506
DETERMINATION OF SMALL AMOUNTS OF NON-POLAR
HYDROCARBONS (OIL) IN SEA WATER
By
Stig R.
CARLBERG
and
C. Bo SKARSTEDT
Industrial Water and Air Pollution Control Company,
Box 4052, S 400 40 Gateborg 4, Sweden.
Small amounts of mineral oil (defined as non-polar hydrocarbons) have been determined
by infrared spectrophotometry after extraction from water samples and separation from
polar hydrocarbons by column chromatography on aluminium oxide. No distinction has
been made between non-polar hydrocarbons originating from mineral oil and those from
marine life. A simple standard mixture of hydrocarbons used as source for calibration
provides a good compromise between the varying specific absorptivities of different types
of mineral oil. The method has been succesfully applied for monitoring in oil pollution
studies.
INTRODUCTION
The need for methods of determination of small amounts of oil has increased
during recent years. The main reason for this is the swiftly rising consumption
of oil throughout the entire world. This means more and larger transports by
land and sea and increased storage in big dumps and in smaller fuel-oil tanks
in all civilized areas.
The increased handling of oil results in increased oil-spillage, which today
amounts to about one per cent of the oil consumption. The spillage of oil stems
from leaking tanks and pipe-lines, over-loading and/or broken tubes when
pumping, transport accidents, discharge of oil-polluted sewages from industries,
cleaning of ship-tanks and dumping at sea.
The type of damage that primarily called for more sensitive methods of
determination of the existence of oil were damages on ground- and surfacewaters used as supplies for drinking-water. In recent years more attention have
been paid to the possible biological effects of oil (BLUMER et. al. 1971).
Very small amounts of oil in water are capable of causing deterioration of
smell and taste. MELPOLDER et. al. (1953) claims 0-005 mg/1 as a threshold
concentration for petrol and GIBBON (1940) 0-2 mg/1 for petrol and 0-7 mg/1
for fuel-oil. The National Institute of Public Health in Sweden gives the threshold concentrations 005-0-1 mg/1 for fuel-oil no 1 as well as petrol.
Several methods for isolation and determination of oil are described in the
literature. Diethylether, petroleum spirit, chloroform and carbon tetrachloride
I
J. Cons. int. Explor. Mer
I 34 | No. 3 | 506-515 I Copenhague, octobre 1972 I
Downloaded from http://icesjms.oxfordjournals.org/ at Penn State University (Paterno Lib) on March 4, 2016
Fishery Board of Sweden, Hydrographic Department,
Box 4031, S 400 40 Goteborg 4, Sweden.
Determination of oil in sea water
507
MATERIALS A N D METHODS
Sampling equipment and techniques
Because of the risk of contaminating the sample it is essential that the equipment used should never need lubrication. Hence Perspex, PVC, glass etc. are
suitable construction materials.
After sampling, the water is transferred to glass bottles with grounded
stoppers. Avoid touching the ground surfaces.
All equipment used should be carefully washed and thoroughly rinsed with
large amounts of ion-exchanged or distilled water. Whenever possible the different items should be rinsed with carbon tetrachloride after washing and
rinsing.
We have used a bucket of stainless steel to obtain surface samples. This
sampling of the sea-surface is always made from the moving ship, just before
it stops at a sampling station. This method seems sufficient to prevent contamination from the ship itself. The samples from the deeper water-layers we have
obtained with TPN (Total Plastic Nansen) water bottles from Hydro-Bios
Apparatebau GmbH, Kiel, W. Germany. The TPN is never fastened on the
Downloaded from http://icesjms.oxfordjournals.org/ at Penn State University (Paterno Lib) on March 4, 2016
have been used as extracting agents. For the quantitative determination several
methods have been used: gravimetry, measurement offluorescence,determination of specific weight, mass-spectrometry, thin layer- and gas chromatography.
SIMARD et. al. (1951) proposed a method with extraction by carbon tetrachloride
and subsequent quantitative determination with infrared spectrophotometry.
The absorption due to hydrocarbons is measured in the spectral range 3333 cm" l
-2500 cm- 1 (3-4 urn).
Oil consists of a mixture of hundreds of hydrocarbons, the proportions of
which strongly vary with the type of oil. In order to overcome these difficulties,
SIMARD et. al. (1951) made a standard-mixture of w-hexadecane (cetane), isooctane, and benzene and compared the unknown samples with this standard.
This standard is now commonly used in infrared analyses of oil.
In the extraction of the sample are isolated not only petroleum hydrocarbons
but also organic substances of other origins such as humic compounds, lignin,
animal and vegetable fats and oils and surface-active agents. Hence what is
really determined is the total amount of carbon tetrachloride extractable substance.
LINDGREN (1957) has taken up this phenomenon and in his exellent article
proposes chromatography through a polar column to remove these interfering
organic substances from the petroleum hydrocarbons. The hydrocarbons passing
the column are regarded as mineral oils and consist mainly of alkanes. This
way of looking at the problem is now commonly accepted. Aluminium oxide is
normally used in the column.
An interesting and precise method for determination, involving thin layer
chromatography has been worked out by GIEBLER, KOPPE and KEMPF (1964).
BEYNON, KASHNITZ and RIJNDERS (1968) have published some analytical methods
for determination of oil in water and soil.
The analytical method described below is an attempt to put the finishing
touches to, and adapt the techniques for determinations of small amounts of
oil in seawater on a routine basis and with a reasonable amount of effort.
508
S. R. CARLBERG and C. Bo SKARSTEDT
\J
b
Figure 1. a. Glass bottle with stopcock adapter in the stand with rings; b. chromatography
column; c. adapter for Gooch crucibles; d. stopcock adapter.
hydrographic wire but on a nylon rope attached to the wire and about 5 m
below the end of the wire. The TPN is messenger-operated. This is performed
with help of an ordinary messenger-operated water-bottle, placed so it overlaps
the joint between the wire and the rope.
The water is poured into 2-5 I glass bottles (Fig. 1 a). To prevent biological
activity in the sample 1-5 ml of saturated mercuric chloride solution (HgCl2)
is added per sample.
Reagents
1. Hydrochloric acid, reagent quality. The concentrated acid is diluted with
an equal volume of distilled water.
Downloaded from http://icesjms.oxfordjournals.org/ at Penn State University (Paterno Lib) on March 4, 2016
5 cm
Determination of oil in sea water
509
Glass-ware and apparatus
1. We extract the samples in the glass bottles mentioned above. To collect the
carbon tetrachloride after extraction we replace the stopper with a TS-cone
provided with a stopcock (e.g. Quickfit Stopcock Adapter MF 11 Fig. Id).
Thus the bottle acts as a separating funnel. (When extracting with 25 ml it
is not necessary to provide the adapter with air-inlet.) But the presence of
any kind of (fatty) lubricant, except the sample itself, will disturb the analysis, giving high, positive values. So we recommend the use of stopcock
keys made of PTFE.
2. Reciprocating motion agitator. We use an apparatus supplied by Edmund
Biihler AG, Tubingen, W. Germany, model Sm 2. The speed used is about
180 strokes (5-6 cm) per minute.
3. Chromatography columns. Easily made from glass tube 10 mm inner diameter, 150 mm long, formed to a narrow neck at one end, (Fig. lb).
4. Glass wool.
5. Filtering paper. Munktell no OOR or equivalent.
6. Funnels for filtering. To reduce the losses of carbon tetrachloride due to
evaporation a high, narrow type should be used. We suggest Pyrex no 3960
(Adapters for Gooch Crucibles) or similar, (Fig. 1 c).
7. Centrifuge. Capacity 50 ml in each tube and 3000 rpm.
8. Infrared spectrophotometer (IR-photometer). Any low- or medium cost
instrument will do if it has double-beam operation and capacity for 40 mm
cells. An instrument with a grating monochromator is preferred.
9. Cells 40 mm Infrasil or equivalent quartz material.
Extraction
To the sample (approximately 2 1) add 2 ml hydrochloric acid 1 + 1 . Check
that the pH is about 3 or less. From a measuring glass add 25 ml carbon tetrachloride. Stopper the bottle and shake it by hand for a few seconds. Open the
bottle to release the pressure and re-stopper it again. Repeat this procedure
until the sample is totally gas free.
Place the bottle (bottles) in the agitator and shake for 60 min. (The Biihler
machine is capable of shaking four 2-5 1 bottles at a time).
After the extraction stand the sample aside to let the phases separate. If an
Downloaded from http://icesjms.oxfordjournals.org/ at Penn State University (Paterno Lib) on March 4, 2016
2. Carbon tetrachloride. It is not important to use expensive spectroscopic or
certified grades. It is important that the quality used should be free of any
significant absorption-maximum in the infrared region 3333 cm"1 to
2500 cm" 1 (3-4 urn). The same quality is used through all steps in the analysis, from cleaning of equipment to spectrophotometric determination. For
the present we use carbon tetrachloride BP 1953 grade from Albright and
Wilson Ltd, London.
3. Aluminium oxide, neutral, activity step 1 due to Brockman. We have found
that type 507 C from Fluka (Switzerland) is quite suitable.
4. «-Hexadecane (cetane), reagent quality.
5. iso-Octane, reagent quality.
6. Benzene, reagent quality.
510
S. R. CARLBERG and C. Bo SKARSTEDT
Column chromatography
This step is necessary in order to remove the polar hydrocarbons from the
sample. A small piece of glass wool is placed in the bottom of the column and
the aluminium oxide is poured above this to a height of about 10 cm. Do not
tap the column or make any other attempt to settle the powder, because this
will inevitably result in an extremely slow-running column. Place a small piece
of glass wool on top of the oxide. This will prevent the powder from whirling
up when pouring in the sample.
Saturate the column with carbon tetrachloride and let approximately 6 ml
flow out. This is about equal to the volume retained by the oxide. When the
column has stopped dripping, carefully pour a part of the sample on the column.
When another 6 ml has passed out (use measuring cylinder) collect the remaining part of the sample as it passes the column. Ground 25 ml conical
flasks are very suitable for this purpose.
Infrared spectrophotometric determination
The reference- and sample cells are carefully cleaned with carbon tetrachloride, filled up with the same liquid and the zero-absorption is checked up
in the region 3333 cm*1 to 2500 cm"1 (3-4 um). Then the sample cell is filled
up with the sample and the spectrum is recorded in the region mentioned above.
If the absorption of the sample appears to be higher than suitable due to the
calibration curve, the sample is diluted with pure carbon tetrachloride and the
spectrum is recorded once again.
If the absorption of the sample is measured directly after extraction and
filtering, the quantity measured is the total amount of carbon tetrachloride
extractable organic substance. If however the absorption of the sample is measured after the chromatography the quantity measured is the total amount of
carbon tetrachloride extractable non-polar hydrocarbons, "oil".
Calibration and evaluation of results
Obviously the most correct way to make the calibration curve is to add
known amounts of the standard mixture to a series of water samples (1 litre,
synthetic seawater or distilled water) and extract the samples, followed by the
column chromatography.
Downloaded from http://icesjms.oxfordjournals.org/ at Penn State University (Paterno Lib) on March 4, 2016
emulsion has formed or if the phases have not separated completely after 2 h
it is necessary to centrifuge.
Replace the stopper with stopcock adapter described above. Secure it to the
bottle with a clamp or rubber band. Invert the bottle and place it in a stand as
shown in Figure 1 a. Draw off the organic phase directly into the narrow filtering
funnel. The filtering paper should previously have been washed with 10-15 ml
carbon tetrachloride. The filtered sample is collected in a 25 ml conical flask
with a ground glass stopper.
In case it is necessary to centrifuge the emulsion and organic phase, draw
these off directly into the centrifuge tube. After centrifuging, collect the organic
phase with a pipette; the filtering may not always be needed. The extract is
now ready for column chromatography. But if the oil content of the sample is
very high it may be necessary to extract twice or more.
Determination of oil in sea water
511
ABSORPTION
0.80
0.70
2
mg standard mixture per
25 ml carbon tetrachtoride
Figure 2. Calibration curve for 37-5 % n-hexadecane (cetane), 37-5 % iso-octane and 25-0 %
benzene (measured in 4 cm cell).
We have simplified that procedure to a great extent. From a solution with a
known concentration of standard-mixture (37-5% H-hexadecane, 37-5% isooctane, 25 % benzene) in carbon tetrachloride, we prepared a series of dilutions
in 25 ml measuring flasks. Thus the concentrations (in mg/25 ml) of these solutions correspond to the concentrations of samples in mg/1 if one litre of water
is extracted with 25 ml carbon tetrachloride. (e.g. 0-34 mg/25 ml = 0-34 mg/1
sample). The correctness of this calibration procedure was confirmed by application of this method when analyzing effluents from some factories. The
concentrations of oil in these samples were so high that we had to dilute them
about 100 times. In parallel we made gravimetric determinations on several of
the extracts after IR-photometric determination.
The following figures are representative for the results we obtained:
Industry
Oil mg/1 IR-photometric
Oil mg/1 gravimetric
A
B
C
324
138
142
319
140
142
Thus the simplified procedure might be regarded as sufficient.
LINDGREN (1957) made a lot of experiments with the extraction procedure
and found that generally one extraction was sufficient to recover 93-99% of
the oil, (where 93% is for petrol). GIEBLER et al. (1964) showed that extracting
Downloaded from http://icesjms.oxfordjournals.org/ at Penn State University (Paterno Lib) on March 4, 2016
0.10 •
512
S. R. CARLBERG and C. Bo SKARSTEDT
2.5
T V.
3.0
I
4.0
18 20
25
30
40 ,» m
100
III
• 100
90
90
80
80
-1
3500
3000
2500
500
400
300
Figure 3. Infrared absorption spectra of a sample before and after column chromatography.
The transmission of the sample increases after removal of the polar hydrocarbons (upper
curve).
for one hour is sufficient to get a recovery of 98-100% of drilling oil, which
seems to be one of the most difficult oils to extract.
To evaluate the spectra obtained a straight line is drawn between the two
absorption minima (that means about 100% transmission) surrounding the
region 3000 to 2800 cm" 1 (Fig. 3). This line is normally horizontal. It is not
unusual however that a slight absorption remains in the region of 2500 cm" 1
even if the recorded curve is a straight line. This is due to increasing absorption
in the carbon tetrachloride and the cells. In such a case a horizontal line is drawn
from the minimum in the 3300 to 3000 cm" 1 region. From the straight line the
absorption maximum at 2930 cm"l is measured. Most instruments are graduated in per cent transmission and in that case the transmission value obtained
is converted into absorption value with means of a standard table or the formula:
A = log io h/I
where A = absorption; 7n = intensity of incident radiation (100%); I = intensity of the radiation passing the sample (in per cent).
To make a general formula for calculating the final concentration of oil, it
is suitable as mentioned above to make a calibration curve with absorption
plotted versus mg oil/25 ml carbon tetrachloride (Fig. 2) and regard this as a
result of an extracted 1 litre sample.
In that case the formula will be:
c=
ml CC14 |000
25
where C = mg/1 hydrocarbons, oil; a = mg/25 ml from the calibration curve;
ml CCU = total volume of carbon tetrachloride used for the extraction (or
multiple extraction); b = ml sample used; and d= possible dilution factor
in the photometric step.
Precision and accuracy
For the time being we regard 005 mg/1 standard mixture to be the lower
limit for the method.
Because different types of oil have different absorptivities ranging from 1 -75
Downloaded from http://icesjms.oxfordjournals.org/ at Penn State University (Paterno Lib) on March 4, 2016
4000 cm
II
Determination of oil in sea water
513
TABLE 1. Concentrations of polar and non-polar hydrocarbons in the Kattegatt and the
Baltic, August-October 1970. Preliminary maximum values.
Station
Date
12. Oct
19.Aug
26. Aug
SW Vinga
N57°33', Ell°31'
L. Middelgrund
N56°58',E11°45'
Bornholm Deep
N55°15', E15°59'
F8
N64°41', E22°44'
29. Aug
F26
N61°59', E20°04'
30. Aug
SS29b
N6P06', E19°57'
30. Aug
F64
o
o
N60 13')E19 04
31. Aug
/
L5
N59°55', E25°36'
1. Sept
F72
N59°17', E21°34'
3. Sept
Gotland Deep
N57°20\ E20°03'
0
5
60
0
5
40
0
5
60
0
5
70
0
5
100
0
5
90
0
5
100
0
5
50
0
5
70
0
100
Total amounts
hydrocarbons
(polar and nonpolar) in mg/1
Non-polar
hydrocarbons
(mineral oil)
in mg/1
008
017
<005
<005
<005
<005
011
019
009
008
<005
005
009
006
<005
009
006
005
012
<005
<0-05
<005
<005
<005
<005
<005
006
<005
<005
<005
013
007
009
005
< 005
<005
< 005
<005
011
008
-
<005
011
012
<005
009
010
005
<005
0-12
<005
<005
<005
005
007
<005
<005
008
to 2-99 1 . g" 1 . cm"1 for lubricating oil, drilling oil and fuel oil respectively
(SCHOLL and FUCHS 1968) they would give calibration curves with different
steepnesses. Obviously the standard mixture method is a (good) compromise
between the extremes mentioned. The specific absorptivity of this mixture is
2-37 1 . g - 1 . cm" 1 and thus the resulting accuracy is ± 26%. This value can
of course be improved if it is known what type of oil is present in the sample
and the calibration is made with the same or a similar product.
RESULTS AND DISCUSSION
In Table 1 are shown some results, representative for what we have found in
the Baltic. As can be seen most of the concentrations are below 0-05 mg/1 and
only a few are higher than 0-10 mg/1.
Recent findings indicate however, that even low concentrations of oil are
toxic to marine life (BLUMER, 1971; BLUMER, SOUZA, and SASS, 1970). This fact
should call upon a more sensitive means for the estimation of the oil content.
Application of gas-liquid chromatography (GLC) seems to be a proper way,
(BRUNNOCK, DUCKWORTH, and STEPHENS, 1968; RAMSDALE and WILKINSSON,
1968). This also makes it possible to distinguish between non-polar compounds
originating from oil and non-polar compounds originating from marine plants
Downloaded from http://icesjms.oxfordjournals.org/ at Penn State University (Paterno Lib) on March 4, 2016
12. Oct
Depth
in m
514
S. R. CARLBERG and C. Bo SKARSTEDT
TABLE 2. Concentrations of polar and non-polar hydrocarbons in the archipelago and
harbour of Goteborg, February 1971.
The distances between the three stations are about six and seven nautical miles respectively.
Station
Hallan
(Outer archipelago)
Total amount of hydro- Non-polar hydrocarbons
carbons (polar and non(mineral oil) in mg/1
polar) in mg/1
0
5
0
5
18
0
5
31
0-71
0-47
0-28
008
007
010
010
005
0-20
013
006
<005
<005
<005
<005
< 005
and organisms (BLUMER et. al. 1971, BLUMER, GUILLARD and CHASE 1971,
YOUNGBLOOD et. al. 1971).
As was pointed out in the introduction, this method is capable of separating
polar compounds from non-polar, but it is not capable of distinguishing between non-polar compounds of different origins. Thus our quoted values
should be regarded as maximum concentrations of oil.
It is not unusual that one has to start up investigations in newfields,without
having the most suitable equipment available. We believe that those who have
access to an infrared photometer will be able to achieve good results in oil
pollution studies with a reasonable amount of effort, by applying this method.
We also regard this method as very suitable for monitoring, perhaps especially
in coastal areas. This is illustrated in Table 2 where some results from the
archipelago and harbour of Goteborg are shown. The spread of oil could
easily be followed. More details of that investigation will be published elsewere.
Those who plan to start investigations in this field, especially studies of
toxicity of oil to marine organisms, may find it advisable to apply a gas chromatography technique however.
ACKNOWLEDGEMENTS
Thanks are due to Dr. Stig H. FONSELIUS (Fishery Board of Sweden) for
valuable discussions and for stimulating us to write this paper.
This investigation has in part been supported by the Swedish Environment
Protection Board, Contract no 7-113/69 and 7-82/70, which is gratefully
acknowledged.
SUMMARY
1. The method involves extracting of oil and other hydrocarbons from water
with carbon tetrachloride.
2. The non-polar hydrocarbons are regarded as mineral oil. They are separated
from the polar hydrocarbons by means of column chromatography on
aluminium oxide.
3. The concentration of non-polar hydrocarbons is determined by measurement of infrared absorption.
4. Non-polar hydrocarbons of different origin (that means from mineral oil
and marine plants and organisms respectively) are not distinguished from
each other.
Downloaded from http://icesjms.oxfordjournals.org/ at Penn State University (Paterno Lib) on March 4, 2016
Rosenlund
(Harbour of Gdteborg)
Dynan
(Inner archipelago)
Depth
in m
Determination of oil in sea water
515
REFERENCES
BEYNON, L. R., KASHNITZ, R. & RIJNDERS, G. W. A., 1968. "Methods for the analysis of
oil in water and soil". Stichting CONCAWE, doc. (3408) pp. 1-V.4.
BLUMER, M., GUILLARD, R. R. L. & CHASE, T., 1971. "Hydrocarbons of marine phy-
to plankton". Mar. Biol., 8: 183-9.
BLUMER, M., SANDERS, H. L., GRASSLE, J. F. & HAMPSON, G. F., 1971. "A small oil spill".
Environment, 13: (2) 2-12.
BLUMER, M., SOUZA, G. & SASS, J., 1970. "Hydrocarbon pollution of edible shellfish by
an oil spill". Mar. Biol., 5: 195-202.
BRUNNOCK, J. V., DUCKWORTH, D. F. & STEPHENS, G. G., 1968. "Analysis of beach
pollutants". J. Inst. Petroleum, 54: 310-25.
GIBBONS, M., 1940. "Water pollution by petroleum oils". J. Am. Wat. Wks Ass., 32:
465-77.
GIEBLER, G., KOPPE, P. & KEMPF, H. T., 1964. "Methode zur Bestimmung von Heiz- und
Schmierolen in Wasser und Boden". Gas- u. WassFach, 105: 1039-42, 1093-97.
LINDGREN, C. G., 1957. "Measurement of small quantities of hydrocarbon in water". J.
Am. Wat. Wks Ass., 49: 55-62.
MELPOLDER, F. W., WARFIELD, C. W. & HEADINGTON, C. E., 1953. "Mass spectrometer
determination of volatile contaminants in water". Analyt. Chem., 25: 1453-56.
RAMSDALE, S. J. & WILKINSON, R. E., 1968. "Identification of petroleum sources of beach
pollution by gas liquid chromatography". J. Inst. Petroleum, 54: 326-32.
SCHOLL, F. & FUCHS, H., 1968. "Bestimmung von Mineralolspuren in Wasser". Bosch
Techn. berichte, 2: 235-44.
SIMARD, R. G., HASEGAWA, I., BANDARUK, W. & HEADINGTON, C. E., 1951. "Infrared
spectrophotometric determination of oils and phenols in water". Analyt. Chem. 23:
1384-87.
YOUNGBLOOD, W. W., BLUMER, M., GUILLARD, R. L. & FIORE, F., 1971. "Saturated and
unsaturated hydrocarbons in marine benthic algae". Mar. Biol., 8: 190-201.
Downloaded from http://icesjms.oxfordjournals.org/ at Penn State University (Paterno Lib) on March 4, 2016
5. The results obtained can thus be regarded as maximum concentrations of
mineral oil.
6. The calibration can be made with a standard mixture of 37-5% iso-octane,
37-5 % n-hexadecane and 25 % benzene.
7. With this calibration the accuracy of the photometric determination varies
with the type of oil that is present in the sample. In extreme cases it can be
as low as ± 26 %, but usually it is much better.
8. The method is especially suitable for monitoring purposes in pollution studies and is applicable down to 005 mg/1.