Fingerprinting petroleum pollutants in the mediterranean sea

Fingerprinting petroleum pollutants in the mediterranean sea Fingerprinting petroleum pollutants in the mediterranean sea Fingerprinting petroleum pollutants in the mediterranean sea Fingerprinting petroleum pollutants in the mediterranean sea Fingerprinting petroleum pollutants in the mediterranean sea Fingerprinting petroleum pollutants in the mediterranean sea Fingerprinting petroleum pollutants in the mediterranean sea

Fingerprinting Petroleum Pollutants in

the Mediterranean Sea

J ALBAIGES

Institute de Quimica Organica (CSIC), Jorge Girona Salgado,

Barcelona-34, Spain

ABSTRACT

To confront the increasing and diversified petroleum pollution problems in the Mediterranean Sea there is a need to develop ef-ficient analytical methods for the source identification of oil pollutants The paper describes an analytical approach for finger-printing crude oils from different geographical areas, which are

of particular incidence in the Mediterranean, and gives special attention to the identification of chronic pollution samples from the open sea

The approach involves, in a first step, the determination of the

S, Ni and V contents of the samples as well as the phytane/pris-tane ratio from the HRGC profiles (FID) With this information a general assignment of the type of the pollutants and the area from where they come can be established The precise identifica-tion of the samples source is attempted by a multi-fingerprinting procedure which is carried out by the use of selective detectors

in GC Four profiles are considered, corresponding to total and polyaromatic hydrocarbons (FID), sulphur (FPD) and nitrogen (NPD) compounds Alternatively, the use of mass-fragmentography (GC-MS-COM) to obtain profiles for specific series of hydrocarbons of geochemical significance, such as C^n-C^Q acyclic isoprenoids,

C97+ steranes and triterpanes is highly stressed

KEY WORDS

Isoprenoids Mass-fragmentography Mediterranean Sea Oil finger-printing Petroleum pollution Selective GC deterctors Steranes Tar balls Triterpanes

INTRODUCTION

The Mediterranean Sea is among the first marine regions to show the symptoms of oil impact In fact, the observed concentrations

of petroleum tars on its surface are one order of magnitude higher

69

70 J Albaiges

than those generally found in any other regional seas (Nat Acad

of Sei., 1975) The factors that contribute to this situation are several Thus, the particular hydrogeological conditions of the basin are such that oil entering or discharged there has lit-tle chance of leaving; it will stay and accumulate until it is degraded Moreover, the cyclonic drift of the water circulation tends to deposit the oil on the shores or to accumulate it at cer-tain exposed points The oil handling activities brought about

in the area, and specially the tanker traffic, are, in turn, quite important, as it is shown in Fig 1 Finally, a lenient le-gislation has been unable to prevent intentional pollution

Fig 1 Location of the different sources of petroleum

pollution in the Mediterranean Sea (courtesy

of P Le Lourd, 1977) From all of the above, the oil pollution control in the Mediterra-nean has become a problem of mayor concern and consequently the development of analytical methods for the source identification

of petroleum pollutants has received particular interest

However, apart from the primary characterization of acute inci-dents , much attention has been recently devoted to the ultimate fate of spilled oils, specially for the assesment of their per-sistance into the sea and their contribution to the chronic pol-lution Chronic pollution by oil is, in the Mediterranean Sea, far more important than accidental pollution and it is generally assumed that is equally due to marine operational losses and to land-based discharges, including petroleum, petrochemicals and fossil fuel combustion products In these situations the chemi-cal composition of the samples found in the sea can be consider-ably altered by environmental conditions Therefore, the ident:L

Fingerprinting Petroleum Pollutants 71 fication methods of oil pollutants should be able to deal with

an entire spectrum of products of different origins and ages

The present paper gives an account of the analytical approach

used for fingerprinting oil products of different sources

car-ried throughout the Mediterranean, with special reference to

the identification of highly weathered samples This has been

attempted by mass-fragmentographic fingerprinting of

geochemical-ly significant series of hydrocarbon components (biological

mar-kers)

EXPERIMENTAL

Reference samples of crude oils from Venezuela, Spain, Nigeria,

Gabon, Libya, Algeria, Kuwait, Saudi Arabia, Iran, Irak, Oman

and AbuDhabi were supplied by several refineries and production

companies Tar ball samples were collected during a survey cruise (July 1977) between Cartagena (Spain) and Civittavechia (Italy), with a neuston net Samples were dissolved in toluene to

elimina-te extraneous maelimina-terials and stored at -4QC until analysis

Total sulphur content was determined by combustion (ASTM D-129;

Bomb Method) and Ni and V by flamless atomic absorption

spectro-metry (Perkin Elmer 4000, equipped with a graphite HG-74 furna-ce)

Dual FID/FPD and FID/NPD chromatograms were obtained by

split-ting the column effluent The former profiles were run on the

deasphaltened oil residues in n-pentane (40 volumes) and the lat-ter on the "polar oil fraction", which was isolated by partition

of the deasphalted fraction into a cyclohexane-nitromethane

mix-ture (1:5) The gas-Chromatograph (Perkin-Elmer 990) equipped

with FI, FP and NP detectors was operated either with 9ft x 1/8" packed columns (1% Dexsil 300 or 3% OV-101 on Gas-Chrom Q 100-120) from 150 to 300QC at 6QC/min or with 200ft x 0.02" capillary

columns (OV-101) from 120 to 280QC at 6QC/min., using He as carrier gas

Mass-fragmentography (MF) was performed on a LKB 9000 S/PDP 11 E

10 computerized GC-MS system The injector and jet separator

we-re maintained at 29 0QC and spectra wewe-re we-recorded and disk stowe-red

at 4 sec intervals In this case, the "branched-cyclic fraction" was preferably used This was isolated throuhgout the previous

recovery of the saturated fraction by conventional silica-gel ad-sorption chromatography, with an absorbent-sample ratio of 20

(eluting solvent: n-penatne), and subsequent inclusion in 30 fold excess of 5Â molecular sieves (solvent: iso-octane)

RESULTS AND DISCUSSION

All identification methods of oil pollutants are essentially a

matching of samples based on the geochemical principle that no

two oils have identical compositions unless they have identical

histories Thus, in theory every oil product is unique; however, oil is a very complex mixture and anlysis is never complete, so

72 J Albaiges

that very similar oils may appear to be identical Furthermore, exact correspondence of compositions of samples exposed to dif-ferent environmental histories cannot be expected

As there is no single analytical technique that can fully charac-terize an oil product, identification must be established by a series of analyses involving gecohemically characteristic and

weathering-resistant sample parameters From the several

tech-niques used for that purpose and according to our previous expe-rience (Albaigés and others, 1976, 1979) we have selected the pa-rameters indicated in Fig 2 for building-up an inventory of pos-sible source-oils in the Mediterranean, with which to compare the pollution samples All of these parameters have been suggested as the most suitable for a quick screening procedure of the pollutants (Brunnock and others, 1968; Zafiriou and others, 1973; Shekel and Ravid, 1977) A major advantage is that the information required

is already available for most of the crude oils or is easily ob-tainable by routine analytical methods, and although the values reported in the literature are referred to unexposed oils, their correlation with pollution samples can be attempted, because the effects of weathering can be predicted (Brunnock and other,1968; Blumer and other, 1973) The HRGC pattern (FID) from which the phytane/pristane ratios are obtained permit, in addition, to

assess the type of pollutant, namely, oil sludge, crude oil,tech-nical fractions, etc

As it can be seen in Fig 2, this characterization procedure al-lows a primary classification of the samples into several geogra-phical groups In this manner, we have assigned to Middle East (M.E.) sources 86% of the pelagic tar ball samples collected du-ring a survey cruise in the Western Mediterranean (Albaigés and others, 1979) However, this classification apparently leads to

a certain overlaping and, in agreement with Jeffrey and others (1974), it is unlikely that with such parameters one could be

able to ascertain the specific origin of the samples Their varia-tion between crudes of the same area and their resistance to the sea weathering processes are not enough, in many cases, for provi-ding the unequivocal identification of the pollutant Furthermore, this procedure cannot be applied to hydrocarbon samples found in highly dispersed forms, namely dissolved in water, adsorbed in se-diments or biological samples, etc Thus, to provide furhter in-sights into these problems the composition of the samples should

be better defined, and since the analysis of such chemically com-plicated samples is so extremely demanding, the great potential

of GC can be particularly useful

Fingerprinting Petroleum Pollutants 73

ϊιηυ

V/NI T.t-10.·

Ph/Pt a « - v t

Fig 2 Chemical charaterization parameters of

crude oils handled in the Western Mediter-ranean Numbers in brackets indicate the number of oils examined

It is known that the GC traces of oil residues are characterized

by a large unresolved envelope above the baseline, with small re solved peaks superposed Incidentally, the occurrence of this un resolved "hump" in sediment or biota samples has been generally considered as an indication of petroleum contamination, being

however its origin unknown In order to get more information on the compounds included in this "hump" different types of

detec-tors capable to enhance the response of specific series of

com-pounds can be used Fig 3 shows a set of profiles actually mea sureble on petroleum residues In addition to the already mentT oned hydrocarbon (HC) fingerprint (FID), profiles of sulphur and nitrogen containing compounds have been obtained with the ^use of

FP and NP detectors, respectively In this manner, a multiprofile characterization procedure can be set up

While the sulphur and HC profiles can be concurrently obtained

straight from the samples by FPD and FID (Adlard and others,

1972), the smaller abundance of nitrogen compounds in unused oils requires their previous concentration For this reason, since

the applied concentration procedure consisted in a simple liquid-liquid extraction of the "polar oil fraction", polyaromatic hydro carbons (PAH) were also extracted together with the N-containing compounds This allowed the obtention of another representative FID profile (labeled FID(PAH) in Fig 3) which will^be useful in assigning the origin of the pollution in some chronic situations

If the source of HC is predominantly petroleum, the chromatogram will contain a large number of peaks arising from alkylated PAH, whereas if the source is pyrolytic (air pollution) or coal tar

74 J Albaigés

or creasote in nature, a more simple profile will be observed, corresponding to a higher predominance of the unsubstituted spe-cies over their alkylated homologs (LaFlamme and Hites, 1978)

Fig 3 GC profiles displayed by a Venezuelan

crude oil residue (b.p 200Ω c ) Num-bers over the peaks indicate the n-pa-raffin carbon atoms FID (HO/FPD pro-files were obtained in a OV-101 capil-lary column, whilst the FID(PAH)/NPD

in a 3% OV-101 packed column

This fingerprinting procedure appeared quite succesful for the correlation of recent oil spillages with their suspected sources Nevertheless, there were some question about the general specifi-city of these profiles and their stability throughout the envi-ronmental exposure of the product at the sea Both aspects are of

Fingerprinting Petroleum Pollutants 75

interest in relation to pollution problems in the Mediterranean

Sea, where, e.g., a particular incidence of very similar oil pro-ducts from the M.E area and a wide occurrence of weathered pelagic tars resulting from tanker washings are expected Therefore a

further evaluation of the method would seem in order

The most significant Chromatographie profiles are

by the FID(HC) and FPD Nevertheless Figs 4 and

pie of oils of similar geological origin, exhibiti

similar patterns to render difficult the precise

of the pollutant source The Aramco and Kuwait oi

ly indistinguishable on the basis of the FID(HC)

their FPD profiles exhibit only slight difference

ter very similar to other M.E oils The NPD and

are of more limited value for oil identification

smaller variability among products, apart from th

suming procedure, as it has been recently noticed

others (1979)

those displayed

5 show the

exam-ng sufficiently identification

Is are apparent-chromatogram and

s, being the lat-FID(PAH) patterns because of the

e more time

con-by Frame and

juUl

" I I I

_JL £> * * 11 IXUJUUWj

4»«& WMMA

■wJ^^

Fig 4.HR GC profiles (FID) of Kuwait and Aramco

crude oils (x : pristane and phytane)

In the case of the identification of oil spillages

washings several additional difficulties arise In

be seen how the fingerprints of the original sampl

modified by the typical sludge peaks (n-C2c^.) a n d

fractions have been altered (affecting the phytane

tios) by the long exposure to sea weathering condi

worth to mention here that the classification crit

ed in Fig 2, according to characteristic chemical

from tanker Fig 6 it can

es have been the lower /pristane ra-tions It is eria establish-parameters,

76 J Albaiges

ROSTAND

Fig 5 HR GC profiles (FPD) of Middle East crude

oils

could also give unconclusive results on such samples, owing to their variable enrichment in paraffins during tanker transport

or deposition at sea

A final observation to be made on the applicability of this GC fingerprinting procedure is that the lack of knowledge about the nature of every resolved compound doesnft allow to

elucida-te the significance of qualitative variations among profiles , thus difficulting to ascertain when the slight differences ab-sorved are attributable to different origins or weathering his-tories For these reasons we turned our attention to a more specific sample characterization procedure, which involves the fingerprinting of HC series of geochemical significance (biolo-gical markers) The occurrence and distribution of these series, which are included in the unresolved "hump" of the chromatogram, are related to the particular genetic history of the samples

Fingerprinting Petroleum Pollutants 77

Fig 6 GC profiles of pelagic tar ball samples

collected at the Western Mediterranean

Weathering increases from S.N.16 to No.23

I )

113

I I I )

I I )

191

Fig 7 Petroleum biological markers I: acyclic

isoprenoids Ilrsteranes Ill: rearranged

steranes IV: hopanes Numbers indicate

preferred fragmentation ions in the mass

spectra

78 J Albaiges

Among these series of HC we have the acyclic and polycyclic isopre-noid alkanes drawn in Fig 7, which seem to be resistant enough

to sea weathering (Reed and Kaplan, 1977), thus being suitable for identification purposes Some recent reports (Rubinstein and

others, 1977; Seifert and Moldowan, 1979) have specifically dealt with the biodégradation of these "markers1' and the results repor-ted confirm that the profiles are severely modified only when conditions for total degradation of acyclic isoprenoids have pre-vailed; a rare ocurrence in the marine environment

Taking into account that these compounds exhibit common ions in their mass spectra (see Fig 7 ) , the M.F paterns from computeri-zed GC-MS provide a powerful tool for the determination of distri-butions of homologous series, affording the relative abundances

of individual members and epimeric and ring skeletal isomers Hence, crude oils quoted in Fig 2 were characterized by the mass-chromatograms of m/e 183, 191, 217, 231 and 259 Although not all of them provide significant fingerprints for each one of the referred ions, in spite of this, relevant differences were still observed Fig 8 shows a few examples from these crude oils The most abun-dant series is that of the triterpanes of the hopane type(IV,m/e 191) This family is formed by a series of C?7-C.r members The stereochemistry of the C-17 and C-21 in petroleum and matured HC samples is 17 e* (H), 21 fi> (H) and 17 fi (H), 21 ex (H) with the

22R 4- 22S isomers, whilst in the precursor biological materials only the 17ß(H), 21*(H) with one diastereomer at position 22 is found This stereochemical fate has been considered as a definite test for oil pollution monitoring in sediments (Dastillung and Albrecht, 1976) In addition, two C^7 members are present, the 17cx (H) and 18c* (H)-trisnorhopanes, Their relative abundance being able to differentiate the previously reported Aramco and Kuwait crude oils

The distribution of the individual members of this series was pre-viously used, after isolation, by Pym and others (1975) to fin-gerprint M.E crude oils However, the advantages of the present

MF procedure are obvious as far as the analysis time is concerned,

as well as to the new identification possibilities offered by the multiparametric profiles hereing described Moreover, another iden-tification parameter displayed by the m/e 191 MF is that corres-ponding to the C?n-Coc tricyclic diterpanes (Reed, 1977)which

elu-te before the hopanes and appear with an aselu-terisk in Fig 8

In the sterane and methylsterane families two series of compounds can be expected: the normal (II, m/e 217, 2 31) and the rearranged steranes (III, m/e 259, 273), the latter occurring exclusively in petroleum and other geochemically matured samples Variations in the stereochemistry of carbons 5, 14, 20 and 24 afford a very com-plex pattern (Ensminger and others, 1978)and give complementary evidence of fossil fuel origin, because the chiral centers are built biosynthetically in only one stereochemical configuration However, steranes seem to disappear earlier with maturation, hence N.A crude oils exhibit the common characteristic of having very small concentrations of these HC (Fig 8)