QUANTITATIVE DETERMINATION OF HEXAGONAT ANDMONOCTINIC PYRRHOTITES BY X.RAY DIFFRACTION

A. R. GRAHAM

Fal,conbr,id.ge Ni.ckel, M,ines Lirnited, Metal'lnr gi,catr Labor atories,Tkornk'i)|,, Ontari,o

Ansrru.cr

Measurement of the difference in amplitude of components at about 51.750 and51.40"20 (Co Ka radiation) of the asymmetric x-ray diffraction peak resulting frompartial superposition of strong diffractions from corresponding planes in hexagonal andmonoclinic pyrrhotite structures allows rapid quantitative determination of amountsand proportions of these structures in certain sulphide ores carrying at least 3/6 ofeither pyrrhotite type. The differences in amplitudes from synthetic mixtures of thepure structures are divided by the amplitudes of a nearby diffraction peak from aninternal standard added in constant proportion; and the resulting ratios are plottedagainst the known proportions of either structure. The plot may be ch;ecked againstweights of magnetic concentrates of monoclinic pyrrhotites from the ores under study.Determination rates of at least two to three samples per hour with an average error ofabout +L2/6 of the amounts present may be readily attained with suitable instrumen-tation. Economical and relatively precise contouring of sulphide bodies lor the amountsand distributions of hexagonal (paramagnetic) and monoclinic (ferromagnetic) pyrrho.tites is thus feasible.

INrnonucrroN

In late 1962 and early 1963 ,t-ray powder diffraction examinations ofmagnetic separates from test samples of nickel ore from the Strathconamine, Sudbury District, revealed varying proportions of two structuraltypes of pyrrhotite. A structural type with hexagonal symmetry showedmuch less magnetic susceptibility than a type with apparent monoclinicsymmetry.

Determinations of proportions of pyrrhotites by magnetic separationin stages seemed feasible. The grain sizes of the two structural types werecoarse enough to allow fairly complete separation after grinding to about80% - 200 mesh. However, Lhe procedure was laborious and also in-accurate because pyrrhotite recoveries calculated from chemical analysesof products contained errors due to incomplete separations and variableconcentrations of four iron-bearing sulphides and several iron-bearingoxide and silicate minerals.

In February of 1964, design of an x-ray diffraction method for quanti-tative determination of amounts and proportions of hexagonal andmonoclinic pyrrhotites was begun, and a satisfactory routine methodwas in use by June of 1964. By December of 1964, more than two

DETERMINATION OF PYRREOTITES 5

thousand determinations on 1200 samples had been made, and datawere available for interpretation (cowan, 1g68; Naldrett and Kullerud,1967). The methods employed are described below.

Queurrr-lrrvr X-Rav PowoBn Drr.rnacrron

Determination of crystalline phase proportions in mixtures Dy tt-raypowder diffraction has been practised for many years, but only a fewroutine industrial applications have been reported. Those unfamiliarwith the principles and practice are referred to standard texts such asKlug and Alexander (1954). Details of practical methods vary con-siderably with the kind of material and desired accuracy, and we do notpropose to review the literature on this. However, a good example ofsuccessful application is by Black (1958) on the polymorphs of alu-minium hydroxide (and the common mineral impurities) in bauxites.Evaluation of millions of tons of aluminium ore has been aided by thismethod. This notable success inspired many workers over the next fewyears to test and use quantitative x-ray methods for a variety of phasedeterminations in rocks, ores, and industrial products. Few have been assuccessful for a variety of reasons, mostly connected with the difficulty ofprecisely repeating preparation of samples of materials that are not insuch a randomly-oriented and a uniformly cryptocrystalline state asin bauxite.

Quantitative tc-ray diffraction depends on the precise measurement ofthe relative intensities of selected x-ray diffraction spectra produced byparticular crystalline phases present in mixtures. These intensities, withsuitable corrections for matrix and other effects. are functions of theamounts of the crystalline phases present. The instrument used in thepresent study was a Norelco x-ray diffractometer, purchased in 19b5,fitted with a gas-filled proportional counter feeding a ratemeter andBrown recorder through a modified 1g58 model electronic circuit panel.

Difficulties in standardization of sample preparation, presentationmethods and a variety of interferences from the complex powder diffrac-tion patterns of low symmetry crystalline phases are the major obstaclesto precision which have to be overcome in establishing quantitativemethods for determination of mixtures of several phases. Anisotropy ofgrain dimensions, cleavage, differential hardness and malleability ofphases may result in undergrind, yielding irregular diffraction responses;or overgrind, yielding diffraction broadening through structural deforma-tions which reduce effective crystallite-size below b00 A. Fibrous andplaty grains result in preferred orientation on the pressed surfaces of drysample mounts. Moisture adsorption by fine mineral powders after

6 'TEE CANADIAN MINERALOGIST

grinding may lead to complex changes of phase, porosity, and density at

ambient laboratory temperatures. All these factors affect reproducibility'

and tend to discourage the investigator with ambitions to develop

quantitative methods by r-ray powder diffraction. However, considerable

progress in increasing the accuracy of this kind of analysis has been

made in the past so that now the method can be used with some con-

fidence if the net effect of a number of variables is minimized.

Most workers control many of the variables through the internal

standard method. The physical characteristics of the internal standard,

such as grain and crystallite size, and its crystal structure should be such

that one of its major diffraction spectra should be of similar shape

characteristics, and adjacent to, a major diffraction spectrum of the

unknown. The amount of standard added should be adjusted to yield

approximately the same. intensities for the two diffraction spectra

measured. Silicon metal, metallic nickel powder, fluorite, and quartz

have been used successfully as internal standards for quantitative

analyses of pyrrhotite, pyrite, and other sulphides. Suitable fluorite was

at hand. and therefore was used in the technique described below.

Sample preparation included mixing thoroughly the internal standard

with the sample after grinding separately to yield a powder 100/6 less

than 40 microns in size, and at least 8O/6 between 1/10 micron and

15 microns. The mixture was presented for r-ray diffraction in such a

way that its irradiated surface contained a consistently random distribu-

tion of grains. Packing and then scraping with a non-magnetic spatula or

glass slide, to present a flat and relatively random surface in the cylin-

drical sample tray provided for the Norelco rotating holder, has been for

us the most satisfactory practical method. Other methods have been

found to provide greater intensities, or more consistent randomness, but

for speed, convenience, and adequate reproducibility of results, this

simple mounting method has proved acceptable.The grinding method requires especially careful attention for relatively

soft minerals such as sulphides. "shatterbox" type automatic grinding,

employing heavy slugs in hard alloy grinding chambers, while admirable

for rapid reduction to micron sizes, apparently destroys crystalline

continuity in certain crystal structures, and the resulting line-broadening

invariably reduces resolution of closely spaced diffraction spectra. We

used 80/6 less than 100 mesh samples as discharged from a plate pul-

verizer, and employed routine ore-sample "bucking" procedure to cut out

2 gm portions, which were then ground by hand in a mullite mortar,

with or without acetone or alcohol. Less than two minutes was usually

sufficient to reduce the ore sample to an impalpable powder. The ore

powder packs we|l, and gives sharp diffraction resolutions accompanied

DETERMINATION OF

by quite reproducible intensities. Wesamples, the rotating sample holderL0-257a, with little or no reduction inwith the stationary type.

PYRRHOTITES 7

have also found that with suchimproves reproducibility about

effective resolution, as compared

X-Rav Drn'rn-q,crroN CHARAcrERrsrrcs or. PvnnuorrrBs

Much has been written on the x-ray crystallography and chemistry ofpyrrhotite over the forty-four years since Alsen (1925) proposed aniccolite-type structure and a composition near FeS. H?igg and Sucksdorf(1933) compared pyrrhotite and troilite structures, proposed a defectstructure to account for non-stoichiometry of pyrrhotite, and described asupercell in hexagonal synthetic compounds. Bystrdm (1945) measuredseveral pyrrhotites from Sweden and showed conclusively that somepossessed monoclinic structures and were deficient in iron. Buerger (L947)proposed another superstructure type for two natural pyrrhotites. In1948, I discovered a third superstructure in non-magnetic, euhedral,artificial material (Graham, 1g4g). Grlnvold and Haraldsen (1g52)determined compositions of material showing simple hexagonal struc-tures, proving that the more iron-deficient artificial structures weremonoclinic, and that the latter compared favourably in lattice dimensionswith Bystrdm's natural monoclinic pyrrhotites.

Interpretation of the very complicated detailed crystal chemistry of thepyrrhotites continues in many laboratories. Several articles, notablythose of Andresen and Torbo (1967), Arnold (1966), Corlett (1907),Morimoto and Nakazawa (1968) and Fleet (1g68a and b), should beconsulted in the originals. All these have appeared since the abstract ofthe present work was submitted in March of 1960, pending oral presen-tation of the research as a paper to the September 1966 meeting of theMineralogical Association of Canada. Only the 1g66 article by Arnoldrelates a method for quantitative determination of proportions inmixtures of pyrrhotites, however.*

Mention of two-phase intergrowths of pyrrhotite in the literature,according to Carpenter and Desborough (1964), began with Scholtz(1936), and continued through Kuovo et al,. (L963). These authors

*In the spring of 1965, after completion of the analytical campaign of June toDecember 1964 on strathcona ore samples, Dr. Arnold visited the author's laboratoryand discussed the quantitative determination of pyrrhotites in natural mixtures, bothby microscopy and by r-ray diffraction. Dr. Arnold's quantitative r-ray method,described in his 1966 article, differs from the ratio method described here only in thechoice of ratio to be plotted against proportion of pyrrhotites. It is therefore equallyvalid for determination of relative proportions in mixtures, but for determination ofabsolute proportions in ores, the procedures described in the present article are required.

8 THE CANADIAN MINERALOGIST

described pyrrhotites from Griqualand, Africa, and Finland respectively,

Von Gehlen (1963) described intergrowths in pyrrhotite from Dracut,

Connecticut, and Ramdohr (1960) mentioned them in his textbook.

Other authors among the large number that have discussed these inter-growths are listed in the references. Most refer to the hexagonal phase as

troil,i,te, Until direct determinations of element proportions are available,

and can be related conclusively to structure, this treatment of Strathcona

material will refer to hexagonal,, lcw-suscept'ibil'i'ty pyrrkot'i,te, and mono-

cl,inic, h'igh-suscepti,bi,l'ity pyrrhot'i,te as the main iron sulphide phases

present.Pyrrhotites from tJre Falconbridge and Hardy mines, Sudbury area,

were shown by P. G. Thornhill in an unpublished report early in 1952 toyield *-ray diffraction doublets with equally intense components, where

single peaks would be expected from hexagonal structures. Doublets

from Falconbridge pyrrhotite were also observed by me in 1955, and

were attributed to monoclinicity. No evidence of mixed pyrrhotites had

been detected in routine tc-ray diffraction examinations of samples from

other mines of our Company, and few, to our knowledge, had been

reported from the Sudbury District until the Strathcona samples were

submitted for study as related above (Hawley, L962).Optical detection of pyrrhotite phase intergrowths in freshly-polished

surfaces is often difficult. Even after staining, etching, or the application

of a magnetic colloid, quantitative measurement of proportions by

point-count or other optical method is tedious and unreliable. The main

reasons are the extreme irregularity of certain of the finer intergrowths

and differential effects of orientation of grains upon the surface reactivities

of the same structural types. X-ray diffraction was therefore the sole

remaining physical technique known to provide unequivocal identifica-

tion, and which could possibly be adapted for usefully quantitative

measurements. Calibration would depend on partial magnetic separations,chemical analyses, and computations of modal compositions.

In developing the r-ray method, a number of assumptions concerning

the compound nature of the diffraction were adopted. The first and most

important is that only two types of pyrrhotite are present in important

quantities in the Strathcona deposit. So far as is presently known, this

assumption is justified by the subsequent results, but is by no means

established as fact.The second important assumption involves the relative intensities and

resolution of the major composite diffractions from atomic planes in the

two closely similar crystal structures. With monochromatic cobalt Ka

radiation, self-filtered by the iron in the sample, strong multiple diffrac-

tions appeared from samples of mixed pyrrhotites. The hexagonal

DETERMINATION OF P)TRREOTITES 9

pyrrhotites yielded single peaks at about bI.95"20 (d : 2.066A;, whilemonoclinic pyrrhotites separated from the same samples gave doublets,measured at about 5L.75"20 (d: Z.IIEA) and bL.4A"20 (d:2.064A).so far as could be determined with the materials and equipment at hand,the hexagonal peak was about twice the height of one peak of thedoublet, for the same weight percent of the two concentrates separatedfrom the same sample by magnetic methods.

It was therefore assumed that the slight distortion of atomic positionsfrom the strictly hexagonal lattice array, resulting in monoclinic sym-metry, had not changed greatly the total intensity diffracted fromanalogous planes in the hexagonal structures. The six equally-spacedplanes of the hexagonal pyramid, responsible for the strong singlediffraction from the hexagonal pyrrhotite, correspond in the monoclinictype to the planes of two monoclinic pyramids (hkt) and (kk\), and twomonoclinic orthopinacoids (h01,) and (h}i), with multiplicities of twoand one respectively. Each peak of the monoclinic doublet is thuscomposed of two diffractions, each from an orthopinacoid paired with apyramid of nearly equivalent spacing.*

The total intensity of the doublet is approximately equivalent to thatof the single hexagonal diffraction, as observed above. suitable com-promises in instrumental settings can minimize the effect of lack ofcoincidence of the monoclinic diffractions, and resolution of the mono-clinic peaks into a quadruple peak can be avoided.

Adjustment of diffractometer scan-rate and slit-system, and carefulattention to degree of grind thus yields an asymmetric doublet frommixed pyrrhotites, the degree of asymmetry of which is a measure of theamount of hexagonal pyrrhotite present, relative to the amount ofmonoclinic pyrrhotite, and the total area under the compound is pro-portional to the sum of the concentrations of both pyrrhotites.

since the diffractometer properly set yields Gaussian-type curves ardiffraction positions, the amplitudes of the difiraction peaks can be usedas direct measures of proportions of the responsible lattice-spacings in thesample. Regular Gaussian curves partially superposed can be completedsubjectively, and the amplitudes of the components can be measured

*From this argument' it is seen that the distortion from hexagonal to monoclinicsymmetry may take place by slight parallel displacement of originally strictly hexagonalarrays_of. lattice parnts (0001) occupied by iron atoms, in the diiection tlool oT ttremonoclinic lattice. This is an effect which might be expected if enough iron deficiencyallows crimping of the structure through development of a higher digree of covalentbonding. Increased bond strength may also explain the observedlower surface reactivityof monoclinic pyrrhotite compared to that of the hexagonal variety. This lower reactivityto etch reageltg (esPecially to acid pbtassium dichromate) is inexplicably the reverseof that recorded by Arnold (1966), although checked by adsorption of magnetic colloidto the more magnetic monoclinic pyrrhotite.

IIH.t.

I

ANADIAN

TrH"*.

+II

STnono.

II

_I_

10 THE C MINERALOGIST

Frc. 1. Examples of interpretation of difiraction profiles traced from actual charts.

20 of CoKa increasing from right to left.

-t x, +t +/oWt. % of Hexagonal Pyrhotite

Frc. 2. Plot of the ratio of the differences between peak amplitudes of the pyrrhotite

doublet divided by the peak amplitude from internal standard CaFz, as a function of

the unlnown tota[Vo oliexagonal pyrrhotite in the test samples (r*5, r*10' r*15' "')'See Table 1, last column, for the data.

IIHu*,

+I

klrono

L

DETERMINATION OF PYRRHOTITES 11

rapidly. Thus the difference in amplitude is proportional to the amountof hexagonal phase, the amplitude of the lower peak is proportional tohalf of the amount of monoclinic phase, and the sum of the amplitudesof both peaks is proportional to the total pyrrhotite present. An exampleof the procedure for measurement is shown in Fig. 1.

QuaNrrrerrvE PRocEDURES

Calibration plots (Figs. 2,3, 4,5, 8) were estimated by eye and drawnfrom data points of synthetic mixes of relatively pure non-magnetic andmagnetic pyrrhotites prepared by Franz Isodynamic separation of oresamples. The methods and procedures used for these plots are describedin detail in the Appendix. The plots were checked by running unknowns

wl % ofHexagonalPVrrhotite

,,

0.6 d75rrr.n- r rr.r,

/ .r .0

t.@ t.ztr

Frc. 3. Calibration plot of weight percentage hexagonal pyrrhotite as a function of the

observed amplitude .utio @/aa. o

L2 THE CANADIAN MINERALOGIST

Wt. % ofMonoclinicPyrrhotite

Flc.4. Calibration plot of weight percentage monoclinic pyrrhotite as a function ofthe observed amplitude ratio I a,ts/ Iaa.o. Open circles are points calculated by subtractingthe weight percentage of hexagonal pyrrhotite (Fig. 3) from the total pyrrhotite (Fig, 8).Crosses are points located by using sulphur assays to calculate total pyrrhotite, thensubtracting the percentage of hexagonal pyrrhotite obtained using Fig. 3 to yieldpercentages of monoclinic pyrrhotite by difference.

both by r-ray diffraction and by magnetic separation methods, andcomparing the results (Figs. 6, 7). Various forms of these plots were useddepending on the relative precision required; the internal standardmethod given in the accompanying appendix is of course the mostreliable. Triplicate determinations were used to attain an average errorof about +LA/o of tJre amount present in important groups of test

o

+

0,7r T @

DETERMINATION OF P)TRRHOTITES

Proportion of Hexagonal Pynhotite

Frc. 5. Calibration plot of ratio of amplitudes f a.r/Is.ts as a function of weight ratioof hexagonal pyrrhotite in synthetic mixtures of hexagonal and monoclinic pyrrhotites,

samples. Table 1 shows as an example the effect of additions of a samplecontaining 8L/shexagonal pyrrhotite to a sample containing an unknownminor proportion (r weight per cent) of hexagonal pyrrhotite, to establishthe calibration plot for hexagonal pyrrhotite.

The most rapid procedure used assay information already derived fromcomposite drill core samples, each representing an average of several feetof pyrrhotite-pentlandite-chalcopyrite ore. The assay rejects were simplyground by hand for a minute or two, mounted, and scanned on thediffractometer over 20 angles from 51.0o to 52.5" at the rate of L/2o perminute. The ratio of the amplitudes of the individual diffractions in thedoublet gave a ratio of hexagonal to monoclinic pyrrhotite from theappropriate calibration (Fig. 5). The sulphur equivalents of the nickeland copper assays were subtracted from total sulphur in the same sampleto yield sulphur associated with iron in pyrrhotite (both pentlandite andchalcopyrite have about 1/3 sulphur by weight and a nickel or copper toiron to sulphur ratio of about L:1:1). The effects of low proportions(<2%) of pyrite were neglected. Application of the hexagonal mono-clinic ratio to the assay-computed total pyrrhotite gave the absoluteproportion of hexagonal pyrrhotite appropriate to that sample, with anaverage error of about L5/6 of. the amount present. The percent hexagonalpyrrhotite was in turn weighted by footage for computation of total

13

8060

l4 THE CANADIAN MINERALOGIST

go

\Nt. %Hieh ,rA

SusceptibilityPvrrhotite

(Davis Tube)

20

Wt. % of Monoclinic Pyrrhotite (X-Ray)

Frc, 6. Correlation plot of weight percentage of high susceptibility (monoclinic)pyrrhotite magnetically separated at low flux densities from ore samples by Davis tubemanipulation, as a function of percentage monoclinic pyrrhotite determined usingFig. 4. Data are from Table 2, columns 5 and 6.

hexagonal pyrrhotite represented by the samples. Tables 2 and 3 givecomparisons of some of the results selected at random from some of theearly data which had been checked by magnetic separations and calcula-tions from assays.

With the above procedure, and two operators, the maximum speed ofanalysis of twelve samples per hour could be attained for short periods,and an average speed of five to seven samples per hour could be main-tained over periods of weeks. Sensitivity estimated from the resultsindicated that a usable result for proportion of hexagonal pyrrhotitecould be determined to one significant figure in samples containing aslittle as five percent total pyrrhotite.

g

60tot O,0too

DETERMINATION OF PYRREOTITES l5

wt. %Low

SusceptibilityPyrhotite

(Davis Tube)

20Wt. % of Hexagonal Pynhotite (I-Ray)

FIc. 7. Correlation plot of weight percentage of low susceptibility (hexagonal)pyrrhotite obtained as Davis tube separates, as a function of percentage hexagonalpyrrhotite determined using Fig, 3. Data are from columns 2 and B, Table 2.

Concr,uotnc REMARKS

The present application of r-ray powder diffraction technique wasassisted in several ways by much background data in the scientificliterature on methods and upon the specific properties of the minerals tobe determined. It can be used therefore as an example of the utility ofbasic research on mineralogical properties and investigative techniques,in the assembly of information eventually vital to efficient mineralextraction and processing.

The procedures outlined for quantitative pyrrhotite phase determina-tions have allowed the contouring of an important ore-body for con-centrations of high- and low-susceptibility iron sulphides (which differchemically by a maximum of only four atomic per cent in elementalcomposition), in the presence of at least four other iron-bearing phases,two of which are sulphides. The variability of the amounts and distribu-tions of the pyrrhotite types, both locally and on a large scale, whencombined with exact elemental distributions determined by microprobe,will probably contribute to our knowledge of the history of ore deposition

6040 5030t0o

L6 TEE CANADIAN MINERALOGIST

wt. %Total

Pyrrhotite(Davis Tube)

Fro. 8. Calibration plot of weight percentage total pyrrhotite in ore samples separatedmagnetically at low and high flux densities by Davis tube as a function of the peak

f . . , - J - f . .

amplitude rutio'3fft Data are from Table 3.

in the area. We have made no attempt to interpret, from the physicalchemistry of the solid-state reactions responsible for the appearance oftwo pyrrhotites, the reasons for this variability but we hope to do so inthe future.

Proper application of quantitative r-ray powder diffraction analysis,a relatively neglected tool, can provide useful information on majorphase proportions in average or individual samples of rocks, (Bristol

/"

L7

o

a

oFT

o

Xa

obn

s)a

o o * t o < l\+r (o b.- 6) i

q q q q "(0 t - t- t- co

t - c \ 3 t c o o !61 <r rO cot-

OD.il r!) 0O (f

t- cO t'- co (o

9 9 " ' i : "o 6 1 a t 6 6 a 6

( t @ t - c 0

b\€) 66 o t-o i o . l N

+++++n S S t t t t

DETERMINATION OF PYRREOTITES

* a \,::"^

\ v

. t H :

i a {

* ^4< r

YrO

NQ

Ebej?.-€EE i l

:i

C . ) F

; : 1 \

x 6 l

\* 6 \gtr

::iiEi rO

\s( tc , :coNCo

' i , i cd;6ii N N 6 A t r

bo

r o r o o r o r ocl co .+i co (o6r c\i ro b.- (>

O i 6 A r O 0 O

F&

of . -< f ooO ( O t - $ Oe < ) @ i r o'j '.;'..ic.i c\'

b0c) t- co t.. or l ) (o@roroF - H ( 0 6 6 N6 0 < < r t o r O

6io t - o o oA c O F - ' + r Oo i m ( o o

N o o o oO r C \ c o < r

+.++++s l i l t l t t t

\! ,:. .PA \E;i. i . o l l

5-o

^ g ^H n Q'o -6\

o v n

q x

<,x.:t-

*dr, i *€

G h : <

"' ,nO : . >

,f;no; b"?> , : -

k.,ta- H - ( ,

E 9 \ !<R5

)io

bs<i00

. 9 e' o+

;;{ -i

g i b4. n ' r o

A\:.;B*

t-

\1

. J

zN

Afrl

^j

zp

-t2

z

rltsh :

|]v)

tsaz

tq

{r

18

H

a

FF.l

Fa

F ^

| x FH h

9 E2 lx vo u )z 9Y I>8b \c^ A2 P< H. \ }? cZ )x n< r ix t rr r oH O

bs614

z

.4

4

oi

F]

tr

b!

5P,

)i

!q

t;

bqu)

FE

bsE; dr ; oP E

g

bq(n. B

a<0.! r i

.9

EZa

TEE CANADIAN MINERALOGIST

6ac \N :aRD ' - -= * s R* o0o 6o ro i : t . '

++++ *+++++++ | I I l++

ssRJSdAFgESS$"83R$$3

HSN:sRSRSHSR$* f i5R5e$

oO co N 6l eO O <r r(! D- n cO $ rO c\l cO H $ c\ lO co- i T - r l + t i + t+ t+ l++ l I l + l

FB3==SHFNNRRS*E=NFN=

3NTf l 3NNF:RRNK ' * "RS8 :

i o <r 6\ i oi o ca ro i t'- i 4 6r cJ N 6-q o c o: i - i i i c \ l i

+++? t++ t+++++ l l l l ++

F- i (oroF- <r o - R p $ * 3

^ 6 6 N (oo)@

PK=T'msVSsS€SVV'V 'RN

i c\l 60 $r ro coN @ o o i 6l c6 tr ro c9 \oq q i?

DETERMINATION OF PYRRHOTITES

Tesls 3. DsrsRMrNerroxs or Pvnnsortrr rrq ConB Saupr,pReyrcts (Ixrnnnar, Srexoeno Mrrnoo)

19

(1)wt .70Hex. Po(Fig. 3)

(2)WL %

Mon. Po(Fig. a)

(3)wt .70

Total Po(1) + (2)

(4)wL70

Total Po(Fig.8)

(5)wt .7o

Total Po(Calc.)

325459ryf24484L4665653154367220M642338L23766253153

235164od20M4L4L707L25563510L7{+o66,,4Lo

3365232956

2248605620UT39386970225633101646672L43

o3572232962

2242595120M3937b6

D I

2247328

133850L730I

2339202228

06I5000I

1 1130I123844

130

L23337

u

1968), ores (Petruk 1964), and process materials, more rapidly than mostother means. In contrast to chemical analysis, quantitative physicalanalysis by this method yields information bearing on the relationshipsand chemical bonding of elements in the crystalline compounds ofmixtures' as well as the proportions of both elements and compounds,and thus gives another dimension to our knowledge of these materials.

Acxnowr-BncEMENTS

Development and application of this technique has required thecooperation of many different skills, and therefore the support and adviceof many associates. For their help and their interest I thank members ofstaff and management too numerous to list individually, at the HeadOffice, Toronto, in the Nickel Division, Sudbury Area, and especially atthe Metallurgical Laboratories, Thornhill, of Falconbridge Nickel MinesLimited.

20 TEE CANADIAN MINERALOGIST

AppBrtrx

Ourr-nvB or PnocBouRES FoR Qualtrrrarrvs DsrnnMrNATIoN oFPvnnnorrrp Tvpss sv X-RAy PowoBn Drrpnecrroll

A. PreBarat'i'on of Stand'ard's for Synthetic Mires of Pyrrhot'ite Types(1) Establish presence of two pyrrhotite phases in a large sulphide

sample by qualitative r'tay scan. Prepare and stain polished

section and examine under ore microscope to ensure grain sizewill permit mechanical separation.

(2) Crush and grind carefully, avoiding overgrinding and over-heating (which may cause oxidation), to appropriate grain sizefor free milling of pyrrhotite phases.

(3) Prepare deslimed size fraction of about 100 gm by screening' (-200 *325 mesh gave good results with Strathcona material)'

(4) Clean sample of residual gangue by heavy liquid, flotation or

other means.(5) Care{ully remove any magnetite by low intensity magnetic

separation methods, avoiding pyrrhotite loss.(6) Using residual magnetic flux only on the Ftanz Isodynamic

Separator (no magnet current) and appropriate end and side

slopes, separate high magnetic susceptibility pyrrhotite' Checkfor purity using r-ray diffraction, and repeat if necessary'

(7) Increase magnetic flux in small steps, until all possible high

susceptibility (monoclinic) phase has been removed. check eachproduct by x'ray diffraction.

(S) When the monoclinic (20 : 5t.75") peak has disappeared,separate remaining pyrrhotite from paramagnetic and non-magnetic residual minerals as completely as possible by further

manipulation of the separator.(9) Check both products for purity by reflection microscopy in

polished grain mounts. Assess proportions (if any) of adulterantsremaining by point counting after staining, or magnetic colloid

adsorption, and make adjustments in weights for mixtures

accordingly.

B. Preparati'on of Cal,'ibrat'ions fram Stand'ard, Sampl,es: Ad'd'i'ti'on Meth'od'(1) Select high purity hexagonal pyrrhotite concentrate (>80%

hexagonal and <.5/6 monoclinic) as in A(9) above'(2) To 1.5 gm portions of sample carrying a high proportion

(>50% est.) of monoclinic pyrrhotite and low proportion

G2A% est.) of hexagonal pyrrhotite, make a series of additionsof hexagonal pyrrhotite concentrate. (An example is in Table 1)'

DETERMINATION OF PYRRHOTITES 2L

Compute theoretical additions in terms of weight /s hexagonalpyrrhotite in final mix, assuming a series (x : 5,10, lb, . . .) ofproportions of hexagonal pyrrhotite in the original sample.To each of these mixes, add internal standard as in C(4).Irradiate samples and interpret results as in C(5) to C(7). Makerepeat runs at least five times on newly mounted samples toincrease precision of the mean.Make addition plot as in Fig. 2 and repeat using the computedpercentages of hexagonal pyrrhotite f.or x : b, 10, 15, . . . , andextrapolate each plot to cut the abscissa.Interpolate best fit figure for actual percentage of hexagonalpyrrhotite in original sample (15/s in Fig. 2).Recompute actual percentages of hexagonal pyrrhotite in theseries of mixes.Plot calibration from these figures, as in Fig. B.

C. P r epar at ion, Irr a.d,,iat'i,on and, Inter pr etation(1) Crush and grind ore sample to -907o less than 100 mesh, as in

A(2).(2) Cut representative samples of L.b gm.(3) Grind to impalpable powder by hand under acetone, or (with

care) dry.(4) Mix 1.5 gm thoroughly with internal standard (LO/p by weight

nickel powder, or 2O/s by weight fluorite, or L|/s by weightsilicon powder in terms of final sample weight, all of which yieldappropriate peak heights for nickel ores). Use mechanicalshakers, but do not overheat or overgrind. Internal standardshould have approximately the same grain size distribution asunknown sample, to avoid segregation effects during mixing andmounting, and should not be ground with the unknown sample,as diffdrential grinding may ensue.

(5) Pour excess sample into sample tray mounted on platformscales. Use consistent pressure of about 20 lb. on the piston topack the sample. Carefully scrape excess sample from trayusing edge of a glass slide or a non-magnetic spatula, leaving auniform surface.

(6) Irradiate sample using rotating holder on the r-ray difrtac-tometer. Scan over appropriate angular range at not more thanI/2"/minute, adjusting counting rates and time constants toyield sufficiently smooth curves on the chart for interpretation.Our usual settings on the Norelco unit are: Cobalt Ka radiation

(3)

(6)

(4)(o/

(7)

(8)

(e)

22 THE CANADIAN MINERALOGIST

(unfiltered) at 30 KV and 9 ma, 320 C/S for full-scale deflection,

time-constant 4.(7) Resolve the superimposed diffraction peaks on the chart by eye

and then measure amplitudes (I) above background of the peaks

at 5L.4o and 51.75o and subtract. Measure amplitude of internal

standard peak (either nickel at 52.2", fluorite at 33.0o or silicon

at 56.3o), depending on which standard is in use, and divide into

above result.(8) Read per cent hexagonal pyrrhotite from calibration plot pre-

pared from synthetic mixes (Fig. 3).(9) Add amplitudes of the 51.4o and 5L75" peaks and read total

pyrrhotite from the calibration plot previously prepared from

synthetic mixes (Fig. 8).(1,0) Repeat steps (5) to (9) for increased precision of the mean.

Preparation of Cal,i,brat'i,on curaes from Stand'ard' Sompl'es:Ratio Method,

(1) Prepare mixes of high purity hexagonal and monoclinic con-

centrates to yield suitable range of known percentages.(2) Mount and irradiate as in C(5) and C(6).(3) Complete superimposed diffraction peaks by eye.(4) Divide amplitude of the 51.4o peak by amplitude of the 5L.75"

peak.(5) Plot as hyperbolic function /s Hex Po : /(/n.r / In.tu).

E. Preparat'i,on of Cal,'ibrat'i,on Plots for Total' Pyrrhot'i'te(1) Prepare and irradiate a number of mixes carrying known

amounts of pyrrhotites, adding common gangue or other ore

minerals as in C(1) to C(6).

(2) Computefat.q" * Iil.tr"

, plot against known proportions of pyr-/ur*u.

rhotite in any available samples and derive the best-fit curve.

RrronBrcns

Ar-srn, N. (1925): Rontgenographische Untersuchung der Kristallstrukturen vonMagnetkies, Breithauptite, Pentlandite, Millerite und verwandten Verbindungen.Geol,. Foren. Forh. 47, 19-72.

ANonosrt, A. F. & Tonno, P. (1967): Phase transition in Fe"S (X : 0.90 - 1.00)studied by neutron diffraction. Act. Chern. Scand',21,2U1-2U8.

Anr.ror-o, R. G. (1962): Equilibrium relations between pyrrhotite and pyrite from 325"to 743'C. Econ. Geol. 57, 72-90.

Cor,puaN, R. G. & Fnvrr-uxo, V. C. (1962): Temperature of crystallization ofpyrrhotite and sphalerite from the Highland-Surprise Mine, Coeur D'AleneDistrict, Idaho, Econ, Geol'. 57,IL6&-I174.

D.

DETERMINATION OF PYRRHOTITES

& Rrrcunx, L. E. (1962): Measurement of the metal content of naturallyoccurring, metal-deficient, hexagonal pyrrhotite by an x-ray spacing method.Amer. Mineral,. 47, L962, 105-111.

(1966): Mixtures of hexagonal and monoclinic pyrrhotite and the measurementof the metal content of pyrrhotite by x-ray diffraction. Amer. M.ineral,. 5t, L22I-1227.

BnnNon, R. A. (1962): Tetragonal iron sulfide. Sc.i,ence I37, 669-670.BrntAut, E. F. (1953): Contribution a l'dtude des structures lacunaires: la pyrrhotine.

Acta. Cryst. 6, 557-561.(1956): Structure de FeS stoechiomdtrique. Butrl,. Soc. Franc. M'i,nera|,.79,276-

292.Br,ecr, R. H. (1953): Analysis of bauxite exploration samples, Anal,. Ckem,25,745148.Bntstol-, C. C. (1968): The quantitative determination of minerals in some meta-

morphosed volcanic rocks by r-ray powder diffraction. Can. Jour, Earth Sci. 5,235-242.

BuEncen, M. J. (1947): The cell and symmetry of pyrrhotite. Arner. M'ineral, 32,4II4T4,

Busocr, P. R. (1964): Discussion of "pyrrhotite measurement" by Groves & Ford,Arner. M'ineral. 48, 911-913. Amer. M,ineral,.49, L49I-L455.

Bvs::n0u, A. (1%5): Monoclinic pyrrhotites. Arki,v. Kem'i. M'ineral. Geol,. l9B (8).Canrrnren, R. H. & DrsBonoucr, G. A, (1964): Range in solid solution and structure

of naturally occurring troilite and pyrrhotite. Amer. Mineral. 49, 1350-1365,ClAm, A. H. (1964): Pyrrhotite-sphalerite relations in the Ylojarvi deposit, Southwest

Finland: A Summary, Geol,ogi 10, 145-150.(1965): Preliminary study of the temperatures and confining pressures of

granite emplacement and mineralization, Panasqueira, Portugal. Bull,. Inst. M'in,anil. Metall. 704, 663-67 2.

Conlrrr, MesBr, (1967): Low-iron polymorphs in the pyrrhotite group. Z. Kr.ist. 126(I4), t24-rA.

CowaN, J. C. (1968): Geology of the Strathcona Ore Deposit. C.I.M.M. Bul,l,eti.n,January, 1968, 1-17.

DosaonoucE, G. A, & Cenrowren, R. H. (1965): Phase relations of. pyrrhotite Econ.Geol. 60, 1431-1450.

Eusnrv, E. M. et al. (L951): X-ray investigation of pyrrhotite. Vestn. Leni,ngrail, Uni,v,12, 18, Ser. Geol,. 'i Geograf.3, 68-78.

(1958): A study of the chemical composition of the natural pyrrhotites. Vestn.Leningro.d, Univ. 13, 6, Ser. Geol.'i, Geograf., 1,24-35.

Flrrr, M. E. (1968.4'): The superstructures of two synthetic pyrrhotites. Can. Jour.Earth Sci. 5, 1183-1185.

(19688): On the lattice parameters and superstructures of pyrrhotites. Amer.Mineral. 53, 1846-1855.

Gnaneu, A. R. (1949): Artificial pyrrhotite. Contrib. to Can. Mineral. 10,462-4; Amer.Mineral,. 34, 462464.

Gnfxvor-n, F. & Henar-osEN, H. (1952): On the phase relations of synthetic andnatural pyrrhotites (Fer-"S). Acta. Cheru. Scanil,.6, t452-1469.

Gnoves, D. I. & Fono, R, J. (1963): Note on the measurement of pyrrhotite compositionin the presence of both hexagonal and monoclinic phases. Amer. M'ineral,. 48, 911-913.

& Fono, R. J. (1964): Reply to discussion of pyrrhotite measurement. Amer.Mineral,.49, 1496.

HAcc, G. & Sucrsooru', I. (1933): Die Kristalstruktur von Troilite and Magnetkies.Z. Phys. Chem.B22, M52.

Hewr-rv, J. E. (1962): The Sudbury ores, their mineralogy and origin. Can. Mineral'.7,2, L-207.

HavAsa, K., OtsurA, R. & Manrco, T. (1963): On the magnetic properties of naturalpyrrhotites, Mineral,. Journ. 4, 1, 41-56.

23

24 THE cANADIAN MINERALOGIsT

Kr-uc, H. P. & Ar-rxal,'orn, L. E. (1954): X-ray D.ifraction Procetl,ures. John Wiley andSons, New York.

KullrnuD, G., DoB, B. R., Busrcr, P. R. & TnomsN, P. F. (1963): Heating experi-ments on monoclinic pyrrhotites. Carnegie Inst. Wash. Year Book 62,2IH.

Kuovo, O., Vuoner-Lruau, Y. & Loxc, J. V. P. (1963): A tetragonal iron sulphide.Amer. Mineral. 48, 5Ll-524.

LArueu, D. M. & Jenow, M. G. (1964): Rapid quantitative illite determination inpolycomponent mixtures, Amer. M.ineral. 49, 272-27 6.

LsowcsnNrosA, E. T. (1963): Nature of pyrrhotite in sulfide copper-nickel ores.(Abstract) Chem. Abs.60,6, 1964, 6627(h).

Montuoto, N. & NarazAwA, H. (1968): Pyrrhotites: synthetics having two newsuperstructures, Sc'ience l6L, 577 -579.

Nar-onnrr, A. J. & Kunnnuo, G. (196:t-F7): Investigations of the nickel-copperores and adjacent rocks of the Sudbury District, Ontario. Ann, Repts., D'i,rector,Geopkys. Lab. Carnegie Inst. Wash. D.C.

(1967): A study of the Strathcona Mine and its bearing on the origin of thenickel-copper ores of the Sudbury District, Ontario. Jour. Petrotr.8, 453-531.

NIsrANnN, E, (1964)l Reduction of orientation effects in quantitative x-ray diffractionanalysis of kaolin minerals. Amer. M'i,neral,. 49,705-214.

Oro, K. et al,. (1962): Mossbauer study of hyperfine field, quadrupole interaction, andisomer shift of Fe67 in FeSr.oo, FeSr.oo and FeSr.oz. Jour. Pbys. Soc. Japan 17, I0,1615-1620.

PaEnMAN, G, (1954): Upon the magnetism of pyrrhotite (German), Acta Acail,em,iaeAboens,is 19, 10, 3-7.

PrtnuK, W. (1964): The analysis of rocks and ores by x-ray difrractometer, Can.M.ineral^8, 1, 68-85.

RAMoorn, P. (1960): D'ie Erwwineral.ien und, 'i,hre Verwachsungen Akademie-Verlag,Berlin.

Salortt, C. A. (1964): The possible role of oxygen in the anomalous pyrite-pyrrhotiterelation. Econ. Geol.59, 8, 1588-1595.

Scuottz, D. L. (1936): The magnetic nickeliferous deposits of East Griqualand andPondoland. Geol. Soc. Soulh Afr'i,ca Trans. 39,81-210.

Sen Gurra, P. R. (1965): Pyrrhotite geothermometry and its application to the sulfideores of the Mosaboni Mines, Singhbhum, Bihar, India. Econ. Geol.6A, 1, 175-180.

Totnunr, P. & Benton, P. B. (1964): A tiermodynamic study of pyrite and pyrrhotite.Geoclrim. et Cosrnocfuim. Acta 28, 64L, 67I.

voN GEEn EN, K. (1963) : Pyrrhotite phase relations at low temperatures. Carneg,i,e Inst.Wash,, Vear Booh 62,2L3-2I4.

Wurrscu, B. J. (1963): On the superstructure and twinning of pyrrhotite. Mineral,ogi,cal,Soc,iety of Amer,iaa Special Paper No. l, Proc. Int. Mineral,. Assoc. SrilGen, Maeting(1962), 157-163.

Manuscri,pt rece'h)ed. December 18, 1968, emand,ed, Apri,l, 16, 1969