Sulfide Mineralogy 0939950014

Table of contents :
Page 1
Titles
COPYRIGHT
PRINTED BY
REVIEWS in MINERALOGY
Page 1
Titles
FOREWORD
FOREWORD to the FOURTH PRINTING
DISCLAIMER
Page 1
Titles
..............
Page 2
Page 3
Titles
........... , ... ,
Page 4
Page 1
Titles
Ch.1
Page 2
Titles
W-2
Page 3
Page 4
Titles
."
/
W-4
Page 5
Titles
w-S
Page 6
Page 7
Titles
-1
Page 8
Page 9
Titles
W-9
Page 10
Titles
1s73.
Page 11
Titles
W-ll
Tables
Table 1
Table 2
Page 12
Page 13
Page 14
Page 15
Page 16
Page 17
Titles
t t
t
t
t t
t
Page 18
Page 19
Page 20
Page 1
Titles
Ch.2
Page 2
Titles
e
Page 3
Page 4
Page 5
Page 6
Titles
Os
b
OC"
.Fo
Os
Qc, e,. Os
()Cln'lI~ted O~
Page 7
Tables
Table 1
Page 8
Page 9
Titles
a
.zn
Os
.Cu
OAs o-
Page 10
Tables
Table 1
Page 11
Page 12
Titles
c
Page 13
Page 14
Page 15
Page 16
Page 17
Titles
b
c
d
Page 18
Page 19
Page 20
Titles
.,
"
'" N
.... '"
..... ,0
'" ""
p.. p..
'"
" "
i 1
.., ..,
M 0 N 0
,..... 11"'1 N 0
N N N N
8
.....
13
......
\0\00 0"I....:t M
1:5
'"
'-' ...
..... '-'
'" '"
N '" ....
p.. p.. p..
" "
.& .&
~ ~
..... N
" 0 11"'1 N
.. . ..
co co co ....:t \0 co
'"
0'\ ...:t \0 0'\ N r""f
.....
r""f ('I") N r-f co ....:t
... '-'
'-' .... 0
"" N ""
N til tIl_
en 0 N ('--
co r-f N""'"
- - _ r-f
00 CD CD....:t
.. H" N" \0
.&l (/) -'l en,.c M
rn ....:t 00 ex) U') ,.Q
'_"c'-"',.Q"_'CI)
M CI) N til ,.....,.....
r-f 11"'1 1'-1 0'\ ......t ......
,0 ,0 ,0 ,0 ,0 ,0
p.. p.. p.. p.. p.. p..
~
'"
'"
'-'
"
'-'. I!.
i
~----
: I ~I I Agt+u S + Ag
-- - :...L --- I
------- I -_-_-_-~::~ ------ I
r--~z_;~-;;---~ I \ I :-r----------- I
1- --~-- ~ J
\ I I I ][
\ I I_----__L----------- I
-------- I I
------; I i---------r-----------
I I I i
I I \ m i :m
! I! i
i
Page 24
Page 25
Page 26
Page 27
Titles
9
Page 28
Tables
Table 1
Page 29
Titles
HIGHER
POLYTYPES
r
-
/'
/'
/'
Page 30
Page 1
Titles
Reprinted by permission.
Ch.6
SULFIDE PETROLOGY'
B-1
Page 2
Titles
PAUL B. BARTON, JR.
B-2
Page 3
Titles
B-3
Page 4
Titles
~
"
Page 5
Titles
SULFIDE PETROWGY
°
~
a-s
Page 6
Titles
·
·
.::
B-6
Page 7
Titles
B-7
Page 8
Titles
x = oIY
B-8
Page 9
Titles
B-9
Tables
Table 1
Page 10
Titles
PAUL B. BARTON, JR.
B-IO
Page 11
Page 1
Page 2
Page 3
Page 4
Titles
-------,
-------,
Page 5
Page 6
Page 7
Titles
-------, -------
Page 8
Page 9
Page 10
Page 11
Page 12
Page 13
Page 14
Page 15
Page 16
Page 17
Page 18
Page 19
Page 20
Titles
-------,
Page 21
Page 22
Page 23
Titles
-------, ------
Page 24
Page 25
Page 26
Page 27
Page 28
Page 29
Page 30
Page 31
Page 32
Page 36
Page 37
Page 1
Page 34
Page 35

Citation preview

COPYRIGHT Reserved by the authors, 1976 Fourth Printing 1982

PRINTED BY BookCrafters, Inc. Chelsea, Michigan 48118

REVIEWS in MINERALOGY (Formerly: ISSN VOLUME 1:

SHORT COURSE NOTES) 0275-0279 SULFIDE MINERALOGY

ISBN 0-939950-01-4

Additional copies of this volume as well as those listed below may be obtained at moderate cost from Mineralogical Society of America 2000 Florida Avenue, NW Washington, D.C. 20009 No.of

Vol.

~

1

SULFIDE MINERALOGY.

2

FELDSPAR MINERALOGY.

P.H. Ribbe, Editor (1974) P.H. Ribbe, Editor (1975; revised 1982)

3

OXIDE MINERALS.

4

MINERALOGY and GEOLOGY of NATURAL ZEOLITES. F.A. Mumpton, Editor (1977)

5

ORTHOSILICATES.

Douglas Rumble III, Editor (1976)

P.H. Ribbe, Editor (1980; revised 1982) R.G. Burns, Editor (1979)

284 ~350 502 232 ~410

6

MARINE MINERALS.

7

PYROXENES.

C.T. Prewitt, Editor (1980)

380 525

8

KINETICS of GEOCHEMICAL PROCESSES. A.C. Lasaga and R.J. Kirkpatrick, Editors (1981)

391

9A

AMPHIBOLES and ,Other Hydrous Pyriboles - Mineralogy. D.R. Veblen, Editor (1981)

372

9B

AMPHIBOLES: Petrology and Experimental Phase Relations. D.R. Veblen and P.H. Ribbe, Editors (1982)

~375

FOREWORD Short courses of mineralogical

interest were begun in 1965 in conjunction with

the annual meetings of the Geological Society of America.

Sponsored by the American

Geological Institute Committee on Education and directed by J.V. Smith of the University of Chicago, short courses of feldspars (1965), pyroxenes and amphiboles (1966), sheet silicates (1967), and resonance spectroscopy mineralogists with expertise in these subject areas. course notes became more comprehensive

(1968) were presented by

With each succeeding year the

and formalized, and AGI published these at

low cost for distribution in the geological community.

Unfortunately,

AGI has been

financially unable to continue this service. In 1973 President J.V. Smith surveyed members of the Mineralogical America concerning the desireability

Society of

of renewing this sort of effort, and in response

the M.S.A. Council appointed a committee to initiate a series of short courses under their sponsorship.

The mineralogy

of sulfides was selected as the topic for the first

of these courses, and the primary lecturers, J.R. Craig, C.T. Prewitt, S.D. Scott, and B.J. Wuensch, gathered in Blacksburg, Virginia in May 1974 to organize their presentation.

The assistance of V. Rajamani as both co-author and lecturer and P.B.

Barton as special lecturer was enlisted, and the work of writing and editing this volume began.

The Short Course on Sulfide Mineralogy was given November 15-17, 1974,

preceding the annual meetings of the affiliated societies of the Geological Society of America, at the Sheraton Four Ambassadors Hotel in Miami, with eighty persons in attendance.

The editor and committee chairman is particularly grateful to Professors J.V. Smith and S.W. Bailey for their strong support of this effort.

As Treasurer of

M.S.A., Dr. Philip M. Bethke volunteered much helpful financial advice, and Mrs. Mary Holliman, Managing Editor of The American time in technical editing.

Mineralogist,

gave considerable

Cheryl Crum was responsible for most of the typing,

assisted by Ramonda Haycocks, Lyn Groover, and Margie Strickler.

John B. Higgins

proof-read the final copy, and S.A. Kissin kindly provided data from his unpublished dissertation. Paul H.

Ribbe

Blacksburg, Virginia October 18, 1974

FOREWORD

to the FOURTH

PRINTING

This book was first published under the title SHORT COURSE Sulfide Mineralogy. and 1500 in 1979.

NOTES,

VOLUME

1:

Fifteen hundred copies were printed in 1974, 1500 in 1976, This, the fourth printing (2000 copies), was completed in Au-

gust 1982 under the new serial title, REVIEWS 1980 by the Mineralogical

Society of America.

in MINERALOGY,

which was adopted in

The 1976 lists of errata, addenda,

supplemental references, and a supplement to the table beginning on page CS-3 are included on page v and pages R-33ff. at the back of this volume.

DISCLAIMER This book was prepared by the authors for the use of participants in the Sulfide Mineralogy Short Course and for others who desire an introduction to the mineralogy of sulfides.

No claim is made that they are complete; they have not

been openly reviewed. (vi)

TABLE OF CONTENTS

..............

ERRATA AND ADDENDA FOR THE SECOND PRINTING (continued at the back of the book on pages R-33ff.) FOREWORD

• . • . . . . . • . . . . . . . . . ..

DISCLAIMER

(v)

(vi) (vi)

Chapter

1.

DETERMINATION, RELATIONSHIPS, AND CLASSIFICATION OF SULFIDE MINERAL STRUCTURES

(B. J. Wuensch) Introduction.

. •

Crystal Structure Problems

W- 1 Determination

in Sulfide

2

Structure Determination

Suitable specimens Chemical and structural X-ray absorption

5 5 6 7

complexity

Packing Considerations .•• Polymorphism Derivative

8

and Polytypism

13

Structures.

"Composite"

16

Structures.

Classification

17

Schemes.

18

Chapter 2.

SULFIDE CRYSTAL CHEMISTRY

(B. Introduction. Elemental

J.

Wuensch)

• •

W-2l

Sulfur.

21

Ionic Sulfides ••

23

Disulfides

23

and Derivatives.

Monosulfides

and Derivatives.

The Nickel Sulfides

25

• • • .

31

Silver and Copper Sulfides.

32

Rocksalt

34

Derivatives •..•

Group V Metal Sulfides

and Related Sulfosalts

34

The Group V sulfides Stibnite derivatives The Bi sulfosalts Lead-Antimony Lead-Arsenic Additional

35 '36 38

Sulfosalts. Sulfosalts

Sulfosalt

Chapter 3.

39 •

41

Structures

ELECTRON

INTERACTIONS

43 AND CHEMICAL BONDING

IN SULFIDES

(C. T. Prewitt and V. Rajamani) Introduction. Electronic

. • • •

Structures

PR- 1 of 'the Elements

2

Wave equations Quantum numbers and the hydrogen atom The periodic system Filling of the d orbitals of the transition metals

2

(i)

3 5 5

Group

6

Theory .•.•••..••.•

6

Basic principles Classes of symmetry operations Representations of groups Orbitals and energy levels

9

10 13

in Sulfides.

14

Crystal field theory Molecular orbital theory Band theory

14 18

Chemical

Bonding

Some Aspects Physical

of Metal-Metal

Properties

22

Bonding

in Fe, Ni, and Ni Sulfides .•

of Transition-Metal

Sulfides

and Band Theory.

Interatomic

Distances

Application

and Covalency

of Bonding

Theories

in Sulfides

to Specific

32

• • •

34

Sulfides.

35 36 38

Pyrite Thiospinels Pentlandite Conclusions

41

Chapter

4.

EXPERIMENTAL (S.D.

METHODS

IN SULFIDE

S- 1 1

of Experimentation

2 3 4

Appearance-of-phase method Equilibrium Applications to nature

5

Methods •••••••

5 9 10

Evacuated silica tube Differential thermal analysis Salt flux Hydrothermal recrystallization Analysis

of Run Products

11

17

•

17

Microscopy X-ray diffractometry Microprobe analysis Presentation Sulfur

of Experimental

Activity:

18 23 Results

Its Measurement

• • • • •

and Control.

of Large

Single

24 26 29 30 30 34 34 35

SL-V buffer Dew point Gas mixing Electrum tarnish Pyrrhotite indicator Electrochemical cells Growth

SYNTHESIS

Soott)

• • •

Some Principles

27 31

Reflectivity

Introduction

23

36

Crystals.

36

Growth from melts Sublimation Chemical vapor transport Growth in halide fluxes Growth in gels Hydrothermal growth

37 37 37 37 38 (ii)

Chapter

5.

SULFIDE PHASE EQUILIBRIA

(Craig and Soott) Introduction

•

CS-l

(Craig)

Survey of Data Sources on Sulfide Phase Equilibria Major

Sources of Thermochemical

3

Data on Sulfides

20

(Soott)••.••••

The Fe-S System

21

Phases

24

Troilite 24, Mackinawite 25, "Hexagonal" pyrrhotite 26, Monoclinic pyrrhotite 28, "Anomalous" pyrrhotite 29, y iron sulfide 29, Smythite 29, Greigite 30, Pyrite 30 Pyrite-Pyrrhotite

Solvus

31

Thermochemistry

32

Pyrite + Hexagonal

Pyrrhotite

Solvus

33

Pyrrhotite field 34, Pyrite field 36, heat of formation of pyrite 36 Effect of Pressure FeS activity solvus 37 The Fe-Zn-S

System

Free energy and

on Phase Relations in pyrrhotite

36 Pyrite + pyrrhotite

36,

(Soott) . • • • •

•••

t

•••

I

•••••

FeS Activity and Sphalerite Composition Sphalerite + Troilite + Iron Sphalerite + Pyrrhotite Sphalerite + Pyrrhotite + Pyrite Iron-rich patches Effect of Pressure on the Sphalerite + Pyrite + Pyrrhotite Solvus: The Sphalerite Geobarometer Phase Relations Below 300°C Partial Molar Volume of FeS in Sphalerite (V-sp) FeS The Cu-S System (Craig)•••• Chalcocite Remaining"

60, Djurleite Covellite 61,

61, Anilite CuS 62 2

61,

Covellite

System

(Craig)

...........

,

51 54 56 58

and "Blue-

Thermochemistry The Cu-Fe-S

41 43 44 47 47 48

...

62 ,

64

Digenite 70, Bornite 71, X-, Anomalous, or Sulfur-rich Bornite 71, Cubanite 71, Idaite 72, Fukuchilite 73, Chalcopyrite 73, Talnakhite, Intermediate Phase I, Intermediate Phase II 73, Mooihoekite and Intermediate Phase A 73, Haycockite 74, CUO.12FeO.94sl.00Phase 74 Thermochemistry The Ni-S System

74

(Craig)•••••••••

Heazlewoodite 79, 80, Vaesite 80

Godlevskite

77 79,

Millerite

79,

Polydymite

Thermochemistry The Fe-Ni-S

System

Pentlandite Monosulfide

80

(Craig)

87, Violarite Solid Solution

82 88, Bravoite (mss) 89

Thermochemistry

89,

(Fe,Ni)l_xS 89

(iii)

(Craig) •.

Sulfosalts

CS-9l

Thermochemistry Stoichiometry

93

of Sulfides

(Soott) . . . . . • . . . . . . . . • . • .•

Silver Sulfide Molybdenite Mercury Sulfide Zinc Sulfide

99 102 102 103 l04

Nomenclature 104, Nonstoichiometry of zinc sulfide 105, Sphalerite-wurtzite inversion 106, Wurtzite polytypes 109. Chapter 6.

SULFIDE PETROLOGY

(P. B. Barton) (Paper reprinted

from MineraZogioal Sooiety of Amerioa Special Paper 3, l87-l98 (l970» B- 1

Introduction. Complexity

of the Problem of Interpreting

Thermodynamic

Approach

to Phase Diagrams.

General Factors Influencing The Sulfidation

Mineral Associations.

1

• • •

4

the Compositions

of Minerals.

State of Natural Environments

9

Goals of Current Research

11

References •••••.••

11

REFERENCES

(Unified bibliography

R- 1

• .

R-32

REFERENCES

R-33

ACKNOWLEDGMENTS SUPPLEMENTAL

except for Chapter 6).

(added for the second printing)

(continued from page (0».

R-34

SUPPLEMENT TO Survey of Data Souroes on Sulfide Phase Equilibria by J. R. Craig ••••.••••••.••.•••.••••

R-36

ERRATA AND ADDENDA

(iv)

DETERMINATION,

Ch.1

RELATIONSHIPS,

OF SULFIDE

AND CLASSIFICATION

MINERAL

Bernardt

STRUCTURES

J. Wuensch

INTRODUCTION

Sulfur occurs sulfides, pounds

in nature as a major

the sulfosalts

exist, of course,

NaCo[NCS]4'8H20,

(Frondel,

metallic

19S2) is a thiocyanate;

Voltzite

species.

The present

Such minerals

series

of wurtzite

are of interest

metals

of lectures

are combined

have a distinct

sulfosalts

role

component

because

present chemical

discussion.

related

metal.

garded as sulfosalts form a trigonal

pyramid

or, alternatively,

of neighbors.

In contrast,

a few minerals

is tetrahedral

bonds -- e.g., parkerite; Marumo,

and Nowacki,

fides if a distinction Sulfur

assigned

to oxygen.

more metallic VI elements

*Dana's minerals

system

configuration

of 2.S is moderate Still heavier

are available

also includes,

in which

re-

with the metal,

[1+2+2] pyramidal

arrangement

Cu3AsS4)

or involves CuAsS

are best included

in

metal-metal (Kulpe, 1961; among the sul-

is to be made.

under

a Group V element

common

in comparison

elements

as their atomic number d orbitals

commonly

together

(Fleet, 1973) or lautite,

The latter minerals

has the s2p4 electron

Its electronegativity

in the

of sulfur about the Group V metal

(e.g., enargite,

Ni3Bi2S2

1964).

Many

included

best based upon crystal-

which,

a square

from sul-

artificial.

in those structures

the coordination

are involved

of the sulfosalts

is probably

three S neighbors

are rare. The

one of the metals

and are accordingly

distinction

in which

Pb, Cu,6r Ag, and less

and is somewhat

The Group V metal

has either

minerals

few metals

- usually

Separation

because

in which more than three

in which

Relatively

concepts

in mineralogy

PbTlAgAs2SS)

of minerals

to sulfides,

Any modern

characteristics.

(Kupcik,

of the rarity of these

with sulfides,

Minerals

in sulfosalts

Zn, Hg, or a transition

are closely

Sb2S20

as ZnSS 0, but has been 4 and a zinc-bearing organo-

(e.g., hatchite

of this family

fides is based upon early chemical sulfosalts

sulfur*.

such as As, Sb or Bi.

as the second metallic frequently

is concerned

with

structural

are a sub-group

is a Group V element

the

Julienite,

kermesite,

compounds in nature. Sulfates are the only salts important S042- represents the final product of oxidation.

one or more metals

groups:

of sulfur com-

was long described

1967) to be a mixture

compound.

of three mineral

Many other classes

and a few are known as mineral

(Preisinger,

1967) is an oxysulfide. shown

constituent

and the sulfates.

to all Group VI elements. to the value of 3.S

in Group VI become

increases.

progressively

In sulfur and subsequent

Group

for hybridization.

"sulfides",

the selenides

is the electronegative

W-l

and tellurides,

element.

and

The crystal

chemistry

of the sulfides

thus differs markedly

from that of

oxygen and is more complex. Several types of bonds might be formed to complete 2 4 2the s p subshell: Two electrons might be added to form S ; one electron might be added and a single

electron-pair

bond formed

bonds might also be formed or, alternatively, a variety

of possible

sulfides

hybrid orbitals.

appreciable

metallic

are known for sulfur mely variable. occurs

Regular

about sulfur

salts, however, atomic

distorted

of nearest-

of a coordination Our discussion

of the structural

procedures

is not intended

and relationships

in crystal

the experimental

the structural

Chapter

classification discuss

The following constants

crystallographic

occupied

of sulfides.

space group

(which embraces

symmetry.

to describe

tensor elements

of the site.

a position

W-2

the unit

is only partially in disorder;

sometimes

involves

displacement

and para-

A more general

to be

representation

second-order

the atom occupies

requires

an assumption

is assumed

of the thermal ellipsoid

some elements

to unit cell

within

of each atom about its equilibrium

When

This condition

or may require

(referred

In this case a symmetric

thermal motion.

materia

about the space lattice

atom contained

of displacemet.t.

the shape and orientation

point symmetry

information

of thermal motion

variation.

ot this chapte

angles of the unit cell); the

by two or more species

of the direction

chemistry.

some basis for

section

the coordinates

independent

crystal

presupposes

of atoms in a crystalline

and interaxial

-- that is, the root-mean-square

an ellipsoidal

of sulfide

with

determina-

DETERMINATION

the arrangement

occupied

discussion

as background

which have been proposed.

STRUCTURE

the thermal vibration

and we will outline

with sulfide structure

of each site in cases where

The description

is required

is divided

The latter

it is presented

of each symmetry-independent

or statistically

structures,

The concluding

schemes

of isotropy

assumes

Inte

of lA or more.

and even the specificati

determination.

of this material

the various

data specify

meters which describe position.

and sulfo-

are encountered.

of sulfide mineralogy

associated

class, and crystal system);

cell; the occupancy

coordination sulfides

arbitrary.

2 is a discussion

(the dimensions

edges as a basis)

is thus extre-

over an interval neighbors

Instead

of the presentation

CRYSTAL

type, crystal

polyhedra

between

structure

difficulties

tion may be contrasted.

will accordingly

or octahedral

somewhat

aspects

to be comprehensive.

The organization

may give

In this chapter we will review some of the basic con-

cepts of atomic packing

lattice

coordination

thus becomes

in structures

In more complex

and second-nearest

polyhedron

into two chapters.

which

tetrahedral

often range almost continuously

The designation

presentday

Its coordination

and symmetric

of bonds involving

separations

All of these modes of bond formation

in many simple sulfides.

extremely

di5tances

Close metal-metal

character.

(see Ch. 3).

(e.g., SH-); two electron-?air

a larger number

must conform

equalities

to be identically

a position

tensor of to the

among certain zero.

Along with

parameters

which depend on the perfection

mary" and "secondary"

extinction),

and microstructure

of a specimen

the above data define completely

tion effects which will be produced by a crystal of a given mineral. ence between observed and computed diffraction for establishing

effects provides

of the diffraction

maxima produced by a crystal depend only upon

the size and shape of the unit cell.

The space group is established

usually with an ambiguity which may be resolved by recourse physical properties,

or the statistical

from the synnnetry of the diffraction

distribution

of diffracted

effect and systematic

intensities

of the diffraction

intensities

ture is determined. for physical experiment

thus constitute

Measured

values for the

the raw data from which a crystal struc-

Before analysis, however,

the intensities

must be corrected

and geometric factors which depend on the type of diffraction

and the shape of the specimen used.

to X-ray polarization, the Lorentz

intensities)

absences among the

maxima depend on the types and

of the atoms contained within the unit cell.

diffracted

(though

to crystal morphology,

intensities.

The relative positions

Correspond-

the usual basis

the veracity of a structural model.

The position

set of possible

("pri-

the diffrac-

The former include an effect due

which depends upon the value of the diffraction

factor, essentially

a measure of the relative

fraction as the crystal is rotated through the diffracting

angle, and

time available position.

for dif-

The factor

dependent upon specimen shape is due to absorption of both the incident and diffracted beam by the specimen.

After making

tions, the square root of each corrected ture factor", F, a sort of normalized depends only upon the diffraction

the physical and geometric

intensity is proportional

amplitude

correc-

to the "struc-

for each diffracted beam which

angle and the type and arrangement

of atoms in

the unit cell. The process

of structure determination

literally and figuratively,

to the process of formation of a magnified

a periodic object with a lens. object combine

In the optical process, rays scattered

to form a diffraction

then proceed on without

from these data may be compared, both

interruption

image of from the

pattern in the focal plane of the lens but to combine to form an image. The X-ray dif-

fraction pattern produced by a crystal is strictly analagous

to the optical dif-

fraction pattern.

of the scattering

Specifically,

density of the object. formation"

In the X-ray experiment,

is interrupted

ties of the diffraction routinely allowed

it is the Fourier transform however,

at the stage of the diffraction

maxima are recorded

The intensi-

or analytically,

source for image formation because

maxima in the pattern do not have the same relative phase.

phases is lost.

pattern.

pattern cannot, either experimentally

to act as a subsidiary

of the diffraction

the process of "image

in an experiment.

be

the

Only the amplitudes The data on relative

This is the infamous "phase problem" of crystal structure

determination. Two basic approaches

have evolved for solution of the phase problem.

concerns itself with the information structure

factor alone.

which is contained in the magnitude

A Fourier synthesis using each of the structure W-3

A first of the factors

(with their appropriate

phase) as coefficients

would be equivalent

tion from the diffraction

pattern.

electrons

per unit volume

at all points within

analogous

series which

uses instead

ture factors

as coefficients

maxima

."

at positions

structure

Summation

translated

the unit cell.

the squares

to the vectors

to a common origin.

a function between

The result

the number

Summation

of the magnitudes

may be shown to provide

which correspond

to image forma

of the series provides

0

of an

of the struc-

which disPlays

all atoms in the

of this summation

is known

as the Patterson function. A crystal which contains N atoms per cell will con2 tain N maxima in its Patterson map, of which N represent the interaction of an atom with

itself

developed

which

crystal

and thus occur at the origin. permit

structure.

A second

structure

The difficulty

body of procedures

factors

and probability through

the unraveling

a "boot strap"

to a sufficient be employed

These permit

Eventually,

or direct-phasing

procedures.

through

successive

puters

in the past deca~eJrefinement

Fourier

methods.

and extinction and calculated

structure

may involve

of a crystal

structure

is the "residual", Rearrangement weighted magnitude

0.001 to 0.0001

value

was done in early days

performed

com-

through

are used as observations,

thermal

motion

parameters,

to give a least-squares

A moderately

measured

the observed

complicated

structure

scale

fit betwee

sulfide

factors

and calculated

and precision

index"

and of

structure

of a crystal

measure

factors

structure

of the overall

agreement

or "R-factor":

l:liFb I-IF 111/l:Fb • o s ca 0 S R = l:F b (lllFl/F b )/l:F b ' shows R to be a as

os

difference

of R in a well-refined

atomic

coordinates

(in terms of fractions

on the scattering

refinement. precision

factors.

and commonly-used

"reliability

of the fractional

This provides

depends

factors

positions,

of

either

os

between

Fobs and F

' where

cal

the

of F is used as the weight.

A typical

7%.

structure

consists

through

of large high-speed

customarily

are adjusted

between

of .this definition,

average

is more

of the correctness

A convenient

determination.

series may

parameters.

of agreement

the measure

structure

the advent

two or three thousand

the order of 200 adjustable The degree

phases may be assigned

have been obtained

With

(e.g., atomic parameters)

observed

provides

inequalities of phases

that a Fourier

The refinement

syntheses.

The measured

and the trial parameters factor,

factors

Certain

the atomic positions.

The final stage in the determination / of the trial parameters ~hich

least-squares

the

rapidly with N.

assignment

probable

structure

improvement Patterson

increases

of their magnitudes.

may be derived.

of the larger

to reveal

have been

to obtain

is based on the fact that the phases of the

procedure.

number

procedures

function

of this procedure

are not independent

relations

Systematic

of the Patterson

The precision in interatomic

with

structure standard

is today of the order of 2deviations

of a cell edge).

power of the atom in question of the atomic distances

which

positions, is typically

W-4

in the range

The precision,

however,

as well as the level of

in turn, translates ±0.003

to ±0.007A.

to a

The above discussion crystal

structure

devoted

to its practical

is only a qualitative

determination. aspects

PROBLEMS The procedures preceding

section

and are not dependent logical involves

experimental

reader will find several books

1960; Stout and Jensen, 1968).

IN SULFIDE

STRUCTURE

determination

apply to analysis

difficulties

DETERMINATION

which were outlined

is organic

Structural

compounds.

has therefore

silicates.

The gaps in our understanding

The acquisition

lagged in comparison

of sulfide

perfect

a detailed

specimen

often

straight

forward.

poorly

low-angle acicular

analysis

sttuctural

chemistry

the

stem in

of these minerals.

specimen

a study which,

Crystals

axes of elongation

sulfides

Certain

sulfides

only as bundles

upon close-packed

arrays

of sulfur atoms.

often cause the true symmetry

of the ideal close-packed

array.

are commonly

and sometimes

for example,

is markedly

single untwinned compounded

fragment

phase transformation

have pro-

knowledge,

only a

of twinning

members

Such analysis

form at a temperature

chalcgcite

form, usua~ly

a crystal

twinned,

are of unequal

of problems.

w-S

is lower to a hexagonal

was thought

structure

to be

determination.

to the separate

volume,

th~ collection

Thus this procedute

is

a rapid

is now known to be monoclinic

may be proportioned

requires

which

(Cu2S) transforms

Low-chalcocite

need not prevent

which are measured

to apply.

To the writer's

Fe7S8,

displays

For example,

that t~

and thus

pyrrhotite,

the mineral

pseudo-hexagonal.

tions superpose.

are pseudosymmetric

Monoclinic

The problem

The low temperature

of a twin provided

twinned.

of the metal atom

to be lower than that

to being pseudosymmetric,

form at ~lOSoC.

Twinning

Small displacements

which are based

has ever been discovered.

ort~orhom9~_and

and is tedious

or

bent or

and sulfosalts

have structures

These structures

invariably

to a more symmetric

than that of deposition.

(Evans, 1971).

invariably

of fibers which have their

of the structure

pseudo-hexagonal.

if, in addition

X-ray intensities

are often massive

in common.

As will be seen later, many simple sulfides

positions

would be relatively

of some species are almost

occurring

must be isolated

The lack of a suitable

in other respects,

grain boundaries. habit,

of a mineral

may procede.

Unlike many other minerals,

crystallized.

nounced

single-crystal

structural

impedes

possess

number

of detailed

crystal

often

much less amenable

Specimens

A relatively before

bio-

however,

to study of, for example,

part from a lack of data, as well as the complexity Suitable

material,

or inorganic,

study of sulfides,

which make these materials

information

in the

of any type of crystalline

the specimen

in origin.

to study than other inorganic

of

The interested

upon whether

or mineralogical

of the process

(e.g., Buerger,

for structure obviously

outline

The

members

or if not all reflec-

of much redundant

has been attempted

data

in only a small

Yet another by exso1ution

problem

encountered

following

may be difficult

deposition

to detect

of intergrowths

are known

have structures

based

slightly

if the intergrowth in sulfides.

upon layers

Pb-As layer of distorted differs

between

Similarly

a large number

compositions

Bi2S3·

Coherent

observed

(Welin,

Chemical

and Structural

may be caus

of the second phas

More insidious

coordination whose

in which

with a

and composition

"Single

crystals"

the rocksalt

layer may

to more than one unique mineral.

with slightly intermediate

intergrowths

types

for example

alternating

thickness

of this group.

1967b)

corresponding

appear to exist as superstructures

different

(Pb+Cu):Bi

to aikinite,

of several

PbCuBiS3,

ratios and bis-

of these phases have been

1966).

Solid solution primarily

structure minerals

of minerals

specimen

is coherent.

of Pb in 9-fold

(Marumo and Nowacki,

fluctuate

a suitable The presence

Many of the Pb-As sulfosalts,

rocksalt-like

among the different

have been observed

muthinite,

in isolating of a mineral.

Complexity

is present

ionic materials

mined by the radius

to which

of the ions involved

(+) and (-) ionic charges ideal composition

in most of the minerals

the extent

balance.

and is subject

This often makes

for an ionic material

found in nature.

solid solution

even though

occurs

In

is deter-

to the restriction

it possible extensive

that

to deduce an

replacement

has

occurred. Sulfides

and sulfosalts

The relative

number

furthermore,

departures

might be given. been described determination unlikely

or metallic

(see Ch. 5).

common

character.

ratios,

and

Many examples

sulfosalts,

had frequently

or Cu SbS + ' The mineral was shown through structure 3 3 x and Neumann, 1934; Wuensch, 1964) to have the rather

CU12Sb4Sl3'

equilibria

Even this composition

in the system

0.03 < y < 0.30.

is approximate!

Recent

(Skinner et ai, 1972; Tatsuke

with 0.11

range Cu12+xSb4+yS13 field does not include

solid solution The stability

and

either

or CU12Sb4Sl3'

Small amounts not clear whether role, or whether hutchinsonite,

of impurity

are often problematical

they represent

substitution,

they are necessary (Tl,Pb)2AsSS9'

1965).

In contrast,

necessary polybasite,

similar

to the formation

0.OS-0.09

Ag16Sb2Sll'

amounts

the structure.

of the phase

Hall

is stable only if an amount

or electron

structural ~or example,

present.

This

(Takeuchi

et ai,

in

have been shown to be

(1967) demonstrated of Cu greater

that

than 3.1 wt. %,

for Ag.

of solid solution

such as wet-chemical

they playa

of the same element

of other minerals.

It is often

in sulfides.

Ag or Cu is normally component

but less than 7.6 wt. % has substituted Even in the absence

whether

to stabilize

has been shown not to be an essential

niques

may occur

one of the most

1973) have shown a wide

< x < 1.77 and

CU33bS3

covalent

unit need not be simple

as CU3SbS3 (Pauling

of phase

Morimoto,

from stoichiometry

Tetrahedrite,

composition

studies

have predominantly

of atoms in a formula

and nonstoichiometry,

microprobe

analysis

analytic

tech-

may not be sufficient]

accurate

to distinguish

uncertainty heavier

elements

determination techniques

may span several

provided

excellent

atoms nor the number

determination

as the appropriate determination

are resolved.

chemical

complexity

of a sulfide

defined

below.

because

information

intensities

Analysis

particularly

which

1973) suggests

The

complexity.

increases

rapidly

as

Large unit cells seem the rule As will be seen later, constant

are superstructures, is especially

are known with poor precision,

in excess of 40A!

a term which will be

detail is contained

models

of com-

with difficulty.

determination

fOe sulfosalts.

of such structures

structural

by the space group

by structural

have one lattice

about structural

of alternative

of metal

the means by which questions

is often obtained

of a structure

structures

Careful

of analytic

the number

were permitted

of atoms in the unit cell increases.

Some of the long-period

formula unit.

a number

Neither

is often paralleled

six of the 17 known Pb-Sb sulfosalts

numbers

through

The

of other much

(Kohatsu and Wuensch,

is frequently

the difficulty

a sulfide.

compotision.

This information

than the exception,

~r

atoms in a corresponding of nuffieldite

of sulfur atoms, however,

position

As noted above,

formulae

for sulfur in the presence

fit to PblOCu4BilOS27'

A structure

Pb Cu(Pb,Bi)Bi S Z 2 7 A structure

the number

alternative

fraction

of the composition

of the mineral.

rather

between

of the weight

difficult,

in part

in the very weak X-ray

and in part because

large

fit the data equally well.

X-ray Absorption Present-day diffracted

counter

diffractometer

X-ray intensities

tural analysis

to a set of structure

accv.racy, for absorpt~on salts commonly

involve

of 1000 to 1500 cm

-1

be transmitted

7

through

tures.

a serious

The magnitude

increasing

crystal

absorption

within

portional

decreases,

diameter.

bounds.

and sulfo-

absorption

coefficients

in CuKa radiation.

This

X-ray beam may

become

contain heavy metal of their crystal

decreases

must therefore

of a reflection sample.

of the specimen

correction

which

increasingly decreases

with

be used to keep are pro-

thus decreases

Further,

as specimen

as the size

shape, which would be negligible difficult

to measure,

correspondingly.

size must be sought but, in any case, the magnitudes W-7

struc-

exponentially

On the other hand, the intensities

The intensity

specimen,

of the absorption

in specimen

intensities

Small specimens

features

Many sulfides

(Pb or Bi, for example)

determination

size in an equi-dimensional

small-scale

to reduce

with comparable

of an incident

by sulfides

in precise

of the diffracted

manageable

in a less-absorbing precision

obstacle

struc-

only 0.1. mm thick.

of X-radiation

to sample volume.

cube of specimen

of the intensity

a specimen

The high absorption

Linear

for Pb-Bi sulfosalts

for measuring

Before

it is necessary

by the specimen.

coefficients.

are common as 3 x 10-

means that as little

however,

of high atomic number

which have high mass absorption

means

factors by correction,

of radiation elements

provide

of a few percent.

with these data may procede,

the intensities

atoms remains

techniques

to an accuracy

and the

A compromise of the

diffracted a result

intensities of poorer

A decade agreement

are small and are subject

counting

ago, sulfide

indices

for absorption

crystal

have increased

and less-absorbing

effort

corrections

is often

and computation

ask whether refinement

in interatomic

motion,

is the determination

to a refinement

standard

deviations

many cases,

to be made. dominated account

such efforts

of bonding

structure

positions

The scattering

sulfide,

atom locations

of locating

hydrogen

if the sulfur

rate positions,

atoms in an organic

atoms are to be located

in certain

in sulfides

have virtually

identical

and Bi (Z = 82 and 83, respectively) (26, 29 and 33). coordination

ples of instances has provided assignment arrangement

of chemical proved

gratonite,

recent

different

Pb As S 9 4 l5

the energy

of the species power.

through

Examples

subtle

of an early

structure

interpretation

the general

nature

galenobismutite, and Zemann,

PbBi2S4 1959; Knowles,

(Rllsch, 1963; Ribar and Nowacki,

forces which

data are

refined.

Accu is

simultaneously

contains

refinement

PACKING The cohesive

to the

or type of metal

The literature

even though

to be correct: TlAsS2(Zemann

atoms

comparable Accurate

scheme

scattering

crystal-chemical

species,

is

are Pb

Ag and Sb (47 and 51), or Fe, Cu, and As

and bond distances.

in which

a rather

1962); lorandite,

structure.

Such atoms must be distinguished

geometry

the metals

and their positions

Several

polyhedra

is completely

is almost

if the bonding

sulfides.

In

in the unit cell. Determi-

in such structures

in turn, are necessary

to be ascertained

sulfides

for example,

not.

if the identifi-

of coordination

matter

re-

other than reduced Sometimes

are required

in heavy-metal

In a Pb-Bi

if the of a study

to a level of 5%

any advantages

quality

variatio

the end of establishing

does refinement

of the linkage

of X-rays

in minor

are required

and bond lengths?

data of the highest

such as site occu-

as reflected

toward

precise

of a structure.

Very often the objec

effects

precision

of 15%, say, provide

by the metals.

required

pro-

compounds

in making

But, if the objective

for as much as 80% of the scattering

problem

refinement

refinement

are necessary.

in a given family,

and interpretation

of sulfur

corrections

for organic

are involved

least-squares

Data of the highest

in atomic

however,

cation of metals

with dis-

Improved

to which

is the study of subtle

of an unknown

minerals

relative

present

expense

or the details

distances.

between

were reported

available

of such studies. are to be meaningful.

lations

nation

but the level

routinely

and in performing

tive of a structure

results

refinements

in the range of 10% to 15%, a factor of 2

is almost

One may justifiably

thermal

errors as

minerals.

Considerable

pancies,

structure

precision,

studies

than that which

absorption

standard

which were often as high as 25 to 40%.

cedes in present-day to 3 higher

to increased

statistics.

maintain

differences several

determination

of a structure

of an atom or ion pair at infinite

(Iitaka and Nowacki, 1965; Fleet,

1973) and

1969).

W-8

in the solid state arise because separation

or

of the atomic

CONSIDERATIONS atoms

in

exam-

decreases

and passes

through a minimum

as the atoms approach

of the potential pair.

at equilibrium

If a third neighbor

pair, a comparable

amount

as that of the initial bonding

between

energy

one another

separation

X is brought of energy

A-X bond jf a repulsive

maximum

packing

in obtaining

size is not unlike

ciency.

One carefully

toothbrush

mensions

X-X interaction

possible

of minimum

maximum

packing

the larger

packing

denSity

items;

for a collection

layer

of an array of atoms of

a suitcase

(Fig. W-1).

to the initial

A subsequent

of two sets of "hollows"

among the spheres of the initial

of a sphere

by A, the coordinates

the location 1/3, 2/3.

layer.

to the original

That is, the third layer may be placed corresponding

directly

sequence

of successive

layers

by a sequence

of symbols A; B,or C.

the array is to be close-packed,

in a close-packed

Assume

that locafor

net) A, 0, 0, or C, over the first, or in The

array of spheres

may

is that a given symbol cannot

directly

if

follow

in the sequence.

Two special

stacking

sequences

lead to simple, highly positions

layers all fall above

of the hexagonal

Sulfides

the long diagonal

have appreciable

covalent

character,

will be decidely

directional.

strictly

As will be seen, however,

arrays

If

The only restriction,

may be noted that the three alternative

*

occupied.

to that which was not used for the second.

thus be specified

one

in the original

Figure W-l shows that the hollows

of a third layer are (relative

the position stacking

for the second

oCC'~y

These sites are

of the two possible

for the next layer are B, 2/3, 1/3, and C, 1/3, 2/3.

tion B is selected

itself

layer.

with a

equivalent

in that they are too close to be simultaneously

is denoted

in two di-

layer if the spheres

of the unit net is taken at the position

layer, and this location locations

spheres

The unit net is hexagonal

approach

exclusive

effi-

things -- socks,

holes.

of identical

a equal to twice the radius of a sphere.

the origin

with maximum

then the smaller

layer will be in closest

mutually

If the

the lowest

number of bonds is formed about

density

that in organizing

arranges

is the close-packed

periodicity

exists.

one can obtain

energy is thus the array with

and the like -- are put in the remaining

The maximum

it may not be as iarge

density.

The strategy different

configuration

The depth

energy of the

of an A atom in an A-X

although

the atoms or ions is non-directional,

The atomic

the binding

into the proximity

is released,

in an array of atoms if the maximum

each atom~

(see p. PR-19 ff.).

represents

apply.

involve neighbors

tant hybrid

The development

at locations

orbitals -- tetrahedral

packed arrays

are thus encountered

coordination in covalent

is directional.

W-9

arrays.

It

in successive

net -- that is, above

and the bonds about a given atom presented

the interstices

consistent

symmetric

for spheres

with

here thus cannot among close-packed

the orientations

for sp3 hybridization. structures

even though

of imporClosethe bonding

Fig. W-l. A close-packed

layer of identical

for a subsequent [11].

equivalent

If the offset of neighboring

+1/3

[1 I] or -1/3

[11],

a sequence

This array has cubic symmetry, the cubic close-packed lattice

constant

[11],

configuration

array.

(ccp).

ABCABC ... results.

cubic lattice, If the spheres

is octahedral.

Therefore layers

the

exist among the spheres

between

has a primitive

[1 1

hexagonal

close-packed

(hcp)

that about the second

and octahedral

in Fig. W-2.

+1/3

ABABAB .•. (or

in the ccp and hcp arrays.

about one type of site is tetrahedral, of the tetrahedral

is

a = 2;-2r.

alternates

This array, which

unit cell of the ccp array is depicted positions

are in contact,

results which may be denoted

pair of symbols).

The location

and is known as

by the fact that a face diagonal

of successive

of layers

Two types of interstices

positions

in the same sense, either

and space group P6 /mmc, is known as the hexagonal 3 The lattice constants are a = 2r, and c = /32/3r*.

The coordination

sites within

The coordinates

a

of these

are:

*This result three-layer Therefore

is always

of the spheres.

of offset

a sequence

by any alternating lattice,

layers

of the array is determined

If the direction

Alternative

which may be denoted

a face-centered

equal to four times the radius

and -1/3

spheres.

layer are indicated.

is easily structure,

c

h cp

= 2/3a

derived

by reference

so that c

h cp

ccp

13

= 2/3

equals

to the ccp array.

The former

2/3 the body diagonal

(2/2)rl3.

1s73. W-lO

is a

of the ccp cell.

a = 2r, from which c /a = hcp hcp hcp

Cubic close-packed

Fig. W-2.

array

The location interstices

These positions

the coordination packed

spheres

1/4

3/4

1/2

0

1/4

3/4

1/4

1/4

3/4

3/4

0

1/2

0

3/4

1/4

1/4

3/4

1/4

3/4

0

0

0

3/4

3/4

1/4

3/4

3/4

3/4

1/2

1/2

1/2

of (a) tetrahedrally

and octahedron

close-Eacked

tetrahedral

The tetrahedral

is placed

interstices:

inter-

of these

at 1/3 2/3 1/2): octahedral

interstices:

0

0

3/8

2/3

1/3

1/4

0

0

5/8

2/3

1/3

3/4

1/3

2/3

1/8

1/3

2/3

7/8

4 f 3m and 2 a 3m in space group P6 /mmc. 3 sites are required by symmetry to be regular.

or octahedral and octahedral

sites in a ccp array are linked by sharing sites are linked by alternate

and faces along c, and by sharing

to c.' A ccp arrangement

structures.

sphere

and octahedral

The coordinates

to equipoints

sites in the hcp array are linked normal

that both

if the array of close

of the tetrahedral

tetrahedral

In the hcp array the tetrahedral

vertices

of these sites it follows are regular

close-packed

arral

coordinated array of spheres.

and 4 b m3m of space group

a unit cell of an hcp array of spheres.

are (if the interior

0

cubic.

the positions

positions

edges.

4 3m

8 c

From the point symmetry

Hexagonal

Neither

and (b) octahedrally

a unit cell of a cubic close-packed

to equipoints

tetrahedron

correspond

interstices:

1/4

is truly face-centered

These positions

octahedral

1/4

Figure W-3 illustrates stices within

interstices:

1/4

within

correspond

Fm3m, respectively.

tetrahedral 1/4

of edges normal

by sharing

of anions

W-ll

more favorable

four octahedral

of of

The octahedral

of faces along c and sharing

is therefore

The ccp array has four spheres,

to c.

sharing

of edges

in ionic

sites, and eight

Fig. W-3.

The location interstices

of (a) tetrahedrally within

and (b) octahedrally-coordinated

a unit cell of a hexagonal

close-packed

array of

spheres. tetrahedral

sites per unit cell.

sites, and four tetrahedral

The hcp array has two spheres,

sites per cell.

fore equal to the number of spheres tetrahedral

interstices

independent

of the details

general

The number of octahedral

in the close-packed

array, while

is equal to twice the number of spheres. of the stacking

two octahedral

sequence

sites is th the number

0

This relation

and may be demonstrated

for

case.

Figure W-4 de?icts

two adjacent

close-packed

layers of spheres.

sphere in the lower layer there exist three tetrahedral

Fig. W-4.

The pOSition adjacent

of (a) tetrahedral

layers in an arbitrary

interstices,

sites and (b) octahedral close-packed

\01-12

sequence

About each plus a

sites betweer

of spheres.

fourth directly

above.

Each interstice

1/4 of each interstice

is coordinated

layer directly

by four spheres,

to" each of the spheres.

per sphere is thus 4(1/4) = 1.

interstices close packed sequence.

"belongs

A similar

The number

geometry

below the first layer regardless

Thus there are always

two tetrahedral

so that

of tetrahedral

pertains

to the

of the stacking

sites per close-packed

sphere.

Figure W-4b shows that in the first layer there exist three octahedral adjacent

to each sphere,

sphere.

A similar situation

and one-sixth

stacking

sequence.

of each octahedral

site "belongs

exists below the first layer, independent

The total number

of octahedral

sites

sites per sphere

to" that of the

is thus

2 x 3 x 1/6 - l. Quite generally,

then, if the sulfur atoms in a sulfide AmBnSp

in a close packed array, the available drally

regardless

tetrahedral

coordinated,

of the stacking

interstices

and a fraction

sequence

are occupied

nip of the available

octahedral

when the B atoms are octahedrally

coordinated.

and occupied

fraction

plays a prominent

of available

interstices

cubic.

This configuration

interstices

will consequently

tetrahedral

interstices

site is regular.

mutually

may again be located,

The octahedral

and at the mid-points orthogonal

and two at a/2).

directions,

but the interatomic

angles

phases

but the coordination

spheres

of cell faces

are located

in

(four are at a/2/2,

on cell faces at positions

are equidistant

are not those of a regular

and

about neither

at the mid-points

are located

of the

Both octahedral

but they are not equidistant interstices

structures.

is body-

rare, and the geometry

in detail.

The six neighboring

All four neighbors

of type

role in most of

of sulfide

sulfide

sites are located

of cell edges.

Tetrahedral

such as 0, 1/4, 1/2.

is relatively

not be examined

interstices

The notion

for the classification

The sulfur atom array in a few high-temperature centered

m/2p of

if the A atoms are tetrahe-

are occupied

the schemes which have been proposed

are arranged

a fraction

(at alS/4)

from this site

tetrahedron.

POLYMORPHISM AND POLYTYPISM The alternative represent chemical

different specie.

stacking

This represents

forms" -- the phenomenon distinct

crystal

sequences

crystallographic

discussed

an example

of a given chemical

structures.

Sulfides

known examples

of polymorphism.

forms of ZnS and the pyrite

is a term having

description

of polymorphism

Polymorphs nearest-neighbor example), second-

Examples

but one

-- literally

existing

with

"many

two or more

many of the early and well-

and marcasite except

section would

below are the wurtzite forms of FeS2

(cf. Ch.5).

that its use is confined

to the

in elements.

have different configurations

crystal

structures;

(the graphite

but often the nearest-neighbor and higher-order

compound

to be discussed

similar meaning

containing

of polymorphism

have provided

and sphalerite Allotropy

in the preceding

forms of a compound

neighbors

the differences

and diamond

coordination

are differently W-13

may be in

forms of carbon,

remains

arranged.

for

the same and only Wurtzite

and

sphalerite exists.

and marcasite

and pyrite

The difference

fore be quite small. the energies

Nevertheless,

cannot be precisely

a given temperature Remaining stable

polymorph

since

"Diamonds

only one polymorph

that the phase

transformation

has its basis

may there-

are distinct,

close, and it follows is the stable

if they exist, must be metastable.

are forever"

the later situation

configurations

the atomic arrangements

does not imply a small difference

tures, but rather slogan

in which

alternative

the same, however

and pressure

polymorphs,

are polymorphs

in energy between

structure.

The existence

in energy between

is hindered

in kinetics

that at

of a met

two struc-

by kinetics.

The

and is thermodynamically

false. The rigorous subject

specification

to subtle questions.

impurities.

the stoichiometry

are not strictly

of one "polymorph"

polymorphous

chemical

composition.

morphism

will be somewhat

loosely

is difficult

in Chapter

5 on sulfide

may be slightly

of crystal

used to describe

differences

exist.

and

by small concentrations

chemistry

phase

different.

since they do not have precisely

In this discussion

for which no gross chemical which

in some systems

In other cases, as will be discussed

equilibria, phases

of polymorphism

A phase may be stabilized

Suci

the same

the concept

a relation

between

There is clearly

of po.

structures

a grey area in

use of the term may be disputed. The permitted

sent a special common

stacking

arrangements

type of polymorphism.

to all the configurations,

is, in their arrangement is termed polytypism. two or more alternate

occur, however,

and they differ

in a third direction. The phenomenon

positions

total of three positions

for layers of close-packed

A two-dimensional

in silicides

This special

is possible

sequence,that

type of polymorphisE

in layer structures

for close-packed

in which

of subsequent structures.

and the class of intermetallic

repr.

unit is

only in stacking

exist for the placement

are involved

spheres

structural

layers.

A

Four positior

compounds

known as

Laves phases. The literature schemes

on polytypism

for denoting

the sequence

involving

stacking

specify

triangles

its crystal

and deltas

faults.

origin

of the cell.

packed

sequences

conventions be determ~ned

That is, the same sequence

a sequence

of symbols

may become

that there exists

The detailed of symbols

of symbols

structure

unwieldy.

for both close

equivalent. of a polytype Secondly,

Both must

the

This may be appreciate

of SiC which has a lattice

W-14

nota-

of the choice of

results

are obviously

of

with

The triangle-delta

may be assigned.

extraordinarily a polytype

concerned

is independent

crystal

but

description,

of the. displacement

in literature

is the same.

ABABAB ••• and BCBCBC .•• , which

before

by realizing

conveyed

differe

structures

the polytype

An alternative

the direction

in that its symbolism

have liabilities.

For close-packed

not only to identify

to indicate

The information advantage

by the fact that several

structure.

layers has come into use, primarily

tion has a slight

sequences

is complicated

have come into use.

of letters ABC •.• etc. serves

also to completely

subsequent

a polytype

constant

of

approximately

l500A;

Ramsdell

594 layers must be placed before

(1947) proposed

poly types which avoids number

specifying

is followed =

referred

by this notation,

This problem

36H, and SIR. subscripts

The distinction

whether

(1954a) assigns

tions along

[110]

1/3).

symbolisms

on either

(h for hcp-like)

is

the notation

and lattice

each are known for SiC 10H, of arbitrary

have come into COmmon use for

polytypes.

A system introduced by

side are displaced

or displaced

BCABC

is that, again,

Further,

To illustrate,

avail-

whether

either the letter h or c to a layer depending

the A layer in the sequence

of the cell.

is not specified

from the information

each is made by the addition

Thus the A layer in the sequence

tage of this notation

The l500A SiC

structure

symbol.

of close-packed

the layers adjacent

like).

between

closely-related

the structures

Two polytypes

an economy

This

R - rhombohedral,

2H.

exist with the same number of layers

a and b to the Ramsdell

Jagodzinski

the crystal

A

(that is,

of one layer) is given first.

One might ask, however,

does indeed occur.

Two additional denoting

while

pattern.

repeats

of two parts:

type (H = hexagonal,

While

sequence

of close-packed

the periodicity

the symbol may readily be established

Might several polytypes

type?

consists

within

is 3C and the hcp arrangement

to above is 549R.

able in a diffraction unique.

giving the lattice

The ccp sequence

the stacking

for designation

The notation

divided by the thickness

by a letter

cubic).

structure

these problems.

notation

how many layers are contained

the c axis dimension

C

a compact

BCACB

in the same direction

would be given the sumbol c.

it is independent

sequence

upon direc-

(c for ccp-

would be given the symbol h, The advan-

of the choice of origin

in symbols may result

the stacking

in the opposite

(by a factor of up to

in the close-packed

structure

of Sm

is ABABCBCAC

ABAB

It might

appear

but Jagodzinski (a)

hhc

that details

of the structure

If the number

of h symbols

is hexagonal

c a number of layers

(space

equal to twice

of symbols.

of h symbols

sign after each h.

is even, assign ±l to each symbol,

(For example,

or -1+1-1).

(b-l)

If the sum of these integers is trigonal,

is not a multiple

of 3 the structure

space group R3m with three times as many layers

c as symbols

(e.g., Sm).

If the sum of these integers the structure

changing

for Sm, hhc, one could write either

+1-1+1

within

this condensation,

this information.

is odd, the structure

and contains within

the total number If the number

(b-2)

are lost through

(1954b) has given rules for resurrecting

group P63/mmc)

(b)

(9R)

-----

c --------

hhchhchh

is trigonal,

is equal to a multiple

of 3, then

space group P3m, with the same number

of layers as symbols. A somewhat

been proposed

more compact notation,

by Zhdanow

(1945).

closely

The stacking W-15

related sequence

to that of Jagodzinski, is represented

by a

has

series

of integers

each of which represents

h in the Jagodzinski becomes

(34).

number

notation.

the number

Thus Sm, hhc, becomes

It should be noted that the Zhdanow

of layers within

using Jagodzinski's

of symbols

following

eacr

(12); SiC 2lR, hcchccc

notation

specified

the total

c -- for example

rules,

the number

SiC 10Hb, hcccc is (55) and not (5), si of layers in the cell is twice the number

of symbols. As with the notion with a precise configuration symmetry

hedral

Should

cases?

Agqin,

structures

of the tetrahedral

about tetrahedral

formed by placing polytypes

we will adopt a rather loose

definition

layer sense its correct far away as l500A?

questions

position.when

Many complex

is not completely

of po1ytypism

fically

devoted

Quite

sequences

with

successful.

large Burgers

Additional

operations

a larger volume

in the parent

intense

structure.

produced indicate

of spheres

the symmetry

The symmetry

structure.

This special

class

in diffraction

(substructure reflections, by the parent the presence

structure)

of a larger W-16

parent

certain

array is

symmetry structure

may

has a unit cell wit

The lattice

constants

of the

sum of the translation is termed a superstrw

by the presence

which

closederivati

If one of the operation

structure

of derivative

patterns

a!rays

of' the derivative

as some vector

are replace

is essentially

suppress

when

such as

spheres

of a simpler

configuration.

the derivative

may then be expressed

recognized

in statements

these more complex

which may sometimes

than that of the basic

reflections

tion effects tions which

structures

may bear a close relation-

that half of the black

arrangement

is translation,

is manifested

structure

and verbalized

of that of the parent

structure

The relation

(1966) speci-

is often intuitively

(1947) termed

in the basic structure.

is suppressed

derivative

Buerger

perturbations

thus be a subgroup

crystal

except

or that an overall

In derivative

through

but this inter

and a general

STRUCTURES

A relation

structures

diamond,

but distorted.

degraded

array.

of crystal

resembles

structures.

which

a complicated

atomic

models

by red spheres"

to be repli-

vectors,

discussion

How can a

layer may be as

are believed

may be found in a book by Verma and Krishna

frequently

"Sphalerite

with polytypes.

to the topic.

ship to a simpler

packed,

in the two the struc-

in the two-

the translation-equivalent

DERIVATIVE

examining

different

and consider

similarity

associated

stacking

cated by growth on screw dislocations

survey

atoms in the tetra

unit.

There are many interesting

pretation

by

sites in an

if minor distortions

layer to be slightly

tures to be poly types if there is close structural dimensional

that the

sites among a ccp array is constrained

sites in the two arrays be considered

cause the structure

in connection

We have seen, for example,

(43m) while the symmetry

(3m).

is lower

there exist fine points

of polytypism.

about the tetrahedral

to be regular

hcp array

of polymorphism,

definition

closely

resemble

plus a collection cell and. which

of a set of

the di.ffra,

of weak reflec-

contain

information

on the nature

of the perturbation

which

There are four basic mechanisms

produced

through

the increased

periodicity.

which derivative

structures

may be

formed: (a)

substitution

(b)

ordered

(c)

addition

of one or more

omission

of atoms

of atoms to a site which

(termed a "stuffed" (d)

distortion

Sulfide

chapter,

derivative

is unoccupied

in the parent

structure

by Buerger)

of an array.

Two, three, or sometimes

derivative

types of atom for another

four mechanisms

and sulfosalt

structures.

structures

Among

the following

may be operative

provide

the structures

examples

in a given derivative.

a rich collection to be discussed

may be selected

to illustrate

of examples

of

in the following each mechanism.

Substitution: ZnZnS2

(sphalerite)

tt CuFeS

FeSS

t FeAsS

(chalcopyrite)

2

Omission

(marcasite) (arsenopyrite)

and distortion:

FeSS8

(troilite)

t

Fe7S8

(pyrrhotite)

Addition

("stuffing")

OBiBiS

3

tt CuPbBiS

3

and substitution:

(bismuthinite) (aikinite)

Distortion: MS

(NiAs-type,

t CrS

hexagonal)

(monoclinic)

Substitution,

omission,

addition

and distortion:

(sphalerite) (tetrahedrite)

"COMPOSITE" Another structures

type of relationship

which

Very complex of simpler

structures

structure.

coordination

of atons

ture type is a common the atoms structures, literature

occurs

is less frequent

in sulfide

in the crystal

are not infrequently

is quite regular example),

such regions

has been proposed one sometimes

to describe

encounters

but these are not satisfactory.

such arrays will be denoted

as composite

structures.

W-17

blocks

or domains

fashion.

(the rocksalt

arrangements

Such structures

sulfosa1t

of other minerals.

in an irregular

irregular

of the domains.

and no terminology of other minerals,

within

but highly

and especially

chemistry

found to contain

These are, in turn, joined

at the boundaries

ning of unit cells",

STRUCTURES

The

struc-

occur about

are not derivative them.

expressions

In the such as "twin-

In the present

discussion

Many examples tain Pb. family

of composite

The plagionite

of homologous

invariant

while

rocksalt

1974).

monoclinic

alternately

lattice

Successive

salt slab.

exist among the sulfosalts

structures

the third increases

structure

the variable

structures

constant

members

in which

with n.

parallel

two lattice

The structures

constants

consist

of the series

1970, 1974; Kohatsu

differ

only in the width

con-

of a remain

of slabs of

to (112) and (112) as one procedes

(Cho and Wuensch,

Other Pb sulfosalts

which

(with n = 0 to 3) consists

group Pb3+2nSb8Sl5+2n

along

and Wuensch,

of the rock-

Pb Bi S (Weitz and Hellner, 2 2 5 1960), ramdohrite (Kawada and Hellner, 1971), lillianite, Pb BiS (Takagi and 3 6 Takeuchi, 1972) and several synthetic Pb-Bi sulfides (Otto and Strunz, 1968) have large regions

of rocksalt

are "polysynthetically

composites

displays

The structure

(Miehe,

the structure

of cosalite,

and Buerger,

region

regions

contain

of two regions,

arbitrary.

to the parent

structure,

to the nature

of the domain:

primarily

symmetry,

fractions

scheme basis

this linkage silicates, crystal

and jamesonite

SCHEMES

used to establish

as a basis,

characteristic

should possess

to classify

crystallographic

such as color.

some predictive

paint,

the many minerals

a set of cubby holes may be quite

composition

for classifying

The sites in tr

of both Sb and Bi.

to have a basis upon which

A scheme

other materials).

resemble,

region primarily

sites common to both the cosalite

or even some physical

a satisfactory

i,

but the disordered

Bi, those in the jamesonite

One could use chemical

a classification

Kobellite

one of which

similar

CLASSIFICATION It is convenient

structures.

Pb _ (Bi,Sb)7+x(Fe,Cu)S17.5 6 x constants a = 22.5, b = 34.0X.

according

comparable

found in nature.

these domain,

of composition

two large lattice

the Group V metal

contain

structures

Pb Bi S , and the other jamesonite, FePb Sb S 2 2 5 4 6 l4 1957). The two domains partially overlap. Each domain

Bi and Sb are partitioned

Sb, while

exist in more complex

1971) is a composite

is not only extraordinarily

cosalite

In the latter

on (311).

Bi-Sb lead sulfosalt

with O (!)

\

/

/

/

/

ORBITALS OF 52

Energy-level diagram showing the valence and conduction bands of FeS,

PR-24

control

the solid-solution

observed

metal-metal

behavior

bond distances

listed

in Table PR-7.

ometry

of a phase,

It is useful

the metal-sulfur

bonds,

so that predictions

pounds

of a given transition

try.

Name,

and the number

in these structures. of bonds

the relation

coordination,

and the number

metal ion, which general

chemis-

which

show

Sulfides.

M-S Coordination

M-M Coordination 4

2.602

3

2.505

3

2.531

M-M Distance,

Fel+xS

4

Co-Pentlandite

C09SB

6, 4

Pentlandite

(Fe,Ni,Co)9SB

6, 4

Millerite

NiS

5

2

2.534

Heazlewoodite

Ni S 3 2 Ni S 7 6

4

4?

2.49

5, 4

2?

2.492.

semi conductivity taining

may be used.

a chalcogen

Pearson's

valence

metal

bonds

these

structures

as the anion

(because

and p electrons

This rule has been used to classify (Takeuchi,

1970).

It should

rule could not be used to classify it took into account

remain

idea that the valence

"unclassified"

shell of anion

and is expressed

only valence

(Pearson,

(in the present

na is the number

cations

in forming

anion-anion

and Na is the number

conthat tt

metal-

and, therefore,

This rule was based

case sulfur)

contains

on the

eight s

b

B,

electrons

valence

honds,

on the anions,

electrons,

b

a

is the number

c of anions;

nc is the number

is the number of electrons

all of these numbers

of electrons

forming

on the involved

cation-cation

are calculated

per forml

in NiS , na = 2 x 6 = 12, nc = 2, b = 2, b 2 a c o and Na = 2. Thus (12 + 2 + 2 - 0)/2 = 8. NiS2 with the pyrite structure is a semiconductor (Hulliger, 196B). In hexagonal NiS with the NiAs structure, na = unit of the compound.

For example,

= 2, ba = bc = 0 and Na = l, (6 + 2 + 0)/1 = B. Although Pearson's valence 0 rule is obeyed, the phase is observed to be metallic above 263 K (Trahan et a'l,

6, nc

1971). CoS

Further,

even in other

(which obey Pearson's

2 (Hulliger,

196B).

~

by the relation:

of valence

less any unshared

containing

electrons)

1972).

structures

be noted here

structures

n+n+b-b a cae N a

bonds

of the com-

Mackinawite

aNi S 7 6

where

the stoichi-

compounds

in Selected

are

of metal-metal

and properties

rule for valence

Coordination

Composition

between

is one of the aims of crystal

valence

Metal-Metal

The

in each structure

to understand

can be made on the structure

To this end, Pearson's Table PR-7.

of these metals

"normal"

sulfides

rule) metallic

It is believed

such as CoS, Co S , Ni S , 3 4 3 4 has been observed

and

conductivity

that the metallic

PR-25

conductivity

in these sulfides

is due to the de localization respect

may have to consider either

directly

additional

or indirectly

evidence

suggests

involved

in the metal-metal

as the number

- 2)/1 = 8.

If cation-cation (Pearson,

=

na

electrons

6, nc

=

d electrons

=

4, b

C

= b

a

a

1968),

IN a

(0* with

sulfides

d

structural

orbitals

equation

could 1

may be definE d orbit

in the antibonding

=

0, b

WE

are involved

Available

in the antibonding

a bonds and anion-anion

1972; Hulliger,

which

interactions.

Then nc in Pearson's

plus any unpaired

in millerite,

d

for

electrons

bonding.

For example,

pair bonds

conditions

d

in the antibonding

Thus, for these transition-metal

in metal-metal

that unpaired

of valence

d electrons

of cations'

d orbitals.

to M-S bonding)

=

2 and Na

l, (6 + 4 + 0

c bonds are considered

as elect

then

and C

b

c

c

IN c

where

C is the number anion-anion bonds per anion and C is the number of catioI a c cation bonds per cation and Nc = number of cations in the formula unit. Pearson's

equation

can be modified

n +n +CN a c aa

N Applying

as n

cc

= S·, Cc

a

+n

c

+CN

N

a

this equation

1) Mackinawite

-CN

a a

-8N

a

c

to C

(Fel+xS):

(6 + 6 + 0 - 8) = 4

c

(Co S)' c = (8 x 6) + (8 x 5) + 0 - (8 x 8) = 88 - 64 = 98' c . 8 8 only tetrahedral cobalt atoms are taken into account because these,

2) Cobalt pentlandite (Note:

the ones involved 3) Millerite

in direct metal-metal

(NiS):

(6 + 4 + 0 - 8) 1

=

C c

interactions.)

=

2

C = (2 x 6) + (33x 6) + 0 - l6 = 14/3 = 2.67 (Ni S ): 3 2 c From Table PR-7 it is evident that, except in heazlewoodite, the above equat

4) Heazlewoodite

seems to apply for all structures powder

diffraction

atoms

(Hulliger,

1968).

heazlewoodite

needs

Nevertheless,

the possibility

that cation-cation

careful

bonds

may not be strictly It is worthwhile necessarily

levels

Factors

exists

which

distances.

electron-pair

to four

the structure

diffraction

that the discrepancy

are essentially

From]

of heazlewoodite

to four other Ni atoms in addition

study using single-crystal

of

techniques.

is due to the assumptic

bonds,

an assumption

which

correct. mentioning

interactions

into bands into which

such d bands

short metal-metal

that in the structure

In view of this discrepancy,

here that metallic

imply the existence

weak metal-metal

d

containing

it was suggested

each Ni is coordinated

(rhombohedral) sulfur

data

through

d

could cause metallic influence

properties

of direct metal-metal anion

electrons

0

11

bonds

PR-26

Partial

Indi

occupancy

a

of

as in C0 S 9 S bonds are not complet

and Pauli-paramagnetism,

of direct metal-metal

understood.

does

could cause broadening

are delocalized.

properties

the formation

or

in a sulfide

bonds and vice-versa.

PHYSICAL The physical properties,

PROPERTIES properties

OF TRANSITION-METAL of sulfides,

have been intensely

studied

especially

trial applications,

these properties

and also are important

Indirectly,

these properties

particular

sulfides.

physical

properties,

of transition-metal

including sulfides

On the basis paramagnetic,

of magnetic

the orbital

and, therefore, known

most shells. magnetic

and electrical

properties,

contribution

sulfides

moments

In transition

metals,

and become

are responsible

for their magnetic

because

Therefore,

cases,

although

the five-fold

degeneracy

is removed

into an a level and a S level corresponding

and e

ion, the sulfides

magnetic

t2

g

orbitals

g low-spin

than 4s and 4p orbitals. localized

of the presence

and antibonding

on the meta of a ligand

Each 3d orbital

orbitals)

to two different

electron

g a level, the d electrons

of the e

number

with

of unpaired

containing

high-spin

An example

before

occupying

e

2 of e

orbitals.

g of unpaired

a minimum

electrons

i:

spins

would be dis-

for a given transition

ions will be paramagnetic

is MnS

is lower than the energy

(Fig. PR-1l).

However,

a levels,

the electrons

g

This electronic

electrons

with high if the energy

will pair il

configuration

and consequently

becomes

low susceptibility

(Fig. PR-12).

In the majority exist

alone

i

susceptibility. S levels

values

in bond'

3d electrons

a and e a levels before pairing takes place in the t S level. 2g g 2g is a high-spin electronic configuration for the ion. Because the high-

The result

g

are involved

ligand

than the energy

spin state has the maximum

of t2

in the outer'

in t

tributed

metal

It is well

electrons

field the t2g a level has the lowest energ S level has the highest energy. If the energy of the t2

In an octahedral

energy)

metal

of an atom is negligible

is important.

essentially

because

split

(potential

as nonbonding

in transition

4s electrons

extent

remain

(referred

S level is higher

Usually

moment

as diamagneti,

of an atom is determine,

band electrons,

field

(Fig. PR-ll).

earlier

and reflectivity

3d orbitals will also overlap sulfur

moments.

but only to a much smaller the 3d orbitals

in

the

filled outer shells will not have

part of the valence

38 and 3p orbitals in many

moment

of ions are due to unpaired

ing with sulfur

of bonding

can be classified

spin contribution

Thus, ions which have completely

moments.

properties

motion.

to the magnetic

aspects

pages to explain

theory.

The magnetic

only the electronic

used in

and geoelectricity.

certain

the help of band

and by their orbital

that the magnetic

could be profitably

in geomagnetism

us to understand

magnetic

with

and electrical

Apart from their indus-

is made in the following

and Pauli-paramagnetic.

by the spin of electrons compounds

enable

attempt

An

of sulfides

AND BAND THEORY

the magnetic

in recent years.

mineral

exploration

SULFIDES

in the high-spin

positive

values

adjacent

cations

exchange

state

of magnetic

coupling

Coupling

directly moments,

(monoclinic

cations

Fe7S , B PR-27

commonly

are paramagnetic

of the magnetic

or by cation-anion-cation especially

'results in anti ferromagnetism

or ferrimagnetism

sulfides,

these sulfides

susceptibility.

of magnetic

usually

transition-metal

and, therefore,

takes place either

this leads to ordering

NiS ,etc.) 2

of the first-row

moments

with

of tl

exchange

at low temperatures.

an,

This

(e. g., Fel_xS ,Col_xS ,Nil_x:

FeCr2S4,etc.)

in many paramagnetic

sulfides.

Only a very few sulfides exhibit ferromagnetism (e.g., CoS and CuCr S ). 2 2 4 and dlO or spin paired d6 (as Fe2+

Sulfides containing transition metal ions with ~

in pyrite) have no net spin moment and thus are diamagnetic.

In certain sulfides

the 3d orbitals of neighboring cations interact to form d-bands.

This effect

becomes particularly significant at low temperatures and in structures where the cation coordination number is small, as in pentlandite (Fe,Co,Ni)9SB' millerite (NiS), heazlewoodite (Ni3S2), and others.

The d electrons in the resultant d-bands

t2g

-- ------------------eg* CI f:::::i

-

IJ

- - - -- - - EFERMI

r

>~ a::

flcov

I&J

Z

flEX

L&J

.t .t

.t

SPIN-UP

SPIN-DOWN

LEVELS

LEVELS

Fig. PR-Il.

Schematic energy-level diagram for MnS2 d orbitals (Bither et aZ, 1968). 2+ is in the high-spin electronic configuration.

Mn

PR-28

-

r

----------EFERMI

,

Acov

I

>(,!) a::

___j

hg.B

I

LLI Z LLI

AEX

!hg

t

L_ L_

1

CI

SPIN-UP LEVELS Fig. PR-l2.

SPIN-DOWN LEVELS

Schematic low-spin

energy-level electronic

are completely

delocalized.

temperature-independent

small susceptibility very useful tronic

in understanding

configuration

between

the bottom

is usually p~omotion

small.

~lectrons

of the conduction At moderately

band.

of current

of magnetic

of bonding

with very would be

and also the elec-

in them. and split 3d orbitals band.

band and the highest

of d electrons

thermal

from d orbitals

electric

The energy

partly energy

filled

level

could cause

conductivity.

of promoted Because

the 'conductivity in such sulfides

PR-29

d

fall gap

(usually a* antibonding)

field the movement

band causes electrical carriers,

then the com-

properties

in sulfides

the localized

high temperatures

In an applied

in the conduction

sma l.l.e r number

the nature

filled,

(Pauli-paramagnetism

band and the empty conduction

of a small number

t~ the conduction

Thus, measure~nt

metal sulfides,

the filled valence

Fe2+ is in the

d orbitals.

are incompletely

paramagnetism

of cations present

between

2

If the d-bands

values).

In many transition

for FeS

configuration.

pounds

exhibit

diagram

of the

is usually

small,

increasing

These sulfides Several resistivity in which tivity

with

temperature

transition

resistivity

with

decreases

that the highest

conduction

(2)

d d

level overlaps electrons

causing

e.g., CoS2 and Co1_xS.

of a cation ~

falls below

of the d orbital

state M(m-l)+

also falls below

removal

from the valence

results

due to the movement

filled valence conductors sulfides

band.

band.

of electrons

within

the transition

by

of this, conduction

the (incompletely)

are called E-type

and this type of conductivity containing

Because

band and metallic

Such compounds

the top

of the same

band, then the cation will be reduced

the valence

with the

Such compounds

the top of the valence

holes are created within

condu,

will be trans-

they are delocalized

conductors,

band and if the energy

of an electron

to semiconduc'

The metallic

field is applied.

oxidation

rise!

ways:

band in which

cation but in a different

as opposed

temperature).

filled

of the d orbital

If the energy

of the valence

promotion

FeS2, NiS2, etc. conduction (with low

metallic

some of the

when an electric

are called ,!!-typemetallic

of electron

Fel_xS,

temperature

in different

band and hence

ferred to the conduction metallic

exhibit

increasing

with increasing

arises

It is possible

empty conduction

e.g., hexagonal

metal sulfides

which increases

in these sulfides

(1)

as the probability

are semiconductors,

is usually

metallic

exhibited

metal ion in which

by

the energy

of the

d orbital is low because of a high effective nuclear charge, e.g. Cu2+ and Cul+. It should be pointed out here that in both ~-type and E-type (3)

metallic

the d orbitals

conductors

A third type of metallic

action of the d orbitals intervening

sulfur.

and, therefore, metal atom. movement

electrons

either

within

also exhibit

on cations.

of the inter-

directly

leads to the formation

are no longer

localized

the band also causes metallic

Pauli-paramagnetism,

or through of

d

bands

to any specific

cOlllillonly these d bands are incompletely

of electrons

Such sulfides Ni S , 3 2

d

arises because

cations

This interaction

the

More

conduction

of adjacent

are localized

filled,

and

conduction.

e.g., C0 S , 3 4

C09SB, NiS,

and ot~ers.

Jellinek

(1970) divided the transition

on the basis of their magnetic (1)

Semiconductors

(2)

Metallic

conductors

(3)

Metallic

(4)

Metallic

metal sulfides

and electrical

with paramagnetism

properties.

into four main classe These include:

or diamagnetism.

of the E-type

with paramagnetism.

conductors

of the ~-type

with paramagnetism.

conductors

with temperature-independent

(Pauli-paramagnetism)

or diamagnetism.

PR-30

paramagnetism

Reflectivity Reflectivity minerals

is one of the important

are commonly

reflected related

identified

to two important

optical

(n) and absorption

medium

and is defined

(I) to that of incident

light

is related

to a parameter, Band theory

in determining

To illustrate

neff'

shows that minerals

The absorption

effective

between

minerals

the significance

(Burns and Vaughan,

reflectivity

In a crystal

of sodium

and bonding,

the 3s orbital

metal,

of neighboring

each Na atom has only one 3s electron,

the valence

exactly

half-filled

each orbital

at absolute

spins).

scale at a particular tals are empty.

temperature

Because

the band can be easily sorbing

excited

of available

return

to the ground

metal.

of excited

In sulfides formation transition orbitals

the Fermi level.

localized

is somewhat

the transition-metal

sulfides,

Thus, in stoichiometric could not account occur between

complex.

conduction.band.

their metallic

for their reflectivity.

Because between appearance

of

3d

in many sulfides

3d

orbitals

transitions

the

the

upon the

In most of

the valence

filled. band

transitions

could

3d

and the

orbitals

are localized

on cations,

band could account

When there is extensive electrons

of

3d

to be completely

within

orbitals

of broad d bands where

t'R-31

the Fermi level a

orbitals.

electronic

and the conduction

and reflectivity.

to cause the formation

for Na

d bands depending

andlor between

3d

electron

in any

The antibonding

band is thought

the

above

on the

of 4s and 4p orbitals

of sulfur.

Sowever,

orbitals

the excited

We have seen earlier

with the sulfur

electronic

depends

and reflectivity

levels

within

the band by ab-

absorbed

When

the orbi-

the electrons

levels.

or form narrow

the valence

sulfides,

the two groups

only transitions

on cations

two electrons on the enp.r~y

high reflectivity

bands by the overlap

of 3d orbitals

mixing

lustre

of empty energy

metal atoms and 3s and 3p orbitals

of covalent

of radiation

in these energy

and conduction

energy

empty levels within

here that metallic

electrons

band.

thus formed will be

can have

potential

light is re-emitterl, causing

the situation

of valence

are either

magnitude

bonding

above

configuration

as the Fermi level above which

The amount

are due to the availability

delocalization

let us first

of each Na atom

band

band is only half-filled,

into overlying

levels

state,

It must be emphasized

compound

is known

radiation.

energy

(Note:

of, the electron

the valence

electromagnetic

number

zero.

The pOSition

pE

of nefl

1970).

Na atoms to form the valence

Because

with opposite

with

coefficient

nlmIDer of free electrons

one to understand

case such as sodium metal which has the electronic

with the 3s orbitals

overlaps

representing

of sulfide

the relation

consider a simple 2 2 6 l 18 28 2s 3s •

is

index of the

equation:

This equation

of solids enables

the reflectivity

of

Reflectivity

such as the refractive

(k) by Fresnel's

is 100 percent.

molecule.

sulfide

as the ratio of intensity

k will have high reflectivity.

of nand

by which

(n - 1)2 + k2 (n + l)2 + k2

R

large values

properties

(Io), R = I/Io'

light

constants

coefficient

When R = 1, reflectivity

physical

for

covalent

could be effectivel)

delocalized,

then electronic

transitions

give rise to high reflectivity.

bonding

*

e

autibonding

g

between

into the e become

For example,

filled and the six 3d electrons

pletely

g

*

orbitals

flectivity

orbitals

2g free electrons

of pyrite.

d

the valence

t2

the nonbonding

of extensive

1970).

by absorbing

Electrons

radiation

of available

for the metallic

lustre and high re-

In the pyrite series,

the reflectivity

decreases

3d

in the se2 Because C0 +(L.S.

unoccupied

into the e

band

g

ations in sulfide minerals to mention

reflectivity. CuTe

that the values of energy

decrease

would

2 is in the order of increasing

When working with silicate

another. compile

coordination

interatomic

being reasonably

a set of effective

which,

It if

in turn, increases

CuS2, CuSe2, and CuS2 < CuSe2 < CuTe2· This sequE of anions.

AND COVALENCY

IN SULFIDES

distances

uniform

This feature of oxide structures

values.

of the metal-chalcogE

and other oxide crystal structures,

ogist is able to depend on average

of vari·

Cu-dichalcogenides

size and polarizability

DISTANCES

electrons

substitution

Thus, compositional

character

delocalization

are

g

Therefore,

their reflectivity

in the sequence

INTERATOMIC

particular

1970).

covalent

in the isotypic

increases

to t2

levels available

affect markedly

of electron

For example,

the reflectivity

of neff in these sulfides

its reflectivity.

here that increasing

bond causes an enhancement

g * ban,

in CoS2, NiS2 and CuS2. Therefore, the levels above the Fermi level decreases in tl

(Burns and Vaughan,

by Co and Ni in pyrite would

important

energy

to the number

*

the e

electrons

It was observed

roughly proportional excited

an,

accounting

will have one, two and three

above sequence.

The

covalent

are excited

electromagnetic

quence FeS > CoS > NiS > CuS (51.6:34:27:17, respectively). 2 2 2 2 Ni2+, and Cu2+ have seven, eight and nine 3d electrons, respectively,

number

COl

band is ,

orbitals.

g

bands because

(Burns and Vaughan,

band from the t

effectively

in FeS2,pyrite

occupy

are empty and form

iron and sulfur

the two groups of 3d orbitals

between

ion in

from one crystal structure

enabled

ionic radii which,

the mineral-

for a particular

Shannon

and Prewitt

when the radius

to

(1968) to

for a cation coordi-

nated by N oxygens

is added to the radius of oxygen coordinated by M cations, usu: o give an interatomic distance within ~O.OlA of that observed. This has proved to 1

extremely

useful in a variety

of applications

in oxide crystal

chemistry.

Shannol

and Prewitt's radius

work showed that there is a linear relation between cell volumes an· cubed (r3) of the varying ion in a series of isotypic compounds. However,

such relationships

are not found in sulfide

with other anions in the periodic for our inability

to use the simple

bonds in sulfides

are not ionic.

in silicates

either,

so what's

tween observed Second,

and, in fact, are not foun, for fluorine.

concept with sulfides

The answer

is that tho

to this question

sure about all parts of it.

radii table was derived

PR-32

ioni is co

The first part i

did find that small inconsistencies

(from radii) bond distances

the Shannon and Prewitt

The reason

One might say that they are not completely

(and others)

and calculated

possibly

ionic radius

the difference?

plex, and we are not yet completely that Shannon and Prewitt

structures

table except

in oxides

exist b

and fluorides

from more than 700 refine

oxide and fluoride

crystal

structures

and because

oxygen

and fluorine

are similar

both electronegativity

and size, the amount of covalence in a given bond, say Si-O 4 a radius to Si + which will "work" in most silicate tetra2 the outer electrons in sulfur are 3s 3p4, their orbitals are more

does not prevent

giving

hedra.

Third,

diffuse

and extend

are close enough

farther

to be involved

What does this mean for sulfides? the inconsistencies

observed

the electronegativity

and the d orbitals

from the atom than those of oxygen,

in energy

for oxides

in bond formation.

First, Shannon and sulfides

(197l) showed

that many of

could be explained

by taking

e.g., Co in CoGe0 . The effec1 3 of a third atom A in compound ApBqXr (AxByX ) is to increase the interatomic disz tance of the B-X bond, when the electronegativity of A is greater and to decrease when

of a third atom into account,

the electronegativity

with the covalence

than we do in oxides. exclusively

is less.

The magnitude

of this effect

of the B-X bond and hence we see greater Second,

it might be possible

for use with sulfides,

to derive

but there are not nearly

data to do so even if there are no other problems. exist, but which has not been well that we find in sulfides. cation-anion

distances,

determine

a correction

a sulfide

may change

factor

for metal

depending

Recently

covalent

Gamble

ratios,

shorter

One result

than calculated

covalent

it is, the shorter

catatoms

in these structures

shows

of sulfur

atoms,

the sharp division

discusses

the reasons

ture calculations

in halides

if covalence

when

or in corrections Shannon of different

and Mg2+.

iTI

on stoichiometry,

the covalence'is

the bond.

Another

are coordinated

depending between

(1974) published interatomic

and chalcogenides. metal sulfides

interesting

are taken into account

result

ratio

are propor-

is that the prism or an octa-

(r+/r-).

Figure

Gamble

these involve

What is important sulfide

either

com-

such as Ti, V, Hf,

the two types of coordination.

for the two types of coordination:

Gamble

That is, the more

a trigonal

on the (ionic) radius

giving ionicl

with ionic char-

distances

greater.

by either

us here.

papers

distances,

crystal

PR-l

(1974) band struc-

is that the chemistry,

in the model being

but onl

investigated

to the radii.

and Vincent

(1974) in a somewhat

types of structures,

of isotypic

the reverse

the

to

the type of bonding

present,

from ionic radii in compounds

and do not concern

effects

bondin~

will affect

is yet known

Third,

is that catatom-anatom

ionic radii can be used to investigate

published

volumes

properties

between

of IVb, Vb, and VIb transition

Zr, Nb, and Ta chalcogenides

hedron

and Vincent

relationships

and magnetic

acter and bond lengths. tionally

interaction.

that does

of the compound.

(1974) and Shannon

ratios

good structure

metal-metal

interactions not enough

on the types of cations

data and discussing

pares radius

enough

is the extensive

However,

in sulfides

a set of radii

One other problem

We know that metal-metal even in oxides.

or even on the stereochemistry

important

documented,

seems to increasE

variations

compounds.

In some structures is true.

investigated

similar,

but more

the effects

comprehensive

of covalence

study

on cell

We have long been concerned with the radii of Ni2+ 2 Ni + seems to be larger than Mg2+, and in others

A few cell

PR-33

TRIGONAL PRISMATIC REGION

L...

0.40

0.38

•

0.36

,.

a a

+' 6

0.34

.;Y. o

~ ~

0.32

7i

Fig. PR-13. fractional

0.30

0.28

ionic character

fi of the

bond in Group Ivb,

metal-chalcogen

OCTAHEDRAL REGION

ratio r+Ir- vs. t

Raduis

and VIb transition-metal (after Gamble,

vI

dichalcogenj

1974).

0.26

0.24

0

10

15

30

25

20

fj X 100

volumes

(~3) from Shannon

and Vincent

assume

that the larger volume

radius

for the substituting NiO

72.4

NiF2

33.4

NiI2

MgO

74.7

MgF

32.6

MgI

2

in volume which

307.9

tivity

difference

Vincent

appear

increases.

and shorter

to be a function

of Ni-X becomes

the magnitude

APPLICATION

stand bonding erals,

88.3

570.3 of covalence

and solid-solution sulfides

THEORIES

orbital

of the bonds at

beh&vior

or indirect

metal-metal

there are indirect

metal-metal

interactions

interactions

series

in which

the Fe-X bonds

which

interaction, through

there are both

or bonding. PR-34

are more

Shannon

and

contraction

effect.

TO SPECIFIC

of transition

have been chosen

than the Mg-A

as the electronega-

increases.

and band theories

'there is no direct

the pentlandite

greater

and devise a covalency

of the covalency

OF BONDING

how molecular

three specific

We

a larger

530.7

two pairs,

to Mg-X as the covalency

many mo~e examples

to represent

To illustrate

For the latter

relative

(1974) provide

parameter

represents

l02.l

2

Fe GeS 2 4 Mg GeS 2 4

290.0

in each pair and the shortening

covalent

for illustration.

In the first three pairs the Ni-X bonds are more covalent

dramatic. bonds

compounds

atom.

Fe Si0 Z 4 Mg Si0 2 4 The reversals

(1974) are given below

of a pair of isotypic

SULFIDES

are used to help under-

metal include

ions in sulfide min(1) pyrite

(2) thiospinels

in which in which

sulfur intermediaries,

direct and indirect

and (3)

metal-metal

(cf. Ch. 2, pp. W23-W25)

Pyrite

The structure the NaC1-type present

of pyrite is based on a face-centered

structure.

as S2 units,

The cations

are located

cubic array of ions with

on Na positions.

and the center of each S-S bond is located

Each sulfur atom is bonded

to three metal atoms and one sulfur

in the form of a distorted

tetrahedron.

atoms and the sulfur corners

octahedron

with neighboring

atom

(in the S2 unit)

Each cation is coordinated

is compressed

octahedra.

The anions are

on a Cl position.

along the trigonal

The metal-sulfur

distance

to six sulfur

axis and shares

in pyrite,

for exam-

ple, is 2.26A, which obtained pyrite

is significantly shorter than the theoretical distance 2.62A VI 2+ IV 2the radii of Fe and S This indicates that in the

by adding

structure

Fe-S bonding

the interatomic hybridization

distances

is essentially

of sulfur

in the structure 3s and 3p orbitals

al, 1968; Burns and Vaughan, One of the four hybrid

orbitals

ions) of the cations the hybrid

orbitals

ular orbitals bonding metals PR-9,

The above authors

1970). ion and

also assumed

towards

Covalent

sulfur

mixing

of

give rise to a set of bonding

molec-

la _ + 6a _ ' totalling seven) and another set of antiS S M S Because sulfur is more electronegative than the transition

(a*).

and are completely

the bonding

filled with

ion); the antibonding

The three t2g orbitals

therefore,

remain

these t2

orbitals

g

sulfur

(those that are directed

and cations

(Bither et

and Kjekshus,

with another

of the type d2sp3.

orbitals

of both anions

ions.

and

to assume

(cr orbitals:

orbitals

empty.

proof

with three metal

arrangement

orbitals

1972; Brostigan

in bonding

group of 3d orbitals

(Fe, Co, Ni, etc.),

the metal

is involved

~orm hybrid

The atomic investigators

to form sp3 hybrid

1970; Goodenough,

the other three are involved that 4s, 4p and the e

covalent. led several

nonbonding.

of

orbitals

11

orbitals

are primarily

anionic

cationic

in a bonding

Burns and Vaughan

and are usually

with

sulfur

ions and,

(1970) considered

in 1I-bonding with sulfur 3d orbitals,

bonding

(Fig.

(12 from the S2 unit and 2 from

are primarily

are not involved However,

are involved

for the existence

(a)

14 electrons

is by no means clear.

that

although

It should be noted

here that the t2 orbitals are no longer degenerate in the pyrite structure because g of the trigonal distortion of the sulfur octahedron. For FeS2 the six 3d electrons of divalent Fe(d6) occupy the three t2 orbitals with their spins paired. Thereg fore, iron in pyrite is a diamagnetic semiconductor. The covalent mixing of e 3 g 2 orbitals of Fe + with the sp hybrid orbitals 0f sulfur results in the destabilization

of these orbitals

conductivity this

e * band.

Because

form a narrow

e*

(a*) band.

antibonding

compounds

Any metallic

could be attributed to partial all the six 3d electrons of Fe 2+ are present

g* band is empty.

in ,pyrite, the e

orbitals CuS2' the

filling

of

in the t2

In the series FeS , CoS , NiS ' an~ 2 2 Z respectively. Increa-

band is filled with 0, 1, 2, and 3 electrons,

g sing the electron because

which

in pyrite-type

of greater

in the eg electronic repulaion

population

*

band results

in increasing

and, therefore,

of eg orbitals with sulfur orbitals. increase in the series, in the order listed

mixing

a reduction

bond distances in the covalent

Consequently,

all parameters

in Table PR-8.

Furthermore,

PR-35

also electrical

and magnetic g*

the e

magnetic

properties

band.

change with

For example,

metallic

M

2

the increasing

- S distance

Number

(1)

e

*

band

It is evident from the foregoing 2+ 2+ Co and Ni in pyrite will increase

FeS

is contrary

- CoS

2 Figure

2 PR-14

to what we would

Furthermore, - NiS

2 shows

FeS2

CoS2

NiS

2.26

2.34

2.40

5.42

5.53

5.69

5.79

0

1

2

3

discussion

the cell parameter expect

the limits

on the basis

2

can also be explained

between

between

Fe-S

(2.32A)

is intermediate

(2.26A)

and Ni-S

solution

(2.401)

between

scheme

considerasystem

discussed

above.

FeS2 and NiS2 is not complete. in the interatomic distances

in the pyrite

structure.

The Co-S distance

Fe-S and Ni-S and, therefore,

with both FeS2 and NiS2

This Obser-

radii

in the ternary

by the bonding

that the solid solution

2 of Fe + by

of pyrite. of ionic

of solid solution

this could be due to the large difference

solid

CuS

2

that substitution

Perhaps

complete

in

CoS2 is a ferro-

of electrons

in the e

tions alone.

of electrons

~emiconductor,

Interatomic Distances and Cell Parameters in Transition Metal Disulfides

Cell edge, a (A)

vation

population

is a diamagnetic

conductor.

TABLE PR-B.

2+

FeS

(Nickel,

could

account

for

1970).

Thiospinels Common spinel

sulfide

structure

minerals

include

with

the

carrolite

(CuC0 S ), linnaeite (C03S4), siegenite 2 4 [(Co,Ni)3S41, polydymite (Ni3S4), violarite

[(FeNi2)S41,

daubreelite spinel

structure

close packing cations

system

FeS2-CoS2-NiS2

(after Nickel,

1970).

in the

B AB2S4

at 700°C

cations). occupied

of the tetra~

in the unit cell which

by one kind cation) kind

for diva-

for trivalent

the tetrahedral

a divalent

The

of the available

and B stands

When

and

The thio-

on the cubic

(where A stands

by another

PR-36

(Fe3S4)

atoms.

and one eighth

lent cations solution

one half

interstices

contains

Solid

is based

of sulfur

occupy

octahedral hedral

Fig. PR-14.

greigite

[(Fe,Mn,Zn)Cr2S41.

of cation

site is (i.e., by

and the octahedral

(i.e., by trivalent

sites cations)

the spinel occupied

is referred

to as normal.

by A and one B cations

type is designated

indirect

cations

of the presence interactions

octahedral

cations

when

the octahedral

and the tetrahedral

as inverse.

to three octahedral indication

Whereas,

Each sulfur

site by another

atom in the structure

and to one tetrahedral of direct metal-metal

are believed

and between

cation.

sulfur

and tetrahedral

B, the spinel

is coordinated

Although

interaction

to exist through

octahedral

sites are

there is no

in the structure, intermediaries

between

(Vaughan et al,

cations

1971). Goodenough thiospinel

(1969) and Vaughan

using molecular

us take a common

thiospinel,

sites are occupied As in pyrite,

orbital

et al (l97l) explained and band theories.

the bonding

aspects

To illustrate

in

this, let

C0 S , which is a normal spinel whose octahedral 3 4 cobalt and tetrahedral si~e by divalent cobalt.

by trivalent

the sulfur

ions are assumed to form four sp3 hybrid orbitals. The 6 of octahedral trivalent cobalt (d ) along with the e group g 3 mix with the sp hybrid orbitals of sulfur to form a set of stable

46 and 4p orbitals of 3d orbitals bonding

orbitals,

antibonding of trivalent paired

0B (which are essentially

orbitals, cobalt

divalent

the three t2

occupy

(i.e., low-spin

be empty.

state).

In the tetrahedral 7 cobalt (d ) mix with

(oA*) orbitals.

anionic)

0B* (which are essentially

e

and another

set of unstable

The six 3d electrons

cationic).

3d orbitals with their spins

nonbonding

*

Therefore, the antibonding e band in Co S will g 3 4 site, the 46, 4p and t2 group of 3d orbitals of sulfur orbitals

The two e group 3d orbitals

to

form bonding

(oA) and antibondin

remain nonbonding

and are completely

filled with four electrons with their spins paired. The other three electrons of IV 2+ * . Co occupy the three t2 group of antibondlng orbitals with their spins unpaire, The bonding

orbitals

are completely

in both cases

(oB and 0A) form part of the valence

filled with electrons.

The antibonding

0B

*

and 0A

*

band and

orbitals

form

the conduction band and are usually empty. The antibonding e * group of octahedra, 2+ g 3+ * form narrow bands and fall Co and t2 group of orbitals of tetrahedral Co below

the main

conduction

hedral

cobalt

ions result

form a single, completely

broad,

band.

Interactions

partially

delocalized

filled

(Fig. PR-l5).

temperature-independent

bands of 3d orbitals

type of cations

occupying

of metal-sulfur

covalent

* g

and t2* bands

localized antiferro-,

explained

mayor

may not overlap,

the two crystallographic bonding

in each site.

and exchange

and ferrimagnetism physical

by the bonding

linnaeite-polydymite

and tetra-

depending

exhibited

scheme outlined

series,

the

sites and also upon the extent

If there is an energy

gap between

in FeCr2S4, then the d electrons tend to be magnetic interactions result in ferro-,

and the sulfide

properties

to

upon the

as, for example,

on the cation

The various

the octahedral

Therefore, C03S4 exhibits metallic and (Pauliparamagnetic) properties. However,

paramagnetic

two antibonding

the e

between

of these e * and t * narrow bands g 2 d band in.which three 3d electrons are

in the coalescence

substitution

above

becomes

a semiconductor.

by the thiospinel

of Co by Ni results

PR-37

group could be

(Vaughan et al, 1971).

In the

in an increase

in

of 3d electrons

the number

in the antibonding d band. In polydymite, an inverse 3+ 2+ 3+ cation distribution Ni [Ni ,Ni ]SB' there will be six 3d

spinel with a formal electrons

series

d

in the

substitution

band as opposed

is believed

"nickel

to only three electrons

of Co by Ni in the linnaeite-siegenite-polydymite to increase

substituting

the cell parameter.

in Co S • Therefore, 3 4 solid-solution

Furthermore,

it is stated

that

for cobalt

structure"

and, therefore,

explained"

(Vaughan

in C0 S does not fundamentally alter the band 3 4 "the existence of this solid solution series is readily

et al, 1971, p. 374).

In the case of the Fe-thiospinel,

grei-

gite (Fe S ), trivalent high-spin iron occupies the tetrahedral and one of the two 3 4 octahedral sites, whereas divalent high-spin iron occupies the other octahedral site.

The presence

lap between

of high-spin

* g

iron in the structure

and t 2* antibonding

the e

conducting.

Because

structures,

the solid

of the difference solution

PR-16).

Similarly

in which

one of the octahedral

greigite

is again

between

the lack of complete

explained

results

in negligible

d bands, and, therefore, in the spin states these two members solid solution

sites is occupied

overis semi,

of Fe and Co in the two is very limited

between

by divalent

as due to the difference

greigite

(Fig.

vio1arite

(FeNi S ) 2 4 low spin iron, and

in the spin states

of iron in

the two structures.

0':

(1':

17:

e9C:-Jt2~e9c=d_E'

e

_!L

_1L

ce'" IN

COB

TET-SITE

Fig. PR-15.

L·S

IN

OCT-SITE

Schematic

energy

Fig. PR-16.

Solid solution

in

level diagram

for the 3d orbi-

the thiospinel

tals in C03S4

(after Goodenough,

Co S4-Ni S • (The diagram is 3 3 4 schematic, indicating observed

1969).

natural

system

Fe3S4-

compositions.)

Pentlandites The composition where M stands amounts

of natural

for Fe, Co, and Ni.

(up to lO wt.%)

pseudo-cubic

pentlandites

closest

of Ag.

packing

is given by the general

Occasionally

The structure

of sulfur

atoms. PR-38

pent1andites

of pentlandite

contain is based

formula M S

9 B

substantial on the

In the unit cell of pentlandite,

there are four formula available The

units of M95 ' The cations occupy one-eighth of the S and one-half of the tetrahedral interstices (see Ch. l).

octahedral

54 tetrahedra

two equipoints (coordinated

are trigonally

distorted.

in the unit cell S(c) by five cations).

The sulfur

(coordinated

The bonding

atoms are located

by four cations)

can be explained

with the help of the

C0 S structure. ~he interatomic distances in C0 S 9 S 9 B (also in natural and magnetic and electrical properties suggest extensive metal-sulfur bonding

and metal-metal

of 3d orbitals bonding

interactions

of tetrahedral

(a) and antibonding

3d orbitals

are filled with

: three d electrons octahedral orbitals

of Co.

four electrons

with their spins paired.

The nonbonding

t2 * band.

the antibonding

results

in the formation

covalent

of S to form

orbitals.

of Co occupy

Similarly,

e group of The remaining mixing

of a 'narrow band of e

of

g

*

t2 orbitals will be filled with six electrons g one d electron (assuming Co is divalent)

The nonbonding

spins paired

3s and 3p orbitals

(a*) molecular

Co and S orbitals

with their

with

pentlandites)

The 4s, 4p, and t2 group

in the structure.

Co overlap

on

and 24 (e)

and the remaining

theeg * band. Although it is assumed here that Co in the pentlandite is dival~nt, the assumption is not strictly correct when we consider 2the structural formula, Co S . If S can be considered as divalent (S ), the

will occupy structure

9 B

structural

formula

(the apparent

valency,

by the expression µ'

=

16/9 ~ l.7S).

conduction

gives the apparent

mµ'

In the pentlandite actions,

complicate

tetrahedral dination

leads

should

in

structure

other

be empty) effects,

to the formation

Co's in the cluster. band which

metallic

especially

is completely

tetrahedral

indicate

tetrahedral

in the

inter-

that each

located

at

The three unpaired in bonding

(with respect

filled with electrons

structure

with

to metal-metal (aM(T)

in

Part of the pentlandite

showing metal

the "cube cluster" atoms which

dina ted to one sulfur

PR-39

is given

Therefore,

Co atoms and this coor-

Co are involved orbitals

sulfur

1.7B

conductivity.

of eight Co atoms,

Fig. PR-l7.

S2

XXX

metal-metal

distances

the unit cell (Fig. PR-l7).

The bonding

, as,

of Co could be located

of a cube cluster

of a small cube within

9 x µ'

causing

The interatomic

to three other

of the type Mµ m = B x 2.

c09SB,

4s and 4p electrons

otherwise

the picture.

form a broad

of Co in the structure

in a compound

For example

in the t2 * band of each tetrahedral

three other bonding)

Xx'.

Co is coordinated

the corners electrons

µ', of a cation

=

The excess

band which

valency

(S2) atoms.

of

are coor-

(Sl) and three

Fig. PR-lB)

and the antibonding

been considered

fixed to a constant

a~(T)' remain empty.

orbitals,

of 3d electrons

that the number

value of 56 electrons

(Rajamani

1

is

1973).

a-*M(T)

121

it hal

cube cluster

and Prewitt,

a-*M(T) a-*M(O)

Therefore,

in the metallic

I

121

eg~

>-

(!)

II::

w

Z

w

e..1L

Schematic

In addition metal hedral

energy

between

and octahedral

for the 3d orbitals

level diagram

to the formation

interactions

Co in M(T)

Co in M(T)

Co in M (0) Fig. PR-IB.

e _tl_

_1L

of cube cluster,

the cube clusters

Co atoms

9 B

the structure

of tetrahedral

S intermediaries.

through

in C0 S • also permits

Co and between

met

tetra-

All these interactions

of eg* and a~(T) bands to form a single, broad, d band as shown in Fig. PR-IB. Therefore, C09SB exhibits metaland Pauli-paramagnetism.

could lead to the coalescence partially-filled lic conductivity The presence has an important end members, homogeneous recemtly, pattern

of the cube-cluster effect

bonding

atoms

in the pentlandite

of this mineral.

the synthesis similar

of a thin film of cubic iron sulfide

to that of pentlandite

conveniently

that because

electron,

(1961) also observed

Fe and Ni in the structure

bonds,

is limited,

PR-40

More

with a diffraction

et aL, 1973).

in the t2

and even if formed,

which might

that the individual whereas

to form a stable 1961).

(Nakazawa

Ni 2+ has four electrons

form three metal-metal

a~(T) will have an unpaired

Knop and !~rahim

has been described

structurE

Of the three possible

Fe S , C0 S , and Ni S , only C0 S was observed 9 B 9 B 9 B 9 B phase in the system Fe-Co-Ni-S (Knop and Ibrahim,

It is easy to visualize cannot

of metal

on the chemistry

*

band,

it

the anti-

destabilize' the phase. substitution

when they substitute

of Co by

simultaneousl

for Co the solid solution landites

was found to ,be complete

also have a restricted

(Fig. PR-19).

range in composition

Natural

pent-

as shown in Fig. PR-19.

Fig. PR-19.

Natural

compositions

(Rajamani

1973).

Dashed

estimated

pentlandite and Prewitt,

lines represent

solid-solution

the

limits out-

lined by Knop and Ibrahim

(196l) in

The M9SB section of the system Fe-Co-Ni-S. Dashed-dot line represents the compositions

for which

Fe:Ni = 1:1.

A possible

reason

the structure

for this peculiar

fixed the number

of

d

chemistry

is as follows:

electrons

at a constant

the cube cluster value.

in

Therefore,

we

could expect

that pentlandite compositions fall close to the join Co S -Fe Ni4 5S 2+ 6 2+ B 2+ 7 9 B 4. 5 • This is because [Fe (d) + Ni (d )]/2 is equivalent to Co (d). Increasing the Ni content electrons

in pentlandite

over the ratio Ni:Fe : 1 will increase

in the unit cell.

cation vacancies electrons

in the tetrahedral

at a constant

increasing

constant

value.

the Fe content

the addition

of excess

number

trons to atoms

Ordering

of electrons,

of pentlandite

Similarly,

where

cations

compound

the structural

in Fe-rich

is due to metal

addition

keep the number

tetrahedral

0 represents

d

of

of

of d

Fe has only six 3d electrons,

because

in the unoccupied

Therefore,

the number

site and creation

over the ratio Fe:Ni : 1 will result

as in an electron

be [Fe,Ni,Co] (Fe,Ni,CoA)BSB compositions or excess

sites could effectively

in pentlandite

cations

is fixed).

of Ni in the octahedral

sites

formula

compositions.

a

(where the ratio of elecfor pentlandite

tetrahedral

and omission

in

to maintain

vacancies

could

in Ni-rich

Thus, the nonstoichiometry solid solution.

Conclusions These examples be described, Furthermore, physical

these theories

and chemical

be pointed pletely

show that aspects

to a first approximation,

properties

exhibited

theory,

in transition

by molecular

offer satisfactory

out that no single

satisfactory

of bonding

orbital

explanations

by these sulfides.

when applied

picture.

PR-4l

metal

sulfides

can

and band theories. for the various However,

to all sulfides,

it should

gives a com-

Ch.4

EXPERIMENTAL

METHODS

IN SULFIDE

SYNTHESIS

S. D. Scott

INTRODUCTION Mineralogists

have two fundamental

sulfide minerals. and chemical

homogeneity

for electrical,

The other is to determine variables

circumstance

has become

possible

This chapter

through

investigation

techniques

described

the phase equilibria the following of growing

the general

of chemical

interactions

to be practical

experiments produce

gradients

result

in growth

homogeneous

crystals

that is not covered given us important significant

here is solution insights

achievements

in synthesizing

The basic gUidel1.nes for studies

of sulfide

of petrology.

Indeed,

from silicate

petrologists,

although

sulfides

have led to some unique

context

requires

procedures.

as emphasized 1972).

of their encompassing

approach phases

are petrology

studies.

and sublimation from equilibrium Another

to

topic

which not only has

but also has led to

crystals

phase

in hydrothermal

special

S-l

problems

A significant bv Barton

solutions.

are about the methods

in

poi~t is that all

(Sulfide Petrology,

stability

have been

encountered

are best examined

be they sulfide

to determine

of interest.

equilibria

some experimental

As such, sulfides

rock systems,

experimentation

transport,

OF EXPERIMENTATION

same as other branches

(Ore Petrology,

of sulfides

sulfide

borrowed

such endeavors

transport

processes

vapor

that are too high

solid solutions.

chemistry

into ore-forming

SOME PRINCIPLES

by Stanton

by chemical

that is often too far removed

of

on sulfides.

phase equilibrium

used in vapor

technology

of a Short Course

at temperatures

of multicomponent

The

by Craig and Scott in

to phase studies

for most geochemically-pertinent

the steep thermal

guiding

sulfides.

on the specialized

have been grown from melts, operate

recently

and were used to obtain most of

This is a topic worthy

but these methods

the best which

and techniques

among metallic

data discussed

large crystals

Of course,

principles

I will touch only briefly

are unsuited

studies.

thermodynamic

research.

and thermochemical

crystals.

of

size, perfection,

Simultaneously

in sulfide

examines

its own but many of its techniques

BeSides,

processes.

both objectives

practice

large single

and by sublimatio~

and x-ray diffraction

and underlying

are all in current

chapter.

For example,

optical,

some advances

primarily

laboratory

of sufficient

natural

is to satisfy

syntheses

crystals

phase relationships

with a view to unravelling

possible

aims in their experimental

One is to grow single

or silicate.

relations

1970) and in the A first

among the

Appearance-of-Phase

Method

In its simplest location

form, the construction

of all phase boundaries

of which phases

are in equilibrium

and bulk composition

(X).

chosen;

at a particular

or, conversely,

the

(T), pressure

are necessarily

under study might

some compounds

requires

that is, the determination

temperature

Not all such solid phases

all or part of the P-T-X conditions in nature

of a phase diagram

for the system

lie outside

have been synthesized

(P),

minerals;

of those found

before

they were

found as minerals. The usual way of determining phase" method.

Carefully

P and T and the products different

phase

assemblages

appear

2, then, assuming

boundaries

must

reducing

the incremental

boundary

is closely

equilibrium

change

expected.

and Skinner

there are two possible

temperatures

arrangements

Within

where

than using

of tie-lines

choosing

terrestrial context

For example,

environments

experiments

discussed

higher

~G;,is negative

P-T conditions.

fixed confining

time can be

assemblages

for study

assemblages

intervals

pyrrhotite

could be most

are commonly

and iron are not.

Therefore,

found in

within

mineralogically

must also be included

during

retrograde

in the Fe-S system,

+

pyrrhotite

this represents S-2

the

to

On the other hand, if thermodynamic

troilite

nucleated

at,

can sometimes

(as

Care must be taken not to over-interpret

phases

For example,

monoclinic

pressure,

and

at all

proceeds

all bulk compositions

Natural

on the former phases.

as many contain

example

pair, not Ag2S + Pb.

further

or divariant

it would be more pertinent

in the next chapter).

ores often reveals

available,

or composition

are sought,

the following system Pb-Ag-S

the reaction

of examining

pyrite

but troilite

data such as FeS activities

assemblages

key univariant

to which phases

studied.

from thermo-

among the phases PbS, Ag2S, Pb

Therefore,

of the components.

of the Fe-S system,

concentrate

interest.

a "shot gun" approach

be used as a guide

the

represents

Wllich is the stable assemblage?

some data are already

say, IO wt. % increments

More

thermochemical

the ternary

Ag2S + Pb t PbS + 2Ag, the free energy,

saved by carefully

profitably

by calculating

to the right and PbS + Ag is the equilibrium

In systems

until

actually

but, for now, consider

(1967, p. 279).

of mineralogical

spontaneously

rather

to experiment

the boundary

Craig and Scott discuss

chapter

and Ag; these are Ag2S + Pb and PbS + Ag. For the reaction,

I and 2.

by successively

below.

in the following

from Barton

of experiments

in X from experiment

If

I and

one or more phase

can then be determined

time can be saved in experiments

the equilibria

calculations

was achieved,

Tests of whether

at a particular

of the experiments.

of, say, experiment

the bulk compositions

bracketed.

are discussed

Occasionally, dynamics

in the products

of a phase boundary

is by the "appearance-ofare heated

at the completion

equilibrium

lie between

location

bulk compositions

are examined

experiment

precise

phase relationships

weighed

hexagonal

an invariant

cooling

natural

from

close examination

pyrrhotite

+

assemblage

pyrite. suggesting

of At a that

one phase--monoclinic concomitant

pyrrhotite

readjustment

as it turns ou~nucleated

Thus far our appearance-of-phase are compatible.

To derive maximum

know the compositions system.

for different

pyrrhotite,

pyrrhotite

(As,S)-liquid

It was known

arsenopyrite

displayed

understanding

arsenopyrite

+ arsenic,

this might be possible

with

(plus vapor)

and Clark,

approximately

of T, P, and the assemblage of the functional

compositions

those expected

relationships

in nature

to Thus

of arsenopyrite That

(1973) who found that the

for different

from experimental

it occurred.

and geothermometry.

by Kretschmar

+

1961) that

from FeAsO.9SI.1

in which

for geobarometry

loellingite and pyrite +

(As,S)-liquid

Morimoto

that

can be in univariant

+ loellingite,

pyrite+

was demonstrated

range of arsenopyrite

(1960a) demonstrated

arsenic

(Clark, 1960a,b;

to P and T would be useful

compatible

A case in point is the Fe-As-S

a range in composition

as a function

composition

have told us only what phases

from a system we also need to

by Clark

+ (As,S)-liquid,

pyrrhotite.

a complete

experiments

T and X conditions

with

FeAsI.ISO.9

experiments information

of its solid solutions.

Appearance-of-phase

equilibrium

upon cooling without

of the other phases.

assemblages

was

data.

Equilibrium EqUilibrium

has been variously

system has no spontaneous "a state

defined

tendency

[such] that after any slight

it returns

rapidly

or slowly

The first definition

as "a state

temporary

to the initial

[of rest] from which

a

(Barton et a1, 1963, p. 171) and

to change"

disturbance

state"

of external

(Lewis and Randall,

has given rise to two tests for equilibrium

conditions 1961, p. 15).

in experimental

studies: I.

The system

periodically.

tj.me, the system 2.

is assumed

The experiment

If the results Neither

is held at constant

If, after initial

using different

equilibrium

any possibility

reliable.

materials,

no change will be observed

If, instead, materials,

+ pyrite

iron and sulfur

some pyrite will

both cases,

the system

consequence

of nonreactivity,

at 300°C.

Kinetics

the system

materials

For example,

may be or else

consider

the

of heating.

1:2 and FeAs2 are the starting

but loellingite

will not change.

In

in a state of rest, but this is only a

not of equilibrium.

The best test of equilibrium, to perturb

of a reaction

in them even after many months

form eventually

materials.

If FeAs2 and FeS2 are used as starting

in the proportion

is seemingly

starting

in the starting

change may occur in some phases but not in others. loellingite

are examined

are noted with

is assumed.

of change

assemblage

changes

equilibrium.

perhaps

of these tests is completely

and the products

no further

to have reached

is repeated,

are reproducible,

such as to preclude

conditions

reaction,

following

in some manner

causing

S-3

the Lewis and Randall

definition,

change

and then return

in the phases,

is

the system state,

to its original

the reaction

One major formation

cause of incorrect

of metastable

is an equilibrium metastable

conditions.

phases

interpretation

and failure

state and, particularly

and most stable states

and go undetected.

reaction

the metastable

reaction

to flux the reaction, hexagonal

Recognition

mineral

of equilibrium

processes.

equilibrium

Cas.

I

-:; Co .. 2

E

Cose

:3

0

fi3 0

Cose

4

0

Case

5

81

Case

6

§

cr

M·~

"I.;;

",'"

"

0

~ ~

provide

o

Fig. S-l.

B

0 (\

c

"g .~

(1970a) was able to

+

hexagonal

pyrrhotite

(1974), using a hydrothermal pyrrhotite

inverted

results

at

technique

reversibility

to

some useful information.

A A

.. A. IA

Some relations

it usually

encompasses

Some components

several

Equilibrium

Z on.d equilibrium. superposition

of

severe!

case I oss.mblages

many more components

equilibrium

Ideol for us. of phose information

equilibrium

01

Use restricted to motched poirs of zones

A~"~f~~=w':~~f~b~iJmus.'utunder

Synchronous

than

low

phase

:iti~

was n,orly

bulk equilibrium

Even when bulk

Application

(s;mp/~J

but

types of

will be in sufficiently

Oescription

problem

to the interpretation

(Fig. S-l), most of which do not represent

A 0

0;

Kissin

Taylor

pyrrhotite

Barton et a1 (1973) have recognized

experiments.

c

0 '0; 0

above for

are encountered

in nature is an even more cantankerous

does occur in nature

our laboratory

"'C

pyrrhotite.

if we are to apply experimental

associations

"'

of the type described

The real difficulties

found that monoclinic

but may, nevertheless,

'!.

the

to Nature

is no less important of natural

between

the test of reversibility

+ pyrite at 254°c.

pyrrhotite

Applications

barrier

is the

Metastability

the stable phase field for an assemblage.

monoclinic

292°C in the solid state whereas

is assumed. results

them as such.

if the activation

few problems.

A case in point involves monoclinic reverse

of experimental

is large, it will survive

occurs outside

their original

and equilibrium

to recognize

In this sense, metastability

+ pyrite presents

loellingite

when complete

If the phases reattain

is said to have been reversed

some circumstances

st~tic deposition

but one or both metasfabi.

Not useful

if precision

is required

Equilibrium because of re -equilibration from Gives conditions of re-equilibration any previous association Internol homogenization of zoned equilibrium in one or both without mutual re-equilibrctton

between

coexisting

Would give unreliable

minerals.

Mineral Soc. Amer. Spec. Pap. 1, 171-185).

S-4

results

(From Barton et a1, 1963,

concentration possible which

that they can be ignored

for those components

case their thermodynamic

concentrations.

to experiment

to preserve

and temperature

which are sufficiently

are useful

or subsequent

metamorphism

sulfides

pyrrhotites

at 250°C and Cu-Fe sulfides

pyrrhotite

annealed

as pOinted

out by Barton

minerals pyrite,

which

arseno-

of the reactivity

For example,

of

experiments

equilibrium

after heating

processes

solid-state

in the laboratory

at 580°C

+

after only seven days.

sulfides

by

+

of sphalerite

(1966), the experimental

on ore-forming in standard,

range below 600°C.

overcoming

of

(1968) found that troilite

equilibrium

and Toulmin

to establish

conditions

that it takes as long to

reaction

on the one hand, refractory

information unreactive

temperature

Yund and Hall

at 90°C reached

faced with a problem:

are difficult

at IOO°C.

solid-state

failed to reach complete

pyrrhotite

in

pressure

at 650°C as it does the nonrefractory

(1966) involving

for more than a year, whereas

upon cooling

and include

In his comparison

the refractory

relatively

gain a better

Such "refractory"

(1970) estimated

equilibrate

most useful

and

at elevated

tools in deciphering of ores.

and sphalerite.

Barton and Toulmin

obtained

are few in number

in the solid state, Barton

hexagonal

Law in

to their

but by systematically

non-reactive

compositions

geochemical

only sluggishly

molybdenite

sulfides

with all of them simultaneously

their equilibrium

react with others pyrite,

will be proportioned

of nature.

Only those sulfides

formation

The latter is

two, then three, then four, etc. we will eventually

understanding

nature

contributions

for.

which obey Raoult's

We may still be left with a large number of components

may be tempted dealing with

or else corrected

of a solid solution

Thus,

mineralogist

will provide

is

the

but, on the other hand, being experiments,

within

The next section

their phase relations

the geologically-important

includes

some methods

of

this difficulty.

METHODS Evacuated

Silica Tube

Dry reaction synthesizing

in an evacuated

sulfides,

in this way than any other. procedures

in considerable

Fused silica

Its low thermal without

sealing

sulfides

tube is the time-honored phase equilibria

(1971) described

so I will discuss container

enables

Silica,

if thin enough,

reactions

into a capillary

the equipment

and

only the main features

sulfur below

expansion

is somewhat

in a high-temperature

(Moh and Taylor,

S-s

1971).

here.

It does not devitrify this temperature.

the tube to be sealed in an oxy-gas

the charge and its low thermal

to follow

means of

have been determined

for sulfides.

IIOO°C and does not react with

rapidly.

so it is possible

detail,

conductivity

disturbing

be quenched

Kullerud

is an excellent

until apprOXimately

silica

and more sulfide

enables transparent x-ray

flame

the runs to to x-rays

camera by

The preparation by Fig. S-2.

two tubes by melting torch

of a standard

An eight-inch

evacuated

the glass in a small, hot oxy-gas

(Fig. S-2,A and B).

This is accomplished

the blue cone of the flame until melting of the tube should be avoided during become

too thin.

collapses

the walls. otherwise

wearing welder's

goggles.

weighed

into the tube.

can be placed directly

in the bottom

actually

piece

starting

in the tube is

during loading.

The materials

of the tube using a narrow piece of creased

paper as a slide or by means of a funnel formed by drawing It is common practice

After all the materials

(~l") of tightly-fitting

they

on them

by first

of pre-weighed

out a piece

to add the sulfur first and gently melt it

in the bottom of the tube in order to avoid losses later during process.

flame

Next, the starting materials

This is best accomplished

In this way, the amount of each material

of soft glass.

the ends

the walls will

When the tubes have been separated

known despite any losses which might have occurred

weighing

the tube over

Pulling

them into cold water and a number is scratched

the tube empty and again after each addition

components.

into

flame from a welding

of the silica but the bright-white

with a diamond scribe for later identification.

weighing

is illustrated

tubing is separated

The two ends of the tube can be safely held with your fingers

are chilled by plunging

are carefully

experiment

by slowly rotating

this process,

thanks to the low thermal conductivity necessitates

silica-tube

length of thoroughly-cleaned

have been successfully

silica rod is inserted

j

VSilica

the evacuation

transferred,

(Fig. S-2,C)

a short

to eliminate

wool wrapping

~ChOrge

t

I 0

u'' ·~

E

wool

.

Sulfur ),:,,:' ....

-

F

A

B

Silico glass rod

,

;I;'./;'

G

!=Charge

H

~Ii~:~ psad

Weld

J Fig. S-2. Various

Charge

K

L

M

types of tubes used in sulfide experiments:

silica tubes with a glass rod; F-G, tube-in-tube; precious

metal

and High

Temperature

tube.

(From Kullerud,

1971, Research

(Ulmer, ed), Fig. I, p. 291).

S-6

A-E, simple evacuated

H, DTA tube; I-M, collapsible Techniques

for High

Pressure

much of the vapor

space and the tube is ready for evacuating

process is tricky but, with practice, the tube is necked down to a capillary again being careful wrapped

around

to avoid thinning

just above the filler-rod of the walls.

Second,

the tube is connected

to O'.02mm Hg or better.

line.

to prevent

the tube is melted

the capillary

flipping

the fine-grained

the

charge

It also helps to have a large capacity

in che line to act as a "vacuum buffer."

vacuum pump running,

water,

the charge and fingers

to a vacuum pump and evacuac~u

pump switch on and off a few times initially

vacuum dessicator

(Fig. S-2,D),

This should be done in stages by rapidly

from being sucked into the vacuum

This First,

A wet strip of cloth

the bottom end of the tube will protect

from the heat.

and sealing.

can be done in just a few minutes.

is sealed

around the filler-rod

Finally,

with the

off in the flame, the top end of

(Fig. S-I,E),

the tube is quenched

in

and we are ready to start the experiment.

The size of tubing used is dictated reacted,

although

materials

can be prepared

6mm I.D. x 8mm O.D. of time. usually

in part by the amount of charge

the larger the tube the more difficult in batches

of several hundred

Larger quantities

Also, for this reason charges sufficient

milligrams

may not homogenize

within

Starting

in tubes of reasonable

should be as small as possible:

and can be conveniently

contained

to be

it is to seal.

within

lengths

30 mg is

a tube 3mm I.D. x 4mm

O.D. x 3cm long. Starting

materials

the native elements 6N (99.9999%)

pure.

are either native

by the evacuated Most metals

This is accomplished

iron, develop by heating

point in a hydrogen

gas stream.

hydrogen

it through a dessicant

by passing

A train for this process hydrogen

of interest

the powdered

alumina

windings

by Kullerud

and then over zinc metal

a nickel

can be controlled

proportional

of an inexpensive

or a time-proportioning

by thermocouples

alumel thermocouples

at 500°C, is ±2°C.

in direct

are reliable

The usual error quoted

tank for

for quality

tubes are reacted

inches by carefully

sheath between

$200-$500

spacing

the inner and outer

±loC or better by means of

or within

controller.

Temperature

contact with the silica tubes.

thermocouples

from a large

±SoC by means

to 900°C and Pt-Pt + Rh at higher chromel-alumel

in

in Fig. S-3.

which are available

in the price range

variac

silica

within

controllers

number of manufactureres

monitored

at 400°C.

oxide boundary

(1971) and illustrated

is spread over several

and inserting

Temperatures

stepless

from the

to the experimenter.

of the furnace

tubes.

solid-state,

sponge,

metal below its melting

is easily set up and will lower f02 of normal

of the type described

the resistanc~

from

can be obtained

Water and oxygen can be scrubbed

The charges which were sealed into evacuated

The hot-spot

presynthesized

Sulfur

an oxide coating which must be

to 10-46 atm which is well below the metal-metal

most metals

furnaces

or sulfides

tube method.

are now available' 4-SN pure as wire, sheet,

or powder but many, particularly removed.

elements

silica

is

Chrome ltemperatures.

is ±3/8% which,

I

Fig. S-3. Standard 1971, Research p. 301).

tube furnace

Techniques

600°C to combine temperature.

in which

can contain

liquid

the pressure

silica

silica

into closer

and regrinding

in a hydraulic

the charge

and gradually

Advantages I.

The equipment

2.

The system

lowering

3.

Clear glass

However,

within

either

through

encapsulating

sulfides,

lengths

can sometimes

to the desired

tube method

for evacuated

or by forcing

in a silica

the

Rates of

Refractory

reasonable

Reactions

the temperature

of

be speeded

the grains

by pelletizing tube or by melting value.

are:

and easy to operate

permit visual

known bulk composition.

examination

of the charge without

termination

of the experiment. 4.

Runs can be quenched

consideration

rapidly

when dealing

the

vessel.

The latter is accomplished

is clean and of precisely tubes

at 1000°C.

the tube with a pressurized

slow and it is not unusual

silica

problem-

and 3mm I.D.

or in the solid phases.

temperatures.

press before

is inexpensive

sulfur

the charge periodically

of the evacuated

I.Smm walls

to reach equilibrium.

contact with one another.

the charge

always wear

are particularly

tubes rely on diffusion,

to eqUilibrate

below

run

of explosion

tube with

pressure

elements

to take months

time at geologically-significant

the desired

point at

tubes from a furnace.

is to support

can be exceedingly

are very difficult

up by quenching

attaining

(~lOO bars) of saturated

in evacuated

above its boiling

is one of the phases

cold-seal

(From Kullerud,

(Ulmer, ed~, Fig. 3,

Temperature,

of the danger

such an experiment

tube experiments

in particular,

tube experiments.

the charge is preheated

or removing

that a silica

for the more volatile

solid state diffusion

because

sulfur

gas such as argon in a standard Reactions

before

inspecting

(1971) claims

safest way to conduct

vapor phase

tubes unless

the sulfur and metals

Kullerud

silica and High

of sulfur rises very rapidly

A word of caution:

Experiments

Pressure

thin-walled

a face shield when inserting,

atical.

for evacuated

for High

The vapor pressure 445°C and will explode

montle

by dropping

with most sulfides.

S-8

the tubes into water,

an important

Disadvantages I.

Solid-state

within 2.

reasonable

Because

in some sulfides

lengths

The run products

for single-crystal 3.

are:

diffusion

are usually

x-ray

Thermal

Analysis

When a phase change surroundings.

occurs,

By detecting

of the phase

into a recess a reference assemblies temperature

confining

and unsuitable,

pressure

in a evacuated

silica

slight

vapor

phase.

change

(endothermic

In practice

from or added to the

temperature

increment

or exothermic)

a test thermocouple

tube containing

the charge

into a block of inert material

thermocouple

are recorded

Typical

and the

(Fig. S-2,H),

such as alumina.

and its difference

as the furnace

is slowly

with

and the

respect

heated

to

or cooled.

is caused by the a-S transition

the reference

large peak beginning

in quartz

863°C represents

to

thermocouple.

at 609-610°C

of pentlandite

heazlewoodite.

are

The small peak at 573°C

which was added to the pentlandite

the breakdown

and

Both

thermo grams for pentlandite

shown in Fig. S-4.

calibrate

we can

is inserted

a few mm apart in the hot spot of the furnace

of the reference

the test thermocouple

for example,

cannot be varied

heat is either absorbed

the resulting

boundary.

thermocouple are placed

equilibrium

encountered.

(DTA)

the sign of the enthalpy

temperature

fine-grained

but will always be that of the ever-present

Differential

determine

very

stable

is frequently

diffractometry.

the silica tube is rigid,

independently

is too slow to attain

of time and metastability

The

represents

to (Ni,Fe)l_xS

The peak beginning the intersection

and

at 862of the

solidus. 4.~

The chief advantage

of DTA is the speed

with which phase boundaries compared

with appearance-of-phase

However,

DTA is plagued

problems

as normal

Fig. S-4.

Differential

thermal

curves for synthetic

Fe4.5Ni4.SS8 and cooling

on heating (top).

(bottom)

(From

silica

tube experiments

and,

can be interpreted

only in the light of some prior knowledge

of

the phase relations

A

under investigation.

heat effect by itself phases were consumed

'(ullerud, 1963, Canadian Mineral. "

experiments.

by the same kinetic

in most cases, its results analysis

Cqn be detected

358).

S-9

does not tell us what or produced.

Salt Flux An obvious

experiments

way of overcoming

reaction

is to add a flux or catalyst

reactivity

among the sulfides

without

shifting

in which the sulfides

are slightly

been used for decades

in the preparation

recently,

Boorman

to extend

the sphalerite

evacuated

silica

facilitated

by its eutectic although

tube method.

temperature

and coarsening

At present

Kretschmar

Binary

(Moh and Taylor,

>360

c: ioh

PyJTho.tite lndicator

S-3.

is to at a

constant end.

temperature

at one end and the charge

This configuration

and temperatures

permits

at a higher

investigation

S-3.

Equilibrium

Constants

Sulfur Species

=

Sn(g)

=

log K Gaseous Soecies S

at the other

range of sulfur activities

liquid + vapor

than given by the univariant

Table

temperature

of a wider

curve for sulfur.

for Gaseous

(Mills, 1974)

(n/2)S2(g)

AT-l +

B

A

+ Clog

T C

B

ll,219

- 2.026

-0.308

S2 S3

- 2,432

3.807

-0.l38

S4

- 3,607

8.293

-0.586 -0.747

Ss

-10,880

l4.933

S6

-14,790

20.347

-1.166

S7

-17,930

24.874

-1.47

Sa

-21,690

30.58

-1. 968

Dew Point As described method

utilizes

vapor

to compute

activity

of S2'

is connected equilibrium

by Dickson

et al (1962) and illustrated

the temperature

of the meniscus

total sulfur

pressure

The reaction

tube containing

to a silica

capillary

temperature

in Figure liquid

and, from the equations

which

sulfur vapor over the charge

sulfur at a point whose

between

the charge

is interpolated

and sulfur

in Table S-3,

at a constant

lies in a temperature condenses

S-16, this

sulfur

in the capillary between

temperature

gradient.

The

to liquid

closely-spaced

thermocouples.

Gas Mixing Mixing practice

of gases to control

in petrological

the equilibrium,

2H S 2

t

studies

activities

of O , CO , S02' etc. is a common 2 2 and can be equally applied to sulfides. For

2H2 + S2' the sulfur

S-30

activity

is given by

THERMOCOUPLE

WELLS

I

STEEL

TUBE

TO

GLASS

CAPILLARY

HOLD

AND

GLASS

ASSEMBLY

SAMPLE

CHAMBER

800

1'00

• %

400

CS-9

(1967)

Schenck & von der Forst (1939) Fleet (1973) DuPreez (1945) Peacock & McAndrew (1950) Brower et al (1974)

System Bi-Pb-S

Mineral Name

Compound 6Pb1_xBi2x/3S,

Bi2S3

heyrovskyite

829

3Pbl_xBi2x/3S.Bi2S3

li11ianite

816

PbBi2S4

galenobismutite

750

cosalite

Pb2Bi2S5 Pbl-xBi2x/3S'

Bi-Sb-S

Max. thermal stability, ·C

2Bi2S3

bursaite

?

bonchevite

?

PbBi6S1O

ustarasite

s.s. Bi2S3-Sb2S3

at >200·C;

a gap from Bil.16SbO.84S3

"'Bi2Te2+xSl_x

Contains minor Ag & Cu

Schenck et al (1939) Van Hook (1960) Craig (1967) Salanci (1965) Otto & Strunz (1968) Salanci & Moh (1969) Klominsky (1971) Chang & Bever (1973)

680-730

PbSBi4S11

indicate

References

14 kbar Buzek & Prabhala (1965) Vogel (1968) Bell et al (1970) 1.
400·

Walia & Chang (1973)

>400

alloclasite COO.7SFeO.25AsS (extensive s.s. between CoAsS and FeAsS)

Klemm (1962) Klemm (1965) Gammon (1966) Kingston (1970)

As-CuFe-S

Gustavson (1963) McKinstry (1963)

As-CuPb-S

PbCuAsS3

seligmannite

Wernick & Geller (195S)

As-CuSb-S

(complete s.s. between tetrahedrite and tennantite)

Feiss (1974) Skinner (1960) Barton & Skinner (1967) Sakharova (1966)

(extensive s.s; between enargite and famatinite)

Radtke et al {1974a}

As-HgTl-S As-PbSb-S

>400

madocite

(variable Pb/(As,Sb) and AsISb composition) 2PbS'(Sb,As)2S3

veenite

>400

(complete s.s. with dufrenoysite, Pb2As2S5) PbS' (As.5Sb.5)2S3

guettardite

l6PbS·9(Sb,As)2S3

playfairite

l2PbS'(Sb,As)2S3

sterryite

17PbS·ll(Sb,As}2S3

sorbyite

PbS· (Sb,As)2S3

twinnite

27PbS·7(As.45Sb.55)2S3

geochronite CS-16

>400'

Walia & Chang (1973) Jambor (1967,6S) Roland (196S) Burkart-Baumann et al

(1966)

all formulas questionable

(complete s.s. with jordanite)

Bi-Cu- CuS.12Bill.54FeO.29S22 Fe-S

Bi-CuPb-S

BiCuPbS3

Bi-CuSb-S

(extensive

hodrushite

aikinite

s.s. between

CU3BiS3

525

Sugaki & Shimall (1970) Onteov (1964) Sugaki et al (197Z) Kodera et al (1970)

540

Springer

- CU3SbS3

and CuSbSZ - CUBiS ) Z Chen & Chang

Ca-FeMg-S

Ca-Fe-

Mn-S Ca-MgMn-S

Cd-MnZn-S

Cd-PbZn-S

Co-FeNi-S

(extensive s.s. among mono and di-sulfides) (complete s.s. between (Fe,Ni)9SS and C09SS)

Co-NiFe-S

(Co,Ni)SbS . (cons~derable

Cu-FeGe-S

CU2FeGeS4

(1971)

Skinner

& Luce (1971)

Skinner

& Luce

Skinner

& Luce (1971)

Kroger

(1939)

Bethke

& Barton

(1971)

(1971)

Knop & Ibrahim (1961) Springer & Schachnerkorn (1964) Demirsoy (1969) Nickel (1970) Klemm (1965)

willyamite >550 name Cabri et al (1970) s.s. toward CoSbS and NiSbS) retained h C N')Bayliss (1969) were 0> 1

briartite

"'640 "'990

;u-FeNi-S

:u-FePb-S

(1971)

inverts a-form

to Francotte et al (1965) Bente (1974)

Craig & Kullerud (1969) Kushima & Asano (1953

betektinite

CS-17

Schuller & Wohlmann (1955) Craig & Kullerud (1967a,b)

Cu-FeSn-S

Il-Cu2FeSnS4 a-Cu2FeSnS4

stannite

680

--

878

Inverts to Springer (1972) Bernhardt (1972) a-form Bente (1974)

mawsonite CU7Fe2SnS10 (considerable s.s. of CU2FeSnS4 in CuFeS2 above 462·C) Toulmin (1960) Fuji (1970) Buerger (1934) Jankovic (1953) Wiggins (1974)

Cu-FeZn-S

(limited s.s. between CuFeS2 and ZnS)

Cu-PbSb-S

CuPbSbS3

bournonite

>530

Harada et al (1970)

CuPb13Sb7S24

meneghinite

>615

Fredricksson & Anderson (1964)

Cu-SnZn-S

CU2SnZnS4

kesterite

1002

Springer (1972) Roy-Choudhury (1974)

Fe-MgMn-S

Skinner & Luce (1971)

Fe-NiZn-S Fe-PbZn-S

Scott et al (1974) Scott et al (1972) Czamanske & Goff (1973) Aveti6yan & Gnatyshenko (1956)

Ir-OsRu-S

Ying-chen & Yu-jen (1973)

Mn-PbZn-S

Bethke & Barton (1971)

Pb-SbSn-S

Pb5Sn3Sb2Sl4

franckeite

>500

Pb3Sn4Sb2Sl4(?)

cylindrite

>500

Pb.6Sn.23Sb.34Sl.57

Sachdev & Chang (in prep) Fe necessary?

617

III

Bethke & Barton (1971)

Pb-SeZn-S As-CoFe-Ni-S

(extensive

Ca-FeMg-Mn-S

(extensive s.s. series)

B.S.

among FeAsS-CoAsS-NiAsS)

Nickel (1970) Klellllll (1965)

Skinner & Luce (1971)

Bente (1974)

Cu-FeGe-Sn-S CS-18

Cu-FeSn-Zn-S

(complete s.s. of cu2FeSnS4 and Cu2ZnSnS4 >680·C) Springer (1972) (complete s.s. of a-Cu2FeSnS4 and ZnS above "'860·C) Lee (1972) Bernhardt et a1 (1972) Petruk (1973) Harris & Owens (1972)

Ag-As-Cu- (considerable s.s. of Cu and Ag, Fe and Zn, and Fe-Sb-S As and Sb in the tetrahedrite-freibergite series)

CS-19

Riley (1974)

MAJOR SOURCES

A thorough knowledge

understanding

of the behavior

of their thermochemical

of these parameters temperatures complete

OF THERMOCHEMICAL

varies

to nonexistent

listing

discussions

from highly

Presently

refined

for numerous

to the brief

ternary

sources

sulfide

and quaternary

is beyond

requires

sulfides

at,high

sulfides.

A

the scope of the present

considerations

systems.

of thermochemical

minerals

the state of our knowledge

for many binary

thermochemical

of some of the specific

of some other major

of the sulfide

parameters.

of all of the data available

work; we are limited

DATA ON SULFIDES

included

The following

in our

is a short list

data on sulfides:

Adami, L.H., and E.G. King (1964) Heats and free energies on formation of sulfides of manganese, iron, zinc, and cadmium. u.s. Bu~eau of Mines, Report of Investigations,

6495.

As takhov , K.V. (1970) Thermodynamic Publishing House, Moscow. Barton,

P.B.,Jr.,

and B.J. Skinner

Geochemistry

of Hydrothermal

and Winston,

pp. 236-333.

Bulletin

of Thermodynamics

Freeman,

R.D.

Foundation

(1962) Rept.

and

and

Inorganic

Constants.

Science

(1967) Sulfide mineral stabilities. In Ore Deposits, H.L. Barnes, ed. Holt, Rinehart,

Thermochemistry

Thermodynamic properties of binary sulfides. 60, Oklahoma State University, Stillwater.

Garrels, R.M. and C.L. Christ (1965) Harper and Row, New York. Karapet'yants,

Thermochemical

Solutions,

Minerals,

Research

and Equilibria.

M.K. and M.L. Karapet'yants (1970) Thermodynamic Constants of and Organic Compounds. Humphrey Science Publishing, Ann Arbor.

Kelley, K.K., (1960) Contributions to the data on theoretical metallurgy XIII. High-temperature heat-content, heat-capacity, and entropy data for the elements and inorganic compounds. u.s. Bureau of Mines Bull. 584. Kubaschewski, 0., E.L. Evans, and C.B. Alcock 4th ed. Pergamon, New York. Mills,

K.C.

(1974)

Reed, T.B. Charts

Thermodynamic

Butterworth,

Tellurides.

Data

(1967)

for Inorganic

Metallurgical

Sulphides,

Thermochemistry,

Selenides,

and

London.

(1971)

Free Energy of Formation of Binary Compounds: An Atlas of for High-temperature Chemical Calculations. Mass. Inst. Tech.,

Cambridge

Press.

Richardson, F.D. and J.H.E. Jeffes (1952) The thermodynamics of substances interest in iron and steel making, III -- Sulphides. Jour. Iron Steel (London), l71, 165-175.

of Inst.

Robie, R.A., and D.R. Waldbaum (1968) Thermodynamic properties of minerals and related substances at 298.15°K (25·C) and one atmosphere (1.013 bars) pressure and at higher temperatures: U.S. Geol. Survey Bull. 1259. Timmermans, J. (1959) New York.

Physico-Chemical

Wagman, D.D., et al (1968, 1969, 1971) properties: NBS Tech.Notes 270-3,

Constants

Selected 270-4,

CS-20

of Binary

values

270-5,

Systems.

of chemical

270-6.

Interscience, thermodynamic

THE Fe-S SYSTEM

Besides pyrrhotite,

containing

two of the most common sulfide minerals,

the Fe-S system is a corn~rstone

and thermochemistry Fe-Ni-S

(Scott)

of many other important

and Fe-As-S.

unresolved,

to the understanding systems

The many complexities

make it a good place

including

pyrite

and

of phase relations

Zn-Fe-S,

Cu-Fe-S,

of the system, some of which

to begin our systematic

are yet

review of sulfide

phase

relations. The basic phase diagram high temperature Fig. CS-3. whereas

above 400·C is shown in Fig. CS-I.

Note that 'the compositions

of the phases

thermodynamics solid phases

Fe.

2 in the condensed

system

on phase relations.

superlattice

are presented

subcell with a NiAs

and

of

The properties

~ and ~ of pyrrhotites the dimensions

of the simple

(l~) structure.

1539°

1500 Liquid 1400

,,------ ...

,,'

1300

.. c:.. ..e.. ..'"

+ /,""

Fe1_,S

V of 2 cc will have its equilibrium shifted by 2XI500XO.0239=71.7 cal. Volume changes for reactions between condensed phases are usually small but may range up to 2 or 3 cc/g atom, amounting to changes in the free energy of reaction of several tens of cal/kbar. Most 1 Some criteria for doing this are suggested by Barton, Bethke, and Toulmin (1963).

JR.

ore deposits that were formed at depths greater than 5 or .10 km,' and perhaps many that were formed at shallower depths, will likely be cooled so slowly that the initial record will be completely erased, thereby making any initial state calculations of moot value. Therefore, neglecting pressure will not usually result in an uncertainty of greater than about a hundred calories/g atom in the free energy for a reaction of significance of sulfide petrogenesis. This uncertainty should be compared to the 100 to 1000 cal (or more) uncertainty in standard free energies. Of course, if one is dealing with sulfides in the mantle or lower crust, pressure will be much more important. Two further concerns in evaluating the role of pressure are thermal expansion and compressibility. Skinner's (1966) compilation of thermal expansion data for sulfides shows average volume increases of 1 or 2 percent (a few tenths of a cc) from room temperature to 400°C; moreover, thermal expansions on opposite sides of a reaction tend to compensate. Of even less importance is compressibility, for Birch's (1966) compilation shows that a pressure increase of 2 kbar achieves only 0.2 to 1 percent volume decrease.

A summary of measured univariant P-T curves where vapor is not present (Fig. 3) shows very little effect of pressure on the equilibrium temperature. In summary, phase relations of sulfides are relatively insensitive to pressures of the magnitude found in the upper crust, and most phase relations have much more promise as thermometers than barometers' The remaining variables of state are temperature and composition, both of which are very important. Compo. sition can be an awkward variable and we, therefore, find it convenient to discuss the variation in phase assemblages for fixed compositions. Another parameter which is particu, A hydrostatic load gives about a maximum of about 10 km/ kbar; lithostatic gives about 3.5 krn/kbar. 3 For possible exceptions see Scott and Barnes (1969) and Clark

(1960)

~ "

~IO

FIG. 3. Depth-temperature plot for vapor-free univariant equilibria. Depth coordinate also shows pressure in khar. Data are from various sources as cited by Barton and Skinner (1967).

SULFIDE PETROWGY lady useful is the activity of a component common to several phases. In the case of the sulfides we find that the activity of sulfur serves as a unifying variable with which to compare different bulk. compositions. The convenient standard state for sulfur is the ideal diatomic gas,s., at a fugacity of. one atmosphere and at the temperature of consideration. This state is used even though it is physically unattainable (due to the condensation of solid or liquid sulfur) below 614°C (the point at which Ps.= 1 atm). The activity of S., as;, is thus numerically equal to the partial pressure of S. in atmospheres, but bear in mind that the presence or absence of a gas phase is inconsequential. The S. gas standard state is convenient because curves for sul1idation reactions are not required to bend arbitrarily at the melting, transition, and boiling points of sulfur as would be the case if the standard state for· sulfur were chosen, in the conventional manner, as the stable form at one atmosphere at the temperature of interest. Compilations of data in the literature frequently use the latter standard state, so some extra care in calculation is warranted. One of the most useful ways to present sulfidation data is by plotting reactions so as to generate a metallogenetic grid, the coordinates of which are temperature and activity of S•. As will be discussed below, there is a sensibly linear relationship between the free energy change, /1G, and temperature for many sul1ide reactions. Because /1G = -RTInK where T is in degrees Kelvin and K is the equilibrium constant, and because most sulfidation reactions can be written so that all of the reactants and products except for S, are in their standard state, it follows that log as, is a sensibly linear function of liT. (See Barton and Toulmin, (1964) and Barton and Skinner, (1967) for further discussion). Figure 4 shows a series of such sulfidation curves for several metals. Many other such curves are compiled by Barton and Skinner (1967) and by Richardson and J effes (1952). The general tendency for /1G versus T, or log as, versus liT; curves to be linear has been pointed out by many, including Richardson and Jeffes (1948) and Kubaschewski,

FIG. 4. Log aB,-temperaturegrid showingtypical sulfidation reactions.

°

-;J+-,substitute for two ions in the host phase, e.g. 2 Pb>+-,in such a way as to maintain the charge balance. Because minerals are

B-7

PAUL

B. BARTON,

so often compositionally complex at the trace level, the activities of components participating in coupled substitution are so involved in the total array of multiple-charge substitution that quantification is virtually worthless. As an example, what kind of components could usefully be extracted from a galena composition in which the substituuon is of the form (Ag, Cu, TI) (As, Sb, Bi, Ga, In) S·, for 2 PbS'

Factors controlling the activity and activity coefficient. In considering the uptake of a minor component by a growing crystal it is convenient to separate the two factors of the equation,

x

=

oIY

The activity, a, deals with the chemical environment imposed on the growing crystal by its surroundings, most specifically by the fluid phase from which crystallization may be occurring. In order of importance, the activity is a function of the composition, temperature and pressure of the environment. All of these factors are external to anything going on within the crystal. The activity can be locally buffered, as described previously; or it may be controlled remotely, as aNiS might be controlled through the interaction of H,S-bearing fluids with nickel-bearing silicates far removed from the site of deposition. Whether or not a component is buffered locally determines whether it is termed "inert" or "perfectly mobile" following the terminology of Korzhinskii (1959). The activity coefficient, )" is determined solely by the temperature, composition, and pressure of the growing crystal itself. The solvent in a dilute solid solution obeys Raoult's Law in that the activity of the solvent component is equal to its mole fraction. The compositional range over which "dilute"

behavior

is maintained

varies from system

to system, but in general, the greater the degree of solid solution, the greater the r'lnge of "dilute" behavior. The compositional range of Raoult's Law also tends to be much wider for simple substitutional solid solutions, such as (Zn, Fe)S, than for omission type solid solutions such as Fel_l:S

or CUl+xFel+XS2.

The Gibbs-Duhern relationship requires that so long as the solvent obeys Raoult's Law, the activity of the solute is proportional to its mole fraction. However, the proportionality constant (= activity coefficient) is generally not unity. Previous arvuments regarding the minimal role of pressure apply he:'" also and we shall probably be safe in assuming that the .i fluence of pressure is negligibly small. The role of compo .i , on is not so minor, but it appears to be small so long a', coupled substitution is excluded. The effect of temperature on activity coefficients is variable, but not of large magnitude. Summarizing

available

data,

variations

in composition

and temperature can produce effects of up to one order of magnitude, and rarely more, on the activity coefficient. In contrast, the activities of many components may vary by not just one log unit, but by many! For example, Figure 8 shows the variation of aF,S over a range of 6 log units

JR.

while being in equilibrium with either pyrite or pyrrhotite. The series of mineral assemblages along the univariant curves superposed on the diagram show that natural environments do indeed span most of this range. The figure also shows how the composition of sphalerite will vary over the aFos-temperature range covered by the diagram. For components such as SnS, MnS, or In,S, that seldom appear as major constituents of ore minerals the variation in activity may be even greater than that for FeS. Thus the range in variability of activity is drastically greater than that for the activity coefficient, and it is obvious that the concentration of a nonessential constituent in a mineral is influenced far more strongly by a than by )'.It is, therefore, futile to try to use the trace component composition of a single phase (such as the silver content of galena or the mercury content of sphalerite) to try to define some parameter such as the temperature of mineral deposition unless the activity of that component is somehow fixed. It is possible that some geochemical reason might prevail to limit variability in a. For example, there are no feasible geochemical processes for effectively separating Zn from Cd; therefore, the Cd/Zn ratio is relatively uniform in base metal deposits and the ccas in sphaleritedepositing environments rarely varies by more than an order of magnitude (which is still far too large a variation for useful thermometry.) Now let us consider the uptake of a component which is not observed as a separate entity in nature, and in fact, need not even have a stable existence as a pure phase. The entrance of gold into a simple sulfide such as galena might be an example. Based on only the most preliminary sort of experimental data, let us consider the gold content of galena in equilibrium with free gold. Gold might enter as the un-ionized

metal atom in interstitial

positions,

or it

might be present as a gold sulfide component, e.g., Au,S, AuAuS" or Au,S, none of which is known as a compound stable relative to gold plus sulfur. Very preliminary experiments (Barton, unpubl.) show that gold enters galena only in the presence of excess sulfur (the quantitative relationship is still obscure) and that silver decreases and bismuth increases the solubility of gold in galena. Therefore, the gold is probably present, at least in part, as the Au,S component whose solubility is increased by the coupled substitution of AuBiS, (analogous to the enhanced solubility of argentite in galena by the substitution of AgBiS" Van Hook, 1960). The reaction 4Au+S,= 2Au,S must lie in the metastable region beyond the reach of pure sulfur vapor as shown schematically in Figure 9. From the stoichiometry of the reaction we can contour the log as, versus T grid in terms of aAu,S. If the activity coefficient for Au,S in galena were known as a function of temperature we could contour the diagram in terms of gold content of gold-saturated galena. Granted that this part of the discussion is purely schematic, it nevertheless illustrates two points: (I) As temperature changes, the behavior of a component in a saturated solid solution is not simple; it

B-8

SULFIDE

PETROLOGY

may decrease on cooling as would be 'the 'case along the pyrite-l-pyrrhotite curve, or it might increase on cooling as along the sulfur condensation curve. The difference between some roasting and free-milling gold ores might well be the effective sulfur buffer system that functioned during the post-depositional history of the ore. Of course, other gold-bearing solid solutions (such as pyrite or arsenopyrite) may not behave as dges galena, but the principles should be similar. (2) It is entirely possible to work satisfactorily with components which may not be seen as minerals. The copper content of pyrite might be expected to have a similar dependency on as, provided that the copper-rich pyrite lies on the FeS,-CuS, join.' A further extension of the discussion of the behavior of a component in solid solution is that of the distribution of a component between two or more phases as discussed below. If we consider the equations for the same component in two different phases and then divide one expression by the other, i.e., Xl

ad'Yl

Go

~

Go 0

s: Q.

"' S 'J)

'0 :.J

c

., 0

u

E Go

(;

:Iii

')'2

Mole

--=--=-=D X2 ad'Y2 /'1

FIG.

the activity terms cancel out and the distribution coefficient, D, is equal to the inverse ratio of the activity coefficients. The activity coefficients are functions of the temperature, pressure and composition of the host phase, but we have noted already that the role of pressure is minor. Two typical isotherms, (P. M. Bethke and Barton unpubl. data), for the distribution of CdS between sphalerite and galena are shown in Figure 10. So long as we are dealing with dilute solid solutions the deviation of the activity coefficients from a constant value should be trivial, and temperature alone exerts a significant control on the 1

.1

The studies of copper-rich pyrite by Frenzel and Ottemann

(1967), Einaudi (1968), and Shimazaki (1969) demonstrate the

possiblesignificanceof this example,

fraction

galena

10. Two isotherms showing the experimentally determined distribution of CdS between sphalerite and galena.

distribution coefficients. The method appears to provide promising geothermometers, but the difficulty of finding and separating for analysis samples that were deposited in mutual equilibrium presents a serious problem for successful application, as is evident from consideration of the highly complex ore textures shown in Figure I. The distribution of sulfides of monovalent and trivalent metals between coexisting sulfides of divalent metals, for example, Ag,S, TI,S, In,S" or Sb,S. between sphalerite and galena, will be extremely difficult to quantify in such a way as to be useful because they inherently become involved in coupled substitutions. THE

SULFIDATION

STATE OF

NATURAL ENVIRONMENTS

The metallogenic grid of sulfidation reactions shown in Figure 11 covers a large range of sulfur activities, and some

r.mptratur.

incrlCI.1n9 -

FIG. 9. Hypothetical log a",-temperature grid suggesting the wide range of variability

of

GAU2B

in equilibrium with native gold.

B-9

FIG. 11. Log a",-temperature grid showing the region of principal ore-forming environments.

PAUL

B. BARTON,

deposits may have sufficient mineralogical variation to be represented by one-third or more of the total ilR' range. As noted earlier the sulfide-forming environments usually do not buffer themselves on a given sulfidation curve. Instead the sulfides seem to precipitate under arbitrary conditions that may either vary systematically or apparently irregularly, showing that the solid phases being precipitated do not buffer aR" but function as indicators of aR, in the depositing solution, and that the solutions themselves are not of constant composition. Something determines the as, of solutions; if not the local precipitates, then what? The source of ore fluids is responsible for the initial state of the fluids, and, although the specific volume of rock responsible for a given ore fluid cannot often be identified, much less examined, the following observation is pertinent. ;\ either igneous nor metasomatic metamorphic rocks commonly have sulfide assemblages that alone would control aR,; instead the buffer systems appear to be such as: (I)

2FeS+8FeSiO"

(in pyroxenej-j-z Fe.O, = RFe,SiO, (in olivine) +S, (2) 6FeS+8K.-\ISi30, (in feldspar)+RH,O+6Fe"O, = 8KFe3AlSi/\o(( lHJ, (in biotite) +.lS, (3) 4FeC03+5FeS,= 3Fe30.+K+5S,

The "main line" sulfidation state of the most ore deposits tends to run from the pyrrhotite field at high temperature well into the pyrite field at 10\\- temperatures, as suggested in Figure 11. The reason, of course. is the general position of the multiple equilibria of the sort just discussed. There is considerable variation within this trend, and an understanding of the specific reasons for a given pattern is a major goal of current research.

Low SUlfid