Journal of Atmospheric Chemistry 17: 353-373, 1993. © 1993 Kluwer Academic Publishers. Printed in the Netherlands.

353

Near UV Absorption Spectra and Photolysis Products of Difunctional Organic Nitrates: Possible Importance as NOx Reservoirs I A N B A R N E S , K A R L H. B E C K E R ,

and T O N G Z H U *

Bergische Universitdt- Gesamthochschule Wuppertal, Physikalische Chemie/Fachbereich 9, Gauflstr. 20, 42119 Wuppertal, Germany (Received: 5 November 1992; in final form: 13 May 1993)

Abstract. Difunctional organic nitrates are important products of the atmospheric reaction of NO3 radicals with unsaturated hydrocarbons about which relatively little is known. In a continuation of the investigation of the atmospheric chemistry of such compounds, the UV absorption spectra of the following organic dinitrates and keto nitrates have been quantitively measured in the gas phase at 298 _+2 K and atmospheric pressure: 1,2-propandiol dinitrate, CH3CH(ONO2)CH2(ONO2); 1,2-butandiol dinitrate, CH3CH2CH(ONOz)CH2(ONO2); 2,3-butandiol dinitrate, CH3CH(ONO2)CH(ONO2)CH3; cis 1,4-dinitrooxy-2-butene, CHz(ONO2)CH=CHCHz(ONO2); 3,4-dinitrooxy-l-butene, CH2(ONOzCH(ONOz)CH~CHz; a-nitrooxy acetone, CH3COCHz(ONO2); 1-nitrooxy-2-butanone, CH3CH2COCH2(ONO2); 3-nitrooxy-2-butanone, CH3CH(ONOz)COCH 3. Although the UV spectra of the nitrates are all very similar in shape those of the keto nitrates are red-shifted compared to the dinitrates and in the spectral range of atmospheric interest (2 > 290 nm) their absorption cross-sections are approximately a factor of 5 higher. The cross-sections of the dinitrates are a factor of 2 higher than those reported in the literature for the corresponding alkyl mononitrates. The UV absorption cross-sections of the difimctional nitrates were used in combination with solar actinic flux data to estimate photolysis frequencies and consequently atmospheric lifetimes for these compounds. The results indicate that for the saturated difunctional nitrates studied in this work photolysis will generally be somewhat some important than reaction with OH radicals as an atmospheric removal process. However, for unsaturated nitrates loss due to reaction with OH will dominate over photolysis as an atmospheric sink. Preliminary FT-IR analyses of the photolysis products of a-nitrooxy acetone, 3-nitrooxy-2-butanone and 2,3-butandiol dinitrate using both mercury and fluorescent lamps indicate that N O 2 is released in the primary step. The further reactions of the radicals thus produced result in the formation of CO, aldehydes and PAN. The possible significance of the results for difunctional organic nitrate as reservoirs for reactive odd nitrogen NO r in the atmosphere, especially during the night, is briefly discussed. Key words. UV absorption spectra, photolysis, organic difunctional nitrates.

1. I n t r o d u c t i o n

The oxides of nitrogen, NO x (NOx = N O + NOz) , are actively involved in tropospheric oxidation cycles which lead to ozone formation and acid precipitation. The * Present address: York University, 4700 Keele St., Toronto, Ontario, M3J 1P3, Canada.

354

IAN BARNES, KARL H. BECKER, AND TONG ZHU

major ultimate fate of NOx is conversion to HNO 3. However, a number of other processes result in the formation of a variety of organic nitrates (Atkinson and Lloyd, t984; Calvert and Madronich, 1987; Atkinson, 1990). The photochemical lifetime of NOx is relatively short, although, organic nitrates can act as temporary reservoirs for reactive nitrogen, enabling it to be transported over long distances and to influence the chemistry in remote regions (Crutzen, 1979; Singh and Hanst, 1981; Atherton, 1989). The organic nitrates include, in probable order of importance, PAN(CH3C(O)OONO 2 and higher analogues, simple alkyl nitrates (RONOa) and complex multifunctional nitrates (Buhr et al., 1990; Ridley et al., 1990; Roberts, 1990). In field measurements, deficiencies as high as 45% have been observed between total reactive nitrogen NOy (NOy -- NOx + N O 3 + N205 + HONO + HO2NOz + HNO 3 + PAN + aerosol nitrate (NO~) + other organic nitrates) measured using an Au/CO converter-equipped chemiluminescence instrument and the sum of [NO], [NO2] , [PAN], [NO?] and [HNO3] which are generally assumed to constitute the majority of all reactive nitrogen in the atmosphere (Atlas et al., 1992a; Buhr et al., 1990; Fahey et al., 1986). Model studies indicate that a considerable fraction of this so-called 'nitrogen shortfall' may well be due to unidentified organic nitrates (Stockwell, 1986; Calvert and Madronich, 1987; Atherton and Penner, 1988). A recent model study by Madronich and Calvert (1990) suggests that multifunctional nitrates could play a more important role compared to simple nitrates. Results from recent field measurements show that simple alkyl nitrates cannot account for the 'shortfall' observed in NOy (Atlas et al., 1992b). Of the organic nitrates, peroxyacetyl nitrate (CH3C(O)OONOa) is by far the most atmospherically abundant and contributes significantly to the total reactive nitrogen NOy budget (Buhr et al., 1990; Atherton and Penner, 1988). The atmospheric abundance of the other organic nitrates is not well known. In recent years, there have been reports of the detection of alkyl nitrates in ambient air (Josson and Berg, 1983; Jiitmer, 1988; Altas, 1988; Grosjean et al., 1990; Flocke et al., 1991) and circumstantial evidence has been reported for the presence of difunctional nitrates in carbonaceous aerosol (Schuetzle et al., 1975; Appel et al., 1980). Alkyl and acyl peroxynitrates are formed in the atmosphere by the addition of NO2 to alkyl and acyl peroxy radicals (Atkinson et al., 1989; Zabel et al., 1989; Tuazon et al., 1991). The reaction of alkyl peroxy radicals with NO has two channels: the major channel results in the formation of alkoxy radicals and NO 2 while the minor channel leads to the formation of alkyl nitrates (Atkinson et al., 1984; Becker and Wirtz, 1989; Roberts, 1990). The importance of the alkyl nitrate channel increases with increasing chain length of the alkyl moiety and decreasing temperature. In recent years it has been firmly established that the reaction of N O 3 radicals with unsaturated hydrocarbons results in the formation of aldehydes, ketones and difunctional nitrates (Akimoto et al., 1978; Hoshino et al., 1978; Bandow et al., 1980; Shepson et al., 1985; Barnes et al., 1990; Hjorth et al., 1990; Wayne et al., 1991). Depending on the structure of the hydrocarbon difunctional nitrates such

NEAR UV ABSORPTIONSPECTRA

355

as keto nitrates (RCOCHONO2R'; R and R ' = organic moieties), dinitrates (RCHONO2CHONO2R') and hydroxy nitrates (RCH(OH)CHONOzR') can be formed in significant yields. There is not a large body of quantitative data on the atmospheric removal processes for organic nitrates although reaction with OH radicals and photolysis are thought to be the most important sinks. The atmospheric chemistry of alkyl peroxynitrates and alkyl nitrates has been reviewed by Roberts (1990). For PAN and its higher analogues decomposition and photolysis are the main atmospheric loss processes; except for very low temperatures reaction with OH is too slow to be of any significance (Wallington et aL, t984; Tsalkani et al., 1988). For most alkyl nitrates, photodissociation is thought to dominate over reaction with OH radicals as atmospheric sink for these compounds; only for nitrates with C numbers >C 5 does loss due to reaction with OH start to become of significance (Gaffney et al., 1986; Becker and Wirtz, 1989; Roberts, 1990; Turberg et al., 1990). For organic difunctional nitrates, the only available information is the recent measurement of Henry's law coefficients (Kames and Schurath, 1992), rate coefficients for the reactions of OH with a number of keto nitrates and dinitrates measured in this laboratory (Zhu et at., 1991), and the ultra violet absorption spectra of a-nitrooxy acetone (CH3COCH2ONO2) and nitrooxy ethanol (HOCH2CHzONO2) reported by Roberts and Fajer (1989). In a continuation of the work in this laboratory on the atmospheric chemistry of difunctional nitrates, UV spectra have been recorded for 1,2-propandiol dinitrate, CH3CH(ONOe)CHz(ONO;); 1,2-butandiol dinitrate, CH3CH2CH(ONO2)CH2(ONO2); 2,3-butandiol dinitrate, CH3CH(ONO2)CH(ONOz)CH3; cis-l,4-dinitrooxy-2-butene, CHz(ONOa)CH~CHCHz(ONO2); 3,4-dinitrooxy-l-butene, CHz(ONOzCH(ONOz)CH~-CH2; a-nitrooxy acetone, CH3COCH~(ONO2); 1-nitrooxy-2-butanone, CH3CH2COCH2(ONO2); 3-nitrooxy-2-butanone, CH3CH(ONO2)COCH > Preliminary investigations have also been carried out on the products formed from the photolysis of some of the difunctional nitrates ha order to better assess the contribution of this type of compound to the total reactive nitrogen budget.

2. Experimental Near-UV absorption spectra of the nitrates were recorded in a cylindrical Duran glass vessel of 480-L volume (300 cm long, 45 cm i.d.) closed at both ends by Teflon® coated aluminum end flanges. The vessel is equipped with a built-in White mirror system which was usually operated at a total pathlength of 51.6 m in combination with a 22 cm spectrometer (SPEX) and a diode array detector (PAR 1412). The experimental set-up is shown schematically in Figure 1. A spectral resolution of 0.7 nm was achieved with a grating of 1200 line/ram covering a spectral width of 70 nm on the diode array detector. The exposure time for a single spectrum was typically 3 s. Three to five individual spectra were acquired and averaged for each compound for several different concentrations. The concentrations of the nitrates

356

IAN BARNES, KARL H. BECKER, AND TONG ZHU

VacuumS

Ga~Sampl

Fig. 1, Schematic diagram of the 480-L Duran glass reaction vessel.

were varied between 3 to 30 ~bar (N4.4 to 44 pmv) and the total pressure was 1013 mbar N 2. A grease free vacuum system was used to subject the organic nitrates to a freeze and thaw purification procedure at liquid nitrogen temperature before weighed amounts were flushed into the reactor by a stream of pure nitrogen. The difunctional organic nitrates were prepared using modified literature methods by treatment of an appropriate alcohol with strong acid ( H N O 3 with H2SO 4 and/or HNO 3 with acetic anhydride, A C 2 0 ) at ice-bath temperatures (McKay et at., 1948) or from an appropriate halogenated hydrocarbon with silver nitrate in acetonitrile (Fishbein and Gallaghan, 1956). The methods applied for the preparation of the individual nitrates can be found in a recent publication from this laboratory on the reactions of these compounds with OH radicals (Zhu et al., 1991). Since many of the nitrates decompose at or near their boiling points, they were purified by vacuum distillation. Their structures and purities were verified by FT-IR, NMR and GC-MS. In all cases the purities were always better than 96%. The concentration of the nitrate in the reactor was calculated on the basis that all of the weighed sample entered and remained in the gas phase. The procedure was performed for concentrations in the range 1-9 x 10 4 molecules cm -3 and at least three experiments were performed per concentration. Wall losses of the nitrates were rotund to be negligible on the timescale of the experiments. Only in the case of a-nitrooxy acetone was slow thermal decomposition observed. Absorption cross-sections a (cm 2 molecule-1) were calculated from the absorbance A, the gas phase concentration n (molecule cm -3) and the pathlength I (cm) according to the expression A = l n ( I o / I ) = nat. Preliminary studies on the photolysis products of the difi.mctional nitrates in 1013 mbar synthetic air at 298 _+2 K were carried out in a 420-L cylindrical glass reaction chamber with a built-in White mirror system for long-path IR absorption

357

NEAR UV ABSORPTION SPECTRA

measurements. The IR spectra were analyzed with a Fourier transform spectrometer (Nicolet 7199) at a spectral resolution of 1 c m -1. This experimental set-up has been described in detail elsewhere (Barnes et al., 1983). Both a low pressure mercury vapor lamp emitting mainly at 254 nm (Philips T U V 40) and 24 fluorescent lamps (Philips T1 4 0 W / 0 5 , 3 0 0 ~<2 ~<450 nm, )~ma~= 365 nm) were used as photolysis sources. 3. Results and Discussion

3.1. UVAbsorption Spectra The U V absorption spectra of the keto nitrates and dinitrates recorded in the range 2 4 0 - 3 4 0 nm are shown in Figure 2. The measured cross-sections in units of 10 .20 cm 2 molecule-: are given for 5 nm intervals in Table I. The spectra of all the dinitrates are very similar in form and, for the keto nitrates, the influence of the additional carbonyl chromophore on the shape of the spectrum is evident. Turberg et al. (1990) have reported that the U V spectra of the mononitrates are comprised of two Gaussian-shaped absorptions, one centered at about 260 nm with low intensity and the other at - 190 nm with high intensity. Although in this study the spectra

100

-

I

I

I

g

o

•

i

llli

i

b

g._=

[]o

°

I o

E

=

~ •

o (3 ~o

nitrooxy

[]

0% ¢n

1-nitrooxy.2-bu

D

"i~+

acetone

I

tanone

0

3-nltrooxy-2-butanone

x

1,2-propandlol

C, •

II~ +

I

O

8~

I

dlnitrate

•

0.1

-I-

1,2-butandlol

dlnltrate

t~

2,3-butandlol

dinitrate

0

1,4-dlnt trooxy-2-butene

•

3,4-dinitrooxy-l-butene

,

0.01 240

,

I 260

, '....,..

,

I., 280

-~

0

~ I A I

0

N

I

..-. + - - + ~ 1 t 300

320

340

360

nm

Fig. 2.

A b s o r p t i o n cross-sections of organic keto nitrates and dinitrates versus wavelength.

358

L~N BARNES, KARL H. BECKER, A N D TONG Z H U Table I. U V absorption cross-sections of difunctional organic nitrates in the region 2 4 5 - 3 4 0 nm at 5 nm intervals Wavelength (nm)

245 250 255 260 265 270 275 280 285 290 295 300 305 310 315 320 325 330 335 340

Absorption cross-section (10 -20 cm 2 molecule-i) a D1

D2

D3

D4

D5

K1

K2

K3

10.46 9.24 8,54 7.98 7.30 6.49 5.54 4,49 3.47 2.60 1.87 1.32 0.96 0,62 0.42 0.32 0.25 0.20 0.14 0-08

10.90 9.46 8,69 8.07 7.34 6.53 5.68 4.59 3,56 2.65 1.90 1.39 1.04 0.75 0.55 0.45 0.34 0.28 0.22 0.14

12.69 10.88 9.89 9.23 8.47 7,55 6.40 5.19 3.98 2.91 2.04 1.42 1.01 0.67 0.43 0.34 0.24 0.18 0.14 0.10

8.31 7.05 6.39 5.97 5.49 4,89 4.29 3.46 2.68 1.98 1.40 0.98 0.68 0.46 0.29 0.22 0.14 0.10 0,07 0,02

9.28 7.63 6.63 5.92 5.29 4,66 4,13 3.38 2.66 2.03 1.52 1.14 0.92 0.65 0,42 0.33 0.14 0.02

22.90 16.58 12.90 11.06 9.95 9.32 8.82 8,13 7,30 6,46 5,56 4.86 3.82 2.87 1.92 1,41 0.85 0.49 0.22 0.16

33.60 24.00 18.10 14.00 11.20 9,48 9.68 8.69 7.59 6.55 5.19 4.02 2.91 1,92 1,16 0.72 0.29 0.24 0.12 0.07

29.34 21.14 16.44 14.24 13.32 12.98 12,73 12.29 11.55 10.36 9.11 7.36 5.85 4.18 2.82 1.85 1.07 0.68 0.49 0.44

" D1 = 1,2-Propandiol dinitrate; D2 = 1,2-Butandiol dinitrate; D3 = 2,3-Butandiol dinitrate; D4 = 3,4-dinitrooxy-l-butene; D5 ~ 1,4-dinitrooxy-2-butene; K1 = a-Nitrooxyacetone; K2 = l-Nitrooxy-2-butanone; K3 = 3-Nitrooxy-2-butanone.

have only been recorded down to 240 nm the form of the spectra implies that this is also the case for the keto nitrates and dinitrates. The spectra in Figures 2 and 3 certainly show centers at about 260 nm and an apparent increase to a second center at shorter wavelengths. The low- and high-intensity absorptions have been assigned to n --, ~* and Jr -, Jr* transitions (Csizmadia et al., 1973). The spectra have no spectral features above 290 nm. The absorption spectra of the keto nitrates extend further into the near U V than those of the dinitrates and their absorption crosssections are also much higher which is probably caused by the presence of the additional carbonyl chromophore. The absorption spectra of organic nitrates in solution and in the gas phase can be fitted satisfactorily by the sum of several Gaussian curves (Csizmadia et al., 1973; Roberts and Fajer, 1989). Least-squares fits of the spectra measured in this work were performed in the range 2 >/250 nm using the equation: a0. ) -- exp(a22 + b2 + c). The measured values for the dinitrates lie slightly higher than the calculated curves at wavelengths above 320 nm. The results of these calculations are listed in Table II. In calculating the absorption cross-sections, some corrections had to be made

359

NEAR UV ABSORPTION SPECTRA

t

100

100

0---- nitr00xy acetone co

.

.

.

.

.

.

.

.

.

89)

• "

U ¢1) =(D O3

0

o

10

10 0

E

,--~E 0 o,

o

0~1

0.1 240

260

280

300

320

340

360

nm

Fig. 3. Comparison of the absorption cross-section of a-nitrooxy acetone measured in this work with that reported in the literature by Roberts and Fajer (1989).

Table II. Least-squares fits of the difunctional organic nitrate absorption spectra to equations of the form or(2) = exp(a2 z + b2 + c) Compound

a (x 104)

b

c

Wavelength range (nm)

Uncertainty"~

1,2-Propandiol dinitrate 1,2-Butandiot dinitrate 2,3-Butandiol dinitrate 3,4-Dinitrooxyo1-butene 1,4-Dinitrooxy-2-butene a-Nitrooxy acetone 1-Nitrooxy-2-butanone 3-Nitroo~3'-2-butanone

-5.990 -6.217 -5.740 -6.217 -5.432 -9.804 -10.i1 -10.44

0.2915 0.3025 0.2771 0.3025 0.2631 0.5404 0.5437 0.5780

-79.24 -80.41 -77.47 -80.41 -75.92 -118.3 -116.9 -123.5

250-320 250-320 250-320 250-340 250-325 270-340 270-340 270-330

25 25 25 25 80 80 80 20

a Relative uncertainty at 310 nm in%.

for baseline drift which, when taken in combination with the low"vapor pressures of the nitrates, has led to a relatively high uncertainty of the results at longer wavelengths. The uncertainties listed in Table II are relative uncertainties (1 c~) at 310 nm (center of the actinic region, 2 9 0 - 3 3 0 nm) which were calculated using a propaga-

360

IAN BARNES, KARL H. BECKER, AND TONG ZHU

tion of errors analysis in combination with the standard deviations about the mean of 3 or more measurements. 1,4-Dinitrooxy-2-butene and 1-nitrooxy-2-butanone have much higher uncertainties at 310 nm than the other nitrates which is mainly due to the very- low vapor pressures of these compounds. The relatively high thermal instability of a-nitrooxy acetone is the main reason for the high uncertainty for the measurement of this compound. For difunctional organic nitrates, only UV absorption spectra for a-nitrooxy acetone and nitrooxy ethanol have been reported in the literature (Roberts and Fajer, 1989). In Figure 3 the spectrum of a-nitroox3, acetone recorded in this study is compared to that reported by Roberts and Fajer (1989). Their spectrum agrees well with this work in the actinic UV region (290-330 nm), however, below 300 nm the spectra begin to diverge, the measured absorption in this work being much higher than that of Roberts and Fajer. At 300 nm, there is a difference of a factor of 1.25 in cross-sections which rises to a factor of 1.94 at 270 nm before falling back to 1.25 at 250 nm. It is presently not clear why the difference in this spectral region is so large. Possible contributions from thermal decomposition products such as HCHO, NO2 or PAN would not be expected to affect the form of the absorption in the region 300-270 nm. Turberg et al. (1990) have reported the absorption cross-sections of alkyl mononitrates. The dinitrates, 1,2-propandiol dinitrate and 1,2-butandiol dinitrate studied in this work have approximately double the absorption cross-sections of the related alkyl mononitrates, 1-propyl nitrate and 1-butyl nitrate measured by Turberg et al. (1990). 3.2. Photolysis Frequencies and Estimated Atmospheric Lifetimes

A number of studies on the photolysis of nitrate compounds have been reported in the literature and the results are briefly summarized (Gray and Style, 1953; Gray and Rogers, 1954; Rebbert, 1963; Maria et al., 1973; Johnston et aL, 1974; Taylor et aL, 1980; Renlund and Trott, 1984, Luke and Dickerson, 1988; Becker and Wirtz, 1989; Luke et aL, 1989; Roberts and Fajer, 1989; Turberg et aL, 1990). Johnston et aL (1974) measured a quantum yield of unity for nitric acid vapor dissociation between 200 and 315 nm. Gray and Rogers (1954) concluded that the photodissociation of methyl nitrate (CH~ONO2) by low-pressure Hg lamps preceded by Ct-I30--NO 2 bond fission but no quantum yield was reported. Taylor et al. (1980) measured the UV absorption cross-section of CH3ONOz and suggested that its photodissociation takes place in a similar fashion to pyrolysis, i.e. by breakage of the CH30--NO~ bond forming methoxy radicals (CH30) and nitrogen dioxide. However, no product analysis was made. Gray and Style (1953) found a quantum yield of 0.31 for the 253.7 and 265 nm Hg line photolysis of ethyl nitrate and concluded that CH3CH2ONO 2 + hv -~ CH3CH20 + NO 2

361

NEAR UV ABSORPTION SPECTRA

was the only reaction pathway. Rebbert (1963) has estimated the overall quantum yield of ethyl nitrate to be 0.5 and proposed that there are three photolysis pathways: CH3CHzONO 2 +

hv

~

CH3CH20 + NO2,

CH3CH2ONO 2 +

hv

~

CH3CHO + HONO,

CH3CH2ONO ~ +

hv

~

CH3CHzONO + O,

O >i 0.24, ® >i 0.09, O t> 0.14.

Recently, Luke and Dickerson (1988) have reported a unit quantum yield for the CH3CH20--NO 2 fisson pathway in the direct photolysis of ethyl nitrate under atmospheric conditions. In a later paper, photolysis frequencies consistent with a quantum yield of unity are reported for n-propyl, n-butyl and 2-butyl nitrates in addition to ethyl nitrate (Luke e t al., 1989). Roberts and Fajer (1989) reported that the direct measurements of ethyl nitrate photolysis frequencies made by Luke and Dickerson (1988) coupled with their absorption spectrum result in a quantmn yield close to unity for CH3CH20--NO 2 cleavage process at a temperature of 25 °C. Becket and Wirtz (1989) have investigated the photolysis of isopropyl nitrate under atmospheric conditions and found acetone to be the major product and concluded cleavage of the RO--NO2 bond was the main photolysis channel: CH3CH(ONO2)CH 3 +

hv ~

CH3CH(O)CH 3 + NO2,

CH3CH(O)CH3 + 02 -~ CH3COCH3 + HO2 CH3CH(O)CH 3 -* CH3CHO + CH3

(major pathway),

(minor pathway).

Since the RO--NO2 bond energy is approximately equal to the activation energy observed in the thermal decomposition of alkyl nitrates (171.5 kJ = 700 nm), Turberg e t al. (1990) have suggested that the photodissociation of organic nitrates probably occurs mainly by the breakage of the RO--NO2 bond at all wavelengths throughout the absorption spectrum. RONO 2 +

hv

~

RO + NO 2.

Since the UV spectra of the difunctional nitrates studied here, like the spectra of mono-substituted alkyl nitates reported in the literature, show little structure at long wavelength, quantum yields of unity have been assumed throughout the absorption spectra in order to calculate their photolysis frequencies. The photolysis frequencies were calculated using the following relationship: t"

j =

where j is the photolysis frequency (s-l), o(2) is the absorption cross-section as a function of wavelength (cm 2 molecule-I), 1(2) is the spherically integrated solar flux as a function of wavelength (photons cm 2 s-l), and 0 ( 2 ) is the quantum yield as a function of wavelength. Because O(2) was assumed to be unity, the calculated values of] represent upper liredts for the nitrates studied in this work.

362

b \ N B A R N E S , K A R L H. B E C K E R , A N D T O N G Z H U

The solar actinic flux as a function of wavelength, I(2), varies as a function of solar zenith angle, altitude, earth surface and cloud albedo. Demerjian et al. (1980) has estimated the actinic flux at the earth's surface as a function of wavelength and zenith angle within specified wavelength intervals for best estimate surface albedos. These values, tabulated for 5 nm intervals from 290 nm and above, were multiplied by the nitrate cross-sections calculated from least-squares fits for the central wavelength of the interval and summed over all intervals. For wavelengths outside the least-squares fit range (e.g. 330-335 nm for 1,2-propandiol dinitrate) median values calculated from the starting and end values of the intervals were used as the central wavelength values. Figure 4 shows the calculated photolysis frequencies j for the difunctional nitrates versus solar zenith angle at the earth's surface for best estimated surface albedos. According to tables of the solar zenith angle as a function of true solar time and month (Demerjian et al., 1980), the solar zenith angles for 40 + latitude at noon on 1 January and at noon on 1 July are 63 + and 16.9 °, respectively. The corresponding photolysis frequencies of the organic nitrates at these zenith angles were estimated from the relationships between j and zenith angles as shown in Figure 4. Photolysis frequencies for maximum (at 0 ° zenith angle), at noon on 1 Januai 3'

i

10 "4

+ - -

l

0

0

0

0

0 El +

0

0

0

D

El

D

+

+

+

~K

~K

~

II

•

II

I

10 .5

0

l

0

o 0

[] + ;K

[] -t~

o 0 t3 +

0 0 o

tl_

10 -6

II

acetone

0

nttrooxy

[3

1-nltrooxy2-bulanone

0

3-nltrooxy-2-butanone

x

0

| 0

•

0

dinllrate

×

1,2-propandlol

+

1,2-butandlol

dlnltrate

2,3-butandlol

dlnltrate

m

o

g. 10 -7

10 .8 -20

t

1,4-dlnltrooxy-2-butene

•

3,4-dlnltrooxy-l-bu

t

I

0

20

o

tene

f ..... 40

T

I

60

80

100

Angle

Fig. 4. Photolysis frequencies of organic nitrates versus solar zenith angle, calculated for the earth's surface for best estimated surface albedo, based on t h e actinic flux extimated by D e m e r j i a n et aL

0980).

363

N E A R UV ABSORPTION SPECTRA

Table III. Calculated photolysis frequencies ] at 40 ° latitude for noon 1 January and noon 1 July at 298 K for organic dif~nctional nitrates in comparison with estimated first-order rate losses for the nitrates due to reaction with typical noonthne OH radical concentrations. I~br comparison estimates of the globatly-averaged 24 h mean loss of the nitrates due to photolysis and reaction with OH radicals are also presented Nitrate

]~ (10-6 s -I)

1 January ] OH b (10 4 S-1)

1 July ] OH ~ (10 -6 S-I)

Global average ]° OH ~ (10-6 S-'I)

1,2-Propandiol dinitrate 1,2-Butandiol dinitrate 2,3-Butandiol dinitrate 3,4-Dinitrooxy- 1-butene 1,4-Dinitrooxy-2-butene a-Nitrooxy acetone 1-Nitrooxy-2~butanone 3-Nitrooxy-2-butanone

10.6 14.2 10.8 6.3 6.3 35.4 20.8 53.6

3.61 4.99 3.61 1.98 1.72 10.9 6.01 16.6

9.76 12.9 9.79 5.88 5.62 32.5 18.7 47.2

4.2 4.8 4.2 2.4 2.0 15 8.0 20

<0.31 1.70 1.07 10.10 15.10 <0.43 <0.91 1.27

<1.24 6.80 4.28 40.4 60.40
<0.3 1.3 0.8 7.8 11.6 <0.3 <0.7 0.98

a Maximum j-values for 0 ° zenith angle. b The first order OH loss rate was calculated fl-om k[OH]; values of [OH] of 4 x 106 and 1 x 106 molecules c m "3 were used as noomime values for summer and winter, respectively (Logan et aL, 1981); the rate coefficients k are from Zhu et al. (1991). Values of the phmolysis frequency for a zenith angle of 60 ° have been taken as being representative of a globally-averaged 24 h mean phototysis frequency (Luke and Dickerson, 1988). d The first order O H loss rate was calculated from k[OH]; a value of [OH] = 7.7 x 105 molecules cm ~3 was used for the globally-averaged 24 h mean O i l concentration (Crutzen and Zimmermann, 1991; Prinn et al., 1987); the rate coefficients k are from Zhu et al. (I991).

and at noon on 1 July are listed in Table Ill along with the estimated first-order rate losses for the nitrates due to reaction with OH. The first-order rate losses k were calculated for winter and summer using the relationship k = kr[OH]: O H concentrations of 1. x 106 and 4 x 106 molecule cm -3 have been used which are typical noontime values for a zenith angle of 40 ° for winter and summeb respectively (Logan et al., 1981), the rate coefficients k~ for the reactions of OH with the nitrates determined were taken from an earlier study from this laboratory (Zhu et at., 1991). Luke and Dickerson (1988) have reported that globally-averaged 24 h mean photolysis frequencies for simple alkyl nitrates can be approximated by the photolysis frequencies corresponding to a 60 ° zenith angle. Assuming that this approximation holds for the difunctionat nitrates studied here, values of the photolysis frequencies calculated for a zenith angle of 60 ° have been taken as representative of a 24 h global average and are also listed in Table III where they are compared with loss by reaction with OH radicals calculated using a globally-averaged 24 h mean value for [OH] of 7.7 x 105 molecules cm -3 (Crutzen and Zinmlermann, 1991, Prinn et al., 1987). The values in lhble III suggest that, in general, photolysis will probably be somewhat more important then reaction with O H radicals as a removal process for unsaturated nitrates in the troposphere for both smmner and winter conditions. For the unsaturated nitrates, 1,4-dinitrooxy-2-butene and 3,4-dinitrooxy-l-butene the first order reaction rate constants with O H are appreciably larger than the

364

IAN BARNES, KARL H. BECKER, AND TONG ZHU

photolysis frequencies in both winter and summer making reaction with OH radicals the dominant loss process for these compounds in the troposphere. Using the globally-averaged photolysis frequencies in Table III the following photolytic lifetimes (j-l) have been estimated for the nitrates: 1,2-propandiol dinitrate, 2.8d; 1,2-butandiol dinitrate, 2.4 d; 2,3-butandiol dinitrate, 2.8 d; 3,4-dinitrooxy-l-butene, 4.8 d; 1,4-dinitrooxy-2-butene, 5.8 d; a-nitrooxy acetone, 18.5 h; 1-nitrooxy-2-butanone, 1.5 d; 3-nitrooxy-2-butanone, 13.9 h. The photolysis lifetimes of the nitrates range between 14 h and 6 d. Although these lifetimes are relatively short, they may be sufficiently long, in the case of some of the saturated nitrates, to enable them to act as reservoirs for NOx and thus contribute to the long-range transport of NOx in the troposphere. A strong temperature dependence for the photolysis frequencies of simple alkyl nitrates has been reported (Luke and Dickerson, 1988; Lueke et al., 1989); a factor of approximately 2 increase in the rates was found between 279 and 299 K. Taking this into account along with the fact that the present measurements are for 298 K and that the mean surface temperature of the Earth is about 288 K, implies a further slowing of the rate of loss of the difunctional nitrates. It should be noted that in the above discussion of atmospheric lifetimes of difunctional nitrates, no reference has been made to loss of the compounds by wet deposition. The Henry's law coefficients recently determined by Kames and Schurath (1992) for difunctional nitrates are significantly higher than those of the alkyl mononitrates and suggests that washout may also be an appreciable atmospheric sink for difunctional nitrates, in particular hydroxy and keto nitrates.

3.3. Photolysis Products of Difunctional Nitrates The products formed in the photolysis of a-nitrooxy acetone, 3-nitrooxy2-butanone and 2,3-butandiol dinitrate in 1013 mbar of synthetic air using a mercury vapor lamp and 24 fluorescent lamps as light sources are listed in Table IV together with their estimated yields. Figure 5 shows the IR product spectrum for 3-nitrooxy-2-butanone after 10 min irradiation with the mercury lamp. The photolysis of a-nitrooxy acetone was only investigated using the fluorescent lamps and the observed products were HCHO, CO, CH3C(O)O2NO 2 (PAN), NO 2 and HONO. The carbon balance is about 88% and a considerable portion of the missing carbon is thought to be CO2, which cannot be accurately determined with the present chamber facility. The nitrogen balance accounts for 72% of the reacted nitrogen. Weak spectral features were also observed at about 2900, 1700, and 1400 cm -z which indicate the presence of methyl glyoxal ( C H g C O C H O ) . The 3-nitrooxy-2-butanone photolysis with the mercury lamp resulted in the formation of PAN and CH3CHO both with yields of N 80%. However, the yield of PAN using the fluorescent lamps was relatively small and the major products observed were CO, HCHO, CH3CHO and NO2. In the mercury lamp product spectrum absorptions were observed at 3438, 1733, 1426, 1367, and 1116 cm -1

365

NEAR UV ABSORPTION SPECTRA

Table IV. Measured and estimated product yields for the photolysis of some organic difunctional nitrates Nitrate

Products (% yield on a molar basis) Mercury lamp

Fluorescent lamp HCHO (41), CO (157) VAN (33), NO 2 (31)

a-Nitrooxy acetone (CH3COCH2ONOa) 3-Nitrooxy-2-butanone (CH3COCH(ONO2)CH3)

CO (21) CH3CHO (90) PAN (81), NOz (35)

CO (49), CO2 (131) HCHO (13), CHaCHO (36) PAN (5), NO2 (63) CH3COCOCH3

(9)

HONO (2, est) CO (38) HCHO (70) CH3CHO (53)

2,3-Butandiol dinitrate (CH3CH(ONO2)CH(ONO2)CH3)

CO (30, est) tICHO (32) CH3CHO (11, est) NO 2 (17, est) HONO (9, est)

N O 2 (167)

HONO (26)

0.1[5

t739 * PAN (CH3C(O)OONO2)

0.10 •

<

tl

1837 |1

l 0,05

.

11~,

1302

!I

.

795

NO2

0,00

L

4000

I

1

1

3600

3200

2800

I ............... .1

2400

2000

I

I

I

I

1600

1200

800

400

wavenumber (cm"1)

Fig. 5. IR product spectnlm of 3-nitrooxy-2-butanone in 1013 mbar synthetic air after i0 min irradiation at 254 nm in which residual 3-nitrooxy-2-butanone has been subtracted for clarity.

which have been attributed to 2,3-butanedione ( C H 3 C O C O C H 3 ) . There are also indications in the spectrum for the presence of low levels of 3-hydroxy-2-butanone (CH3COCH(OH)CH3). Quantitative determinations have not yet been made for these two compounds. CO, HCHO, CH3CHO, NO 2 and HONO were identified as products of both

366

IAN BARNES, KARL H. BECKER, AND TONG ZHU

the 254 nm and fluorescent lamp photolysis of 2,3-butandiol dinitrate. Because of the relatively slov¢ photolysis rate in comparison to loss to the reactor wall, the uncertainty of the product yield determination for the dinitrate using the fluorescent lamps as light source is as large as 100%. NO 2 is a major product in all of the systems studied. This leads to complications in the interpretation of the product data since NOJhydrocarbon photolysis systems are known to produce OH radicals in reaction chambers at concentrations of the order of 107 molecules cm-k H O N O was also observed as a product in the reaction systems and its photolysis above 300 nm will also result in the formation of OH radicals. Based on the rate coefficients for the reactions of the compounds with OH (Zhu et al., 1991), an OH radical concentration of 107 molecules cm -3 and the photolysis frequencies measured here it has been estimated that for the photolysis studies using the fluorescent lamps for a-nitrox37 acetone photolysis will dominate over reaction with OH in determining the product distribution, whereas for 3-nitrooxy-2-butanone and 2,3-butandiol dinitrate both processes could well be of importance and reaction with OH could even dominate. With the mercury light source photolysis is expected to dominate over reaction with OH radicals in all cases because of the high absorption cross-sections at 254 nm. As will be discussed below, the involvement of OH in the oxidation can change the product distribution as is particularly evident in the case of 3-nitrooxy-2-butanone. 3.4. Pathways Leading to Product Formation As discussed earlier, the photolysis of nitrates proceeds mainly through cleavage of the R O - N O 2 bond: R O - N O 2 + h v ~ RO + NO 2. Therefore, in the study on a-nitrooxy acetone using the fluorescent lamps where photolysis is expected to dozmnate over the slow reaction with OH (Zhu et al., 1991), the initial oxidation step will probably be CH3COCHzO-NO 2 + hv ~ CH3COCH20 + NO2, followed by C H 3 C O C H 2 0 + 0 2 ~ CHsCOCHO + HO2,

CH3COCH20 -" C H 3 C O + HCHO, CH3COCHO + h v --" CH3CO + HCO. The observed PAN can then be formed by the well-established mechanism (Finlayson-Pitts and Pitts, 1986): C H 3 C O + 02 ~ C H 3 C ( 0 ) O 2 , C H 3 C ( O ) O 2 + N O 2 + M ** C H 3 C ( O ) O a N O z + M.

NEAR UV ABSORPTION SPECTRA

367

There is evidence in the IR product spectrum for the presence of methyl glyoxal (CH3COCHO). However, the yield is low and an accurate determination was not possible. The low yield of this compound can possibly be explained by a combination of its photolysis and rapid reaction with OH radicals and would also account for the observed high yield of CO which is an expected oxidation product of this reaction (Plum etaI., 1983, Atkinson, 1986, 1990). In the 254 nm Hg line photolysis of 3-nitrooxy-2-bumnone the high yields of NO2 and the approximate 1 : 1 ratio of PAN and CH3CHO suggest that the main initial reaction is CH3COCH(O-NO2)CH

3+

hv (2 = 254 nm) -~ CH3COCH(O)CH 3 + N O 2

:followed by thermal decomposition of the energy rich CH3COCH(O)CH 3 radical to CH3CHO and CH3CO radicals the further reactions of which would result in the formation of PAN: CH3COCH(O)CH 3 ~ CH3CO + CHsCHO, CH3CO + 02 + M --" CH3C(O)O2 + M, C H 3 C ( O ) O 2 + N O 2 + M ~ C H 3 C ( O ) O 2 N O 2 + M.

Alternatively, the C H 3 C O C H ( O ) C H 3 radical could react with 02 to form 2,3-butanedione which could photolysis to CH3CO and eventually PAN CH3COCH(O)C/-I 3 + 02 --, CH3COCOCH 3 + HO z CH3COCOCH 3 +

hv(2 = 2 5 4 nm) ~ 2 C H 3 C O .

However, no evidence was found for the formation of 2,3-butandione in this system and also the photolysis of this compound would be too slow, on the timescale of the experiments, to account for the high yields of PAN. In contrast, in the fluorescent lamp photolysis ( 2 = 3 0 0 - 4 5 0 nm) of 3-nitrooxy-2-butanone, 2,3-butanedione has been tentatively identified as a product and the yield of PAN is very low. This suggests either a change in the photolysis mechanism at these longer wavelengths or the possible involvement of OH in the oxidation as discussed above. Certainly, the high yield of NO 2 shows that it is released during the oxidation. This can either occur directly via photolysis,

hv (A = 300-450 rim) CH3COCH(O)CH3 + NO2,

CH3COCH(O-NO2)CH

3+

CH3COCH(O)CH3 + O 2 --" CH3COCOCH3 + HO2 or indirectly as a result of the reaction with OH CH3COCH(ONO2)CH 3 + OH --, CH3COC(ONO2)CH 3 + H20, C H 3 C O C ( O N O 2 ) C H 3 + M ~ CH3COCOCH 3 + NO2 + M.

368

IAN BARNES, KARL H. BECKER, AND TONG ZHU

Both pathways can lead to the formation of 2,3-butanedione. The observation of HCHO, CH3CHO and PAN show that the pathway, CH3COCH(O)CH3 + M --, CH3CO + CH3CHO + M must also be occurring. Because of the presence of the carbonyl chromophore in the keto nitrates the photolytic cleavage of the RCO-CH(ONOz)R bond leading to RCO is also another possible pathway which must be considered, e.g. for a nitrooxyacetone and 3 -nitrooxy-2-butanone: CHsCO-CHz(ONO2) +

hv ~

CH3CO + CHz(ONO2) ,

CI-I2(ONO2) + M ~ HCHO + N O 2 + M,

CH3CO-CH(ONOz)CH 3 +

h v ---" CH3CO

+ CH3(ONO2)CH ,

CH(ONO2)CH 3 + M --" CH3CHO + NO 2 + M. The above reaction pathways can certainly occur in the 254 nm experiments, however, based on a comparison of the photolysis frequencies measured for glyoxal, methyl glyoxal and 2,3-butanedione at this wavelength with those for the keto nitrates, it is considered that rupture of the RCOCH(O-NO2)R probably dominates over RCO-CH(ONO2)R. It should be noted, however, that both photolysis pathways would result in high yields of PAN and CH3CHO. Because of the low"photolysis rate of 2,3-butandiol dinitrate with the fluorescent lamps and the associated high uncertainty in the product yields, it is not possible to derive reliable mechanistic information from these data. However, from the study using the mercury lamps, it is evident from the high NO 2 yield that both NOz groups are being released during the course of the photooxidation. The primary~ photolysis step is probably, CH3CH(O-NO2)CH(ONO2)CH 3 + + hv

(2 = 254 nm) -* CH3CH(O)CH(ONO2)CH 3 + NO2

followed by thermal decomposition, CH3CH(O)CH(ONO2)CH 3 + M ~ CH3CHO + CH3CH(ONO2) + M, CH3CH(ONO2) + M ~ CH3CHO + NO 2 + M, or reaction with O 2 to form 3-nitrooxy-2-butanone: CH3CH(O)CH(ONO2)CH 3 + O 2 ~ CH3COCH(ONO2)CH 3 + HO 2. The high observed NO 2 yield tends to suggest that the thermal decomposition pathway probably dominates. In the present work, it was not possible to positively establish whether or not 3-nitrooxy-2-butanone is being formed during the photolysis, because of the overlap of its IR absorption bands with those of 2,3-butandiol dinitrate. The conspicuous absence of PAN in this photolysis system suggests that

NEARUV ABSORPTIONSPECTRA

369

CH3CO is not produced. Therefore, it is likely that the HCHO and CO observed in the system is coming from the photolysis of CH3CHO, which is known to produce these compounds (Finlayson-Pitts, and Pitts, 1986), and not from its further reaction with OH radicals. H O N O which was observed in all the systems studied is probably being formed heterogeneously at the reactor wall. The formation of H O N O in chambers has been described in many papers (Carter et al., 1982; Sakamaki et al., 1983; Akimoto et al., 1987). However, Rebbert (1963) has proposed that photolysis of organic nitrates to form an aldehyde and H O N O is also a possible source: R'CH2-ONO 2 + h v -+ R'CHO + HONO. The present study is not extensive enough to allow any conclusions to be made as to whether photolysis or wall processes are the major source of HONO in the present work.

4. Conclusions The near UV absorption cross sections of organic difunctional nitrates which are important products of NO3/atkene reactions were determined for the first time and used in conjunction with solar flux data to estimate photolysis frequencies and atmospheric lifetimes of the nitrates. Summertime photolysis frequencies of between 5.6 x 10 .6 S - 1 (1,4-dinitrooxy-2-butene) and 4.7 × 10 -s s -~ (3-nitrooxy-2butanone) were obtained. A comparison of the photolysis frequencies and the estimated first-order toss of the nitrates due to reaction with OH radicals, suggests that removal by photolysis is somewhat more important than loss via reaction with OH radicals for the saturated nitrates. For unsaturated nitrates, however, toss due to reaction with OH radicals dominates over photolysis. Using photolysis frequencies calculated for summer conditions, lifetimes of between 14 h and 6 d have been estimated for the nitrates at 298 K. However, since the estimations represent only the upper limits for the photolysis of the organic nitrates and the photolysis frequencies are expected to be strongly temperature dependent, their photolytic lifetimes in the troposphere may be much longer. Using a mercury vapor lamp and 24 fluorescent lamps as light sources, it was shown that PAN, NO2, CO, CH3CHO and HCHO are the photolysis products of keto nitrates. For the dinitrate 2,3-butandiol dinitrate CO, HCHO, CH3CHO, H O N O and NOa were identified as photolysis products. Possible mechanisms for the formation of the products have been suggested. Since the nighttime reactions of NO3 with hydrocarbons are expected to be one of the major sources of difunctional nitrates, it can be expected that such nitrates will be transported considerable distances from their point of origin before daybreak. After sunrise, their photolysis and/or reaction with O H will produce PAN (or analogues) and its precursors NO2 and RCHO. Therefore, although the daytime oxidation of the difunction nitrates will release NO~., a considerable proportion may be reincorporated into another

370

IAN BARNES, KARL H. BECKER, AND TONG ZHU

reservoir, namely PAN. Since PAN is relatively stable, it can further contribute to the long transport of reactive nitrogen. Field measurements are required to determine the atmospheric concentrations of di- or multifunctional nitrates in the atmosphere. Further, detailed studies on the chemistry of this class of compound is required in order to better assess their contribution to the total reactive nitrogen budget and also their role in tropospheric chemistry.

Acknowledgement This work was financially supported by 'Der Minister fiir Umwelt, Raumordnung und Landwirtschaft des Landes Nordrhein-Westfalen, Forschungsprogram Luftverunreinigungen und Waldsch/iden'.

References Akimoto, H., Hoshino, M., Inoue, G., Sakamak, E, Bandow, H., and Okuda, M., 1978, Formation of propylene glycol 1,2-dinitrate in the photooxidafion of a propylene-nitrogen oxides-air system, J. Environ. Sci. Health A13, 677-686. Akimoto, H., Takaki, H., and Sakamak, E, 1987, Photoenhancement of the nitrous acid formation in the surface reaction of nitrogen dioxide and water vapor: extra radical source in smog chamber experiments, Int. J. Chem. Kinet. 19, 539-551. Appel, B. R., Wall, S.M., and Knights, R. L., 1980, Characterization of carbonaceous materials in atmospheric aerosols by high-resolution mass spectrometric thermal analysis, Adv. Environ. Sci. Teehnol. 9, 353-364. Atherton, C.S., 1989, Organic nitrates in remote marine environments: Evidence for long range transport, Geophys. Res. Lett. 16, 1289-1292. Atherton, C. S. and Penner, J. E., 1988, The transformation of nitrogen oxides in the polluted troposphere, Tellus Ser. B. 40,380-392. Atkinson, R. and Lloyd, A. C., 1984, Evaluation of kinetic and mechanistic data for modeling of photochemical smog, J. Phys. Chem. Ref Data 13,314-444. Atkinson, R., Aschmama, S. M., Carter, W. P. L., Wirier, A. M., and Pitts, J. N. Jr., 1984, Formation of alkyl nitrates from the reaction of branched and cyclic alkyl peroxy radicals with NO, Int. J. Chem. Kinetics 16, 1085-1101. Atkinson, R. and Aschmann, S. M., 1989, Rate constants for the reactions of the OH radical with the propyl and butyl nitrates and 1-nitrobutane at 298 _+2 K, Int. J. Chem. Kinetics 21,1123-1129. Atkinson, R., 1986, Kinetics and mechanisms of the gas-phase reactions of the hydroxyl radical with organic compounds under atmospheric conditions, Chem. Rm: 86, 69-201. Atkinson, R., 1990, Gas-phase tropospheric chemistry of organic compounds: A review, Atmos. Environ. 24A, 1-41. Atlas, E. L., 1988, Evidence for ~>C3 alkyl nitrates in rural and remote atmospheres, Nature 331, 426428. Atlas, E. L., Ridley, B. A., Hubler, G., Walega, J. G., Carroll, M. A., Montzka, D. D., Huebert, B. J., Norton, R. B., Grahek, E E., and Schauffier, S., 1992a, Partitioning and budget of NOy species during the Mauna Loa Observatory photochemistry, experiment, J. Geophys. Res. 97 (D10), 10,449-10,462. Atlas, E., Schanffler, S. M,, Merrill, J. T, Hahn, C. J., Ridley, B., Walega, J., Greenberg, J., Heidt, L., and Zimmerman, R, 1992b, Alkyl nitrate and selected halocarbon measurements at Manna Loa Observatory, Hawaii, J. Geophys. Res. 97 (D10), 10,331-10,348. Bandow, H., Okuda, M., and Akimoto, H., 1980, Mechanism of the gas-phase reactions of C3He and NO 3 radicals, Z Phys. Chem. 84, 3604-3608.

NEAR UV ABSORPTION SPECTRA

3 71

Barnes, I., Becker, K. H., Fink, E. H., Reimer, A., Zabel, E, and Nik, H., 1983, Rate constants and products of the reaction CS2 + OH in the presence of O~, Int. J. Chem. Kinet. 15, 631-645. Barnes, I., Bastian, V., Becker, K. H., and Zhu, T., 1990, Kinetics and products of the reactions of NO3 with monoalkenes, dialkenes and monoterpenes, J. Phys. Chem. 94, 2413-2419. Becker, K. H. and Wirtz, K., 1989, Gas phase reactions of alkyt nitrates with hydroxyl radicals under tropospheric conditions in comparison with photolysis, J. Atmos. Chem. 9, 419-433. Buhr, M. R, Parrish, D. D., Norton, R. B., Fehsenfeld, E C., and Sievers, R. E., 1990, Contribution of organic nitrates to the total reactive nitrogen budget at a rural eastern U.S. site, J. Geophys. Res. 95 (DT), 9809-9816. Calvert, J. G. and Madronich, S., 1987, Theoretical study of the initial products of the atmospheric oxidation of hydrocarbons, J. Geophys. Res. 92 (D2), 2211-2220. Carter, W. R L., Atkinson, R., Winer, A. M., and Pitts, J. N. Jr., 1982, Experimental investigation of chamber-dependent radical sources, Int. J. Chem. Kinet. 14, 1071-1103. Crutzen, E J., 1979, The role of NO and NO 2 chemistry of the troposphere and stratosphere, Ann. Rev. Earth Planet. Sci. 7, 443-472. Crutzen, E J. and Zimmermann, R H., 1991, The changing photochemistry of the troposphere, Tellus 43 AB, 136-151. Csizmadia, V. M., Houlden, S. A., Koves, G. J, Boggs, J. M., and Csizmadia, I. G., 1973, The stereochemistry and ultraviolet spectra of simple nitrate esters, J. Org. Chem. 38, 2281-2287. Demerjian, K. L., Schere, K. L., and Peterson, J. T., 1980, Theoretical estimates of actinic (spherically integrated) flux and photolytic rate constants of atmospheric species in the lower troposphere, Adw Environ. Sci. Technol. 10, 369-459. Fahey, D. W., Hiibler, G., Parrish, D. D., Williams, E. J., Norton, R. B., Ridley, B. A., Singh, H. B., Liu, S. C., and Fehsenfeld, E C, 1986, Reactive nitrogen species in the troposphere: Measurements of NO, NO2, HNO3, particulate nitrate, peroxyaeetyl nitrate (PAN), 03, and total reactive odd nitrogen (NOy) at Niwot Ridge, Colorado, J. Geophys. Res. 91, 9781-9793. Fishbein, L. and Gallaghan, J. A., 1956, The preparation of cis- and trans-l,4-dinitroxy-2-butene, J. Am. Chem. Soc. 78, 1218. Finlayson-Pitts, B. J. and Pitts, J. N. Jr., 1986, Atmospheric Chemistry." Fundamentals and Experimental Techniques, Wiley, New York. Flocke, E, Volz-Thomas, A., and Kley, D., 1991, Measurements of alkyl nitrates in rural and polluted air masses, Atmos. Environ. 25A, 1951-1960. Gaffney, J. S., Fajer, R., Senum, G. L., and Lee, J. H., 1986, Measurement of the reactivity of OH with methyl nitrate: impfication for prediction of alkyl nitrate-OH reaction rates, Int: J. Chem. Kinetics 18,399-407. Gray, E and Rogers, G. T., 1954, The explosion and decomposition of methyl nitrate in the gas phase, Trans. Faraday Soc. 50, 28-36. Gray, P. and Style, D. W. G., 1953, The photolysis of ethyl nitrate, Trans. Faraday Soc~ 49, 52-57. Grosjean, D., Parmar, S. S., and Williams II, E. L , 1990, Southern California air quality study: A search for methyl nitrate, Atmos. Environ. 24A, 1207-1210. Hjorth, J., Lohse, C, Nielson, C. L, Skov, H., and Restelli, G., 1990, Products and mechanisms of gas phase reactions between NO 3 and a series of alkenes, J. Phys. Chem. 94, 7494-7500. Hoshino, M., Ogata, T., Akimoto, H., Inoue, G., Sakamaki, E, and Okuda, M., 1978, Gas phase reaction of N205 with propylene, Chem. Lett., 1367-1370. Johnston, H. S., Chang, S.-G., and Whitten, G., 1974, Photolysis of nitric acid vapor, J. Phys. Chem. 78, 1-7. Josson, A. and Berg, S., 1983, Determination of low-molecule-weight oxygenated hydrocarbons in ambient air by cryogenic sampling and two-dimensional gas chromatography, Z Chromatogr. 279, 307-322. Jiittner, E, 1988, A cryotrap technique for the quantitation of monoterpenes in humid and ozone-rich forest air, J. Chromatogr. 442, 157-163. Kames, J. and Schurath, U., 1992, Alkyl nitrates and bifunctional nitrates of atmospheric interest: Henry's law constants and their temperature dependencies, J. Atmos. Chem. 15, 79-95. Logan, J. A., Prather, M. J., Wofsy, S. C., and McElroy, M. J., 1981, Tropospheric chemistry: a global perspective, J. Geophys. Res. 86, 7210-7254.

372

IAN BARNES, KARL H. BECKER, AND TONG ZHU

Luke, W. T. and Dickerson, R. R., 1988, Direct measurements of the photolysis rate coefficient of ethyl nitrate, Geophys. Res. Lett. 15, 1181-1t84. Luke, W. T., Dickerson, R. R., and Nunnermacker, L. J., 1989, Direct measurements of the photolysis rate coefficients and Henry's law constants of several alkyl nitrates, J. Geophys. Res. 94, D12, 14,905-14,921. Madronich, S. and Calvert, J. G., 1990, Permutation reactions of organic radicals in the troposphere, J. Geophys. Res. 95, 5697-5717. Maria, H. J., McDonal, J. R., and McGlyIm, S. E, t973, Electronic absorption spectrum of nitrate ion and boron trihalides, J. Am. Chem. Soc. 95, 1050-1056. McKay, A. E, Meen, R. H., and Wright, G. E, 1948, levo-2,3-Dinitroxybutane, J. Am. Chem. Soc. 70, 430. Plum, C. N., Sanhueza, E., Atkinson, R., Carter, W: E L, and Pitts, J. N. Jr., 1983, OH radical rate constants and photolysis rates of a-dicarbonyls, Environ. Sci. Technol. 17, 479-484. Prinn, R., Curmotd, D., Rasmussen, R., Simmonds, R, Alyea, E, Crawford, A., Fraser, R, and Rosen, R., 1987, Atmospheric trends in methylchloroform and the global average for the hydroxyl radical, Science 238, 945-950. Rebbert, R. E., 1963, Primary processes in the photolysis of ethyl nitrate, Z Phys. Chem. 67, 19231925. Renlund, A. M. and Trott, W. M., 1984, ArF laserAnduced decomposition of simple energetic molecules, Chem. Phys. Lett. 107, 555-560. Ridley, B. A., Shetter, J. D., Walega, J. G., MadroItich, S., Elsworth, C. M., Grahek, E E., Fehsenfeld, E C., Norton, R. B., Parrish, D. D., Hiibler, G., Buhr, M., Williams, E. J., Allwine, E. J., and Westberg, H. H., 1990, The behaviour of some organic nitrates at Boulder and Niwot Ridge, Colorado, J. Geophys. Res. 95 D9, 13,949-13,961. Roberts, J. M., 1990, The atmospheric chemistry of organic nitrates, Atmos. Environ. 24A, 243-287. Roberts, L M. and Fajer, R. W., 1989, UV absorption cross sections of organic nitrates of potential atmospheric importance and estimation of atmospheric lifetimes, Environ. Sci. 7Ochnol. 23, 945951. Sakamaki, F., Okuda, M, Akimoto, H., and Yamazaki, It., 1982, Computer modelling study of photochemical ozone formation in the propene-nitrogen oxides-dry air system, generalized maximum ozone isopleth, Environ. Sci. TechnoI. 16, 45-52. Schuetzle, D., Crorm, D., Crittendon, A. L, and Chartson, R.J., 1975, Molecular composition of secondary aerosol and its possible origin, Environ. Sci. Technol. 9, 838-845. Shepson, P. B., Edney; E. O., Kleindienst, T. E., Pittman, J. H., Namie, G. R., and Cupitt L T., 1985, The production of organic nitrates from hydroxyl and nitrate radical reaction with propylene, Environ. Sci. TechnoL 19, 849-854. Shagh, H.B. and Hanst, E L., 1981, Peroxyacetyl nitrate (PAN) in the unpolluted atmosphere: an important reservoir for nitrogen oxides, Geophys. Res. Lett. 8, 941-944. Stockwetl, W. R., 1986, A homogeneous gas phase mechanism for use in a reNonal acid deposition model, Atmos. Environ. 20, 1615-1632. Taylor, W. D., Allston, T. D., Moscato, M. J., Fazekas, G. B., Kozlowski, R., and Takacs, G. A. (1980), Atmospheric photodissociation lifetimes for nitromethane, methyl nitrite, and methyl nitrate, Int. Z Chem. Kinet. 12,231-240. Tsalkani, N., Mellould, A., Poulet, G., Toupance, G., and Le Bras, G., 1988, Rate constant measurement for the reactions of OH and C1 with peroxyacetyl nitrate at 298 K, Z Atmos. Chem. 7, 409419. Turberg, M. E, Giolando, D. M., Tilt, C., Soper, T., Mason, S., Davis, M., Klingensmith, P., and Takacs, G.A., 1990, Atmospheric photochemistry of alkyl nitrates, Z Photochern. Photobiology A: Chemistry 51, 281-292. Tuazon, E. C., Carter, W: R L. and Atkinson, R., 1991, Thermal decomposition of peroxyacetyl nitrate and reactions of acetylperoxy radicals with NO and NO a over the temperature range 283-313K, J. Phys. Chem. 95, 2434-243Z Wallington, T. J., Atldnson, R., and Wirier, A. M., 1984, Rate constants for the gas-phase reaction of OH radicals with peroxyacetyl nitrate (PAN) at 273 and 297 K, Geophys. Res. Letters 11, 861-864. Wayne, R., Barnes, L, Biggs, P., Burrows, J. P., Canosa-Mas, C. E., Hjorth, J., Le Bras, G., Moortgat, G.,

NEAR UV ABSORPTION SPECTRA

373

Perner, D., Poule, G., Restelli, G., and Sidebottom, H., 1991, The Nitrate Radical: Physics, chemistry and the atmosphere, Atmos. Environ. 25A, 1-206. Zabel, E, Reimer, A., Becker, K. H., and Fink, E. H., 1989, Thermal decomposition of alkyl peroxynitrates, J. Phys. Chem. 93, 5500-5507. Zhu, T., Barnes, I., and Becker, K. H., 1991, Relative-rate study of the gas-phase reaction of hydroxy radicals with difunctional organic nitrates at 298 K and atmospheric pressure, J. Atmos. Chem. 13, 301-311.