NGU-BU LL 436, 200 0 - PAGE 189

Mineral characterisation of industrial mineral deposits

at the Geological Survey of Norway: a short introduction

NIGEL J. COOK

Cook, N.J., 2000: Mineral characte risation of indust rial minerals deposit s at th e Geologic al Survey of Norway: a short
int roduct ion. Norg es geologiske und ersekelse Bu lle tin, 436, 189-192.

Mineral characte risat ion involves t he st udy of minerals in ter ms of th eir size, habit, chemical compositi on,
morphology, textural po sit ion, associat ion and ot her att ribut es. There is an estab lished and increasingly imp ortant
need for th is type of st udy applied to deposit s of indust rial minerals. This is du e to increasing specification
requ irements for raw materials and min eral produ ct s to compete in th e market place and th e realisat ion t hat mineral
character isat ion can ensure op t imisat ion of mineral processing, t hus maximising prof it. Development s in micro-
analyt ical technolog y in recent years have made it possib le to characterise minerals in an increasingly
comp rehensive manner. This paper b riefly summarises some of th e metho ds used by t he Geological Survey of
Norway in routine investigation s carried out in partn ership w ith th e indust rial minerals indu stry. Charact erisat ion of
minerals has become an increasingly im port ant part of Survey act ivit ies, simultaneous w it h a relative decline in field-
based st udies. X-ray diffract ion analysis (XRDl, image analysis, quant itative elect ron probe microanalysis (EPMA) and
scanning electron micro scopy (SEM) are becomin g routi ne methods and supplement t radit io nal mi croscop ic st ud ies
and geochemical analysis. Ot her innovative micro-beam t echni qu es such as laser ablation induct ively coupled
plasma mass spect roscopy (LA-ICP-MS) are likely t o become important in th e futu re, as th e need for microanalysis at
lower detect io n limit s int ensifies. The resource indust ry is becoming increasingly aware of t he importance of th is
type of information and th e value of mineral characte risat ion data. This tr end is likely to increase in fut ure.

Nigel John Cook, Geolog ica l Survey of Nor wa y, N-749 7 Trondh eim, Norway.

standing of form ation and distribu ti on of indu strial min eral

Introduction
The charact erisatio n of minerals in t erm s of th eir size, habit,
chem ical com po sitio n, morphol ogy, t exture, associat ion
w it h ot her min erals and ot her ph ysical attributes is im por-
tant in st ud ies of min eral deposit s of all typ es. The grow ing
need for d etailed informat io n abo ut th e min eralogi cal co m-
posit io n of a min eral d eposit d etermines th at mineral charac-
terisat ion st udies form an integral and ofte n crit ical part of
investigati on s of deposits.This is t rue bot h for metal oresand
d eposit s of indu stri al min erals and cove rs both deposit s
und er act ive explo ita t io n and prospects being evaluated fo r
future exploi tat io n. As far as in dustrial minera ls are con-
cerned , th e reason s fo r th is are many-fold. Firstly, th ere is a
recog nised increasing need for specif icat ion of raw mate rials
and pr odu ct s t o meet market dem ands, a co nt inued need to
achieve opt im isat io n of min eral processing and mot ivati on
t o achieve maximum econo m ic benefit fro m resour ce explo i-
t ati on. Second ly, rapi d deve lop ments in mi cro-analyt ical
t echn ol ogy have made it po ssib le to characte rise materia ls,
and th erefore id enti fy and pote nt ially add ress and solve
probl ems un th in kabl e a ge nerat io n ago . The Geol ogical Sur-
vey of Norw ay (NGU) is extensive ly inv olv ed in partn ership
projects wi t h an expand ing domest ic indu stri al minera ls
indust ry in Norw ay and th e characte risatio n of min erals is
beco mi ng in creasingly im port ant in our investigation s of
min eral resour ces. This is t rue in cases in w hich th e investiga -
ti on is carried out on a co ntrac t ual basis for indust rial end-
users and also in research proj ects aimed at a deeper und er-
resources in Norway.
Appli cation of research tool s such as X-ray diffracti on
analysis (XRD), im age analysis, qu anti t ativ e elect ro n p robe
microanalysis (EPMA) and scanning electron mi croscopy
(S EM) on a rout ine basis are valuable supp lements t o t radi-
t ion al microscopic st udies and geoch emic al analysis,
alt ho ug h it should be stressed t hat t hese tr adit ion al met hod s
remai n essent ial in any study. Mo re recent ly, ot her in novative
micro-b eam t echniqu es such as laser ablat io n ind ucti vely
coup led p lasma mass spectro scopy (LA-ICP-MS) are p laying
an incr easingly im port ant rol e. Such tec hniqu es permit
detectio n lim it s for tr ace elements in mi nerals not believed
possibl e some years ago, and w it h inc reasing ly good resolu-
ti on, p recision and accuracy. This tr end is likely t o int ensify in
com ing years as new techn iqu es are develop ed and becom e
available com mercially. Also, as indu stry becomes incr eas-
ingly awa re of th e kind of informa tio n th at can be generat ed,
still g reate r em phasis w ill be placed upon min eral character -
isat io n in st udies of indu strial m ineral depo sits.
Th is paper briefly review s some of th e t echniqu es used in
mineral characte risat io n proce dure s current ly used by NGU
and ap plied in st udies in th e indu str ial minerals sect or. The
reader is refer red t o th e references on each techni q ue for
det ail s of pri nciples and app licatio ns. In an effo rt t o meet th e
growing needs of indust ry and th e researcher, NGU is conti n-
uou sly develop ing meth odologi es and upg rading ap plica-
t ion s of routine characte risation t ech niqu es. Mo reover, we
are co nti nui ng to develop our min eral characteri sat ion capa-
biliti es st ill fur th er and to design approa ches for specific end-
 NGU -BULL 43 6, 2000 - PAGE 190
users of the dat a. Several oth er papers in th is volume discuss
th e typ e of data generated and its interpretati on in mor e
detail.
The need fo r mineral
characterisation data
In order to exploit any mineral deposit , it is necessary to pro -
vide comprehensive data on all mine rals present and th eir
respect ive pro po rti ons in th e ore as we ll as in waste and con-
cent rate products, in addition t o th e spatia l distri but ions of
th ose minerals on the scale of the deposit. A comprehensive
mineralogical invest igation should encompass not onl y
th ose mine rals of economic value (e.g., carbon ate, t alc, rut ile,
ilmenite; i.e. the ore it self), but also harmful minerals, whose
presence may negatively impact upon exp loitat ion, proc ess-
ing or saleabil ity of end prod ucts. Furth ermore, t he chemical
compo sit ion of th e minerals, th eir size, morpholog y and
associat ion are all factors in flu encing th e att racti veness of a
depo sit or prod ucts prod uced from it, and therefore it s suc-
cess as a commercial venture. These and oth er factors, such
as local variations in compo sitio n, grain size, mo dal mi neral-
ogy or the mineralogical distr ibut ion of harmful t race ele-
ments, can influence profit ability or can affect th e selection
of processing rout ines t o be employed. In short, a compre-
hensive mineral character isation of a deposit is an indi spen-
sable advant age, which can g uide exp loitation of t he deposit
for maximum profit. The requ ired investmen t for unde rtak-
ing such a study can readily pay fo r itself by allo wi ng for an
opti misat ion of recovery or, for example, by th e early recog-
nition of prob lems of an env ironmental or market ing charac-
ter not realised by chemic al analysis alone.
Methodology of mineral
characterisation
The starti ng point in investi gations of all types of samples is
an adequate characte risation of bu lk chemist ry t o assesscon-
centrat ions of majo r and minor elements present . The analyt -
ical met hods used should be selected on th e basis of estab-
lished practice and suita bility t o the typ e of sample and
desired accuracy and precision and minimum levels of detec-
tio n. In som e rare cases, sub sequent mi neralog ical analysis
may provide data th at makes repetit ion of geoch emical anal-
ysis necessary; for examp le, if min erals are found to be
present w hich are insol ubl e in t he acids used fo r sample
digest ion . Equally important in th e early stages of any min-
eral charact erisatio n is examination of representative sam-
ples using conv entio nal optical microscop y in tr ansmi tte d
light. Reflected ligh t microscopy shou ld be carried out in
cases w here oxides, sulphi des or othe r opaque minerals are
present. Alth ough comm on sense, th e impo rt ance of suffi-
cient optical microscop y at applicable magn ifications , in any
mineral characterisation, is commo nly und erplayed. Muc h
fund amenta l mi neralogic al informati on can be gained at
min im al expense from optical examinatio n of suita ble sam-
ples. Such information may be critica l in subsequent int er-
pretation of micro-analytical data.
NIGEL J. COOK
Image anal ysis
Image management systems open up a wide range of new
possibili ties in the field of mineral characterisation and are
now relatively commonplace in mineralogical laboratories.
The capabilit y to analyse images from SEM or microscope
imag es on a routine, systematised basis has been possible
through th e developm ent of user-friend ly and com pati ble
soft wa re coup led wit h increasingly powerful computers. The
type of inform ati on provided can be confi gur ed to meet th e
needs of th e individu al researcher, but can include details of
grain size and dist rib ut ion of grain size, grain morphology
and association and the relative abundance of minerals in the
investigated sample. Providing that individual minerals have
different appearan ces in the optical image, the ir physical
attribu tes can be qualitat ively or quantitatively identified.
The image can range from an optical transmitted-light micro-
scope im age to a back-scatt ered electron SEM image in
w hich individu al minerals appear at different brightness
int ensities depend ing on th e average atomi c weigh t of com-
ponent element s. False-colour tagging of minerals of inter est
(for examp le, deleter ious amo unts of quartz in carbonat e or
pro d ucts, or minor quant it ies of sulphide mine rals in a mag-
netite concent rat e) allows presentation of t he scale of the
problem and also offers an effective means to commun icate
t he prob lem to interested parties in repor ts. Image analysis is
parti cularly effective when coupled with quantitat ive or
semi-quanti tative mi croanalysis, allowing rapid estimation of
elemental distrib ut ions betw een mineral phases present in a
given sample, and fo r assessment of character istics and di s-
t ribut ion of compo siti onally distinct varieties of t he same
mineral in a sample (e.g., two co-existing white micas in a
concentrate).
X-ra y diffracti on analys is
X-ray diffract ion (XRD) analysis is a mineralog ical identi fica-
tio n method that permits semi- to full-quant itat ive assess-
ment of the minerals present in a given sample and in what
relativ e prop ortions t hey occur. The sample may be a pulver-
ised bu lk sample, a processing produ ct or a pure, separated
mineral fraction. The method is especially suitable for t he
st udy of material conta in ing signifi cant small-scale variation
in mi neralogy (e.g., carbon ates). or in mi neral chemistry of
compo nent minerals or an exceptio nally coarse-grained
sample, inhibit ing the reliability of imag e analysis studies of
thin-sect ions. Relia ble informat ion can be obtained from XRD
analysis of a split of a larger representative sample. XRD is
t hu s a standard met hod and one that is of key importance in
many investigations of industrial minera l deposits and prod-
ucts. Furt hermore, changes in mineral compos it ion are
reflected in the diffraction properties, allowing small changes
in comp osition within, for example , carbonate from different
parts of a given deposit to be identified. Information on crys-
tal st ruct ure can be obta ined if t he necessary software is
available, allowi ng, fo r example, rapid discriminat ion of cal-
cite from aragon ite or iden t ification of clay minerals or zeo-
lites. The XRD method has its advanta ge in being a rapid
t echni que involving only minimal sample preparat ion and is
t herefore a convenient starting point in mineral character isa-
 NIGEL J. COOK
tion procedures. It may be used to compl ement microscopic
study of th in- sections and may be follo wed up by oth er sup-
plementary techniques w here more informat ion is required .
Scanning electron microscopy
Investig ation of samples by scanning elect ron mi croscopy
(SEM) allow s det ailed identification of indi vidual minerals,
eit her in situ w it hin a polish ed thin -section prepar ed from a
rock sample, or a sample mount prepared from concentrate
or oth er proce ssing product. SEM also provides an optical
im age th at can be pro cessed and treated by im age analysis
techniques, permitting characterisat ion of size, morphology,
habit and association . Back-scattered elect ron im aging adds
th e possibility to search for minor or trac e ph ases of inte rest,
whi ch are oth erwi se difficult to recogni se using conventional
micr oscopic procedures, and to map their distribution . Back-
scatte red scanning electron microscopy can also be used to
reveal morphological attributes and mineral fabrics in appre-
ciably mor e detail th an conventional optical micro scop y
(e.g., Kring sley et al. 1998) and is an indispensabl e tool for the
inv estigation of many type s of indu strial mineral deposits.
The applicat ion of cat hode lumin escence is a furth er tool for
th e st udy of microfabric s in st ruct urally complex rocks.
Electron probe microanalysis
Quantitative microanalysis is att ainable by using an energy-
disper sive syst em (EOS) or wavelength-di spersive system
(WOS) as an add-on to an existing SEM or, ideally , by using a
dedicated electron probe and a suitable set of elemental
sta ndards. Such quantitative methods are often the only way
to acquire the necessary data to solve key questions concern-
ing mineralogical di stribution pattern s and to achieve full
mineralogical charact erisation of a given sample. The tech-
niques involved have been described in detail in several ref-
erence texts (Reed 1993, 1995, 1996, Scott & Love 1994). The
current generation of electron probe instrumentation, cou-
pled with user-friendly operating soft ware, has opened up
the possibility to obtain rapid and reliable non-destructive
analyses of mineral phases in situ. Automatic operation, now
a stand ard feature of tod ay's microprobes, allows analysis to
be carried out rapidly and at relatively low cost, enabling a
stat ist ically valid database to be assembled in a short time.
Minerals can be identified and, in cases where the mineral in
question is part of a solid soluti on series, the preci sechemical
composition can be defined with excellent accuracy and pre-
cision . The high spat ial resolution in the range of 1 to 3 IJm
also permits for quantitative measurement of the fine st sam-
ple s, crystal-scale variation s in mineral chemi stry, chemical
mapping of zon ed crystals and identification of small trace
phase inclu sions.
Knowledg e of mineral chemi stry is critical to th e charac-
teri sation of a min eral, whether for qu antification of product
or fo r petrologically oriented research on the genesis of an
indu strial mineral depo sit. The det ection limits offered by
electron probe microanalysis are reasonably low (on th e
ord er of 0.1 wt . %, but depending on op erating conditions
and the min eral being analysed) using norm al count tim es.
NGU -BULL 4 36, 200 0 - PAGE 191
I hi s may, however , be reduced by an order ot magnitud e by
using longer counting t im es, thu s making the method useful
for determination of trace and minor eleme nt concentr a-
t ion s. Robinson and eo-workers (1998) recentl y review ed the
applicat ion of electron probe micr oanalysis t o qu antification
of element s at th e trace element levels. Precision of the elec-
tron microprobe is typically ± 1% w it h an accuracy of about
3% relative, alt houg h a relative accuracy of 1% is readil y
achievable for major eleme nts in silicate min erals.
One of the few disadvant ages of th e microprobe meth od
is th at th e lightest eleme nts (Z<4) cannot be analysed. Fur-
th ermore, analysis of elements w it h Z between 5 and 10 is
not po ssible using many microprobes in service t od ay w it h-
out the additi on of synt hetic multi-l ayer diffr action devices
enabling analysis of some of th ese elements (e.g., 0 ; Nash
1992). Thus, Li and F (bot h con stituents of some micas), B, Be,
H, C. N and 0 are missing in mo st analyses. The H20 content
of minerals is norm ally estim ated as being th e difference
betw een the sum of th e oxid es and 100%, but thi s may be
subje ct to con siderable error. An additio nal probl em is th at
microprobe techniques cannot measure th e oxidat ion-state
of Fe (and other elements such asMn ). This can lead t o uncer -
taint ies in the calculat ion of min eral formul ae, for example in
th e case of mineral ssuch as magnetite. Electron probe micro -
analysis is also dep endent upon the sample material being
mounted and poli shed, making analysis of some typ es of
material s problematic and addin g t o prep aration tim es.
However , such problem sare, in gen eral, not in surmountable.
The electron microprobe allows th e rapid accumulat ion
of a large body of quantitative data in a relatively short time
period . How ever, the manner in w hich the t echnique is
applied will depend on th e typ e and depth of inform ation
required . In simplest t erm s, analysis of repr esent ativ e points
can be used to assess th e range of comp osition s in a given
sample that can then be comp ared to ot her samples. The
elect ron beam may also be enlarged to give an average com-
position across a given area. If the objective is to qu antify
inhomogeneities on the scale of a single crystal or thin -sec-
tion, a scanning technique can be applied, giving a di stribu-
tion map. Altern atively, analyses can be made along a
trav erse across th e area of int erest. If greater det ail is
required, inhomog eneities can be furth er st udied on a point -
by-point basis. Careful optical micro scopy and/ or use of th e
scanning electron micro scope are an essenti al prerequi site
for effi cient use of microprobe tim e.
It should be st ressed that not all elect ron micr op rob e
data obtained and/or published are of acceptabl e qual ity . For
successf ul use of electron probe microanalysis, th e user
should be well informed and tra ined in mineralogy and have
relevant prior knowledg e of th e sample materi al, allow ing
the appropriate type of data t o be obtained and sensibly
interpreted. Successful microprobe analysis also depend s on
an interplay between the user and trained technical person -
nel, whose appreciation of the in strument and inherent limi-
tations is suff icient to select suit able operating condit ions for
the task concerned and thu s reduce th e dang er of producing
poor or misleading data.
 NGU -BULL 436 , 200 0 - PAGE 19 2
Laser ablation inductively-coupled plasma mass
spectrometry
Laser ablatio n indu ctively coupled plasma mass spectrorne -
try (LA-ICP-MS; 'laser ablat ion microprobe analysis' or 'laser
probe analysis') has rapidly evolved dur ing the past decade
as a powerfu l technique that complements, bu t does not
replace t he electron microprobe. Like other relatively new
techniques (prot on and ion microprobes), LA-ICP-MS allo ws
in situ analysis with appreciably lower dete ction lim its (a few
tens of ppm or low er) albeit wi th a reduce d spat ial resolu tion
compared to t he elect ron microp robe . The method is rapid ly
establishing it self asan effi cient method for determinat ion of
trace element concentration s in situ w it hin mineral samples.
Application of th e met hod to industrial mi nerals can include
determinatio n of harmfu l eleme nts in ilmenite, rutile, ta lc or
other minerals, e.g., REE con centr ation s at th e ppm level
within carbonates. Larsen et al. have discussed t he det ermi-
nati on of impuriti es in 'ult ra-pure' qu artz in th is vol ume.
The LA-ICP-MSmethod, described by Perkins et al. (1995),
allows qu antitative analysis of nearly all elements in the peri-
odi c table, including most of th e light elements wit hin all
typ es of mineral s. The met hod also allows determinati on of
th e abundance of indivi d ual isoto pes and has found consid-
erable application among geoscientis ts as a tool for determi-
nation of sta ble isotopes (Wright 1995) and no tably for Ar-Ar
dating of minera ls (Kelley 1995). Matri x effects, although far
from as serious as in ion probe analysis, infer that th e abil it y
to obtain precise and accurat e quantitat ive analysis depends
on careful selection of suitabl e standards wit h comparable if
not exact ly the same matri x. Unlike the electron microprobe,
however, the method is destructive and the volume of sam-
ple analysed is larger; ablation craters have a typical size of
20-80 urn in diameter. Rapid technical develo pment of LA-
ICP-MS methods, as well as an expanding catalogue of suita-
ble mineral standards, is likely in the coming decade and the
met hod will un doub ted ly play an increasingly important role
as a minera l characte risation too l in the fut ure.
Concluding statements
Variou s ot her meth od s are available w hich have application
to min eral deposit s, includ ing stab le and rad iogenic iso-
top es, microanalysis using Micro-PIXE or SIMS method s and
flu id inclusions. Several reviews ofth ese and other micr oana-
Iyt ical metho ds are available (e.g. Cabri & Vaughan 1998,
McKibben et al. 1998). Transmi ssion electron micro scopy
(TEM) is also used extensively asa diagnostic tool in t he study
of mi nerals at t he sub-m icroscopic scale (Buseck 1992), e.g.,
in cases involving mineral polytypism and stacking disorder s.
Application ofTEM is also standard pract ice in the investiga-
tion of asbestos min eralog y.
The number and combination of techniques used in a
given study will vary from case to case and will depend upon
the quest ion s requiring resolut ion. The type of informat ion
obtained can be individually tailored to meet the require -
ments of the end-user. Typical end result s include identi fica-
tion of a part icular mineral or range of minera lsasthe host for
a part icular useful or harmful element, dep iction of com posi-
tional variatio n in a partic ular mi neral across a giv en spat ial
NIGEL J. COOK
area, or a full dete rmination of mi neralogical balances in ore
and th eir correspo nd ing mi neral products. In many cases, an
inte grated approach is necessary, involving several of the
meth ods refer red t o above in order to give the fullest miner-
alogical characterisation possible. For example , electron
prob e microan alysis, LA-ICP-MS and im age analysis can be
comb ined to prov ide informa tion on the minera log ical distri-
bution of harmf ul elements (e.g. Ni) in different types of ore.
In all cases, selection of a st atistically valid number of rep-
resentative samples is a prerequisite for obta ining meaning-
fu l data. Furthe rmor e, since minera l deposits, like other geo-
logical objects, are almost never homogeneous, it may be
necessary to repeat the miner al characterisation procedures
du ring th e lifetime of a mineral deposit, in order to recogn ise
change s in mineralog y w it hin th e deposit as it undergo es
successive exploi tation.
Ackno wledg em ents
Sincere thanks are due to my colle ague s Tor -Arn e Karlsen, Rune Larsen,
Gur li Meye r and Are Korneliu ssen for the ir helpfu l sugges tio ns and on-
st ruct ive com m ents on the manu script.
References
Buseck, P.R.(ed .) 199 2: M iner als and react ions at t he atom ic scale: Trans-
m ission elect ro n m icro scopy. Reviews in Mineralogy 27,508 p p.
Cabri, L.J. & Vaugha n, DJ . (ed s.) 1998: Mo dern Appr oaches to Ore and
Environm ent al Mi neralogy . Mine ralogical Association of Canada
Short Course Notes 27.
Jackson, S.E., Longe rich , H.P., Dunning , G.R. & Fryer, 8.J. 1992: The ap pli -
cat ion of laser-abl at ion m icroprob e - inductively coupled p lasma-
ma ss sp ect rometry to in sit u t race-elem ent de term inations in m iner-
als. Canad ian Mine ralo gist, 30. 1049-1064.
Kell ey. S. 1995: Ar-Ar dat ing by laser m icroprobe. In: M icroprobe Tech-
niq ues in the Ear th Sciences (PJ. Potts et aI., eds.) Chapman & Hall,
Lond on, 327-358.
Krinsley . D.H.• Pye, K., Bogg s. S. & To vey, N.K. 1998: Backscattered scan-
ning electron m icroscopy and image analysis of sediments and sedi-
ment ary rocks. Cambridge University Press, 193 p p.
McKi b ben . M A. Shanks, W.e. III & Rid ley. W.1. (ed s.) 1998: Applications of
m icroa nalyti cal tech niqu es t o understand ing mineral izing p roc-
esses. Review s in Econ omic Geology 7, 263 pp .
Nash, W.P. 1992: Ana lysis of oxygen wi th th e elect ro n m icr op rob e: appl i-
cation s to hydrated glasses and m inerals . American Mineralogist, 77,
45 3-45 7.
Perkin s, W.T. & Pearce. NJ .G. 1995: M ineral m icro analysi s by laserpro be
ind uctively coupled p lasma mas s spec t rome try . ln: Microprobe Tech-
nique s in th e Earth Sciences (P.J. Potts et aI., eds.) Chapman & Hall,
London , 291-3 25.
Reed. S.J.B. 1993: Electron Microp robe An alysis. 2nd edition. Cam b rid ge
University Press, Cambridge, Eng land. 326 p p..
Reed, S.J.B. 1995: Electron probe micro analysis. In: Microprobe Tech-
niqu es in the Earth Sciences (P.J. Potts et al., eds.) Chapman & Hall,
London, 49-89.
Reed, S.J.B. 1996: Electron Microprobe Analys is and Scanning Electron
Microscopy in Geology . Cambrid ge Unive rsity Press. Camb ridge, Eng-
land. 20 1 pp.
Robin son, BW .•Ware, N.G. & Sm ith , D.GW. 1998: Modern elec tron-m icro-
probe tr ace-element analy sis in m ineralog y. In: Modern Approaches
to Ore and En viro nm ental Mineralogy (U . Cabri & DJ . Vaughan eds.).
M in eralogical Asso ciat ion of Canada Short Course Volume 24, 153-
180.
5cott, VD. & Love, G. 1994: Quant ita tive Electron Probe M icroa na lysis, 2nd
edi t ion . Ell is Horw ood , Ch ichester. U.K.
Wright.I.P. 1995: Stab le isot op e ratio measurement using a laser m icro-
p ro be.l n: Microprobe Techniques in the Earth Sciences (PJ. Potts et al.,
eds.) Chapman & Hall, Lond on, 359 -386 .