Kinetic study of the vapour-phase reaction between

0 downloads 0 Views 239KB Size Report
series of aldehydes is needed to develop a reliable structure» reactivity relationship and to get a better understanding of atmospheric chemistry in general.
View Article Online / Journal Homepage / Table of Contents for this issue

Kinetic study of the vapour-phase reaction between aliphatic aldehydes and the nitrate radical Barbara DÏAnna and Claus J. Nielsen* Department of Chemistry, University of Oslo, PO Box 1033 Blindern, N-0315 Oslo, Norway

Published on 01 January 1997. Downloaded on 21/04/2014 03:01:34.

Vapour-phase rate coefficients for the reaction of the nitrate radical with seven aliphatic aldehydes have been determined at 298 ^ 2 K and 1013 ^ 10 mbar by the relative rate method using FTIR detection. The obtained rate coefficients, in units of 10~14 cm3 molecule~1 s~1 and with 3p statistical errors, are : propanal, 0.57 ^ 0.04 ; butanal, 1.09 ^ 0.08 ; pentanal, 1.46 ^ 0.09 ; hexanal, 1.73 ^ 0.18 ; 2-methylpropanal, 1.21 ^ 0.06 ; 2,2-dimethylpropanal, 2.29 ^ 0.09 ; and 3,3-dimethylbutanal, 2.00 ^ 0.14. The existing rate coefficients for NO and OH reactions with aldehydes are compared and it is suggested that the reaction 3 between NO and aldehydes is not a simple hydrogen abstraction. 3

Introduction Aldehydes play an important role in the chemistry of the polluted troposphere. They are emitted as primary pollutants from partial oxidation of hydrocarbon fuels, and they arise as secondary pollutants from the oxidation of volatile organic compounds (VOC). The aldehydes may either photolyse or react further with OH during the day-time or with NO 3 during the night-time. The NO initiated degradation is not 3 the prime loss process for aldehydes in the atmosphere. However, kinetic data for the NO reactions with a larger 3 series of aldehydes is needed to develop a reliable structureÈ reactivity relationship and to get a better understanding of atmospheric chemistry in general. The present study addresses aldehyde reactions with NO 3 and we submit new reaction rate coefficients, determined by the “ relative rate Ï method, for seven simple C ÈC aldehydes. 3 6 Several kinetic studies have been presented of the NO reac3 tions with formaldehyde1h5 and acetaldehyde.1,3,6h8 Only a few investigations have addressed larger aldehydes, i.e. 3methylbutanal,9 hexanal,3 cis-4-acetyl-2,2-dimethylcyclobutylethanal (pinonaldehyde),9h12 acrolein,3,13 crotonaldehyde,13 methacrolein14 and benzaldehyde.3,15,16 Recently, kinetic data has also been presented for two C aldehydes : 3-(110 methylethenyl)-6-oxo-heptanal (endolim)10 and 2,2-dimethyl3-(2-oxopropyl)cyclopropaneacetaldehyde (caronealdehyde).11,12 Arrhenius parameters for the vapour-phase reaction of NO with aldehydes are only available for acetal3 dehyde, k(T ) \ (1.44 ^ 0.18) ] 10~12exp[[(1860 ^ 300)/T ] cm3 molecule~1 s~1.8 No systematic product study of the reaction between aldehydes and NO has so far been 3 published.

and used without further puriÐcation. Typical reactant volume fractions, / were : aldehydes and reference compounds, 2È10 ppm ; N O , 15È30 ppm. 2 5 The stability of the reference compounds and the aldehydes in the reaction chamber was investigated separately. The compounds showed lifetimes in the order of days and wall loss could thus be neglected in the data analyses.

Results Fig. 1 shows the expected mechanism of the NO initiated 3 degradation of aldehydes (here propanal) in smog chamber experiments. The initial step in the NO reaction with ali3 phatic aldehydes is believed to be abstraction of the H-atom from the aldehyde group. The following steps will eventually lead to acetaldehyde as the Ðrst “ stable Ï product. In addition, the NO reactions with the reference compounds, propene 3

Experimental Experiments were performed in puriÐed air at 298 ^ 2 K and 1013 ^ 10 mbar in a 250 l electropolished stainless-steel reactor equipped with a White-type multiple reÑection mirror system of 120 m optical pathlength for on-line FTIR detection. IR spectra were recorded with a Bruker IFS 88 employing a nominal resolution of 0.5 cm~1, HappÈGenzel apodization and adding 100 scans ; the time of registration was ca. 60 s. The NO radicals were produced in situ by 3 thermal dissociation of N O , which was synthesised from O 2 5 3 and NO and puriÐed by vacuum distillation prior to its use. 2 The aldehydes and reference compounds, propene and but-1ene, were all commercial samples with purity better than 98%

Fig. 1 Mechanism for NO initiated degradation of propanal in smog chamber experiments 3

J. Chem. Soc., Faraday T rans., 1997, 93(19), 3479È3483

3479

View Article Online

and but-1-ene, lead to varying amounts of formaldehyde and acetaldehyde or propanal. Thus, during reaction of a C alden hyde in the smog chamber one may, in the worst case, experience varying amounts of additional formaldehyde, acetaldehyde or propanal, and a C aldehyde. Admittedly, n~1 IR spectroscopy is not the best analytical method to quantify a mixture of aldehydes, and overlapping bands may lead to serious systematic errors. Fig. 2 shows the spectral region 2875È2600 cm~1 of the aldehydes studied and including formaldehyde and acetaldehyde. As can be seen, the IR bands of the C ÈC aldehydes display very little rotational Ðne struc3 6 ture, but their band shapes di†er sufficiently to allow quantitative characterisation of formaldehyde, acetaldehyde and any two of the C ÈC aldehydes in a mixture. 3 6

As can be seen in Fig. 1, the acylperoxy radicals formed after the initial H abstraction enter an equilibrium with ald NO to form the metastable acylperoxynitrates. Because the 2 NO, RO and NO concentrations are very low in the reac2 3 tion chamber, the acylperoxy radicals primarily react with NO . It is, in principle, possible to block completely the reac2 tions following, and thereby prevent information of the C n~1 aldehyde, by adding additional NO to the reaction mixture. 2 This, on the other hand, has the disadvantage that the NO 3 concentration (NO ] NO H N O ) and thereby the reac3 2 2 5 tion rate is lowered. We have carried out several experiments both with and without addition of NO to the initial reaction 2 mixture ; we found very small di†erences in the subsequently derived rate coefficients (see later). We therefore believe that the possible interference of aldehydes resulting from the initial reaction is negligible. The NO reaction rate coefficients were determined by the 3 relative rate method, that is by considering two simultaneous bimolecular reactions with the rate coefficients k and k : ref ald

Published on 01 January 1997. Downloaded on 21/04/2014 03:01:34.

kref reference ] NO ÈÈÈÕ products 3

(I)

kald aldehyde ] NO ÈÈÈÕ products (II) 3 Assuming that there are no other loss processes than the above two bimolecular reactions, then the following relation is obtained :

G

Fig. 2 IR spectra of the CwH stretching region, 2870È2600 cm~1, ald the seven aldehydes studied of formaldehyde, acetaldehyde and

H

G H

[ald] [ref] k 0 \ ald ln 0 (1) [ald] [ref] k t t ref where [ald] , [ald] , [ref] and [ref] denote the concentra0 t 0 t tions of the aldehyde and the reference at times zero and t, respectively. A plot of lnM[ald] /[ald] N vs. lnM[ref] /[ref] N will 0 t 0 t give the relative reaction rate coefficient k \ k /k as the rel ald ref slope. The ratio between the concentrations of the aldehydes and the reference compounds was found by spectral subtraction using spectra of the pure starting compounds and spectra of other compounds identiÐed in the reaction mixture. Table 1 summarizes the spectral regions used in the quantitative work as well as the compounds included in the subtraction procedures. Propene was only used as reference compound in the studies of propanal ; in all other experiments but-1-ene was used as the reference compound. The recommended rate coefficients for the NO reaction at 298 K with 3 these compounds are currently 9.45 ] 10~15 and 1.35 ] 10~14 cm3 molecule~1 s~1, respectively.17,18 Several experiments were carried out for each aldehyde. The data from the independent experiments were analysed jointly according to eqn. (1), that is the regression curves are “ forced Ï through the origin, and the analyses are shown in Fig. 3. Table 2 summarizes the results obtained in the present study ; the error limits quoted correspond to 3p from the statistical ln

Table 1 Wavenumber regions and additional reference compounds used in the spectral subtraction procedures

compound

wavenumber region used/ cm~1

additional compounds used in spectral subtraction formaldehyde, acetaldehyde formaldehyde, propanal formaldehyde, propanal, butanal formaldehyde, propanal, pentanal formaldehyde, propanal formaldehyde, propanal formaldehyde, propanol, 2,2-dimethylpropanal N O N2O5 2 5

propanala butanalb pentanalb hexanalb 2-methylpropanalb 2,2-dimethylpropanalb 3,3-dimethylbutanalb

CH wCH wCHO CH3wCH2wCH wCHO CH3wCH2wCH2wCH wCHO CH3wCH2wCH2wCH2wCH wCHO 2 2 2 2 CH3wCH(CH )wCHO CH3wC(CH ) 3wCHO CH3wC(CH3)2wCH wCHO 3 32 2

2775È2675 2750È2650 2750È2670 2750È2670 2750È2680 2770È2660 2760È2670

propene but-1-ene

CH wCHxCH 2 CH3wCH wCHxCH 3 2 2

3150È3050 3150È3050

a Propene used as reference compound. b But-1-ene used as reference compound.

3480

J. Chem. Soc., Faraday T rans., 1997, V ol. 93

Published on 01 January 1997. Downloaded on 21/04/2014 03:01:34.

View Article Online

Fig. 3 Plots of lnM[ald] /[ald] N vs. lnM[ref] /[ref] N for the decays of aldehydes and reference compounds during reactions with NO . Di†erent 0 t 0 tquoted are the 3p statistical errors. (A) Propanal : 37 data points from Ðve experiments 3 symbols indicate independent experiments. Errors were Ðtted to give k \ 0.61 ^ 0.05. (B) Butanal : 52 data points from six experiments were Ðtted to give k \ 0.81 ^ 0.06. (C) Pentanal : 40 data rel rel points from four experiments were Ðtted to give k \ 1.08 ^ 0.07. (D) Hexanal : 36 data points from six experiments were Ðtted to give k \ 1.28 rel six experiments were Ðtted to give k \ 0.90 ^ 0.04. (F) 2,2-Dimethylpropanal :rel39 data ^ 0.14. (E) 2-Methylpropanal : 40 data points from points from three experiments were Ðtted to give k \ 1.70 ^ 0.07. (G) 3,3-Dimethylbutanalrel: 34 data points from four experiments were Ðtted to rel give k \ 1.49 ^ 0.11. rel

analyses and include neither uncertainties in the rate coefficients of the reference compounds nor systematic errors from the experiment. Table 2 also includes other available rate coefÐcients for the reactions of OH and NO radicals with alde3 hydes. Of the seven aldehydes studied here, only data for hexanal has been reported previously in a conference paper by Carlier et al.3 However, they o†ered very little experimental information and it is with this in mind that we state that our rate coefficients are in fair agreement with the values given. In our study the random “ experimental Ï errors arise from temperature Ñuctuations and from the quantitative determination of relative reactant concentrations by spectral subtraction. These errors are reÑected in the statistical error from the leastsquares analysis and amount to ca. ^10%. In addition, we may have systematic errors due to competing reactions in the

smog chamber involving HO or OH radicals arising from 2 HO , Fig. 1. A conservative estimate of the inÑuence of such 2 systematic errors in the derived rate coefficients is an additional ^10%. Thus, the derived relative rate coefficients are believed to be accurate to within ^20%.

Discussion The vapour-phase reactivity of the nitrate radical towards organic molecules spans several orders of magnitude. Typical values for alkanes are 10~17 cm3 molecule~1 s~1 and for aldehydes they are of the order of 10~14 cm3 molecule~1 s~1, a thousand times faster.17,18 Rate data for hydrogen abstraction reactions of NO have been analysed as a function 3 of the bond dissociation energy of the broken bond in the substrate.19 On the basis of the CwH bond dissociation ald J. Chem. Soc., Faraday T rans., 1997, V ol. 93

3481

View Article Online

Table 2 Rate coefficients (cm3 molecule~1 s~1) for the reactions of aldehydes with NO and OH radicals at 298 K 3 k compound 1 2 3 4 5 6

formaldehyde acetaldehyde propanal butanal pentanal hexanal

7 2-methylpropanal 8 2,2-dimethylpropanal 9 3-methylbutanal 10 3,3-dimethylbutanal 11 pinonaldehyde

Published on 01 January 1997. Downloaded on 21/04/2014 03:01:34.

12 caronealdehyde 13 endolim 14 benzaldehyde 15

acrolein

16 crotonaldehyde 17 methacrolein

k

NO3

OH

]10~14

ref.

]10~11

ref.

0.058 ^ 0.005a 0.29 ^ 0.10a 0.57 ^ 0.04b,c 1.09 ^ 0.08d 1.46 ^ 0.09d 1.73 ^ 0.18d 1.28 ^ 0.08 1.21 ^ 0.06d 2.29 ^ 0.09d 1.2 ^ 0.3d

17 17 this work this work this work this work 3 this work this work 9

0.90 ^ 0.27a 1.62 ^ 0.32a 1.96 ^ 0.48a 2.53 ^ 0.06e 2.8 ^ 0.4e

20 20 20 21 21

2.9 ^ 0.6e 2.2 ^ 0.6e 4.0 ^ 0.7 2.89 ^ 0.09e

21 21 9 21

2.00 ^ 0.14d 5.4 ^ 1.8d 2.35 ^ 0.37b 2.71 ^ 10.15 26 ^ 8 0.37 ^ 0.07 0.255 ^ 0.008b O0.96b 0.81 ^ 0.01 0.111 ^ 0.017e 0.512 ^ 0.017b 0.44 ^ 0.17b

this work 9 12 12 10 3 15 16 3 13 13 14

9.3 ^ 1.5 8.7 ^ 1.1 12.1 ^ 3.6 11 ^ 3 1.3 ^ 0.33a

9 12 12 10 20

1.96 ^ 0.5a

20

3.6 ^ 0.9a 3.1 ^ 0.7a

20 20

a Recommended values. b Measured relative to propene, k \ 9.45 ] 10~15 cm3 molecule~1 s~1. c Uncertainties quoted for results from this 3 work represent three standard deviations from the statisticalNOanalysis. d Measured relative to but-1-ene, k \ 1.35 ] 10~14 cm3 molecule~1 s~1. NO3 e Measured relative to ethene, k \ 2.05 ] 10~16 cm3 molecule~1 s~1and k \ 8.52 ] 10~12 cm3 molecule~1 s~1. NO3 OH

energy and those of primary, secondary and tertiary CwH bonds in alkanes, it is obvious that NO , as OH, reacts pref3 erentially with aliphatic aldehydes through H abstraction. ald The position of CwH stretching bands is often associated with the strength of the bond ; the higher the vibrational wavenumber, the stronger the bond. In most aliphatic aldehydes, the aldehyde CwH stretching region contains two bands between 2850 and 2670 cm~1.22,23 This doublet is explained by an interaction of the CwH stretch fundamental with the overtone of the CwH bending vibration near 1390 cm~1. Considering the rate constants in Table 2 and the spectra of the CwH stretching region shown in Fig. 2, it is obvious that any attempt to correlate the reactivity of an aldehyde with the apparent position of its CwH band is futile. ald It has been suggested that rate coefficients for H-atom abstraction reactions by NO and OH should correlate and 3 show a “ linear free energy relationship Ï.19 Fig. 4 shows the logarithm of the rate coefficients for the reaction of NO with 3 aldehydes as a function of the rate coefficient for the reaction of OH with the same substances. The Ðgure also includes the correlation lines for abstraction reactions involving alkanes and addition reactions involving ethenes, etc.19 As can be seen, the aldehydes do not really Ðt the correlation line for hydrogen abstraction reactions. Obviously, endolim (compound 13) which contains an isolated double bond, Ðts the correlation line for addition reactions, but all the saturated aldehydes show a reactivity towards NO that apparently 3 is almost an order of magnitude too large. Atkinson has shown that there is an excellent correlation between the H-atom abstraction rate coefficients per equivalent CwH bond for reactions of NO and OH radicals with a 3 series of alkanes, formaldehyde and acetaldehyde : ln(k ) \ NO3 6.498 ] 1.611 ] ln(k ).17 We have used this relationship in OH combination with the developed estimation technique for OH rate coefficients,20,24,25 to calculate the rate coefficients for the NO reactions with saturated aldehydes. The results show 3 that the C ÈC aldehydes react up to an order of magnitude 3 10 faster with NO than predicted from this simple model. It may 3 3482

J. Chem. Soc., Faraday T rans., 1997, V ol. 93

be argued that such a discrepancy is quite acceptable considering the limited database upon which the model is developed. However, taken together with the systematic deviation from the regression line for abstraction reactions, Fig. 4, it may also be taken as an indication of experimental and/or theoretical errors. We here consider three cases : (1) we, and our colleagues, have made serious errors in our determinations of the NO 3 reaction rate coefficients ; (2) the NO reaction with aldehydes 3 is not just a simple hydrogen abstraction reaction ; (3) the OH reaction with aldehydes is not just a simple hydrogen abstraction reaction. As mentioned previously, the relative rate coefficients are believed to be accurate to within ^ 20%. The correlation curves in Fig. 4 have been derived from rate

Fig. 4 Log of the rate coefficients at 298 K for the reactions of NO 3 vs. the log of the rate coefficients for the reaction of OH with the same substrates. The numbers identifying the individual aldehydes are the same as in Table 2. The full curve is the regression line for H-atom abstraction from alkanes ; the dotted line is that of addition reactions involving alkenes etc. The two curves are data from ref. 19.

Published on 01 January 1997. Downloaded on 21/04/2014 03:01:34.

View Article Online

coefficients consistent with those of the reference compounds used in the present study. Hence, the data and correlation curves in Fig. 4 should show internal consistency, and we therefore discard the possibility of serious, systematic errors in the determined rate coefficients. As can be seen from Fig. 4, the rate coefficient data actually follow the correlation curve for addition reactions better than the curve for hydrogen abstraction. It is well known from basic, “ wet Ï organic chemistry that carbonyl compounds are polar and that they are readily attacked by both electrophilic and nucleophilic reagents in solution. We tentatively suggest that the NO vapour-phase reaction with aldehydes proceeds 3 via an exothermic adduct formation before CwH bond ald cleavage and subsequent elimination of nitric acid. Fig. 5 exaggerates the postulated potential energy of the reaction coordinate. The transition state for the formation of the adduct is very loose with a high density-of-states and a low threshold energy ; the adduct should therefore be formed at almost collision frequency. The measured reaction rate coefficients are four orders of magnitude smaller than this implying that most of the adducts dissociate back to the reactants rather than products. Further, the equilibrium and steadystate concentration of the adduct will be small. The second transition state is expected to be much tighter with a lower density-of-states as the nitrate radical breaks the CwH ald bond. The suggested potential energy of the reaction coordinate is in accordance with a reaction showing a positive temperature dependence, as is observed for acetaldehyde.8 In the high-pressure regime the postulated mechanism via adduct formation is equivalent to : kf aldehyde ] NO ÈÈÈÕ adduct 3 kb adduct ÈÈÈÕ aldehyde ] NO 3

(IIIa) (IIIb)

Financial support from the Norwegian Research Council, Climate Ozone Research Program is acknowledged. B.DÏA. acknowledges a Norwegian Government Scholarship under the Cultural Exchange Programs. References 1 2 3 4 5

kr adduct ÈÈÈÕ products

(IV)

Steady-state conditions applied to adduct formation gives :

6 7 8

d[adduct] \ k [aldehyde][NO ] f 3 dt

9

[ k [adduct] [ k [adduct] \ 0 (V) b r k f [adduct] \ [aldehyde][NO ] (VI) 3 k ]k b r d[products] \ k [adduct] r dt k k \ f r [aldehyde][NO ] 3 k ]k b r k k @ k ] k \ f k \ Kk r b eff k r r b

where K is the equilibrium constant of the intermediate adduct formation. As mentioned before, most of the adducts dissociate back to the reactants rather than products, k A k . b r The rate coefficient will thus depend upon the stability of the adduct, that is, it will depend upon the electronic properties of the alkyl group and its ability to distribute the reaction energy, and not just the CwH bond dissociation energy. ald Finally, we note that a frequently used structureÈreactivity estimation technique for calculating the OH radical reaction rate coefficients, including those involving aldehydes,20,24,25 reproduce the rate coefficients for reactions involving the unsaturated C ÈC aldehydes that are included in the data1 5 base used in its derivation. We also note that the same technique underestimates the recently measured rate coefficients for OH reactions with pinonaldehyde9h12 and caronealdehyde11,12 by more than a factor of four. As the data are scarce and any conclusion therefore premature, we have initiated a systematic investigation of the OH reaction with C ÈC aldehydes to see if this reaction might also be more 5 10 complex than normally thought.

(VII)

10 11 12 13 14 15

(VIII)

16 17 18 19

20 21 22 23 24 25 Fig. 5 Suggested potential energy of reaction coordinate for the reaction between NO and saturated aldehydes 3

R. Atkinson, C. N. Plum, W. P. L. Carter, A. M. Winer and J. N. Pitts, Jr, J. Phys. Chem., 1984, 88, 1210 ; 1984, 88, 4446. C. A. Cartell, W. R. Stockwell, L. G. Anderson, K. L. Busarow, D. Perner, A. Schmeltekopf, J. G. Calvert and H. S. Johnston, J. Phys. Chem., 1985, 89, 139. P. Carlier, H. Hannachi, A. Kartoudis, A. Martinet and G. Mouvier, Air Poll. Res. Rep., 1987, 9, 133. J. Hjorth, G. Ottobrini, G. Restelli and H. Skov, Air Pollut. Res. Rep., 1987, 9, 107. J. Hjorth, G. Ottobrini and G. Restelli, J. Phys. Chem., 1988, 92, 2669. E. D. Morris, Jr and H. Niki, J. Am. Chem. Soc., 1974, 78, 1337. C. A. Cantrell, J. A. Davidson, K. L. Busarow and J. G. Calvert, J. Geophys. Res., 1986, 91, 5347. E. J. Dlugokencky and C. J. Howard, J. Phys. Chem., 1989, 93, 1091. M. Glasius, A. Calogirou, N. R. Jensen, J. Hjorth and C. J. Nielsen, Int. J. Chem. Kinet., 1997, 29, 528. N. R. Jensen, A. Calogirou, J. Hjorth, D. Kotzias and C. J. Nielsen, Air Poll. Res. Rep., 1997, in press. M. Hallquist, I. Wangberg and E. Ljungstrom, Air Poll. Res. Rep., 1997, in press. M. Hallquist, I. Wangberg and E. Ljungstrom, Environ. Sci. T echnol., 1997, submitted. R. Atkinson, S. M. Aschmann and M. A. Goodman, Int. J. Chem. Kinet., 1987, 19, 299. E. S. C. Kwok, S. M. Aschmann, J. Arey and R. Atkinson, Int. J. Chem. Kinet., 1996, 28, 925. R. Atkinson, C. N. Plum, W. P. L. Carter, A. M. Winer and J. N. Pitts, Jr, Int. J. Chem. Kinet., 1984, 16, 887. W. P. L. Carter, A. M. Winer and J. N. Pitts, Jr, Environ. Sci. T echnol., 1981, 15, 829. R. Atkinson, J. Phys. Chem. Ref. Data, 1991, 20, 459. R. Atkinson, J. Phys. Chem. Ref. Data, 1997, 26, 215. R. P. Wayne, I. Barnes, P. Biggs, J. P. Burrows, C. E. CanosaMas, J. Hjorth, G. Le Bras, G. K. Mortgat, D. Perner, G. Poulet, G. Restelli and H. Sidebottom, Atmos. Environ., Part A., 1991, 25, 1. R. Atkinson, Chem. Rev., 1985, 85, 69. J. A. Kerr and D. W. Sheppard, Environ. Sci. T echnol., 1981, 15, 960. N. B. Colthup, L. H. Daly and S. E. Wiberley, Introduction to Infrared and Raman Spectroscopy, 3rd edn., Academic Press, 1990. S. Pinchas, Anal. Chem., 1955, 27, 2 ; 1957, 29, 334. R. Atkinson, Int. J. Chem. Kinet., 1987, 19, 799. E. S. C. Kwok and R. Atkinson, Atmos. Environ., 1995, 29, 1685.

Paper 7/02719B ; Received 21st April, 1997 J. Chem. Soc., Faraday T rans., 1997, V ol. 93

3483