + + + + + +

+-> r\ CM CD LO o 00COfö c CD CD LO LO LO CD LOCD o 1 1 1 * 1 1 1

+-> II II II ii II II ItCD

_CföD Cd Cd Cd cd Cd Cd Cd

1— CT Q o Q Q o o oLU _J _J —I _J —1 —1

13 C

i— oOtn 4-> 0Û <c C <C <c c

C\J

CO o• i— 1—

CO 1Z3 > o no noO 03 CD 1— CD CD CO LO

+->föc CD CD 00 00

r-’

c <C CM i— 1CDO

,__ ,__CD

<T3 Q_ CO

• i— +-> CO

S~JC.CD

>>CM CM CD no 03 CD o

föCD CD c CM CM i— CM 1— 03 no

Z2 C i— i— CM i— r— 14->

fö

s: (Ö CM 1CM

o o Os~ no no o o S-CD CM o OE CD 13 +-> 3 4->E f= CD <C o CD Oo fO tn oo CD CD S-

CJ c: r3 •~D O <c D_ o Q_

cCD CDE

LOL) CD *CD ~aCL o i— r' O 03 no CM 1

uo o CM 1 1 1

Tabl

e V. Final

resu

lts fo

r Zn.

For det

ails s

ee tex

t to Ta

ble II

I.

25

toc Cd o o LO 1--- 00 LO i—o o 1— oo 1— 1--- 1— r— OO

•1— __i--—^ +-> o o o o o o o✓—^ 03> •1--E > OO

CU oo T3 1—o y 1—00 ■a 1 ID oo CXJ UO 00 to

£Z L- C\J >1 03 o E

O T3LU c

•1-- "---^ 03 CXI 1— CXJ (XI CXJ CXJ OOCXI +-> r—

4-> o OO CJ O o o o O O o

03 +

13 r—c_>

crII

O)Cd > > > > > > >co E E E E E E E_J

o o o o O o oCU 0) o o o o O o o

c 00 CO 00 00 00 00 00M— •1—

r— 1 i 1 103

+-> LU LU LU LU LU LU LU1— dc '>— '■— ^ —-- ^^

CD i— r— i— f— 1— r— 1—1 1 1 1 1 1 1

03 > > > >* > > >E E E E E E E

+-> OO oo oo oo OO oo ooCO 1 1 1 1 1

o o O o o o oto 1— 1— i— f— 1— 1— 1—CUi. X X X X X X X03 r^- CO UO ^d" CD CD 00Z3a-to + + + + + + +

+-> r^. ,__ ,_ ^d- UO CXJ oto •03 c r^ to to to I-"* 1"-CU o i i 1 1 1 1 1

-M ii ii II II II II II<U 03

-C =3 Cd QC Cd Cd Cd Cd Cdh- cr CO Q CO co o o oLU _1 —1 —1 _J _l —I

1zs c

1— oo •I—

OO +J CO C <c < < <c

(XI^3"

to «d-•

to ,—c >i 1

M 1—1 OO CD to oo to03 oo UO CD ^3" 00 00+-> c

£Z C |--- r— LO f— r— o oCUo

,__ s- ,__CU

03 Q- to•r— -M to

_£Z >>S- CD r— f— i— OO to oo•i— 03 OO oo to 1--- CD 00cu CU C

+->3 <c •“ r~“ LO O o o

0303 CXJ ,—

•i— r— CXJO o oL- OO oo o o i-CU r->. r^. 1— oo o r— oE CU 13 4-> 13 -ME E CD O CD oo 03 OO oo CU 21 i- CU S-CJ c: '“O r_D O C Q_ O CL

cCU toE r\

•1— LOo CUCU -aCL o 1— O CD OO CXJ p—

OO c_> CXI 1— 1— '—

26

DISCUSSION

As the liquidus and the solidus temperatures of many dental gold alloys are

fairly close these alloys might be expected to be very homogeneous in the

as-cast condition. On the contrary, however, considerable coring and dendritic

segregation can often be seen in the castings. As the presence of an inhomo­

geneous surface structure is a factor which contributes to the occurrence of

corrosion in the oral milieu, various kinds of heat treatments to homogenize

the structure have been suggested (Hedegård 1958, Björn and Hedegård 1965,

Eick and Hedegård 1968, Hedegård 1974).

However, the conditions become more complicated when a cast construction is

to be soldered. If the soldering is properly performed and the surfaces of

the cast parts to be soldered have been wetted by the liquid solder a mutual

saturation of surface bonds occurs and a strong and stable interface bond is

obtained. It is known that the strength of the soldered joint may be influenced

by the composition of the different alloys, particularly when high fusing

solders are used. This indicates that diffusion of elements may occur between

solder and casting alloy. If the solder is overheated the strength of the

joint is seriously affected by diffusion of the solder into the grain bounda­

ries of the cast alloy and by diffusion of the cast alloy into the solder

(Phillips 1973). The compositions of the two original alloys are changed which

means that the mechanical properties as well as the electrochemical corrosion

behaviour of the joint are no longer under control. As was shown by El-Ebrashi

et al (1968), using electron microscopy, a distinct demarcation and no diffusion

between the two alloys is seen in the properly soldered joint.

A mutual diffusion between cast alloy and solder alloy can occur because of

overheating during the soldering process but it may also occur at lower

27

temperatures, although not as readily, if the heating is sufficiently prolonged.

A fact to be taken further into consideration is that many dental gold solder

alloys, particularly the lower fusing types, contain eutectics which contribute

to the inhomogeneity of the solder itself (Phillips 1973).

Thus it is, from a theoretical point of view, extremely doubtful whether the

validity of the theories regarding homogenizing heat treatment of pure casts

should be extented to include soldered constructions as proposed by Hedegård

(1958, 1974). These doubts have been confirmed by the results of Paper I.

According to Brown & Thresh (1970) the electron probe is capable of carrying

out chemical analysis for the complete concentration range from 0.1-100 % with

an accuracy of ± 2 % and sometimes ± 1 %, but for greatest accuracy point ana­

lysis should be used. In the present work point analysis was used to study

the composition of the demarcated domains appearing in the region of the sol­

dered joint after the heat treatment of some of the specimens. During the

line-scanning carried out on all the specimens the main purpose was not to

estimate the alloy composition at a given point but merely to study changes

in the element distribution in the region of the gold-soldered joint.

The curves presented in Figs. 4-23, showing the variation in distribution of

available alloy components in the region of the gold-solder junction, are mean

value curves representing four average curves concerning the treatments denoted

a, b, c, d and two average curves concerning the treatment denoted f. In dental

material research the electron probe microanalyzer is an often used instrument.

Unfortunately this instrument has often been used to investigate one single

specimen representing for example one alloy and one treatment. This must be

considered unsatisfactory if only from consideration of instrument variations

and conceivable differences between specimens.

28

The results of Paper I have shown that indium from the solder alloy rapidly

migrates into the three casting alloys while palladium from the palladium

containing casting alloys A and B rapidly migrates into the solder alloy.

This migration takes place during the soldering process and is in some cases

increased after heat treatment. These circumstances indicate that if an

element is an integral part of only the casting alloy or only the solder

alloy a slight mutual element migration probably occurs immediately during

the properly performed soldering. The results also show that a migration of

alloy elements across the gold-solder junction occurs during the so called

homogenizing heat treatment of the soldered assemblies. Furthermore such

heat treatment may result in the appearance of demarcated domains at the

gold-solder junction. The alloys within these domains have compositions

which are different to a greater or lesser extent from the original alloys

and in addition have unknown physical properties. With no, or at least mini­

mum, element migration across the junction more favourable conditions are

created. The chemical compositions of casting alloy and solder alloy are

indeed different but the possibility of corrosion occurring can to some extent

be estimated. The mechanical properties of the two alloys are preserved and

grain growth, microporosities and oxide formation are diminished or avoided.

With respect to the above arguments regarding the importance of avoiding as

far as possible any element migration between casting alloy and solder alloy

the main conclusion of Paper I in the present work is that "homogenizing"

heat treatment of gold-soldered assemblies at high temperatures should not

be carried out as a routine.

Paper II deals with the dissolution rates of cadmium, copper and zinc from

commonly used commercial dental gold solder alloys. As previously stated some

29

of these alloys are becoming increasingly complex as regards their chemical

composition. This development implies that the composition of the alloy is

often changed to meet a new, emerging requirement as regards for instance

fusion temperature, some mechanical property or corrosion behaviour. In fact

some of the solders studied (II) have at present chemical compositions which

differ from those existing when the experimental study was carried out. This

is particularly true for the cadmium and zinc content of these alloys. On

June 24, 1974 the Swedish National Board of Health and Welfare issued a

statement which definitely advised against the use of dental alloys containing

cadmium (MF 1974:34). However, there are still a lot of commonly used dental

gold-solder alloys on the market, which contain cadmium.

It has been shown in many studies, for example by Friberg (1950), Berlin and

Ullberg (1963), Friberg and Piscator (1972) and Bergman et al (1976) that the

element cadmium can be enriched and stored for long periods in various tissues

of the mammalian body, especially in the liver and kidney. Thus it is important

to find out whether cadmium can be dissolved from gold-solder alloys by electro­

chemical corrosion. To study this problem clinically would be very intricate

and the results would probably be hard to interpret as appropriate corrosion

tests for use in the oral cavity scarcely exist. An in vivo corrosion test in

the oral cavity might also be dubious from an ethical point of view. It was,

therefore, decided that an in vitro potentiostatic study should be made to

estimate the dissolution rate of cadmium from six different dental gold solder

alloys.

In most studies of this type the conclusions are based mainly on the relation­

ship between the electric current and the potential during the progress of

the corrosion experiment. However, this current is a quantity which is diffi­

cult to interpret because it is the result of a number of chemical reactions.

In the present work (II) a specific reaction has been studied by means of

30

direct chemical analysis of the electrolyte solution after the corrosion

experiment.

One special factor which has to be considered as regards the alloy composition

as gained from a chemical analysis of the metallic material is whether there

has been any melting of the alloy before the analysis. This was the case with

the roentgen fluorescence analysis, the results of which are presented under

"Analysis 1" in Tables III-V. For example, during a melting of the material

the cadmium can partly evaporate which results in a lowering of the cadmium

content in the solidified material. Attention must be given to this evaporation

of cadmium during melting of the alloy in technical dental laboratory work.

In a narrow, badly ventilated laboratory the dental technician who solders

dental bridgeworks runs the risk of inhaling the cadmium vapour. As inhaled

cadmium is absorbed to a much larger extent than ingested cadmium (Friberg et al,

1971) this aspect must be taken into consideration when choosing solder alloys

containing this element.

The results of Paper II as presented in Tables III-V show that the dissolution

rates of the alloy elements studied are not directly proportional to the amount

of metal in question. This indicates that other factors influencing the process

must be taken into consideration. Metallurgical variables may change the free

energy of the electrode material and - for a given metal - alter its position

in the galvanic series. From this point of view the most important of such

variables are high energy grain boundaries, presence of different phases,

residual internal energy resulting from cold working of the material, compos­

itional and surface heterogeneities and foreign inclusions.

The corrosion experiments in the present study were performed in a well-defined

gas atmosphere consisting of an oxygen-nitrogen mixture of known composition.

The partial pressure of the oxygen was 0.21 kPa. Under clinical conditions,

31

variations in oxygen concentration between different parts of the same restora­

tion do occur. In regions where plaque deposits exist low oxygen concentrations

are found and these areas will tend to corrode rather than areas of high oxygen

concentration (Greener et al, 1972). According to Hoar & Mears (1966) inorganic

solutions are satisfactory substitutes for extra-cellular body fluids, at least

when the anodic behaviour of passive metals is under consideration. The physio­

logic saline chosen had been buffered to pH = 4.0 (in one series to 6.5) thus

avoiding the withdrawal of the dissolved metal from the chemical analysis due

to hydroxide precipitation. Thus the in vitro corrosion experiments in this

study (II) were carried out under circumstances which may be said to represent

very unfavourable conditions within the oral cavity. It was shown that under

these given circumstances the alloy elements cadmium, copper and zinc dissolved

from the dental gold-solder alloys and that the dissolution rate of each of the

three elements could be determined.

The dissolution rate of cadmium in a potential region of 550-600 mV (LDR = -11)

which can occur in the oral cavity, corresponds to a dissolution rate of about -1 -20.04 yg year mm , see Fig. 3 and Table III. This extrapolation is doubtful,

however, the value of LDR could not possibly exceed -8. Thus a tentative value

of LDR = -8 might be of interest. Assuming a fairly uniform dissolution rate

over a prolonged period a value of LDR = -8 thus corresponds to 0.04 mg Cd -1 -2year mm . For example a patient who has a large dental bridge construction

2with several soldered joints can have an exposed solder area of about 0.5 cm .

Under the defined conditions and according to the calculations above a disso­lution of 2 mg Cd year"^ may occur in the oral cavity of this patient. It

must be taken into consideration that a restoration will be exposed to mechanical

stress and to a certain degree of abrasion, both of which factors may further

proceeding corrosion.

Taking into consideration different sources of cadmium intake (main source: food)

32

the World Health Organization (1972) has proposed a provisional tolerable weekly intake of 400-500 yg Cd per individual. This means 21-26 mg Cd year~^. In compari­

son to this value the contribution of cadmium in man resulting from cadmium

dissolution from dental gold solders in the oral cavity must be judged to be

rather small. However, the long-term biological significance is hard to estimate

at the present time.

A great deal of effort has been expended in providing specifications for dental

materials. Previously the main interest was focused to their physical and

chemical properties although the effects of dental materials on biological

tissue has also received a great deal of attention in the last decades. However,

the extremely complicated and individually variable ecology of the oral cavity

makes it very difficult to interpret the results of various tests as regards,

for example, the corrosion tendency of dental materials, in terms of general

validity. Even if an alloy has been shown to be corrosion resistant in different

kinds of laboratory tests this alloy can corrode when combined with other

materials in oral restorations. Factors such as variations in the properties

of saliva and in the intake of drink, food and drugs, various oral hygiene

measures and different magnitude and distribution of load during function make

the net effects hard to estimate.

CONCLUSIONS OF PRACTICAL IMPORTANCE

1. "Homogenizing" heat treatment of gold-soldered assemblies at high tempera­

tures should not be carried out as a routine.

2. Dental gold alloys containing cadmium should not be used because this element

can be released from the alloy and contribute to damage to biological tissue.

33

3. Because of the risks of corrosion of dental restorations when different

alloys are used in a patient's oral cavity it is important that the dentist

who is able to estimate the oral situation as a whole, should prescribe the

quality of alloy to be used and not leave the selection of the alloy to the

dental technician.

34

PART TWO (Papers III and IV)

Background. The dental gold alloys, mainly because of their complex chemical

composition, after melting often solidify during coring and dendritic segrega­

tion. As the present manufacturing trend often appears to entail the addition

of further elements to the alloys to meet new requirements, principally with

respect to physical properties, it seems to be important to consider the

compositional problems from a basic point of view, starting with the simple

gold-copper and gold-silver-copper alloy systems.

Those dental gold alloys which are capable of age hardening gain their increased

hardness from heat treatment usually carried out at a temperature in the region

of 723-523 K. However, it is important that before such an age hardening the

alloy is first subjected to a heat treatment carried out at temperatures around

1000-1100 K and followed by quenching. In this way residual stress is relieved

and the subsequent age hardening can start with the alloy a uniform solid solu­

tion. The solid state reactions allowed to occur during the age hardening then

increase the hardness, tensile strength and proportional limit of the alloy.

As the dental gold alloys often contain six or more elements several solid

state reactions are possible. Many of the alloy elements can in binary combina­

tions during age hardening furnish precipitation phases e.g. AuCu, AuCu^ and

PtCu. PdCu is not very effective in producing age hardening unless silver is

present when a ternary phase is formed. Other ternary and possibly quaternary

phases may occur although they are difficult to demonstrate (Phillips, 1973).

As concerns the formation of such precipitation phases during age hardening

it has for example been shown that the rate of the AuCu formation depends on

the temperature from which the alloy was previously quenched (Kuczynski et al

1955, Krivoglaz & Smirnov 1964). The higher the quenching temperature the more

35

rapidly was the subsequent ordering attained. It has also been shown that

clusters of CuAu^, AuCu^ and AuCu can exist in the solid solution in the

temperature range about 670-800 K (Germer et al 1942, Ogawa & Watanabe 1952,

Gehlen & Cohen 1965, Greenholz & Kidron 1970). Thus an increased knowledge

of the structure of the dental gold alloys immediately after casting and at

temperatures just above the quenching temperatures can afford improved

possibilities of directing the solid state reactions. In this way the mechani­

cal properties and perhaps also the homogenizing reactions could be better

controlled. This in turn could, in the author's opinion, prepare the way for

dental gold alloys with a decreased number of components which would be most

advantageous e.g. as regards corrosion.

The general purpose of Part Two of the present work was thus to study the

occurrence of ordering in solid solutions of gold-copper and gold-silver-copper

in a state representing the conditions around 1100 K.

MATERIALS

The alloys used in Papers 111-IV were all produced directly in the laboratory.

The raw materials were metal wires with a round cross section and a diameter

of 0.5 mm. The purity for gold and silver was 99.95 % (Nyström and Matthey,

England) and for copper 99.90 % (Cu, type SM-0010-02 Svenska Metallverken,

Sweden).

Specimen preparation. The metals were weighed out in proper proportions

immediately after being treated for 1 min in 4 M H^SO^Ag, Cu) or 4 M HN03(Au)

to eliminate impurities and oxides on the surface. The wires were then twisted

together and melted in an electrically heated furnace with an inert atmosphere

of argon (AGA, quality SR).

36

As regards the material used in Paper III the molten specimens were kept at

1373 K for thirty minutes and cooled in the furnace to 1073 K where they were

kept for another thirty minutes, and then quenched in water at room temperature.

The compositions for these alloys are given in Table VI. Powder specimens were

prepared from the alloys care being taken to extract the powder from the inner

parts of the specimen. To reduce the effect of the lattice distortion, occurring

due to plastic deformation during powder preparation, the powders were heat

treated at 1073 K in an argon atmosphere. The time for this heat treatment was

selected to allow recrystal 1isation without any sintering and thereafter the

crucible with the powder specimen was quenched in water at room temperature.

Concerning the material used in Paper IV the molten specimens were kept at

1373 K for fifteen minutes, cooled down to 1073 K in the furnace and quenched

in water at room temperature. The compositions are given in Table VII. Imme­

diately before a run the solid solutions were heat treated for six hours at

1073 K in an argon atmosphere. After the alloys were quenched powder specimens

were prepared as described above. Copper powder and copper (I) oxide (Merck p.a.,

Germany) and copper (II) oxide (Fisher p.a., U.S.A.) were used in testing the

reference systems Cu-Cu^O and CuO - Cu^O. Platinum foils and wires were obtained

from Platinaverken, Vänersborg (Sweden).

METHODS

X-ray measurements. In Paper III the unit cell length of all the heat treated

powder samples was determined from X-ray powder diffraction photograms from a

Guinier-Hägg camera and mainly the CuKa radiation (A = 1.54051). All the

structures observed on the photograms had a face centered cubic unit cell.

No extra lines could be detected. For calibration of the camera constant silicon

with the unit cell length of 5.43054 ± 0.00017 Å at 298 K (Parrish 1960) was

37

used. The parameter a ± a'(a) for all the powder specimens was calculated

and refined using conventional methods (Hägg 1954). The values obtained are

given in Table VI. Powder lines with differences between the calculated and 2observed values of sin 0 greater than 0.0002 were omitted in the determina­

tion of the lattice parameters.

Emk-measurements. In Paper IV the activity of Cu in Cu-Au alloys was deter­

mined by use of the following cell:

Pt, 02(g), Cu(in alloy), Cu^O || ZrO^ ( CaO) || 02 (air), Pt

In the cell there are two platinum electrodes situated at each side of the

solid electrolyte tube. If the partial pressures of the oxygen are the same

at both electrodes no reaction can occur. However, if the partial pressure2_of oxygen is higher at one electrode, 0 ions will move from this electrode

to the other one, as a potential difference between the two platinum electrodes

has arisen (Kiukkola & Wagner 1957, Sato 1971). By connecting the electrodes

with a potentiometer of high input resistance (Data Precision, series 2000

Digital Multimeter; Model 2500) the existing potential difference, E, could

be measured. This measured value, E, is related to the partial pressure of

oxygen as follows:

E = RT7F In (3)

where R is the gas constant, T the absolute temperature, F the Faraday

constant, Pq the oxygen partial pressure of the sample and p* the oxygen

partial pressure in air (= 21.23 kPa).

The cell arrangement shown in Fig. 24 was used and during each run a continous

stream of purified and dried inert gas (argon, AGA quality SR) passed through

the furnace tube. Within the furnace tube the gas was pre-equi1ibriated by

passing it over a powder mixture of the same type as in the experimental cell

38

before it reached the cell itself. Emf readings were continuously taken in

the region of about 50 K below the solidus temperature of the alloy in question

and down to about 1050 K. Between each reading enough time was left to reach

equilibrium and about ten hours were needed for each series of measures. The

reproducibility of the emf values was ± 2 mV or better. The reference system

Cu-Cu20 was measured several times during the progress of the present study.

The value of E for pure Cu was needed for the calculations but was also used

for testing the experimental set-up which in turn was also tested by the

reference system Cu0-Cu20.

inlet Ar

outlet Ar <r

Fig. 24. The experimental set-up A = Pt wires.B = Pt(10 mass % Rh) wire.C = teflon tape.D = glass tube.E = suspending Al^ capillary.F = Zr02(Ca0) tube.G = pre-reactor with sample.H = Pt electrode wire.I = Al2O3 crucible.K = sample.L = Pt foil.M = external Al^O^ tube; the bottom is kept in the centre of the furnace.

39

The activity of copper, called a^, in the alloy is related to the measured

quantity E as follows:

l°g aCu = - F(RTln 10)_1x AE (4)

where AE is the value of E (pure Cu) - E(alloy).

The alloy was considered to behave as an ideal mixture of free copper, free

gold and the complexes CUpAu^ and hence thè activity of copper, a^, could be

put equal to the mole fraction of free copper, Xçu. Thus, the E-values measured

permit determination of x^u.

RESULTS AND DISCUSSION

The concept order-disorder in alloys. In common linguistic usage the application

of the concepts order and disorder is very subjective and individually variable.

However even on the basis of such subjectivity the terms order and disorder can

attain a certain defined meaning. For example it can be said regarding a mixture

of red and white balls: Ordered distributions are those which are characterized

by a simple succinct description, such as "all the red ones together" or "every

other one red" etcetera. Disordered distributions are those which, for their

exact description, require the specification of each ball.

Thermodynamics is based on such macroscopic observations, that is to say only

summary statements can be made such as, for example, two kinds of molecules

are gathered, each in one region. This implies that there are two pure elements

or that the different kinds of molecules are regularly distributed within a

crystal or that they are more or less regularly distributed. If this reasoning

is applied to the copper-gold phase diagram, see Fig. 25, we find that the pure

phases Au,Cu, CuAu and Cu^Au (marked in the diagram) can be defined as ordered,

40

while the one-phase region, denoted (Cu, Au) is a solid solution and thus has

to be defined as disordered. At lower temperatures the boundaries between the

one- and two-phase regions shown in the diagram are impaired by some degree of

uncertainty.

°C

1000

0 10 20 30 40 50 60 70 80 90 100Cu Au

ATOMIC PER CENT GOLD

Fig. 25. Phase diagram of the copper-gold system (after Hansen & Anderko, 1958).

In solid solutions of this type the Au and Cu atoms are of equal size and can

all be randomly distributed. The degree of disorder will depend on the relation­

ship between the numbers of Cu- and Au-atoms. The number of macroscopically

equal distributions can be used as a measure of the degree of disorder in such

a mixture of Cu and Au atoms. Such a measure will of course depend upon the

measuring implements available and will thus lack exactness to a greater or

lesser extent. In this situation it is usual to choose to consider systems with

a lot of atoms and to use a relatively rough implement, for example observations

41

by a human eye. Under these conditions the number of ordered distributions

together with those which are not "completely disordered" are only a small

fraction of the total number of distributions. The number of disordered

distributions which are not macroscopically distinguishable from each other

is a quantity which can be expressed mathematically, see for example Prince

(1966), but under the conditions mentioned it can also be approximated to the

total number of distributions. The more molecules there are the better this

approximation will be.

However, the solid solution need not always consist of completely disordered

Au- and Cu-atoms büt can, to a certain degree, be ordered and thus contain

regular arrangements in for example the AuCu and AuCu^ structures, see Fig.26.

Such a state is usually described as "ordered structures in a disordered matrix".

These ordered structures can be called clusters or complexes but are also often

referred to as superstructures because they cause extra lines or line broadening

to appear on the powder photograms. Although in a case like this there is a

certain degree of order in the disorder the system as a whole must be considered

to be disordered because the different kinds of molecules (Cu, Au, AuCu, AuCu^)

are mutually disordered.

O Oo o

cr?n&i r*£n^?

}/-' i çg/ U iV '1 g ^

Au AuCu AuCu^ Cu

Fig. 26. Examples of ordered structures

42

As stated above the aim of the present work is thus to try and establish the

degree of this latter type of ordering in the two alloy systems Au-Cu and

Au-Ag-Cu. Beginning with pure Au and Cu the formation of the disordered [Cu, Au]

matrix can be written

[Cu] + [Au] -> [Cu, Au] (HI)

and for ordered Cu Au phases P q

p[Cu] + q[Au] -* [CUpAUq] (IV)

i.e. the solid solution formed will contain [CUpAu^] complexes (clusters)

distributed in a [Cu, Au] matrix. It is open to question whether these two

reactions occur at random or if there is any mutually dependence, according

to the law of mass action for example.

Experimental data. Two different types of experimental data have been collected.

(i) The cubic unit cell length, a, as a function of the mole fractions X^, Xg,

Xq where X^ = X^, Xg = X^ and X^ = XCu is presented in Paper III. Data are

given in Table VI of the present work. The composition range used which covers

the region of interest to dentistry is further illustrated in the diagram of

Fig. 27. Data of this kind are often explained by a simple Vegard's law model

(Hume-Rothery et al 1969) but as can be seen from Fig. 27 this was impossible

in the present case, and thus cluster formation had to be taken into considera­

tion.

(ii) Values of the emf, E, measured using the cell presented on page 37, as

a function of the composition X^, X^ i,e. X^u and X^u are presented in Paper IV.

Fourteen compositions have been studied and the data are given in Table VII.

43

Table VI. Experimental data a (X^u, XCu) with standard deviation o'.

The table also presents the residuals A-j and ^ obtained from two model calcula­tions. A-j = (a(exp) - a(calc))xlo^ Å (a simple Vegard's law model) where a(calc) has been calculated using a^u = 4.078, a^ = 4.086 and a^ = 3.615;

= (a(exp) - a(calc)) xio^ Å where a(calc) has been calculated using the complexes A^B, A^C, AC, AC^ (final results of Paper III). In addition the standard deviation, a'(a) obtained for the determination of the lattice para-

3meter is given in A x10 . The first part of the data given are those obtained from own determinations. The latter part of the data has been used only for calculations in the binary systems. The data within parentheses are data from the literature while the remainder has been obtained by extrapolation from own determinations.

XA5 XB’ XC9 a’ a'(a)’ V A2* °> °> 4-078> 1» 0, 0; 0, 1, 0, 4.086, 1, 0, 0;0, 0, 1, 3.615, 2, 0, 0; 0.83, 0.17, 0, 4.073, 1, -6, 2; 0.69, 0.31, 0, 4.072,1, -8, 0; 0.62, 0.38, 0, 4.073, 1, -8, 0; 0.56, 0.44, 0, 4.074, 1, -8, -1; 0.55,0.45, 0, 4.075, 1, -7, 0; 0.50, 0.50, 0, 4.075, 1, -7, -1; 0.45, 0.55, 0, 4.076,1, -6, -2; 0.74, 0, 0.26, 3.981, 2, 23, 1; 0.65, 0, 0.35, 3.942, 2, 26, -2; 0.51, 0, 0.49, 3.879, 2, 28, -2; 0.48, 0, 0.52, 3.864, 1, 27, -3; 0.43, 0, 0.57, 3.844,2, 30, 1; 0.33, 0, 0.67, 3.791, 2, 23, -1; 0.83, 0.09, 0.08, 4.048, 1, 6, 3;0.70, 0.20, 0.10, 4.040, 2, 7, 1; 0.69, 0.17, 0.14, 4.025, 1, 10, 0; 0.64, 0.06, 0.30, 3.963, 1, 23, -1; 0.62, 0.14, 0.24, 3.988, 1, 20, 0; 0.61, 0.25, 0.14,4.025, 1, 10, -1; 0.59, 0.34, 0.07, 4.054, 1, 6, 2; 0.55, 0.17, 0.28, 3.971, 1,21, -1; 0.54, 0.17, 0.29, 3.969, 1, 24, 1; 0.54, 0.09, 0.37, 3.934, 1, 27, 0;0.51, 0.27, 0.22, 3.994, 1, 16, -3; 0.50, 0.37, 0.13, 4.032, 1, 11, 1; 0.48, 0.45, 0.07, 4.054, 2, 5, 0; 0.48, 0.05, 0.47, 3.891, 2, 30, 1; 0.47, 0.12, 0.41, 3.919,1, 30, 3; 0.45, 0.21, 0.34, 3.949, 2, 27, 3; 0.44, 0.29, 0.27, 3.975, 2, 20, -1;0.42, 0.08, 0.50, 3.875, 1, 28, 0; 0.41, 0.38, 0.21, 4.000, 2, 16, 0; 0.39, 0.05, 0.56, 3.847, 2, 28, 1; 0.38, 0.23, 0.39, 3.924, 2, 25, 1; 0.38, 0.16, 0.46, 3.893, 1, 27, 1; 0.37, 0.04, 0.59, 3.831, 2, 26, 0; 0.35, 0.11, 0.54, 3.854, 1, 25, 0.

XA> XB, Xc, a: (0.10, 0.90, 0, 4.082; 0.20, 0.80, 0, 4.081; 0.30, 0.70, 0, 4.079;0.40, 0.60, 0, 4.076; 0.80, 0.20, 0, 4.072; 0,90» 0J0» 0» 4,074;) 0,80, Q, Q,2Q,4.003; 0.70, 0, 0.30, 3.965; 0.60, 0, 0.40, 3.921; 0.55, 0, 0.45, 3.898; 0.50, 0,0.50, 3.875; 0.40, 0, 0.60, 3.828; 0.35, 0, 0.65, 3.804; 0.25, 0, 0.75, 3.755;0.20, 0, 0.80, 3.728; 0.15, 0, 0.85, 3.700; 0.10, 0, 0.90, 3.672; 0.05, 0, 0.95,3.643.

44

Table VII. Basic experimental data at 1100 K and calculated values for Z(exp) and Z(calc) - Z(exp)

X^u = mole fraction of copper in the gold-copper alloy;

E/mV = emf obtained at 1100 K; a^u = the activity of copper in the alloy;XCu^ca^c^ ” xCu Xfu ” xfuZ(calc) = -----------— and Z(exp) = Lu Lu

Au XAu

XCu E/mV aCu Z(exp)x 102 (Z(calc) - Z(exp))x 102

0.10 171.3 0.07 10.3 6.1

0.20 172.1 0.07 23.2 -1.6

0.25 189.0 0.09 30.5 -4.00.30 232.0 0.14 36.8 1.50.40 265.1 0.20 53.4 7.30.45 256.9 0.18 67.3 -2.90.50 269.7 0.21 79.6 -3.30.55 284.0 0.24 93.7 -5.30.60 317.9 0.35 99.1 5.90.70 356.0 0.52 121.9 4.60.75 370.6 0.60 134.2 1 .80.80 387.5 0.72 140.3 8.40.85 396.2 0.79 157.7 -1.70.90 403.3 0.85 173.5 -11.1

(0.95 409.1 0.91)

45

Cu30a70

3Fig 27. The residuals (a(exp) - a(calc))xlo Å as a function of the composition XAu> XAq, XClj. The quantity a(calc) has been calculated without taking cluster formation into consideration. The diagram illustrates the effects which have to be explained.

Model and calculations. A model was assumed where the alloy was considered to

be an ideal solution with components and complexes. Furthermore it was assumed

that a state of equilibrium existed between complexes and solution that is to

say between ordered species and disordered matrix. This model means that one has

to study equilibrium reactions of the general type

pA + qB + rC ^ ApBqCr (V)

where A, B and C are the different components and p, q and r are integers.

If xA, Xg, xc denotes the mole fractions of free A, B and C the law of mass

action and the conditions for the total mole fractions X^, Xg and X^ will be

46

XA ' XA + p 6pqr XA XB XC

XB = XB + q Bpqr xP x< xj

XC " XC + r epqr XA XB XC

(5)

(6)

(7)

where 3p^r is the equilibrium constant for the formation of the complex in

question. On the basis of this model the aim of the caluclations was thus to

find the set of (p, q, r) values which would best explain the experimental

data. This systematic calculation procedure was carried out by use of a least

squares computer program LETAGROPVRID (Ingri and Sillen 1964, Arnek et al 1969,

Brauner et al 1969). The model which was considered the best was the one that

gave the lowest error squares sum.

U = Z(y(calc) - y)2 (8)

where y = a (the unit cell length) in Paper III and y = Z (the average number

of "bound" Cu per Au) in Paper IV. In both cases y(calc) denotes the corres­

ponding calculated quantity. The errors given in the two papers, a(a), a(apqr)>

a(3pqr) and a(Z(exp)) are defined in accordance with Sillen (1962).

Results of the calculations. The measurements of the cell edge as presented in

Paper III indicated that the cell edge followed the relation

a = aAXA + aBXB + aCXC + 2 aA B C XA B Co P q r p q r

(an extended Vegard's law)and the species which best could explain experimental

data were Au, Ag, Cu, AgAu^, CuAu^, CuAu and AuCu^ (Table VIII).

47

Table Vili. The results of Paper III in terms of the complexes tested and the corresponding best constants. The equilibrium constant is defined by the expressions (5), (6) and (7), and the constant by eqn (9).

Number Error qof points tested

H H ' tested squaresr

sum X 10°CT(a) X TO'3 3pqr ± "‘W1 apqr 1 °(apqr)

15 4 1 0 7.6 2.5 1.4 + 0.7 20.16 ±0.03

15 3 1 0 3.6 1.7 1.1 ± 0.5 16.14 ±0.01

15 2 1 0 1.1 1.0 1.2 + 0.6 12.133±0.004

15 8 2 0 16 3.6 3.6 ± 2.0 39.7 ±0.2

15 6 2 0 8.0 2.5 2.7 + 1.4 31.87 ±0.09

15 4 2 0 3.1 1.6 2.1 + 1.0 24.05 ±0.03

20 3 0 1 198 11 1.5 ± 0.7 16.46 ±0.07

20 3 0 2 78 6.7 2.0 + 0.8 20.50 ±0.07

20 3 0 1 16 3.3 *1 0 3

20 5 0 4 84 7.5 *4 0 5

20 3 0 2 3.4 1.4 0.230± 0.001 32 ±12 0 3 89 + 67 19.26 ±0.01

20 3 0 1 1.8 1.1 632 + 114 15.93 ±0.011 0 1 0.935± 0.004 8.46 ±0.051 0 3 33 + 32 14.99 ±0.01

40 3 0 2 13 2.0 66180 ±1008 19.65 ±0.012 0 3 192 + 49 19.25 ±0.022 1 0 0.12 + 0.10 11.77 ±0.05

40 3 0 1 8.1 1.5 639 + 1 15.92 ±0.011 0 1 0.90 + 0.02 8.68 ±0.061 0 3 3.9 + 2.8 14.66 ±0.083 1 0 0.342± 0.003 15.96 ±0.03

* Negative values of 3pqr were obtained (lack physical meaning).

48

Table IX.equilibrium

Some of theconstants.

models tested in Paper IV and the corresponding best

p qtested

Error squares sum X 10

a(Z(exp))X 103

3pq 0<epq>

0 3 1.9 120 3.0 X 107 1.1 X 107

1 1 2048 3413 1 1704 308

1 3 2.1 139 5.6 X 107 3.0 X 107

1 1 984 533 1 660 80

1 9 0.42 57 4.3 X 1024 242.7 X 10^1 3 1.8 X 108 0.8 X 108

1 1 2251 1503 1 1408 29

The best fit with the experimental data of the emf-measurements in gold-copper

alloys (IV) was obtained for the model with the complexes CuAUg, CuAu^, CuAu

and AuCUg(Table IX). From the residuals AZ given in Table VII it can be seen

that there is a fairly good agreement between Z(calc) and Z(exp). Thus the

results of Paper IV strongly confirm the model proposed in Paper III. In both

studies the formation of complexes seemed to follow the law of mass action. The

models presented in the two studies agree well as regards the main complexes

occurring in the gold-copper binary alloy system. However, the magnitude of the

formation constants are not directly comparable for the following reasons: 1

1. The models used were not exactly the same and this will result in different

values for the formation constants.

49

2. The constants obtained in Paper III are not mutually independent, they

can be said to "eat" at each other.

3. The alloys studied in Paper IV had been heat treated for six hours at

1073 K while the alloys studied in Paper III had been subjected to a much

shorter heat treatment at the same temperature.

Comparing Papers III and IV and summing up all the factors involved, it is

the author's opinion that the determination of the equilibrium constants

presented in Paper IV is to be considered the most reliable one.

In Paper IV not only the temperature 1100 K but measuring points at about 50 K

above and below this central temperature have been studied, see Fig. 28, where

the measured emk, E, was plotted as a function of temperature, T. At the basis

of the following relationship

ln8AS0R

AH0RT (10)

(see for example Prince 1966 and Manenc 1972) where AH0 and AS0 denote the

changes of the enthalpy and the entropy respectively and R is the gas constant

the following interpretation can be suggested. The constancy of the slopes of

the lines in Fig. 28 indicates that within the temperature range studied the

value of the equilibrium constant does not generally change with temperature.

This could mean that the value of AH0 is very small or zero and this in turnAS0implies that the entropy term ^ determines the formation of the complexes.

50

1050

Fig. 28. The emf (E/mV) measured against temperature (T/K) for different mole per cent of copper (X^u) in the alloy.

Comments in relation to previous studies of some physical properties. The solid

state reactions which occur during ordering have marked effects on the struc­

ture-sensitive physical properties. However, most of the investigations within

this field have been carried out at temperatures in the range 400-700 K, i.e.

at temperatures considerably lower than those of the present work.

51

For example the temperature variation of electrical resistance in copper-gold

alloys has been examined in a number of studies (e.g. Kurnakow et al 1916,

Le Blanc & Wehner 1932, Sykes & Jones 1936, Bumm 1939, Hirabayashi 1959,

Teramoto & Tachiki 1968). Sykes & Jones (1936), who studied the alloy AuCu^,

presented resistivity-temperature curves in the region below 673 K and related

the results to X-ray data. They found that the electrical resistance of the

alloy AuCu^ started to diminish when the ordered regions were about 50 A long

and had not reached the equilibrium value even when these regions were 300 A

long. Hirabayashi (1959) studied the alloy CuAu^ by measuring the electrical

resistivity of the alloy which was quenched from 723 K and annealed at various

temperatures from 373-493 K. The results were related to various states of

ordering as revealed by X-ray patterns. It was confirmed that the formation

and growth of small ordered species in the disordered lattice was accompanied

by a monotonous increase of the resistivity. However, a reduction of the

resistivity was observed after prolonged annealing at about 440 K due to an

increased degree of order.

According to the theoretical model given in the present work the formation of

many small ordered species should result in an increased resistivity. Later

on, at lower temperatures, when the species have precipitated as a special

phase the resistivity may be expected to decrease; thus the results of the

investigations discussed concerning resistivity measurements are in good

agreement with the proposed model.

Nowack (1930) studied the variations in hardness of initially annealed and

quenched AuCu as a function of ordering temperature and time. Nowack found,

after thirty minutes at a temperature of 623 K, a marked increase in hardness

followed by a softening at longer annealing times. In a study of order

52

hardening in AuCu, Arunchalam & Cahn (1967) found that quenched alloys

hardened far more than cold-worked ones when heat treated in the region

423-613 K. A maximum hardness was attained followed by decrease in hardness.

The latter observation was also made by Leinfelder et al (1972) who presented

results from X-ray diffraction studies and hardness measurements on the AuCu

alloy and on a series of proprietary dental gold alloys during various stages

of aging in the temperature region around 623 K. These authors concluded that

hardening of the proprietary alloys was probably a dual mechanism involving

precipitation and order hardening. The ordered structure was AuCu and the

precipitate was thought to be a gold-rich or silver-rich phase. As concerns

the binary AuCu alloy specimens it was shown that maximum hardness occurred

at about 75 % full order. Further ordering of AuCu was associated with a

decrease in hardness.

In Paper III of the present work the preliminary hardness tests revealed that

the maximum hardness values were found near the Au-Cu edge of the investigated

part of the ternary Au-Ag-Cu diagram and around the composition range Au-Cu,

50 atomic per cent. The results of the present work, although dealing with

alloys representing high temperature conditions, thus indicate the importance

of the AuCu complex in the hardening of the alloys.

Starting from the Au-Ag edge of the ternary Au-Ag-Cu diagram and proceeding with

decreasing Ag content towards the Au-Cu edge reveals a change in the colours of

the alloys. Along the cuts corresponding to 25 and 50 atomic per cent Au the

alloy colour changes from greenish or yellowish through a continuous colour

series to strongly reddish. However, along the 75 atomic per cent Au cut the

colour changes only from yellow to reddish yellow and in far fewer directly

observable stages (Sterner-Rainer 1926, Leuser 1949). Not forgetting the

decreasing Ag-content during this change of colours, the Cu-content seems to

53

influence the colour properties to a very great extent.

The studies discussed concerning resistivity, hardness and colour in relation

to the presence of ordered species point out the existence of a mainly qualita­

tive relationship between certain physical properties and the amount of

complexes. It is possible that these properties may be more precisely governed

by an individually adapted heat treatment of the alloys in question. However,

at the present time it is difficult to interpret the special effect on a

physical property that is caused by a certain complex. The use of better tools

such as a determination of the molar absorption coefficient could be one way

of obtaining a quantitative relationship between the amount of a complex and

its individual effect on a physical property.

CONCLUSIONS IN SUMMARY (Papers III and IV)

The tentative model obtained from the calculations of Papers III and IV is

associated with the situation existing in an ideal solution with complexes.

In a solid solution with the components A and B the solvent is a matrix of

A and B atoms. In this solution certain complexes ApB^ exist. These complexes

are perhaps more or less solvated with A and B atoms. A state of equilibrium

between ordered species and disordered matrix, that is to say between com­

plexes and solution, is attained. According to the results of the present

work this complex formation seems to follow the law of mass action. When the

thermal motion is no longer able to keep the complexes separated they cluster

together. At temperatures much lower than those considered in the present

work the concentrations of the complexes are high enough to exceed the solu­

bility products and the complexes precipitate. In addition it was found that

the length of the edge of the face centered cubic unit cell was changed by

the influence of ordered species appearing in the disordered matrix. The

results seem to indicate that the cell edge of the cubic unit cell varied in

54

accordance with an extended Vegard's law.

The proposed model, which seems to be confirmed by the experimental results

of the present work, indicates a new way of attacking some ordering processes

occurring in certain alloy systems. The complexes have been expressed in terms

of free units of atoms from each alloy constituent. This simplified picture

may hopefully contribute to the understanding of the ordering processes. The

equilibrium calculations cannot provide any direct information regarding the

actual structure of the complexes although in Paper III an attempt was made

to give a structural interpretation of the results.

The best model obtained from the calculations in Paper IV has been illustrated

in a diagram (Fig. 29) revealing the distribution of free Cu and complexes as

functions of X^, the total mole fraction of copper. The diagram represents

the conditioris at 1100 K. According to this model the fraction of free gold is

very small (x^u < 0.003) and therefore the alloys can be regarded as consisting

of gold-copper complexes in mainly a copper matrix. At < 0.5, that is to

say in the region of the composition of the dental alloys, the dominant com­

plexes are AuCu followed by CuAu^, CuAUg and AuCu^. It also deserves mention

that one of the models presented in Paper IV, See Table IX, suggests the exist­

ence of trimers of gold (Au^-aggregates) in the most gold rich part of the

alloy system. The AuCu, AuCu^ and CuAu^ complexes may, as a suggestion,have

the structures illustrated in Fig. 30. The solution and illustration of the

structure of CuAUg is somewhat more intricate but may be similar to that of

CuAu^ although it is a larger species.

55

OL

Cu Au

Au CuCu Au

AuCu

Fig. 29. Distribution of free Cu and complexes as functions of X^, the total mole fraction of copper. At a given X^, the fraction of a species, a, is represented by the segment of a vertical line falling within the corresponding range.

Fig. 30. Tentative structure units taken out from the fee lattice and represent­ing the complexes AuCu^ (left) and AuCu (right). CuAu^ is the reverse of AuCu^. •= Au,o= Cu. In AuCu3 1/4 and in AuCu 1/8 of each copper atom belongs to the

complex.

56

However, there are various possibilities for determining the structures of

the complexes more definitely in the future.

(i) With a transmission electron microscope of sufficient resolution, using

the so-called lattice-image technique it would be possible to make visible

the exact projected atomic positions. This technique is of topical interest

in the literature.

(ii) By use of electron spectroscopy for chemical analysis (ESCA) it might

be possible to determine the bonding energy of the inner electrons of an

element. As this bonding energy depends on the chemical surroundings of the

element in question the ESCA technique is useful for structural analysis.

In this way the bonding situation within the complexes could be clarified

and the free atoms and the atoms within the complexes could each be studied.

When the solution conditions within the solid gold-copper alloys at elevated

temperatures are known, the alloys ought to be studied at successively lower

temperatures. In this way it could be established whether there is any

defined solubility boundary for a certain complex. A detailed knowledge of the

amounts and structures of the complexes existing in the temperature region 1000-

1100 K in the basic dental alloy gold-copper, will open the way for direct

steering of the solid state reactions in this system and thereby also the

structure-sensistive physical properties. In this way the inherent possibili­

ties of the simple basic alloys could be better utilized. This in turn implies

that dental gold alloys with a decreased number of alloy elements could turn

out to be usable for the different kinds of dental gold restoration. This

will finally be an advantage to our patients as fewer components can be dissol­

ved from the alloys due to corrosion.

57

GENERAL SUMMARY

The purpose of the present work was:

On commonly used proprietary dental gold alloys to study changes of the element

distribution appearing after so-called homogenization of soldered joints and to

determine in vitro the dissolution rate of cadmium from solder alloys (I, II).

To study cluster formation in solid solutions of the basic dental alloys gold-

silver-copper and gold-copper (III, IV).

The most important, results and the conclusions reached were as follows:

The distribution of available alloy components in the region of the gold-solder

junction after homogenizing heat treatment revealed that concentration diffe­

rences had not been levelled out to any great extent. On the contrary certain

negative effects of the heat treatment were established the most important

among which was the appearance of small demarcated domains with alloys of

various compositions precisely at the gold-solder junction. As these alloys

have chemical and physical properties which in each separate case are un­

controllable it must be concluded that "homogenizing" heat treatment of gold

soldered assemblies should not be performed as a routine (I).

The dissolution rate of cadmium from dental gold solder alloys under circum­

stances which may be said to represent very unfavourable conditions within the

oral cavity was determined. This dissolution rate was rather small. The bio­

logical significance of a corresponding dissolution in vivo in the long run is

hard to evaluate at the present time. However, taking into consideration addi­

tional effects due to further sources of cadmium intake and also the risk to

the dental technician from inhaling cadmium vapour during soldering it was

58

concluded that dental gold solder alloys containing cadmium should not be

used (II).

The variation of the lattice parameter of the cubic unit cell of from 1073 K

quenched gold-silver-copper alloys of different compositions was determined.

The cell edge seemed to vary in accordance with an extended Vegard's law. The

results were interpreted in terms of complexes and the corresponding equilib­

rium constants. The proposed model was further and more directly confirmed by

the results of emf-measurements of gold-copper alloys at temperatures around

1100 K. According to the tentative model assuming complexes in a matrix

of A and B atoms there is a state of equilibrium between ordered species and

disordered matrix, that is to say between complexes and solution. The complex

formation seems to follow the law of mass action. It was concluded that a

detailed knowledge of the amount and structure of the complexes existing in

the temperature region 1000-1100 K in the basic dental alloy gold-copper,

will open the way for direct steering of the solid state reactions and thereby

also the structure-sensitive physical properties. In this way dental gold

alloys with a decreased number of alloy elements could turn out to be usable

for the various kinds of dental gold restoration (III, IV).

59

ACKNOWLEDGEMENTS

It is a pleasure for me to express my sincere gratitude to all those who have

guided, assisted and supported me during the course of these investigations.

Special thanks go to Professor Lars Holmlund and Professor Nils Ingri, for

their unfailing interest in this work, for numerous stimulating and fruitful

discussions and for all the facilities placed at my disposal. Thanks are

also due to Docent Erik Rosén and Docent Olle Ginstrup for interesting

discussions, constructive criticism and generous help during the course of

this work.

Fi1. kand. Åke Björnham helped me with statistical problems and his performance

is much appreciated. In addition I wish to express my sincere gratitude to

Fi 1. lic. Staffan Sjöberg, Fil. dr. Lage Pettersson and Fil. dr. Ragnar Tegman

for valuable discussions concerning computer calculations and experimental

problems.

I am also indebted to Miss Ewa Gruffman and Chief technician Rudolf Lindberg

for skilful laboratory assistance and to Mrs Lilly Leijonhufvud for excellent

secreterial help. Thanks are also due to Lector Patricia Shrimpton who revised

the English manuscript and made helpful suggestions on linguistic points.

The work has been financially supported by the Swedish Medical Research

Council, grants K70-24X-3043-01, K71-24X-3043-02K and B74-24X-3043-03 and

by the Faculty of Odontology, University of Umeå.

Umeå in January 1976

Maud Bergman

60

REFERENCES

American Dental Association specifications. 1973. Guide to dental materials and devices, ed. 6, ADA, Chicago, Illinois, pp 182-185

Arnek, R., Sillen, L.G. & Wahlberg, 0. 1969. High-speed computers as a supplement to graphical methods, 8. Some devices to speed up computations on chemical equilibria and simplify programming for LETAGROP. Application to complex formation. Arkiv Kemi 31, 353-363

Arunachalam, V.S. & Cahn, R.W. 1967. Order-hardening in CuAu. J. Mater. Sci.2, 160-170

Asgar, K., Charbeneau, G.T., Myers. G.E., O'Brien, W.J., Powers, J.M. 1975. Restorative dental materials. Craig & Peyton, C.V. Mosby Co. Saint Louis, pp 142, 333-339

Bergenholtz, A., Hedegård, B. & Söremark, R. 1965. Studies of the transport of metal ions from gold inlays into environmental tissues. Acta Odont. Scand.23, 135-146

Bergman, B., Bergman, M. & Söremark, R. 1976. Gold solder implants in rats.To be published

1 DQBerlin, M. & Ullberg, S. 1963. The fate of Cd in the mouse. Archives of Environmental Health 7, 686-693

Björn, E. & Hedegård, B. 1965. Element distribution in gold-platinum alloys.Acta Odont. Scand. 23, 323-333

Brauner, P., Sillén, L.G. & Whiteker, R. 1969. High-speed computers as asupplement to graphical methods. 9. Adjustment for systematic experimental errors and other "group parameters" in LETAGROP. Applications to potentio- metric titrations. Arkiv Kemi 31, 365-376

Brown, L.C. & Thresh, H. 1970. Tools and techniques in physical metallurgy.Voi. 2. Weinberg, F., Marcel Dekker Inc. New York, pp 601-648

Bumm, H. 1939. Die Ausbildung binärer Uberstrukturphasen unter dem Einfluss beschränkt löslicher Zusätze. Z. Metallkd. 31, 318-321

Charpy, M.J. 1952. Eczemas de sensibilisation à divers produits absorbés par la muqueuse buccale. Bull. Soc. Fr. Dermatol Syphiligr. 59, 338-339

Craig, R.C., O'Brien, W.J., Powers, J.M. 1975. Dental materials, properties and manipulation. C.V. Mosby Co% Saint Louis, pp 190-191

El-Ebrashi, M.K., Asgar, K. & Bigelow, W.C. 1968. Electron microscopy of gold soldered joints. J. Dent. Res. 47, 5-11

Eigart, M.L. & Higdon, R.S. 1971. Allergic contact dermatitis to gold. Arch. Derm. 103, 649-653

Eick, J.D. & Hedegård, B. 1968. The effect of heat treatment on cast dental gold alloys. Research report given at the IADR meeting in San Francisco in March 1968. Personal communication

61

Eick, J.D. & Hegdahl, T. 1968. Segregation in dental gold casting alloys.J. Dent. Res. 47, 1118-1127

Friberg, L. 1950. Health hazards in the manufacture of alkaline accumulators with special reference to chronic cadmium poisoning. Acta Med. Scand. 138 (suppl 240), 67-97

Friberg, L., Piscator, M. & Nordberg, G. 1971. Cadmium in the environment.The Chemical Rubber Co. Press, Cleveland, Ohio

Friberg, L. & Piscator, M. 1972. Långtidseffekter av kadmium. Laek. Tidn. 69, 983-987

Frykholm, K.O. & Hedegård, B. 1968. Fall av elektrokemisk korrosion i munhålan. Sven. Tandlaek. Tidskr. 61, 435-442

Frykholm, K.O., Frithiof, L., Fernström, Å.I.B., Moberger, G., Blohm, S.G. &Björn, E. 1969. Allergy to copper derived from dental alloys as a possible cause of oral lesions of lichen planus. Acta derm.-venereol. 49, 267-281

Gehlen, P.C. & Cohen, J.B. 1965. Computer simulation of the structure associated with local order in alloys. Phys. Rev. 139, 844-855

Germer, L.H., Haworth, F.E. & Lander, J.J. 1942. Order in the alloy Cu^Au.Phys. Rev. 61, 614-626

Greener, E.H., Harcourt, J.K. & Lautenschlager, E.P. 1972. Materials science in dentistry. The Williams & Wilkins Co, Baltimore, pp 360-362

Greenholz, M. & Kidron, A. 1970. The structure of short-range ordered alloys.I. Clustering of ordered cells II. Ordered zones in a disordered matrix.Acta Crystallogr. A 26, 306-314

Hansen, M. & Anderko, K. 1958. Constitution of binary alloys. Me Graw-Hill Book Co, Inc. New York, Toronto, London, pp 198-203

Hedegård, B. 1949. Homogenisering av dentala guldgöt. Sver. Tandlaek-Förb. Tidn. 10, 1-4

Hedegård, B. 1958. Homogenization. A clinical study of its significance incrown and bridge prosthesis, Trans. Roy. Sch. Dent. Stockholm & Umeå 1, 3-24

Hedegård, B. 1974. Dental material 1 ära. Sveriges Tandläkarförbunds Förlagsförening u.p.a. Stockholm, pp 131, 147-148

Hirabayashi, M. 1959. Electrical resistivity and superstructure of CuAuvJ. Phys. Soc. Jpn 14, 262-273 J

Hoar. T.P. & Mears, D.C. 1966. Corrosion-resistant alloys in chloride solutions; materials for surgical implants. Proc. R. Soc. London, Ser. A 294, 486-510

Hume-Rothery, W., Smallman, R.E. & Haworth, C.W. 1969. The structure of metals and alloys. The metals and metallurgy trust of the institute of metals and the institution of metallurgists, London, pp 121-122

Hägg, G. 1954. Fasanalys och identifiering med röntgendiffraktion. Compendium. University, Uppsala, Sweden

62

Ingri, N. & Sillen, L.G. 1964. High-speed computers as a supplement to graphical methods. IV. An ALGOL version of LETAGROP VRID. Arkiv Kemi 23, 97-121

Johansson, G. 1965. Analysis of column effluents by controlled-potential coulo- metry. Talanta 12, 163-170

Kawahara, H., Yamagami, A. & Nakamura, M. 1968. Biological testing of dental materials by means of tissue culture. Int. dent. J. 18, 443-467

Kawahara, H., Takashima, Y., Nakamura, M. & Yamagami, A. 1975. Electron micro­scopic study of the cytotoxicity of cadmium and mercury in vitro. J. Dent. Res. 54, 125-130

Kiessling, R. 1960. Elektronmikrosonden, ett viktigt mikroanalytiskt hjälpmedel. Jernkontorets Ann. 144. 11, 847-854

Kiukkola, K. & Wagner, C. 1957. Measurements on galvanic cells involving solid electrolytes. J. Electrochem. Soc. 104, 379-387

Krivoglaz, M.A. & Smirnov, A.A. 1965. The theory of order-disorder in alloys. American Elsevier, London, pp 55-64

Kuczynski, G.C., Hochman, R.F. & Doyama, M. 1955. Study of the kinetics of ordering in the alloy AuCu. J. Appi. Phys. 26, 871-878

Kurnakow, N., Zemczuzny, S., Zasedatelev, M. 1916. The transformations in alloys of gold with copper. J. Inst. Met. 15, 305-331

Lain, E.S., Schriever, W. & Caughron, G.S. 1940. Problem of electrogalvanism in the oral cavity caused by dissimilar dental metals. J. Amer. dent. Ass.27, 1765-1772

Le Blanc, M. & Wehner, G. 1932. Untersuchungen Liber die Umwandlungen in fester Phase beim System Kupfer-Gold. Ann. Phys. 14, 481-509

Leinfelder, K.F., O'Brien, W.J. & Taylor, D.F. 1972. Hardening of dental gold- copper alloys. J. Dent. Res. 51, 900-905

Leuser, J. 1949. Uber die Besonderheiten der Edelmetallegierungen im Schmuck­gewerbe. Metall 3, 105-110

Manenc, J. 1972. Structural thermodynamics of alloys. D. Reidel pubi. Co. Dordrecht-Holland, Boston-U.S.A., pp 12-28

Maschinski, G. 1970. Potentialmessung an Metallen in der Mundhöhle. Zahnaerztl. Praxis 21, 29-30

Nielsen, J.P. & Tuccillo, J.J. 1966. Grain size in cast gold alloys. J. Dent. Res. 45, 964-969

Nowack, L. 1930. Vergütbare Edelmetall-Legierungen. Z. Metallkd. 22, 94-103

Ogawa, S. & Watanabe, D. 1952. Electron diffraction study on the ordered alloy Au^Cu. J. Phys. Soc. Jpn. 7, 36-40

Parrish, W. 1960. Results of the I.U. Cr. precision lattice-parameter project. Acta Crystallogr. 13, 838-850.

63

Phillips, R.W. 1973. Skinner's Science of Dental Materials. 7th ed.,W.B. Saunders Co. Philadelphia, London, Toronto, pp 273-295, 381-388,563-576

Prince, A. 1966. Alloy phase equilibria. Elsevier pubi. Co. Amsterdam, London,New York, pp 3-7', 17-21

Rheinwald, U. 1953. Beiträge zur Elektrobiologie der Mundhöhle. Vili Mund­krankheiten durch elektrische Elemente. Zahnaerztl. Welt 8, 31-32

Sato, M. 1971. Electrochemical measurements and control of oxygen fugacity and other gaseous fugacities with solid electrolyte sensors. Research techniques for high pressure and high temperature. Ed. by Ulmer, G.C. Springer Verlag, New York, Heidelberg, Berlin, pp 43-56

Schriever, W. & Diamond, L.E. 1952. Electromotive forces and electric currents caused by metallic dental fillings. J. Dent. Res. 31, 205-229

Sillen, L.G. 1962. High-speed computers as a supplement to graphical methods I. Functional behaviour of the error square sum. Acta Chem. Scand. 16, 159-172

Spreng, M. 1940. Homogenität und Inhomogenität zahnärztlicher .Legierungen. Schweitz Mschr. Zahnheilk. 50, 158-163

Spreng, M. 1963. Allergie und Zahnmedizin. Auf1. 2. J.A. Barth Verlag, Leipzig, pp 62, 69, 92-104

Sterner-Rainer, L. 1926. Einige Eigenschaften der Legierungen Au-Ag-Cu.Z. Metallkd. 18, 143-148

Sykes, C. & Jones, F.W. 1936. The atomic rearrangement process in the copper-gold alloy Cu^Au. Proc. R. Soc., London 157, 213-233

Söremark, R., Ingels, 0., Plett, H. & Samsahl, K. 1962. Influence of some dental restorations on the concentrations of inorganic constituents of the teeth. Acta Odont. Scand. 20, 215-224

Söremark, R., Freedman, G., Goldin, J. & Gettleman, L. 1966. Structure andmicrodistribution of components of gold alloys. J. Dent. Res. 45, 1723-1735

Taylor, N.O. & Teamer, C.K. 1949. Gold solders for dental use. J. Dent. Res.28, 219-227

Teramoto, K. & Tachiki, M. 1968. Electrical resistivity and thermoelectric power of CuAu alloy. Phys. Rev. 166, 656-666

Törnell, S.O. 1962. Allergiska reaktioner i munslemhinnan av dentalguld. Sver. Tandlaek. Tidn. 54, 151-154

Vetter, K.J. 1967. Electrochemical kinetics, theoretical aspects. Academic Press, New York, London, pp 119-120

WHO Food Additives Series 1972. Cadmium. No. 4, 51-57

Wise, E.M. 1964. Gold - recovery, properties and applications, van Nostrand Co., Inc. Princeton, New Jersey, pp 236-249