Formation of nitrogenous compounds in the photooxidation of n-butane under atmospheric conditions

Article abstract of DOI:10.1007/PL00010127

The photooxidation of n-butane under atmospheric conditions in the presence of NO_x resulted in the formation of the following nitrogenous products: peroxy acetyl nitrate 23, sec-butyl nitrate 16, n-butyl nitrate 1.3, ethyl nitrate 1.3, peroxy n-butyryl nitrate 1.3, and peroxy propionyl nitrate 0.5% of the initially added odd nitrogen. In addition, an electron capturing compound eluting at the retention time of sec-propyl nitrate was also observed accounting for 5% of initial NO_x. Butan-2-one was the major product of conversion of n-butane with a yield of 37%. From product ratios it is evident that the formation of sec-butyl nitrate is favored over that of n-butyl nitrate by a factor of 2.1. The rate of reaction of sec-butoxy radicals with oxygen is equal to their decomposition rate.

Full text of DOI:10.1007/PL00010127

Monatshefte fuÈr Chemie 129, 1151±1159 (1998)
Formation of Nitrogenous Compounds in the
Photooxidation of n-Butane under Atmospheric
Conditions
Ã
Efthalia Evmorfopoulou and Sotirios Glavas
University of Patras, Department of Chemistry, EL-26500 Patras, Hellas
Summary. The photooxidation of n-butane under atmospheric conditions in the presence of NOx
resulted in the formation of the following nitrogenous products: peroxy acetyl nitrate 23, sec-butyl
nitrate 16, n-butyl nitrate 1.3, ethyl nitrate 1.3, peroxy n-butyryl nitrate 1.3, and peroxy propionyl
nitrate 0.5% of the initially added odd nitrogen. In addition, an electron capturing compound eluting
at the retention time of sec-propyl nitrate was also observed accounting for 5% of initial NOx.
Butan-2-one was the major product of conversion of n-butane with a yield of 37%. From product
ratios it is evident that the formation of sec-butyl nitrate is favored over that of n-butyl nitrate by a
factor of 2.1. The rate of reaction of sec-butoxy radicals with oxygen is equal to their decomposition
rate.
Keywords. n-Butane; Photooxidation; Product yield; PAN; PPN; PnBN; Alkyl nitrates.
Zur Bildung von Stickstoffverbindungen bei der Photooxidation von n-Butan unter
È
atmospharischen Bedingungen
È
Zusammenfassung. Die Photooxidation von n-Butan unter atmospharischen Bedingungen und in
Gegenwart von NOx fuÈhrt zur Bildung folgender stickstoffhaltiger Verbindungen: Peroxyacetylnitrat
23, sec-Butylnitrat 1.3, Ethylnitrat 1.3, Peroxy-n-butyrylnitrat 1.3 und Peroxypropionylnitrat 0.5%
È
des zugesetzten NOx. Zusatzlich wurde eine elektronenaf®ne Verbindung mit der Retentionszeit des
sec-Propylnitrats im Ausmaû von 5% bezogen auf NOx gefunden. Butanon-2 war mit einer Ausbeute
È
von 37% das Hauptprodukt. Aus den Produktverhaltnissen folgt, daû die Bildung von sec-Butylnitrat
gegenuÈber der von n-Butylnitrat um einem Faktor 2.1 bevorzugt wird. DaruÈber hinaus wurde
gefunden, daû die Reaktionsrate der sec-Butoxyradikale mit Sauerstoff gleich der Rate der
Zersetzung der sec-Butoxyradikale ist.
Introduction
The development of control strategies for the abatement of formation of photo-
oxidants, mostly ozone, in urban areas is usually carried out with the application of
an appropriate air quality model. A signi®cant component of such a model deals
Ã
Corresponding author

1152
E. Evmorfopoulou and S. Glavas
with the chemical transformation of ozone precursor emissions. The validation of
the chemical mechanisms of air pollution is usually performed utilizing smog
chamber data which can be obtained with hydrocarbon mixtures [1, 2, 3] or single
hydrocarbons [4, 5, 6].
n-Butane is a signi®cant component of urban [7] and rural [8] ambient air
hydrocarbons; in addition, it may serve as a surrogate of the saturated hydro-
carbons of the atmosphere in the photochemical mechanisms [2]. Although the
photooxidation mechanism of n-butane is rather well established, the nitrogen mass
balance is rather low [4]. The speciation of nitrogenous products is very important
because inactivation of NOx ( ˆ NO ‡ NO2) into compounds which remove NOx
from the ozone formation cycle would result in less ozone formation. The
nitrogenous compounds which serve as sinks of NOx are peroxy acyl nitrates (may
be considered as temporary sinks because of their thermal instability), alkyl
nitrates, and nitric acid. Three of the past studies of n-butane reported alkyl nitrate
yields [4, 9, 10], two reported yields of peroxy acetyl nitrate (PAN) [4, 11], and one
quanti®ed the formation of nitric acid [11]. In addition to the products reported in
past studies, two further peroxy acyl nitrates, peroxy n-butyryl nitrate (PnBN) and
peroxy propionyl nitrate (PPN), would be expected to be formed within the
currently acceptable mechanism of n-butane photooxidation. However, although
their precursor aldehydes, n-butanal and n-propanal, have been experimentally
measured [4, 5], PnBN and PPN were not reported.
In the present work, we have quanti®ed all the previously measured alkyl
nitrates. We report a previously not measured alkyl nitrate. In addition, we
determined PnBN and PPN. We also measured butanone, the major product of
transformation of n-butane, under atmospheric conditions. From the measured
product ratios we derive rate constant ratios important for the formation of alkyl
nitrates as well as to the decay steps of butoxy radicals.
Results and Discussion
Two sets of experiments were carried out with six experiments in each set. In the
®rst set the reactants were butane/NO/NO2, in the second one butane/NO2. The
total irradiation time was six hours. Within this time, ꢀ10% of butane reacted
when NO was present and ꢀ20% when NO was absent. Reactant concentrations
and product yields are given in Table 1. The main product was butanone; the
nitrogenous products in descending order of yield were: PAN, sec-butyl nitrate, a
compound eluting at the retention time of sec-propyl nitrate, PnBN, ethyl nitrate,
n-butyl nitrate, and PPN. PPN was observed only in the experiments carried out
with NO2 alone. Although the concentration of formed butanone was signi®cantly
higher in the experiments without initial NO, the resulting yields were the same
due to the larger consumption of n-butane. The remaining products were formed in
similar amounts in both sets, but their yields were signi®cantly different. The stated
uncertainty is Æ one standard deviation.
The only route of ambient air alkane degradation is via their reaction with
hydroxyl radicals. This occurs through H-atom abstraction: 85% from a secondary
and 15% from a primary H-atom [12]. The formed alkyl radical will add quickly to
an oxygen molecule, forming the respective alkyl peroxy radical. In the presence of

Photooxidation of n-Butane
1153

1154
E. Evmorfopoulou and S. Glavas
nitrogen monoxide, the alkyl peroxy radicals will react with NO and will either
oxidize NO to NO2 or will form the respective alkyl nitrate, as shown in Scheme 1.
CH3CH2CH2CH3 ‡ ÁOH ! CH3CHCH2CH3 ‡ H2O
ꢂ1†
ꢂ2†
Á
CH3CH2CH2CH3 ‡ ÁOH ! CH3CH2CH2CH2Á ‡ H2O
CH3CHCH2CH3 ‡ O ! CH3CHCH2CH3
Á
2
j
OOÁ
ꢂ3†
ꢂ4†
CH3CHCH2CH3 ‡ NO ! CH3CHCH2CH3 ‡ NO2
j
j
OOÁ
OÁ
CH3CHCH2CH3 ‡ NO ! CH3CHCH2CH3
j
j
OOÁ
ONO2
CH3CH2CH2CHÁ2 ‡ O2 ! CH3CH2CH2CH2OOÁ
CH3CH2CH2CH2OOÁ ‡ NO ! CH3CH2CH2CH2OÁ ‡ NO2
ꢂ5†
ꢂ6†
CH3CH2CH2CH2OOÁ ‡ NO ! CH3CH2CH2CH2ONO2
Scheme 1
Secondary alkyl nitrate versus primary alkyl nitrate
The ratio of concentrations of sec-butyl nitrate to n-butyl nitrate, obtained from
Table 1, is 11.7 for set 1 and 12.3 for set 2. These values are slightly lower than the
14.3 value previously reported [9]. Ignoring the rate of destruction of the alkyl
nitrate products by the hydroxyl radicals, the ratio of formed butyl nitrates is [sec-
butyl nitrate]/[n-butyl nitrate] ˆ ((k4/(k3‡k4))/(k6/(k5‡k6))) Á f1/f2, where ki are the
rate constants of the reactions shown in Scheme 1 and f1 and f2 are 0.85 and 0.15,
respectively. The correction of the products due to their reaction with hydroxyl
radicals [13, 14], carried out according to Atkinson et al. [9], amounted to only a
few percent; thus, no correction was applied. Substitution yields (k4/(k3‡k4))/(k6/
(k5‡k6)) ˆ 2.1 for data set 1 and ((k4/(k3‡k4))/(k6/(k5‡k6))) ˆ 2.2 for data set 2.
These values indicate that the formation of the secondary alkyl nitrate is favored
over that of the primary nitrate by a factor of 2.1±2.2, in agreement with the value
of 2.5 reported by Atkinson [15] for all secondary and primary nitrates as derived
from comparisons of experimental values of rate constant ratios of secondary alkyl
nitrate formation over the total reaction of the secondary alkyl peroxy radicals with
the respective values for the formation of primary alkyl nitrates [12].
Transformation of sec-butoxy radicals
The butoxy radicals formed in reactions (3) and (5) may react with oxygen,
decompose, or isomerize either by 1,4 or 1,5 H-shift. In the case of the sec-butoxy
radicals, the above steps are shown in Scheme 2.

Photooxidation of n-Butane
1155
CH3CHCH2CH3 ‡ O2 ! CH3COCH2CH3 ‡ HO2Á
ꢂ7†
j
OÁ
CH3CHCH2CH3 ! CH3CHÁ2 ‡ CH3CHO
ꢂ8†
ꢂ9†
j
OÁ
CH3CHCH2CH3 ! CHÁ3 ‡ CH3CH2CHO
j
OÁ
CH3CHCH2CH3 ! CH3CHCH2CH2Á
ꢂ10†
j
j
OÁ
OH
Scheme 2
Reaction (7) leads to butanone, the most abundant product of the photooxida-
tion of n-butane under atmospheric conditions. The obtained yield for all data was
0.37Æ0.02. This is similar to the 0.40 value obtained by Cox et al. [5], but signi-
®cantly larger than the 0.06 value of Carter et al. [4]. Butanone can react with OH
radicals by H-atom abstraction from the secondary hydrogens, eventually leading
to PAN. Biacetyl may or may not be formed according to Scheme 3 as proposed by
Cox et al. [5].
CH3COCH2CH3 ‡ ÁOHꢂO2† ! CH3COCHCH3
j
O2Á
CH3COCHCH3 ‡ NO ! CH3COCHCH3
j
j
O2Á
ONO2
CH3COCHCH3 ‡ NO ! CH CHCOCH3 ‡ NO2
j
3 j
O2Á
OÁ
CH3CˆO ! CH3CHO ‡ CH3CO
j
CH3CHOÁ
CH3CˆO ‡ O2 ! CH3CˆO ‡ HOÁ2 ! 2CH3COÁ
j
j
CH3CHOÁ
CH3CˆO
CH3CˆO ! 2CH3COÁ
j
CH3CˆO
CH3COÁ ‡ O2 ! CH3COO2Á ‡ NO2 ! CH3COO2NO2 ꢂPAN†
CH3CHO ‡ OH ! CH3COÁ !! PAN
Scheme 3

1156
E. Evmorfopoulou and S. Glavas
Reaction (8) leads to acetaldehyde which may form PAN according to the above
scheme and ethyl radicals which may form ethyl nitrate according to Scheme 4.
CH3CHÁ2 ‡ O2 ! CH3CH2O2Á
CH3CH2OÁ2 ‡ NO ! CH3CH2ONO2
CH3CH2OÁ2 ‡ NO ! CH3CH2OÁ ‡ NO2
CH3CH2OÁ ‡ O2 ! CH3CHO ‡ HOÁ2
ꢂ11†
ꢂ12†
Scheme 4
Reaction (9), also a decomposition like reaction (8), leads to methyl radicals
which ± following the analogous reaction scheme of ethyl radicals-lead almost
exclusively to formaldehyde, as formation of methyl nitrate is less favorable
because of the small size of the nitrate [16]. The formed n-propanal may, in
analogy to acetaldehyde, yield PPN. Although n-propanal has been observed in
past studies of n-butane [4, 5], PPN was not reported. PPN has been measured in
ambient air [17]. Thermochemical calculations have shown that decomposition by
reaction (8) is favored over reaction (9) by a factor of 160 [15]. Thus, all products
from reaction (9) will be negligible, and therefore contribution of this step to the
formation of PPN would also be negligible.
The 1,4 H-shift isomerization reaction (10) leads to the formation of hydroxy
alkyl radicals which would be expected to react similarly to alkyl radicals, although
at different rates. No product was detected from this path. Thermochemical
calculations have shown that a 1,4 H-shift is not favorable; the estimated rate
constant is 26 sꢁ1 [13]. Thus, the dominating steps would be reactions (7) and (8).
According to the above reaction schemes, the ratio of formed ethyl nitrate to
butanone would be equal to [Ethyl nitrate]/[Butanone] ˆ (k8/(k7[O2])) Á k11/
(k11+k12). The rate constant ratio k11/(k11+k12) was experimentally determined to
be 0.014 [9] and was calculated to be 0.01 [16]. Substitution of the value 0.008
(®rst set of data) yields k8/k7 ˆ 2.9Â1018 molecules/cm3 and of 0.02 (second set of
data) 8.1Â1018 molecules/cm3. These rate constant ratios are comparable to the
experimentally obtained ratios of 3.1Â1018 [4] and 2.6Â1018 molecules/cm3 [5],
but are signi®cantly larger than the 2.6Â1016 molecules/cm3 reported by Heiss
et al. [18] and 5±10 times larger than the thermochemically estimated ratio of
6Â1017. The thermochemical calculations, however, have an inherent uncertainty
of a factor of 5 [15]. Substitution of the k7 value of 7.5Â10ꢁ15 cm3 moleculeꢁ1 Á sꢁ1
[13] yields k8 ˆ (2.5±5.3)Â104 sꢁ1, a value larger by an order of magnitude than
the thermochemically calculated value [15]. Taking the mean value of the two sets
of k8 we deduce that reactions (7) and (8) occur in a ratio of 50:50. This result
along with the later derived k4/k3 ratio results in a yield of butanone of 0.40 which
is in excellent agreement with the experimental value given earlier.
The ratio of formed sec-butyl nitrate to butanone yields [sec-Butyl nitrate]/
[Butanone] ˆ k4/(k3 Á f7). Substitution of the experimental values of set 1 yields k4/
k3 ˆ 0.04, those of set 2 k4/k3 ˆ 0.07. The value of 0.093 calculated by the fall-off
expression of Carter and Atkinson [16] is similar to our derived values. The rate
constant ratio of 0.04±0.07 indicates that the yield of sec-butyl nitrate would be
0.03±0.06. As derived earlier, the formation of secondary alkyl nitrates is favored

Photooxidation of n-Butane
1157
over the primary ones by a factor of 2.1. Therefore, the total butyl nitrate yield in
our system would be 0.05±0.09, in agreement with the previous value of 0.077 [9].
Transformation of n-butoxy radicals
The n-butoxy radicals formed from reaction (5) may also react with oxygen,
decompose, and isomerize according to Scheme 5.
CH3CH2CH2CH2OÁ ‡ O2 ! CH3CH2CH2CHO ‡ HOÁ2
CH3CH2CH2CH2OÁ ! CH3CH2CHÁ2 ‡ HCHO
ꢂ13†
ꢂ14†
ꢂ15†
ꢂ16†
CH3CH2CH2CH2OÁ ! CHÁ3CHCH2CH2OH ꢂ1; 4 H-shift†
CH3CH2CH2CH2OÁ ! ÁCH3CH2CH2CH2OH ꢂ1; 5 H-shift†
Scheme 5
The n-butanal formed in reaction (13) was observed in the smog chamber study of
Carter et al. [4]. n-Butanal would be expected to react with hydroxyl radicals or
photolyze (the photolysis rate constant is unknown) to yield peroxy n-butyryl
nitrate, similarly to the reaction sequence giving PAN from acetaldehyde. PnBN,
however, like PPN was not observed in the previous smog chamber study of n-
butane [4] As far as we know, PnBN has never been reported in smog chamber
studies, although it has been measured in ambient air [17]. PnBN was observed in
the products with a 2.5 times higher yield than that of PPN.
The decomposition reaction (14) yields n-propyl radicals and formaldehyde.
The n-propyl radicals would be expected to react, as shown earlier for ethyl
radicals, to yield n-propyl nitrate and n-propanal, part of which may be converted
to PPN. However, n-propyl nitrate was not observed in the products. The only three
carbon alkyl nitrate observed was that eluting at the retention time of sec-propyl
nitrate, whose formation, however, cannot be explained with above reaction
schemes. PPN was observed only in the experiments in which the reactants did not
include NO. The presence of NO would destroy PPN through destruction of peroxy
propionyl radicals as was shown to be the case with other similar peroxy acyl
radicals [19, 20]. Thus, the small PPN precursor sources shown earlier to be
negligible for sec-butoxy radicals and those obtained from reaction (14), along
with the destruction of PPN by NO and possibly its coelution with sec-butyl
nitrate, a dif®culty which we have overcome with the use of high resolving power
capillary chromatography, are probably the reasons why PPN was not reported in
past photooxidation studies of n-butane.
If we ignore the formation of n-propyl nitrate, which anyway according to the
fall-off expression of Carter and Atkinson [16] accounts for 5% of the conversion
of n-propyl peroxy radicals with NO, and assume that the rates of formation of
PnBN from n-butanal and PPN from n-propyl radicals are the same, the product
ratio [PnBN]/[PPN] ˆ k13[O2]/k14 yields k14/k13 ˆ 2.0Â1018 molecules/cm3. Sub-
stitution of the k13 value 8Â10ꢁ15 cm3moleculeꢁ1 Á sꢁ1 [15] yields k14 ˆ 1.6Â
104 sꢁ1, a value 670 times larger than the thermochemically calculated one [15].
This difference, although it may be reduced by an order of magnitude due to the
uncertainty in the thermochemical calculations, is still too large to be accounted

1158
E. Evmorfopoulou and S. Glavas
for. No products were observed that could be attributed to the isomerization
reactions (15) and (16). Thermochemical calculations have shown, however, that
1,5 H-shift dominates over the 1,4 H-shift [15]. Thus, reactions (13) and (16)
would be most important.
At the end of the irradiation period, the sum of nitrogenous products of Table 1
accounts for 46±50% of the initially added odd nitrogen in both sets of experi-
ments. In addition at the end of each run on the average 30.4 ppb NOx (mostly
NO2, but also HONO and traces of NO) were left. Thus, the nitrogen mass balance
in these experiments was 74±78%. If we add to the above concentrations of
nitrogenous species the nitric acid concentration of 24 ppb, determined in
experiments carried out under the same conditions and with similar initial reactant
concentrations [11], the nitrogen mass balance makes up to 97±101% of the
initially added odd nitrogen, indicating that all signi®cant nitrogenous products
have been accounted for.
Experimental
All experiments were carried out in 200 A FEP te¯on chambers of ꢀ600 liters volume equipped with
a miniature te¯on coated fan. The Te¯on bags were housed in an aluminum box and were photolyzed
by a mixture of Philips lamps with maximum emissions of 310 and 365 nm. The temperature was
maintained at 25Æ2ꢃC by a commercial air conditioner.
Ambient air was puri®ed by an Aadco air puri®cation system, supplied with a methane reactor so
that the sum of hydrocarbons contained in the puri®ed air was less than 10 parts per billion (1:109)
carbon. The relative humidity of the dried puri®ed air was adjusted to 30±40% by passing part of the
bag ®lling air through pure distilled water. Photolysis of puri®ed air for six hours formed 5 ppb ozone
and zero PAN. NO2 was prepared daily in darkened glass ¯asks by oxidation of 99.85% nitrogen
monoxide, obtained from Messer Griesheim, with a large excess of medical grade oxygen, obtained
from Linde Hellas. n-Butane, 99% minimum stated purity, was obtained from Messer Griesheim and
was used without further puri®cation. GC/FID analysis of n-butane showed only one peak. PAN,
PPN, and PnBN were prepared by nitration of the respective peracids [21]. The alkyl nitrates were
prepared in the dark by addition of 1-propyl, 2-propyl, 1-butyl, and 2-butyl bromides obtained from
Aldrich to a saturated solution of AgNO3 in CH3CN at room temperature. NO and NOx were
determined by the ozone chemiluminescence method. The catalyst used to convert NOx into NO was
FeSO4 Á 7H2O. In separate experiments it was determined that this catalyst converts NO2, peroxy
acyl nitrates, and alkyl nitrates into NO but does not convert nitric acid.
PAN, PPN, PnBN, and the alkyl nitrates were initially separated on a 5 m HP-1 non polar 100%
methyl silicone gum column of 0.53 mm i.d. and were detected by GC/ECD. Because PAN coeluted
with 2-propyl nitrate and PPN with sec-butyl nitrate, quanti®cation of these compounds was carried
out by heating the ECD, in separate duplicate experiments, at 150ꢃC, a temperature at which we
determined that the peroxyacyl nitrates are thermally destroyed and not detected by the ECD,
whereas the alkyl nitrates are not affected. We thus found that on the average 95% of the peak of
PAN and sec-propyl nitrate was due to PAN, whereas only 10% of the PPN and sec-butyl nitrate peak
was due to PPN. In later experiments, the 5 m 0.53 mm i.d. column was replaced by a HP-1 15 m
0.32 mm i.d. one. As has been also shown previously [22], the longer and smaller diameter column
separated well the PPN from the sec-butyl nitrate and PAN from sec-propyl nitrate. The results from
the 0.53 mm column, after correction with the heated detector experiments, were similar to those of
the 0.32 mm column. The calibration of the ECD for the peroxy acyl nitrates was carried out after
their alkaline hydrolysis and determination of the formed NOꢁ2 using Dionex ion chromatography
[23]. n-Butane and butanone were analyzed by GC/FID using a HP-1 10 m 0.53 mm i.d. column.

Photooxidation of n-Butane
1159
Two Hewlett-Packard 5890 gas chromatographs were used. Reactants and products were directly
injected to GC using Precision Sampling gas tight syringes.
Acknowledgements
Financial support of the European Commission under contract EV4V-0070 (D) is gratefully
acknowledged.
References
[1] Lurmann FW, Lloyd AC, Atkinson R (1986) J Geophys Res 91: 10905
[2] Gery MW, Whitten GZ, Killus JP, Dodge MC (1989) J Geophys Res 94: 12925
[3] Carter WPL (1990) Atmos Environ 24A: 481
[4] Carter WPL, Lloyd AC, Sprung JL, Pitts JN Jr (1979) Int J Chem Kinet 11: 45
[5] Cox RA, Patrick KF, Chant SA (1981) Environ Sci Technol 15: 587
[6] Leone JA, Flagan RC, Grosjean D, Seinfeld JH (1985) Int J Chem Kinet 17: 177
[7] Moschonas N, Glavas S (1996) Atmos Environ 30: 2769
[8] Hagerman LM, Aneya VP, Lonneman WA (1997) Atmos Environ 31: 4017
[9] Atkinson R, Aschmann SM, Carter WPL, Winer AM, Pitts JN Jr (1982) J Phys Chem 86: 4563
[10] Atkinson R, Aschmann SM, Winer AM (1987) J Atmos Chem 5: 91
[11] Mineshos G, Roumelis N, Evmorfopoulou E, Glavas S (1991a) J Atmos Chem 13: 343
[12] Carter WPL, Atkinson R (1985) J Atmos Chem 3: 377
[13] Atkinson R (1989) J Phys Chem Ref Data, Monograph 1
[14] Atkinson R (1994) J Phys Chem Ref Data, Monograph 2
[15] Atkinson R (1990) Atmos Environ 24A: 1
[16] Carter WPL, Atkinson R (1989) J Atmos Chem 8: 165
[17] Grosjean D, Williams EL, Grosjean E (1993) Environ Sci Technol 27: 110
[18] Heiss A, Tardieu de Maleissye J, Viossat V, Sahetchian KA (1991) Int J Chem Kinet 23: 607
[19] Roumelis N, Glavas S (1992) Monatsh Chem 123: 63
[20] Grosjean D, Williams EL, Grosjean E (1994) Int J Chem Kinet 26: 381
[21] Gaffney JS, Fajer R, Senum GI (1984) Atmos Environ 18: 215
[22] Mineshos G, Roumelis N, Glavas S (1991 b) J Chromatography 541: 99
[23] Krognes T, Danalatos D, Glavas S, Collin P, Toupance G, Hollander JCTh, Scmitt R, Oyola P,
Romero R, Ciccioli P, Dipalo V, Cecinato A, Patier RF, Bomboi MT, Rudolph J, Mueller KP,
Schrimpf W, Libert Y, Geiss H (1996) Atmos Environ 30: 991
Received March 3, 1998. Accepted (revised) May 4, 1998