EVALUATION OF ORGANIC POLLUTANTS IN THE ~0PEN AIR AND

ATMOSPHERES IN ENDUSTRL4k SfTP.3 WXNG GRAPHETEED CARBON

BLACK TRAPS AND GAS CHROMAT0GRAPKEC-MASS SPECTR0METRTC
ANALYSIS ~WETH SPECXFIC DETECTORS

PAOLO CKCIOLI, GIULIANO BERTONE, ENZO BRANCALEOM, RINALDO FRATAR-

CANGJSL.1 and FABRIZIO BRUNER’
Laboraiorio sdPhquinamenro Attnosferiw ciei C.N.R., Via Monrorio Komano, 36-GilJ31 Rome {Jtai_v)

SUMMARY

A method for evaluating the organic content of the atmosphere, involving the
use of personal samplers, is described. A comparison of the performances of Tenax
GC and Carbopack B is given in terms of sample recovery. Apparatus including
selective columns, specific detectors and combined gas chromatography-mass spec-
trometry is also described.
Practical examples of the analysis of organic air pollutants trapped in the open
air and the atmosphere in a chemical plant are discussed. A comparison of the results
of compIete gas chromatography and of the total hydrocarbon content is also made.

INTRODUCXON

The evaluation of organic pollutants in air involves techniques of various

degrees of complexity, depending on the nature of the compounds involved, from the
total hydrocarbon content including methane (using a simple flame-ionization detec-
tor), through the measurement of organic substances other than methane, the deter-
mination of lower alkanes and alkenes, and finally to the individual analysis of single
organic substances.
The last type of measurement is essential when the situation in industrial sites
has to be evaluated, because one needs to know the health hazards that specific com-
pounds may have for persons continuousIy exposed to polluted atmospheres. Ln this
context, the use of “_personal samplers” has been proposed in order to obtaina better
knowledge of the nature and amounts of actuai and potential poisons it&aired in
particular environments such as chemical plants and laboratories and in industrial
sites where volatile organic substances have to be used.
On the other hand, the occurrence of photochemical oxidants in the open air
and the recognition of the important role played by reactive hydrocarbons and other

* To whom correspondence should be address.

75s - . P. CICCIOLI et al.

organic substances in the formation of photochemical smog make measurements of

speciiic air pollutants important for research purposes and for optimizing air pollution
control.
EIeeuse of the low concentrations of each component usually found in the
atmosphere, sample enrichment is necessary in both instances, and various workers
have sought the best enrichment procedures, ensuring that trapping is as complete as
possible and that there is complete release of the organic compounds from the traps.
Different approaches have been followed, mainly the use of cold traps at very low
temperatures1-6 and of traps working at room temperature7-13. Different kinds of
trapping materials have aIso been studied, such as porous polymers alone6-lo or
coated with liquid phases14**s,strong adsorbents such as active charcoal and successive
solvent extractionr’, or moderately strong adsorbents such as graphitized carbon
black’Z**3. However, cold traps obviously cannot be used in personal samplers, while
traps operated at room temperature are also much more useful for monitoring organic
pollutants in the open air, especially if automatic sampling and analysis is desired.
The main problem is to find an adsorbent that is capable of trapping the highest
number of substances, even though large volumes of air have to be sampled, and at
the same time that is able to release higher boiling compounds on heating at
moderately high temperatures.
With such severe limitations, both strong and weak adsorbents should be
excluded, and only intermediate adsorbents or porous polymers remain to be con-
sidered. From preliminary results, Tenax GC (Enka, Arnhem, The Netherlands) and
Carbopack 2 (Supelco, Bellefonte, Pa., USA.) appeared to be the best materials for
such purposes_

GENERAL CHARACI’EFUSTICS OF TENAX GC AND CAREOPACK B

Tenax GC (lg.6 mZ/g) is a porous polymer based on 2&-diphenyl-p-phenylene

oxide, the most important properties of which are its high operating temperature with-
out appreciable bleeding and the capability of eluting a large number of compounds
with good transmission of even high!y polar or basic molecules. However, the resolu-
tion of Tenax GC as a packing material for chromatographic columns is very low and
the seIectivity is also poor in comparison with other adsorbcnts. Water is readily
eluted from Tenax. The most interesting feature of Tenax, however, is its capability
of eluting many compounds of different natures and molecular weights in a relatively
short time. This porous polymer has been used mainly as a trapping material for
enriching air or water samples 16. It should be noted that, owing to its peculiar charac-
teristics, Tenax was chosen for the chromatographic coiumn in the Viking project”.
Carbopack B is a ,ggphitized carbon black adsorbent (90 m*/g) and has a large
variety of applications as a column material in gas chromato,oraphy. It has been used
in isotope separations’* and in the analysis ofsulphur compounds in air at the parts per
billion level=. When suitably modified with small amounts of liquid phases, it shows
a high selectivity for isomeric hydrocarbons* and complete transmission of alcohols,
carboXyhc acids and amineszo~*‘.Columns made of Carbopack B offer good perme-
ability and extremely high performance. The thermal stability of Carbopack B is
without hrnit in practice and it does not show any bleeding, even at extremely high
temperatures.
 EVALU.4TION OF ORGANIC POLLUTANTS 759

EXPERIMENTAL

Trapping procedure, measurement of breakthrough vohtnes and sample recovery

The traps are glass tubes, Iength 12 cm, I.D. 4 mm and Q.D. 6 mm, with a
g-cm portion occupied by the trapping material, plugged with glass-wool on each
side. The amount of material in each trap is 0.28 g of Tenax GC and 0.45 g of
Carbopack B. Carbopack B is in the size range SO-100 mesh and Tenax GC 60-50
mesh.
Trapping in the open air is carried out with an electrically operated pump. Its
pumping capacity is such that sampling with up to six traps simultaneously is pos-
sibIe, with a flow-rate in each trap of 30 ml/min. Sequential sampling is also possible
for monitoring; a scheme of the trapping system is shown in Fig. 1. The sampling
funnel is placed at about 10 m from the ground in the open air. A PTFE tube is
used to connect the traps to the funnel.

Fig. 1. Scheme of the trapping system. 1 = Traps; 2 = rotameters; 3 = needle valves; 4 = pump;
5 = flow counter.

Personal sampling is effected by using an Aircheck Personal Pump (SKC,

Pittsburgh, Pa., U.S.A., Model 222-351). This pump is equipped with a digital flow
meter and has a regulatable pumping capacity ranging from about 30 to 200 ml/min.
The dimensions of the trap are as described above. Breakthrough volumes of test
compounds are measured by injecting a known amount of the sample into the trap
at room temperature, allowing extra-pure nitrogen to pass through the trap at the
flow-rate desired until 0.5, 1.5 or 5.0 1 have passed through, then closing the trap and
heating it at 300” for 10 min.
A special four-way valve operating at temperatures up to 400” is used (Valco,
Houston, Texas, U.S.A.). Injection is made by passing the carrier gas through the
trap before entering the column. A scheme of the injection device that is also used
when the organic compounds from traps have to be analyzed is shown in Fig. 2.
The chromatograms are compared for quantitative evaluation with that ob-
taiued by injecting the test mixture directly through the trap at 300”. fn this way, one
knows the amount lost or irreversibly retained by the tiap. A Hewlett-Packard Model
33SOA digital integrator is used for calculations.

Ii is important to note that with the device shown in Fig. 2, no loss of chro-
matographic resolution is observed for compounds scarcely retained in the column,
as shown in Fig. 3.
Fig. 2. Iujection and deiection system.

Fig. 3. Chromztogsam obtGxd by injection through the mp, &owing t&hztdead vohun~ do not
aEect resolution. Peaks: 1 = met&me +- etltue; 2 = propane; 3 = isobutie; 4 = butene-i; 5 =
r-butze; 6 = irons-buteue-2.

A Carlo Erba Modei GI gas chr&ato,pph waspod%ed & the injection poti
as described and 2 four-way manifold wssnsed at the ~cohnn exit to divide tie gas
ctre2nr into three pitionS, so t’nat fflaqe-ionization, FID;-e&ctron-
three detectors
capture, ECD; 2nd f&me-photometric, FPD) co*uld &-Used, a&&own in Fig- 2. An
analogous apparatus w& used by McLeod eb al.“: Ams@ik~ ehromatograph;
_-
-EVALUATIN OF ORGANK POLLWA~ 761

eq&e~_ ~witb #hi sa& injection device,_ is coupled- to an AEI MSE2 mass
spectroieter with e FID placed in parallel according to the technique described
previo&Iyz3. The apparatus for trapping and analyzing low-boi&g compounds was
shewn in a.pzevious pap&_
The tot& organic content Iess methane is monitored using a~ automatic ap-
par&uCtinstructed in our laboratory- Et consists of a modification of the apparatus
d-&&bed e2sewherP and is shown schematically in Fig_ C One of the two detectors
is directly feel with ambient air, sampled at the Lx&ion of the traps, wbiIe the other
is placed at the end of a chromatographic coiumrr (stainless steel, 2.0 m long, 4 mm
I.D., 6 mm O_D., packed with Porzpak Q, 80--IO0 mesh) where the same sample of
air is injected every 5 min.

__._________
_________

Fig 4. Scheme of the appvatus for monitoring total hydrocrrrboncontent and methane in air.

The cohunn is operated at room temperature so that the methane peak is

separated from any other peak. Further, owing to the large diEerence in the concen-
tration& of methane and other single organic substances, it is possible to regulate the
electrometer attenu2ttio;r so as to detect the methane pe& only. From calibration
graphs, by subtracting the metbane concentration, that of the total organic conteentis
obtained.
A highly seIe&iVe cofumn was devised for the purposes of this work. The col-
u-mu shoufd be able-to work either at room temperature, yielding good separation
factors for light compounds, or at-higher temperatures to separate aromatic corn-
. pounds up to C16 CU. It should also be able to elute polar compounds.
The coh&~n chosen was made of stainless steel, L.80 m .long and 2 mm I.D.,
packed with Cqbopack C (Supelco), coated with 0.4% SPiW_ Such a stationary
phase, ba&d on the unique pro~rti& of gasiiq~uid-solid chromato,aEaphy, shows a
pronounced selectivity for isomeris hydrocarbons and, owing to the pokxity of
the
762 P. CKCIOLE et &I.

liquid modifier, gives symmetrical peaks for hydrogen-bonding compounds. It shouId

be noted that a polar liquid phase on graphitized carbon black enhances the separation
of isomers and decreases the retention times =. The percentage of SPIOOOis such that
the first monolayer is incomplete, so that the separation is best for the effect of lateral
interactions.

RESULTS AND DISCUSSION

Sample recovery
In Table I, the recoveries of various compounds attained with Carbopack B
and Tenax GC are compared. The choice of the compounds examined was made
with the aim of screening most of the types of organic compounds likely to be present
in air pollutants. The list of compounds is obviously incomplete, but gives a fair idea
of the trapping and releasing properties of the adsorbents considered.
There are two reasons why a particular compound injected in the trap could
be incompletely recovered : (i) the compound is hardly ad$orbed and is duted from the
trap by the air passing through it during the trapping step, or (ii) the compound is
strongly adsorbed and is not released when the trap is heated. Thus, there is a critical
region where the trap is most efikient for each adsorbent_ The results in Table I show

TABLE I
FLJZCOVERY (%) OF ORGANIC COMPOUNDS FROM _4IR SAMPLES USJNG CARBOPACK
B AND I-ENAX GC TRAPS

Carbopack B Tenax GC
0.5-Z 1.5-I 5.0-I 0.5-I 1.5-i 5.0-I
sample sample sample sampie Sample sample
-_
Methanol 3 0 0 1 0 0 -
Ethanol 3 0 0 1 0 0
Merhyl chlori lde 1 0 0 3 1 0
Acetone 5 1 0 68 2 0
Chloroform :z 80
54 501 100 84 5
Diethylxnine 80 50 1
Jsobutanol 100 100 25 100 95 16
rr-Pentane 100 100 100 loo 50 9
Cycfozz.xaIze 100 100 100 100 50 9
rt-Hexane 100 100 LOO 100 100 20
Ethyl acetate 100 100 Jo0 100 JIM 35
zr-Butanol 100 loo 100 loo 100 35
Benzene loo loo 100 100 loo 35
Toluene loo loo loo Jo0 100 Jca
- - - - - -
65 65 65 100 JeO 100
a6 a6 16 100 100 100
25 25 25 100 100 loo
8 s 8 100 Jo0 100
1 1 1 JO0 100 JGU
0 0 0 loo 100 100
--
* S-m, ethyJ&enzene. xyiae, pyridi=e, chlorophenol, dkanes, aJken= from C,Csz had JW%
recovery in any case.
that Carbop2ck B 2nd Tenax GC are compar2ble in this respect. However, benzene
2nd n-hexane are completely retained by C2rbop2ck B in the trapping step, but pass
through Tenax Gc; 2nd the same occurs for the ah&nes from n-pentane to isooctane.
It is worth noting that significant differences are observed for light compounds.
Acetone is recovered better from Tenuc GC, while for butanol 2nd diethylamine
Carbopack B is more efficient.
Such differences may be explained by the strong retention of hydrocarbons on
graphitized carbon black. For polar compounds, the specific active sites present on
the carbon black surface increase the retention. The small number of carbon atoms
and the Iow polarity are responsible for the low retention of acetone on C2rbopack
B. Tenax, which elutes amines linearly17, i.e. without appreciable irreversible ad-
sorption, 2hows these compounds to pass through the trap.
The same reasons that favour Carbopack B for the range C,-C, favour Tenax
GC if compounds above Cl2 are considered. However, it should be noted that such
compounds are seldom present in the open air or in industrial sites, as will be shown
later with examples of field measurements. They are, however, major constituents of
the organic fr2ction of atmospheric dust, but this is 2 different problem.
For the sake of comparison in the real use of the adsorbents, the following
experiment was carried out. Two traps of the same length and diameter, one of Tenax
GC followed by &.rbopack B according to the direction of the air stream, 2nd the
other of Carbopack B followed by Tenax GC, were used for par2Ilel air sampling, 1.5
I of air being sampled in each trap. During the trapping step, the compounds that
were not ret2ined by Tenax GC passed into the second part of the trap 2nd were
retained by the Carbopack B. In a similar manner, in the other trap compounds that
were not retained by Carbopack B were subsequently retained by Tenax GC.
The traps in tandem were then separated 2nd their organic contents injected
into the column, giving the four chromatograms shown in Fig. 5. Calculations were
made by considering the total peak areas of chromatograms A _t B and C + D as
10%. The two tot21 2reas were closely correlated, showing that the sampling 2nd
injection procedure is highly reproducible.
From this experiment, it was also found that Carbopack B behaves 2s a better
trapping 2nd releasing material than Tenax GC (83 T/,recovery compared with 75 %).
Only acetone, as expected from the results in Table I, is better retained by Tenax
GC, while cyclohexane and pentane are partly lost by Tenax GC and completely re-
tained by Carbopack B in the tr2pping step.
In addition, the chromatograms in Fi,.= 5 show that there is 2 discrepancy be-
tween the breakthrough volumes measured with the method described and the actual
retention and releasing properties of the adsorbents in field measurements. In fact,
both of the materials lose some of the C&r, fraction, 2nd this could not be’forecast
from the results in Table I. A possibfe explanation of this effect is that sampling is
carried out under extremely dilute conditions, which are completely different from
those which occur when breakthrough volumes are being measured_ In this instance,
the samples injected in the traps are much more concentrated.

Sampling md amdysis of atmosphetic p0i~zttant.s

A 5.0-1 volume of air from suburban open air &s sampled with a Carbopack
B trap at 2 flow-rate of 30 ml/loin. A cold trap4 was used at the same time for the
0 8 16 24 32 40 46 56

RECOVERED FROM CARBOPACK B

83%

LOST BY CARBOPACK 8
17%
- a&is of-e& and ffie totd non-methane
&lpof.mgs qgmic compounds @KC)
were dso mon%xed. The TEC trace obtainedis shown in E.g. 6 together with tit ob-
I%&& with a sampIefrom an urban site, ‘Fhearea within the broken line conesponds
to the-time-pertod when the traps were O&E&~&
The an&is wzs theti NQ usinngthe FPD, ECD and FID as parallel detectors

2-

f-

*.
‘*__
-*.
/’ .L_
, -_ .
em__ --I__ .. I
___-/
0 ’ , , I ,
2 4 6 8 fo 12 14 f6 18 20 22 :
t (haurs)

Fig. 6. Toid hydrcxzrbon content during t4 h monitored in typical urban (-) 2nd suburban
(---) 23x2s.

I 60 70
0 10 20 30 tim6 ‘+ (*in) So

Fig 7. Trip!e-detector chronz~+togmm showing the analysis of 2 s2mpIe.of air taken in suburban open
&nosphexe. Fe& Nos. aS in T26Te EL S2mpIing m;tM&: Gxbopack B.
and the resulting chromatogmm is shown in Fig. 7. Most peaks were identified by
MS. The potentiality of specitic detection is shown by the identification of benzo-
thiophene and halogenated compounds, which was elected on the basis of retention
times. The amount of these compounds is so low that when other compounds were
present in the same region of the chromatogram mass spectrometric identi&ation
was impossible.
Another feature of the chromatogram is that, in spite of its complexity, major
peaks corresponding to single compounds were obtained, showing a high selectivity.
This can also be seen from peaks Nos. 30, 31 and 32, which gave identical mass
spectra, and which were identified as the three xylenes from the retention times. The
separation achieved for the other aromatic compounds is also worthy of note. Only
two pairs of peaks overlapped.
Table 11lists the compounds identified, together with the approximate amounts
found. The accuracy of the determinations was 10-20 % based on the absolute amounts
present.
Parallel sampling and analysis of the light fraction were also ti& out, using
a technique described elsewhere 4. The analysis was effected with the triple detector
system and the qualitative and quantitative results are reported in Table III.
It is interesting to note the noticeable amount of methylene chloride found,
which indicates the importance of halogenated compounds in air pollution.
Table IV shows the results for the balance of the components. From a com-
parison of the data in Tables II and III, if can be seen that the C,-C, fraction is only
25 % of the total, showing that a complete analysis up to CL0 is needed for a reasonable
screening of air pollutants_ The aromatic fraction is present in about the same
percentage as the C,-C, hydrocarbons, in good agreement with the data available
from the literaturez6. The concentration of the heavier organic compounds is ob-
tained by subtracting the peaks correspondin, = to the light fraction from the total
integration of the chromatogram in Fig. 7.
The unidentified fraction is the sum of the minor peaks, present in low amounts,
for which mass specirometric identification was impossible. The fraction lost probably
corresponds in part to compounds heavier than ClO, but the discrepancy may also be
due to errors in the analytical measurements. However, repetitive measurements al-
ways gave a lower figure for &&e total amount found by gas chromatography with
respect to the direct THC values, showing that the first hypothesis is more feasible.
The non-hydrocarbon fraction comprises about 14% of the total organic com-
pounds, confirming the strong prevalence of petroleum products in organic air pol-
lutants.
No evidence for the presence of terpenes has been found; in fact, the suburban
area where sampling was carried out is a long distance from forest areas. The presence
of some ketones and halogenated compounds is probably due to some small factories
located l-2 km from the sampling point.

Carbopack B traps were used for samp@g organic substances in a chemical

plant where solvents are used in process syntheses. A 1-l volume of air was sampled
in the area of the plant where workers usually &end most of their working day, in
order to obtain a representative sample of inhitIed air. The sampling rate was ap-
 I 0.72 Butene-1 f cis-butene-2 + isobutane 25
2 1.13 R-Bu’a.ne T rrcas-buter,e-2 16
3 I.35 1,3-Butadiene 7
4 2.02 Methyl merczptzn 4
5 2.37 Aa;one i CMrCI,’ 16
6 3-42 C, unaturated hydrocubon 2
7 3.88 i-c, hydrocarbon 1
s 4.4& n-c, hydrocarbon 40
9 6.45 Cyclohexane 17
10 10.0 - so
.
11 11.20 5
12 11.80 Carbon tetrachloride’ 5
13 l4.66 Dietyl ketorze 11
14 15.64 Benzene 74
15 16.88 Cs unsaturzted hydrocarbon 33
16 1uO Trichloroethylene’ <1
17 19.54 n-C, hydrocarbon 8
18 25.03 Butanoi 6
19 27.63 Methylcyctohexane 19
20 27.99 C, unsaturated hydrocarbon 90
21 28.55 C, rmszturated hydrocarbon 21
22 29.06 Toluene 20
23 29.61 C, hydrocarbon 100
24 29.80 1,3-Dichloropropane’ 6
25 33.59 Dimethylcyclohexule 30
26 35.19 c, unsatura1ed 26
27 35.91 Ethyl butyi ketone I5
28 36.25 Styrene 13
W 3696 Ethylbenzene 21
30 37.50 nr-Xylene 11
31 37.98 p-Xylene + isopropylbenzene 41
32 38.44 a-Xyiene 3
33 39.81 - * 4
34 4x.a n-C, hydrocarbon 20
3s 42.07 Ethyltoluene 19
36 42.83 1,3,5-Trimethylbenzene 18
37 44-38 1,2&TrimethyIbenzene 54
38 45.46 1,2,3-Trimethylbenzene 22
39 46.05 Ethyidimethylbenzene 18
40 47.67 Methylisopropylbenzene 18
41 48-67 Benzothiophene’
42 50.77 1,2,3,5-Tetramethylbenzene 1:
43 51.57 1,2.3,4- 4 I ,2,4,5-Tetramethylbenne 4
44 52.59 - 18
4s 57.74 - 18
46 58.99 - 4
47 65.05 - 19

Totzl: 988
-

l Detected with specific detectors and identified with standards.

 _- I
=z+kBE Ez
_4NiiLYSlS OkLEGHFQRGA.tC COMFOT-MDS ~YKTEE OFtiN AZR (COLD l%CAP>
&i per;arsiderised %vi*sl- :

Perk No. Retentiin time(miizj Identijed as Rmoizts rppbcE&> . -

-1 3.00 Et&Z 25
2 3.23 EW 7
3 7.45 Propene 8
4 7.84 Propane 5
5 xoi Isobutane IO
6 35.08 n-Butane 35
7~ 32.12 1.3-Butadiene 10
8 37.0 rrans-Botene 5
9 47.45 DichIoro&ane’ 150
10 50.82 - 35
11 61.25 Z-Methylbutene 14
12 67.54 Isopentane IQ
13 77.70 n-Fentane 40
14 80.25 - 35
I.5 90.20 _- tl
Total: 376
_..
l Detected with ECD.

TABLE IV
BALANCE OF TKE COMPONENTS fN THE OPEN AIR
~-- -
Fraction Amount (ppb)

THC (Don-methane) 1550

Light orgznia (cold trzp) 375 (26%)
Heavier orgaks trapped 2nd recovered 1076(74%)
Tota! orgnics Qapped 1451
Unidentified (minor peaks} 336 (23.1%)
Lost 98 (6.3 Yd

i J
I 0 25 30
L

35
5 10 1.5 20 40
proximatfzly ICklm&nin_ !ZxticnZar emphasis was pptaced on a search fix a.rometic
solvents. The chromatogram obtained is shown in Fig. 8.and peak identification is
reporfed in Table V.
The GC-MS anafysis was made 2 h &er sampling, and a parallel sampling
was made with an identical trap, which was processed I we& later. The same chro-
tiatogram was obtained without sign&ant changes, which indicates that traps can
be stored prior to analysis witbout problems unless very reactive and unstable com-
pounds are present in the trap. The trap was closed at the terminals during storage.
The amount of acetone was estimated knowing that about 95% is lost during the
trapping step.

TABLE V
-ANALYSIS OF AIR SAMPLE FROM A CHEMICAL PLANT
Chroimtogmns in Fig. 7.
Peak No. Iderzifed as Amorolr (ppm)

1 ACetOne a50
2 c, unsaturated 1.0
3 C6 unsaturated 2.2
4 Ethyl isopropyl ketone 2.0
5 Diethyl ketone 3.6
6 Benzene 2.9
7 Hexane 0.7
8 Toluene 0.3
9 Styrene 2.6
10 Ethylbenzene 0.12

CONCLUSION

The results show that the analytical procedure involving the use of Carbopack
B or Tenax GC traps for samp!ing at room temperature ~SII be used to evaluate the
air quality with reliable results. It should be noted that the total amount of organic
pollutants unidentified is rcfatively high (23x), but also that the practical value of
this fi,m is limited, because it includes only very minor components.

REFERENCES

1 R E. Kaiser, Anal. Chem.,45 (1973) 965.

2 C_ Nicotm, A_ Corm, R. Mascot 2nd Perilhon, in S. G. Peny (Editor), GQSChromatography 1972,
Applied science Fubl., Barking, 1973.P. 9.
3 C. Nzrain. P. J. Marron 2nd J. N. Glover, in S. G. Perry (Editor), Gas Chromutogruphy 1972,
Applied Science Publ., Barking, 1973, p_ 1.
4 F. Biuner, P. Ckcioii ami F. di Nude, J. Chronatogr., 99 (1974) 661.
5 N. Bq&r uld L. S. Et&e, Anai. Ciem., 31 (1959) 1815.
6 B. Versino,M. de Groat uld E Geiss, &hrormtographia, 6 (1974) 302.
7 -A. Zfatkis, H. A. Lichtenstein2od A. Tishbee, ~km?QtographiQ, 6 (1973) 67.
8 W. Bertoh, R. C. Cbang and A. Zhtkis, J. Cimmu~ogr. Sci., 12 (1974) 173.
9 R. Je&es, Arrms_ Environ., 3 (1969) 587.
IO J, W. Russel, Emfro~. Sci. Teckml., 8 (1974) 143.
1E K. Grab ad G, Grab, J. cht-Otl?QfO~.., 62 (1971) 1.
12 A. Raymmd ‘and G. Guiochon, AmImis, 5 (X973) 357.
 _>:

16 B.
Versjno, M. de Gro& and Fi G&s Report of the C‘OST-Project 6&Z, European Gxrxmmity,
Brussels; 1975.
17 M. Novoz$, J. M. Efayes, F. &iner and P. G: Simmotids, X&e, 189 f197.$?15..
1s F. Bniner. C. Cmulli, A. di Con% znd A. Libeti, Nature Pkys_.ScL. 231 (1971).175.
19 F. Bnmer, P. Ciccioii uld F. di Nzrdo, Anal. Cirem;, 47 (f975)-141.
20 A. di Cmcia, A. Likrti and R %rriperi, Anal. Ckem, 45 (1973) 1228.
21 A. di Ccmiz and R !hrnperi, Anal_ Chem., 45 (1974) 977.
22 H. A. McLeod, A. G. Buttefie!d, D. Lewis, W. E. J. Philtipsaad D. k. Cohin, AnaL Ckan, 47
(1975) 674.
23 F. Bruner, P. Ciccioli and S. ZeUi, Anal_ Ckm., 45 (1973) 1002.
24 F. Brmer uld P. Ciccio!i, Proceedkgs of TECOMA P, HeLsinki, 30 Ju.I_vA A~&s.st,1973,WMO,
Geneva, 1974, p. 324.
25 F. Brmer, P. Ciccioli, G. ti. C_-hi and M. T. P&ok& And. Ckem, 4.5 (1973) 1851.
2G Air Qua&y Criteria for Hjdroc~rbons, NAPS& Ws&i~gton, D.C., I970.