100 Magne OdegArd GU - BULL 427. 1995

The significance of analytical procedure In geochemicaland environmental studies

MAGNE ODEGARD

Magne OdegArd. Norges geologiske undersoketse, P.O.8oks 3006 -Lade. 7002 Trondheim

IntroductionInformation on the chemical composition of geo­logical materials is often of fundamental impor­tance within many branches of geoscience. In theearly days of the development of mineralogy andpetrology , the classical wet chemical analyticalmethods were the only ones available, and usedmainly for the determ ination of major elementsfor the characterisation and descript ion of mine­rals and rocks. Since then, an intensive researchin analytical inorganic chemistry has led to thedevelopment of rapid and accurate instrumentalanalytical methods with good element coverageand detect ion capabilities. This has been of greatbenefit for research within geology and geoche­mistry, where the interest for chemical data hasbeen strongly expanding, not only with regard toelements and concentration level, but also con­cerning the type of geological materials, whichcan range from solid rocks and sediments towater and air.

The choice of analytical technique , includingsample type and preparation , is in many cases ofvital importance for a meaningful conclusion , andthis choice must therefore be carefully based onthe problem at hand. This is particularly promi­nent within two fields, namely: (1) Geochemicalprospecting. (2) Environmental investigations.

Geochemical prospectingThe application of geochemical principles in thesearch for ore and mineral deposits has long tra­ditions, and techn iques based on such principlesare today valuab le both as independent methodsand as supplements to geophys ical and geologi­cal methods .The simple principle for geochem ical prospect ingis to search for traces of characteristic compo­nents of potent ial ore or mineral deposits, and totrace these findings back to their origin.Presuppos itions are that dispers ion of elementsfrom the depos it has taken place, and that theelements have been taken up or collected bysome medium outside their source. Dispersion of

elements from a source can take place either inthe form of particles or as solutions , but also insome cases in the form of gases, and the mainforms of dispers ion are through transport by ice,water or air.

A variety of media have been studied and usedin geochemical prospecting, e.g. stream water ,stream and lake sediments, mineral soils, rnorai­nic materials, humus and different types of vege­tation. Some of the most frequently used media,however, are different types of sediments.

Environmental investigationsContro l and surveillance of the environment is ofvital importance for all forms of life, and hasbecome an activity with high political priority. Agreat deal of the damage to the environment isrepresented by pollution, which again is a typicalresult of today 's complex, industrialised society.However, pollution of the environment can also,in some cases , take place in nature even withoutany human activity. This phenomeno n is normallycaused by ore or mineral deposits, or otheroccurrences with elevated element concentrati­ons. No matter whether a pollution is caused byhuman activity or by nature itself, the consequen­ces will be the same, and the methods for deter­mining the degree and cause of the pollution willalso be much the same.

Problems common to geochemicalprospecting and environmental inve­stigationsBoth geochemical prospect ing and environmenta lpollution studies have three basic elements incommon , namely: 1. A primary source for thesupply of chemical elements. 2. A mechanism fordispers ion of these elements. 3. A medium forpicking up or collecting the elements .

In contrast to traditional solid rock geochem is­try used in modern petroloqy, the major goal ofsuch investigations will be to determine quant ita-

NGU - BULL 427. 1995

tively and selectiv ely the elements or elementcompounds which have been supplied secondari­ly to the medium from exte rnal sources. Th ismeans that the ideal analytica l techni que wouldbe a method where analyte elemen ts potentiallypresent in the medium beforehand would escapeanalysis. No such method exists. Therefore, com­promise techniques must be chosen.

The way of attacking this problem may havegreat consequences for the final interpretationsand conclus ions and is, in fact, one of the mostimportant parts of investigations of this kind.Good knowledge of geology, geochemistry ,mineralogy and chem istry is a requisite. The idealsituation would be a neutral medium containingnone of the analysed elements, and which, thus,would readily give a correct picture of the take-upof elements. This ideal situation never exists, sin­ce most geological materials contain variou samounts of most chemical elements. This maycomplicate the situation considerably.

There are, in principle, two possible modes ofanalys is: (1) total analysis; (2) partial analysisbased on some form of extraction. This immedia­tely raises many important questions. In the caseof total analysis, what are the absolute and relati­ve amounts of ana lyte elements primarily presentin the medium compared to the amoun ts suppl iedfrom external sources? In the case of partial ana­lysis the same question is signif icant , but twoadditional important questions are: what are theamoun ts of analyte elements brought in solutionby the extraction, and how select ive is the extrac­tion ?

Total analysis: By all methods determ ining thetota l element content it is not possible to dist ingu­ish between analyte elements primarily present inthe medi um and elements supplied to the geolo ­gical system from external sources.

X-ray Fluorescence (XRF) is one of the met­hods used for total analysis, but even though thistechniqu e uses solid samples with pract ically nodilut ion, the detection capab ilities are often insuf­ficient for many geochemical and environmentalstud ies. Many other methods can be used formetall ic elements, e.g. Atom ic AbsorptionSpectrometry (AA), Inductively Coupled PlasmaAtomic Emission Spectrometry (ICP-AES) andInductively Coupled Plasma Mass Spectrometry(ICP-MS). These methods normally apply soluti­ons which, for total ana lysis , are prepared eitherby fusion with LiB02 followed by disso lution indilute acid , or by dissolution in HF in combinationwith other mineral acids , followed by complexingof free HF with H3B03. Both these techn iquesintroduce considerable amounts of salts whichgenerally complicate the subsequent analysis.

Partial analysis: It is evident that the use of

Magne 0degard 101

total analyses can introduce many uncertaintiesand prob lems in geochemical and environmentalstudies, and there are thus many reasons forchoosing an analytical technique based on partialextraction. Two factors will be espec ially impor­tant in relat ion to the extract ion technique, name­ly: 1. The extraction attack must be strongenough to bring the secondarily supplied ele­ments or element compounds to be studied insolution. 2. The acid attack should influence theprimary sample minerals as little as poss ible toavoid introduction of unwanted elements.

Unwanted elements in this context are first ofall potential analyte elements primar ily present inthe geological mater ial, and which at the sametime are completely or partly released during theextraction. Major matrix elements which go intosolution during extraction are also unwanted, sin­ce a high content of dissolved solids in the analy­sis solution always complicates the analys is.

Partial extraction of geologicalmaterialsMost geolog ical materials used as the medium inpollution studies or geochemical prospecting con­sist mainly of rock material rich in silicate mine­rals. The type and amount of silicate mineralscan, however, vary within wide limits, and are lar­gely depen dent on the type of parent rock mater i­al. The rock-forming silicate minera ls as a grouphave traditionally been considere d as largelyresistant against attacks from mineral acids,except HF. This is completely incorrect.

Systematic research at the Geo logical Surveyof Norway (Graft & Roste 1985) has shown that alot of silicate minerals have considerable solubili­ty in mineral acids , such as HCI and HN03. Somesilicate minerals even show a solubility of practi­cally 100 %, while some are nearly insoluble.Graft & Roste have shown that a general para­meter for the solubil ity of silicate minerals seemsto be the ratio between the number of metalatoms in the mineral except Si, and the numberof Si atoms . Thus , with a ratio below 1, the mine­ral has little solubility. With a ratio above 2.4 thesolub ility is high. In the range between 1 and 2.4the solub ility can vary considerably . For instance ,the end member of the plagioclase series, anor­thite (CaAI2Si20 a), with a ratio equal to 1.5, hasa very high solubility , while the other end mem­ber, albite (NaAISi30 a), with a ratio equal to 0.67 ,has very low solub ility (Fig. 1).

Muscovite, with the comm on formula2H20 .K20 .3AI20 3.6Si02' and a ratio equal to1.33 has a very low solubility, while biotite, with acomposition ranging mainly between(H,Kh(Mg,Fe)4(AI,Fe)2Si401 6

102 Magne OdegArd NGU • BULL 427. 1995

so

a rock sample varies strongly within one and thesame sample . 2. The extraction yield for eachsingle element of a rock sample varies stronglyfrom sample to sample. 3. The extraction yieldsfor all elements of a mineral are equal. 4. Theextraction yields for the elements of a mineral in arock sample are independent of the concentrationof the mineral.

Table 1 shows some of the results obtained onthe rock materials. The elements are speciallyselected to cover actual elements in an environ­mental survey of offshore oil drilling. In this activi­ty, Ba is used in the form of barite (BaS04) as aweight agent, and this metal is therefore com­monly used as a tracer in monitoring the distribu­tion of drilling fluids in the marine environment. Sris also a typical componen t of drilling fluids and islikewise used as a tracer for contamination con­trol. The other elements are included becausethey are of particular toxicological interest.

The conclusions related to the extraction ofpure minerals in the work of Faye (1982) arebased on olivine, chromite, beryl, lepidolite andallanite. Thus, extraction tests on several stan­dards of olivine, (Mg,FehSi04 ' showed a con­stant extraction yield of about 85% for the ele­ments Fe, Mg and Ni, while Cr, which is found asthe resistant mineral chromite (FeCr204) ' showeda constant extraction yield of 15 %. The elementNi enters the lattice of olivine. The resistant mine­ral beryl (3BeO'2AI20 3'6Si0 2) showed a con­stant extraction yield of 1.4%, while the Li minerallepidolite {K,Li[AI(OH,FhlAI(Si0 3h } correspon­dingly showed around 70%. Tests with standardsof allanite (a complex rare earth silicate) alsoconfirmed a constant extraction yield indepen­dent of concentration. The situation of Si duringthe extraction of silicate minerals is that even thiselement goes into solution, but will immediatelyprecipitate as insoluble Si02·x·H20 in the prevai­ling acid milieu, and is therefore not found in theextract. It is important to be aware of this behavi­our of Si, which falsely will give a too low solubili­ty if the calculation is based on the weight of theresidue. This may be the case for some of thework of Graff & Roste (1985), and lead to theconclusion that their solubility figures in realitysystematically should be even higher than repor­ted.

Another important discovery in the work ofFaye (1982), and also with particular relevance tooil drilling activity, is that the amount of BaS0 4

which can go in solution during extraction is verylimited, and regulated by the solubility product ofBaS04. The maximum amount of Ba which couldbe measured by the analytical procedure used inthe work of Faye (Odeqaro 1981) was approxi-

' 00

'00 r-- --- - - - - - - - - ---=='"---,

and (H,K)2(Mg,Fe)2AI2Si30 12' and a ratio normal­ly between 1.7 and 2.0, has a high solubility. Thisexplains why biotite is a much better soil fertili­zing mineral than muscovite for the supply ofK20 , even though muscov ite normally has a hig­her K20 content than biotite.

The most important aspect in relation to polluti­on studies and geochemical prospecting will bethe release of analyte elements bound in the latti­ce of the primary silicate minerals. The possibilityof introducing such elements can, in many cases,be evaluated, but this requires: (1) knowledge ofthe composition of the rock material; (2) knowled­ge of the basic geochemical distribution laws; (3)knowledge of the general solubility of rock-for­ming silicate minerals.

In addition to the contribution of trace elementsfrom silicate minerals, such elements can also beintroduced from accessory minerals which usual­ly occur in geological materials. The solubility ofsuch minerals can differ greatly, from acid-solu­ble sulphides such as pyrrhotite(Fe1_xS) and pyrite (FeS2), to more resistant sul­phates, oxides and silicates, such as barite(BaS0 4), chromite (FeCr204) and beryl(Be3AI2Si6018)'Faye (1982) has made a separate study of thesolubility of some resistant accessory minerals,and a more comprehensive solubility study of sili­cate rocks. These studies were undertaken pri­marily to reveal the amount of major and traceelements which can be introduced during acidextraction procedures used in geochemical inve­stigations. The rock material used were certifiedgeological reference samples covering the wholerange from acidic to ultramafic rocks and somesediments . Four general conclusions can bedrawn from the extraction investigations, namely:1. The extraction yield for the various elements of

Fig. 1. Solubil ity of plagiocrase feldspar.

I 50

'"

/ 1

to

0

0 to zc 50 " se '00%Anorthlle

NGU · BULL 427.1995 Magne 0degard 103

Table 1. Range of extraction yield (extracted amount of ele­ment as percent of total content) for some selected elementsbased on ex1raction of 24 international reference materia lswith 7N HN03 (Faye 1982).

Element:Low

Ex1raction yield (%) :High Mean

(1985) are based on both boiling 6N HN03 andboiling 6N HCI. Even if the experimental conditi­ons used in these investigations are not identical,the results are unambiguous and have the samegeneral conclusions on solubility.

* The extraction yield for Cu and Pb is present ed graphically ,indicating that Cu, and probably also most of the Pb, mainlyoccurs as sulphides and largely goes quantitatively into soluti­on.

mately 0.25%. In this procedure a sample weightof 1 g and an extrant volume of 5 ml were used.The procedure being used today applies in accor­dance with directions given in Norsk Standard NS4770, a sample weight of 1 g and an extrant volu­me of 20 ml. This standardised procedure, whichis being used in many laboratories, limits theamount of Ba (in the form of BaS04) which canbe determined to around 1%. This level hasrecently been confirmed in separate tests by thepresent author. Many sample types can also beexpected to contain S in the form of sulphatesother than BaS04, or in the form of sulphides orother compounds which can form S04- · ionsduring the extraction. In such cases the maxi­mum amount of Ba which can be determined willbe less than 1%, since the amount is regulatedby a constant ion product [Ba++j[S04- ]'

With regard to the use of Sr as a tracer for thecontamination of sea floor sediments by drillingfluid, it is important to take into consideration thefact that the background level of Sr is frequentlyrelatively high, and shows great fluctuations. Sr iscommonly found in silicate minerals containing Kand Ca, and in a marine environment also oftenin carbonates. The extractable amount of Sr ine.g. K-feldspar will be very small, while the totalcontent will be released in e.g. CaC03. From anenvironmental survey at the Statfjord field in 1987(!KU-report 02.0840) it was reported that Sr is notas good as Ba in tracing drilling fluid contaminati­on of sediments. This is probably mainly due to amuch higher concentration of Ba than Sr in thedrilling fluid, but may also be due to a relativelyhigh and fluctuating background level of Sr.

Most of the extraction studies of Faye (1982)are based on boiling 7N HN0 3 and the analyseswere done by the ICP-AES technique (0degard1981). The solubility studies of Graft & Roste

FeBaSrCrNiZnc u­Pb*

1.7 85.02.0 88.31.7 80.30.4 87.28.8 1007.6 94.2Not calculatedNot calculated

46.427.838.338.150.356.9

ConclusionsA general conclusion is that the amount of majorand trace elements which can be introduced fromrock material during a standard acid extractionprocedure can vary within very wide limits,depending on the type and amount of both rock­forming silicate minerals and accessory minerals.This may in extreme cases disturb, or even over­shadow, effects which are to be studied; thedanger being especially great when working atlow concentration levels and with marginaleffects. This is often the case when studying theeffects of moderate or marginal environmentalloads.

For the conclusions it is very important that thebackground contribution is stable, which meansthat the medium must be homogeneous. This canbe a problem for sediments which often are inho­mogeneous due to mineral segregations.

The use of reference materials is a central ele­ment in quality control. However, such materialsare as a rule certified for total element content,and are therefore of limited value for extractionanalyses. The best alternative will be to base thecontrol on interlaboratory comparisons, and whe­re identical and reproducible extraction procedu­res are being used.

The main intention of this work has been todemonstrate some of the many problems and pit­falls which can be encountered in geochemicaland environmental studies based on geologicalmaterials, and the few examples outlined showthat such activity requires multiprofessionalknowledge at all stages from planning to the finalinterpretation.Systematic research on the general solubility ofrock-forming silicate minerals has been of parti­cular significance in this field, and has contribu­ted considerably towards providing a safer basisfor drawing conclusions. A great challenge lies inutilising this knowledge in the future developmentof extraction routines with optimum selectivity.

ReferencesFaye, G. Chr. 1982: Metodestud ier i geokjemi - HN0 3-ekstrak­

sjon av geokjemiske prover. Nor. geol. unders. Report 1687C, 19 pp.

Graft , P. R. & Roste, J. R. 1985: Utluting av silikatmineralermed minera lsyrer. Nor. geol. unders . Report 85.105,26 pp.

Odeqard , M. 1981: The use of inductively coupled argon plas­ma (ICAP) atomic emission spectroscopy in the analys is ofstream sediments . J. Geochem. Explor.14 , 119-130.

IKU NS, Continental Shelf and Petroleum TechnologyResearch Institute N S 1988e: Statfjord environmental sur­vey 1987. Report no. 02.0840, 134 pp.