Phototautomerization of Acetaldehyde to Vinyl ... - ACS Publications

Nov 12, 2012 - Alexander E. Clubb, Meredith J. T. Jordan, and S. H. Kable*. School of Chemistry, University of Sydney, Sydney, NSW 2006, Australia. Da...
2 downloads 0 Views 626KB Size
Download PDF
Letter pubs.acs.org/JPCL

Phototautomerization of Acetaldehyde to Vinyl Alcohol: A Primary Process in UV-Irradiated Acetaldehyde from 295 to 335 nm Alexander E. Clubb, Meredith J. T. Jordan, and S. H. Kable* School of Chemistry, University of Sydney, Sydney, NSW 2006, Australia

David L. Osborn Combustion Research Facility, Sandia National Laboratories, PO Box 969, MS 9055 Livermore, California 94551-0969, United States S Supporting Information *

ABSTRACT: The concentrations of organic acids, key species in the formation of secondary organic aerosols, are underestimated by atmospheric chemistry models by a factor of ∼2. Vinyl alcohol (VA, CH2CHOH, ethenol) has been suggested as a precursor to formic acid, but sufficient tropospheric sources of VA have not been identified. Here, we show that VA is formed upon irradiation of neat acetaldehyde (CH3CHO) in the actinic ultraviolet region, between 295 and 330 nm. Besides the well-known photochemical products CO and CH4, we infer up to a 15% quantum yield of VA at 20 Torr acetaldehyde pressure and a photolysis wavelength of 330 nm. The experiments confirm a recent model predicting phototautomerization of acetaldehyde to VA and imply that photolysis of small aldehydes and ketones could provide tropospheric sources of enols sufficient to impact organic acid budgets. We also report absolute infrared absorption cross sections of VA. SECTION: Spectroscopy, Photochemistry, and Excited States

O

beam environment, but its transient formation was inferred from the isotopic abundance of the various photoproducts. The reader is directed to the original modeling paper for a description of the mechanism and key energies.11 In this Letter, we report the observation of phototautomerization of acetaldehyde to VA under collisional conditions in a bulb. Acetaldehyde (pressure, P = 2.5−20 Torr) was irradiated by a laser (295−330 nm, 10 Hz, 5 mJ/pulse) for 4 min. The FTIR absorption spectrum was obtained before and after irradiation, resulting in a difference spectrum that shows loss of parent CH3CHO as negative-going features and photolysis products as positive peaks. Under these conditions, we observed 2−4% loss of CH3CHO over 4 min of irradiation. The negative peaks were removed by adding a scaled CH3CHO spectrum. Figure 1a shows a representative spectrum obtained from 20 Torr of CH3CHO irradiated at 315 nm. In this experiment, 3% of CH3CHO was lost. Features from production of CO, CH4, H2CO, CO2, C2H6, and (CH3)2CO are assigned in the spectrum. Although not apparent in Figure 1a, biacetyl [butanedione, (CH3CO)2] and 2-propanol [CH3CHOHCH3] are also a clear products, especially at longer irradiation wavelengths. The spectrum of 2-propanol underlies many of the features assigned in the figure and only becomes apparent after subtraction, although there is one clear feature, near 975 cm−1, that is identified in the figure. Also indicated in Figure 1a

ver one billion tons of nonmethane volatile organic compounds (NM-VOCs) are released into the troposphere every year, of which 90% are estimated to arise from biogenic sources.1 The number and variety of known NMVOCs has risen rapidly over the past few decades. In 1950, only one such compound was known, formaldehyde.2 In 1978, that number had risen to over 6003 and, in 1986, almost 3000.3 It is now estimated that 10 000−100 000 different NM-VOC compounds are being released into the troposphere.1 Organic acids are significant trace components in the troposphere, with estimates ranging from 60 to 120 megatons formed or released per year.4,5 These acids appear to play a key role in the formation of highly oxidized organic aerosols.6 Current atmospheric models, however, are unable to account for this volume of acids, falling short by about a factor of 2.4 Despite, or perhaps because of, the large number of NM-VOCs present in the troposphere,1 the mechanism responsible for producing the “missing” organic acids is unknown.5−9 Recently, a new mechanism for the production of organic acids in the atmosphere was proposed, which involves the photoisomerization of aldehydes or ketones to the corresponding unsaturated alcohol as the first step. This keto−enol phototautomerization was modeled for acetaldehyde (CH3CHO) to CH2CHOH (vinyl alcohol, VA), based on molecular beam experiments of the photolysis of specifically deuterated acetaldehyde.10,11 Master equation modeling was then used to predict the effect of collisions, and it was concluded that significant collisional stabilization of VA should result.11 VA was not observed in the collision-free molecular © 2012 American Chemical Society

Received: October 21, 2012 Accepted: November 12, 2012 Published: November 12, 2012 3522

dx.doi.org/10.1021/jz301701x | J. Phys. Chem. Lett. 2012, 3, 3522−3526

The Journal of Physical Chemistry Letters

Letter

where M represents a collision partner. Our spectra show evidence of all stable products from reactions 1, 2, and 4−8. The product of reaction 9, methylglyoxal, was not clearly observed, but there are some unassigned features in the carbonyl stretch region (1700 cm−1) that might belong to almost any carbonyl (vide infra). Several of the products in the reaction scheme above, and observed in Figure 1a, have carbonyl chromophores that will also absorb the laser radiation. However, the (n,π*) absorption of this chromophore is similar across a wide range of compounds. Therefore, by analogy with the photolysis loss of parent CH3CHO, we do not expect more than 3% “secondary” photolysis of these products. A feature of our experiment that differs from those reported previously14,15 is the unambiguous assignment of VA as a product, which has been proposed to arise from the following tautomerization:11 CH3CHO + hν + M → CH 2CHOH + M

(10)

Once the formation of VA is established, the observation of 2propanol as a product is reasonably explained. Double bonds are readily attacked by radicals that are produced in the primary process, CH3 and HCO. Methyl attack on VA will produce the hydroxypropyl radical. CH 2CHOH + CH3 + M → •CH 2−CHOH−CH3 •

CHOH−CH 2−CH3( +M)

Figure 1. (a) FTIR spectrum (black) of the photoproducts of CH3CHO (20 Torr, 315 nm). The red spectrum (inverted for clarity) is a composite formed by adding reference spectra of the identified compounds with an appropriate number density. The residual trace at the bottom shows no significant unassigned features; the small remaining structure is discussed in the text. (b) Reference spectrum of VA obtained from photolysis of butanal under the same conditions.

(1)

→ CH4 + CO

(2)

→ H + CH3CO

(3) (1)

although reaction 3 is unimportant (ϕph ≲ 0.01) at λ > 300 nm.14,15 The radical products were proposed to undergo secondary reactions. CH3 + CH3 + M → C2H6 + M

(4)

HCO + HCO → H 2CO + CO

(5)

CH3 + CH3CHO → CH4 + CH3CO

(6)

CH3CO + CH3 + M → (CH3)2 CO + M

(7)

CH3CO + CH3CO + M → (CH3CO)2 + M

(8)

CH3CO + HCO + M → CH3COCHO + M

(9)

(11)

Either of these radicals will abstract a more labile H-atom (e.g., from CH3CHO) to form propanol. Attack on the OH end of VA seems preferred based on our observation of 2-propanol over 1-propanol, although we cannot discount 1-propanol, nor products of attack on VA by other radicals, as being part of the unassigned residual spectrum. The absolute mass (number density) yield of each photolysis product was obtained by simultaneously fitting reference spectra of a known number density of each species, measured in our lab under identical experimental conditions. The yields of biacetyl and 2-propanol were determined using the literature absorption cross sections.16,17 These spectra are broad. Therefore, we expect that the different spectral resolution between our spectrum and the reported ones should make little difference to the measured yield. The absorption cross section of VA, σVA, was determined from a separate set of experiments: the photolysis of butanal at 315 nm, under identical conditions. The Norrish type II mechanism in butanal produces VA and ethylene with 1:1 stoichiometry.18 A reference spectrum of ethylene was used to determine the number density of ethylene produced from butanal photolysis, which therefore provides the total number density of VA initially formed. However, VA was observed to decay in our cell (with a concomitant increase in acetaldehyde) over a period of ∼30 min. The first-order decay rate constant was determined to be (1.0 ± 0.1) × 10−3 s−1. This rate constant models the long-time loss of VA, likely due to wall-catalyzed tautomerization back to CH3CHO. σVA was extrapolated to zero time based on this rate constant to yield the absolute cross sections shown in Figure 1b. A table of σVA versus wavenumber is provided as Supporting Information. We estimate the uncertainty in σVA to be 15% based on the uncertainty of the pressure measurements and uncertainty in the decay constant. However, we note that reaction of VA with radicals formed

are distinctive spectral features assigned to vinyl alcohol, consistent with literature spectra.12,13 Using gas chromotagraphic techniques, Horowitz and Calvert (HC) studied the production of stable, gas-phase products following photolysis of CH3CHO under similar conditions.14 They proposed three primary reactions. CH3CHO + hν → HCO + CH3

or

3523

dx.doi.org/10.1021/jz301701x | J. Phys. Chem. Lett. 2012, 3, 3522−3526

The Journal of Physical Chemistry Letters

Letter

during the laser pulse has not been considered; σVA should therefore be considered as a lower bound. The composite spectrum of identified photolysis products that best fits the experimental spectrum is shown in Figure 1a, inverted for clarity. The absence of notable features in the residual, shown at the bottom of Figure 1a, demonstrates that no significant stable, gas-phase, IR-active molecule remains unaccounted. A few weak features in the residual spectrum require explanation. There are some correlated positive and negative features apparent near 2170 and 3000 cm−1. These are clearly associated with CO and CH4 products and arise because the product spectrum has a slightly higher temperature after photolysis than the reference spectrum. There is also a negative-going spike at 1760 cm−1, which is assigned to imperfect subtraction of the CO stretch absorption feature of the parent acetaldehyde. Absorption at 1760 cm−1 is sufficiently strong at this pressure that there is a slight nonlinearity of Beer’s Law at the absorption maximum. At lower pressure, this negative feature disappears. Finally, there are some trace features that remain in the residual spectrum, between 1000 and 1400 cm−1 and at 1750 cm−1, that resemble those of a carbonyl compound, possibly glyoxal, methylglyoxal (reaction 9), or some similar molecule that could not be unambiguously assigned. Absolute number densities of the products and reactants were determined from the reference spectra of each species. In all cases the mass recovery varied between 60 and 75% of the measured loss of CH3CHO. The remaining 25−40% is likely to be comprised of one or more unassigned species associated with weak residual spectral features or products adsorbed to the walls. H2 is a known minor product14,15 that is IR-inactive but will have minimal effect on the mass balance. The observed yield of each photoproduct was converted into a quantum yield using the known absorption cross section of acetaldehyde at each wavelength15 and the measured laser power (see the Experimental Section). While the CO and CH4 primary products are detected directly in our experiments, HCO and CH3 undergo further reaction with themselves or CH3CHO, as summarized by reactions 4−6. We derive the primary photolysis quantum yield, ϕph, of reactions 1 and 2 from the quantum yield of the CO-containing products, CO + H2CO. The quantum yield of the CH3-containing products, CH4 + 2C2H6 (two CH3 products form a single C2H6) + (CH3)2CO, provides an independent measurement of ϕph. Biacetyl products from reaction 8 were also detected. This product, however, arises from reaction of the primary or secondary products with another CH3CHO molecule. While important for the mass yield, biacetyl does not contribute to the primary photolysis quantum yield. The wavelength dependence of ϕph is shown in Figure 2 for P(CH3CHO) = 20 Torr. The value of ϕph is shown as two independent estimates, arising from measurement of the COcontaining and CH3-containing products, which are in good agreement. Our data in Figure 2 are broadly consistent with the HC results,14 although our quantum yields are lower at short wavelengths and higher at long wavelengths. We can offer no definitive explanation for this difference. However, we note that the HC data were collected using lamps with a much broader band pass than the narrow line width lasers used here, which leads to different apparent CH3CHO absorption cross sections. Nevertheless, our data are consistent with previous results.

Figure 2. Wavelength dependence of the quantum yield for the primary photolysis products, defined as production of (CO + H2CO) or (CH4 + 2C2H6 + acetone), and vinyl alcohol in 20 Torr of pure acetaldehyde. The quantum yield of (VA + 2-propanol) is also shown as a proxy for the primary yield of VA (see the text).

The phototautomerization quantum yield, ϕVA, is reported here for the first time. At P(CH3CHO) = 20 Torr, ϕVA rises from 2.5% at 295 nm to 7.7% at 330 nm. The rate coefficient of the primary photolysis channels in reactions 1 and 2 slows exponentially with decreasing photon energy.11 At longer photolysis wavelengths, there is more time for collisional relaxation to occur, allowing trapping into the VA tautomer well to compete with dissociation. The balance between collisional relaxation, tautomerizarion, and dissociation, however, is complex and requires more detailed master equation modeling.11 The formation of 2-propanol is likely to arise from reaction of VA (eq 11); therefore, the primary quantum yield of VA from irradiation of CH3CHO under these conditions can be estimated as the sum of the VA and 2-propanol quantum yields, also shown in Figure 2. Under this assumption, the primary quantum yield of VA is at least 15% at the longest wavelength studied here. The pressure dependence of ϕph, ϕVA, and (ϕVA+ ϕpropanol) at 320 nm is shown as a Stern−Volmer (SV) plot in Figure 3. The SV equation derived from the photolysis data is 1 = (3.2 ± 0.5) + (0.07 ± 0.03)P(CH3CHO) ϕph for P in units of Torr. The errors are 1σ of the fit plus an allowance for the uncertainty of each data point. HC also measured the pressure dependence of the acetaldehyde photolysis quantum yield at 320 nm; their fitted SV equation was 1/ϕ = 2.13 + 0.0521P(CH3CHO). Although HC do not report errors, the two sets of SV parameters are likely to be within their mutual error bars. The SV plot for the phototautomerization quantum yield in Figure 3 is clearly nonlinear, whether considering VA on its own or (VA + 2-propanol) to represent the primary VA quantum yield. The production of VA is inherently collisional; collisions are required to stabilize vibrationally excited VA with a rate that is competitive with the dissociation rate. ϕVA must therefore go to zero (1/ϕVA → ∞) at zero pressure. At very high pressure, the tautomerization reaction must compete with collisional cooling, and therefore, the VA yield should also approach zero at very high pressure. The shape of the SV plot 3524

dx.doi.org/10.1021/jz301701x | J. Phys. Chem. Lett. 2012, 3, 3522−3526

The Journal of Physical Chemistry Letters

Letter

cm−1), pressure broadening does not contribute to the spectral line shape. Liquid acetaldehyde and butanal (Aldrich, ≥99.5%) were purified by several freeze−pump−thaw cycles. Neat vapor was delivered directly to the evacuated cell. The pressure of the gas was monitored using two MKS Baratron gauges. Irradiation experiments were conducted using a Sunlite EX OPO tunable laser pumped by a pulsed Nd:YAG laser at a 10 Hz repetition rate. The photolysis beam (5 mm diameter, 5−6 mJ/pulse of UV light) was passed twice through the cell. We did not expect multiphoton absorption to be important due the unfocused beam. The reflectivity of the mirrors and windows was measured to determine absolute quantum yields. Laser pulse energy was monitored on every pulse by a calibrated laser energy meter after the beam exited the cell on the second pass. After 4 min of irradiation, the number density of acetaldehyde was reduced by an amount varying from 1% at 330 nm to 4% at 295 nm.

■

Figure 3. Stern−Volmer plots for the production of VA (VA + 2propanol) and all primary photolysis products, defined as (CO + H2CO) or (CH4 + 2C2H6 + acetone) at λ = 320 nm.

ASSOCIATED CONTENT

* Supporting Information S

A table of σVA versus wavenumber and other fit data. This material is available free of charge via the Internet at http:// pubs.acs.org.

in Figure 3 likely reflects these two effects, although the data at 2.5 Torr are starting to approach our signal-to-noise limit. The experimental conditions in this work do not yet reproduce those in the troposphere, where the partial pressure of CH3CHO is extremely low and collisions with N2 and O2 dominate. Future measurements of ϕVA under atmospheric conditions will require a very low partial pressure of CH3CHO yet must be sensitive enough to detect VA at this low concentration. In summary, we have shown experimentally that vinyl alcohol is a primary, collision-induced photoproduct following irradiation of neat acetaldehyde between 295 and 330 nm and 2.5 and 20 Torr, with measured VA quantum yields between 2.5 and 7.7%. When corrected for reaction of VA with CH3 radicals, the inferred primary quantum yield of VA is as high as 15%. These results validate the previous hypothesis11 that aldehydes and ketones may phototautomerize to their respective enols when irradiated in the troposphere by actinic ultraviolet radiation. In the atmosphere, enols such as VA will quickly react with any available radicals, such as shown here in the reaction of VA with methyl radicals. In the atmosphere, reaction with OH radicals is the most likely fate, leading subsequently to the production of organic acids. Phototautomerization pathways do not feature in current atmospheric models,19 and these results suggest that such pathways might be important.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Telephone: +61 (2) 93512756. Fax: +61 (2) 9351-3329. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was funded by the Australian Research Council (Grants DP0560020 and DP1094559). D.L.O. is supported by the U.S. Department of Energy, Office of Basic Energy Sciences, Division of Chemical Sciences, Geosciences and Biosciences. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed−Martin Company, for the National Nuclear Security Administration under Contract No. DE-AC04-94-AL85000.

■

REFERENCES

(1) Goldstein, A. H.; Galbally, I. E. Known and Unexplored Organic Constituents in the Earth’s Atmosphere. Environ. Sci. Technol. 2007, 41, 1514−1521. (2) Glueckauf, E. In Compendium of Meteorology; Malone, T. F., Ed.; American Meteorological Society: Boston, MA, 1951. (3) Graedel, T. E. Chemical Compounds in the Atmosphere; Academic Press: New York, 1978. (4) Ito, A.; Sillman, S.; Penner, J. E. Effects of Additional Nonmethane Volatile Organic Compounds, Organic Nitrates, And Direct Emissions of Oxygenated Organic Species on Global Tropospheric Chemistry. J. Geophys. Res. 2007, 112, D06309. (5) Paulot, F.; et al. Importance of Secondary Sources in the Atmospheric Budgets of Formic and Acetic Acids. Atmos. Chem. Phys. 2011, 11, 1989−2013. (6) Jimenez, J. L.; et al. Evolution of Organic Aerosols in the Atmosphere. Science 2009, 326, 1525−1529. (7) Archibald, A. T.; McGillen, M. R.; Taatjes, C. A.; Percival, C. J.; Shallcross, D. E. Atmospheric Transformation of Enols: A Potential Secondary Source of Carboxylic Acids in the Urban Troposphere. Geophys. Res. Lett. 2007, 34, L21801.

■

EXPERIMENTAL SECTION Experiments were conducted in a 12.4 × 12.4 cm cross-shaped stainless steel cell, internally coated with Teflon. One axis of the cell was used for irradiating with monochromatic UV light through fused silica windows. The cell was placed in an evacuated FTIR spectrometer (Bruker IFS 66 V/S) such that the second axis of the cell was aligned with the IR beam path. FTIR scans were taken through polished KBr windows using a liquid-nitrogen-cooled HgCdTe detector and 1 cm−1 resolution in the range of 720−3949 cm−1. The cell was evacuated and flushed with N2 between experiments. No photochemical deposits on the fused silica windows were observed during these experiments. At the instrumental resolution used (1 3525

dx.doi.org/10.1021/jz301701x | J. Phys. Chem. Lett. 2012, 3, 3522−3526

The Journal of Physical Chemistry Letters

Letter

(8) da Silva, G. Carboxylic Acid Catalyzed Keto−Enol Tautomerizations in the Gas Phase. Angew. Chem., Int. Ed. 2010, 49, 7523−7525. (9) Welz, O.; Savee, J. D.; Osborn, D. L.; Vasu, S. S.; Percival, C. J.; Shallcross, D. E.; Taatjes, C. A. Direct Kinetic Measurements of the Criegee Intermediate (CH2OO) Formed by Reaction of CH2I with O2. Science 2012, 335, 204−207. (10) Heazlewood, B. R.; Maccarone, A. T.; Andrews, D. U.; Osborn, D. L.; Harding, L. B.; Klippenstein, S. J.; Jordan, M. J. T.; Kable, S. H. Near-Threshold H/D Exchange in CD3CHO Photodissociation. Nat. Chem. 2011, 3, 443−448. (11) Andrews, D. U.; Heazlewood, B. R.; Maccarone, A. T.; Conroy, T.; Payne, R. J.; Jordan, M. J. T.; Kable, S. H. Photo-tautomerization of Acetaldehyde to Vinyl Alcohol: A Potential Route to Tropospheric Acids. Science 2012, 337, 1203−1206. (12) Koga, Y.; Nakanaga, T.; Sugawara, K.-I.; Watanabe, A.; Sugie, M.; Takeo, H.; Kondo, S.; Matsumura., C. Gas Phase Infrared Spectrum of Syn-Vinyl Alcohol Produced by Thermal Decomposition of Several Alcohols and Aldehydes. J. Mol. Spectrosc. 1991, 145, 315− 322. (13) Joo, D.-L.; Merer, A. J.; Clouthier, D. J. High-Resolution Fourier Transform Infrared Spectroscopy of Vinyl Alcohol: Rotational Analysis of the ν13 CH2 Wagging Fundamental at 817 cm−1. J. Mol. Spectrosc. 1999, 197, 68−75. (14) Horowitz, A.; Calvert, J. G. Wavelength Dependence of the Primary Processes in Acetaldehyde Photolysis. J. Phys. Chem. 1982, 86, 3105−3114. (15) Moortgat, G. K.; Meyrahn, H.; Warneck, P. Photolysis of Acetaldehyde in Air: CH 4, CO and CO2 Quantum Yields. ChemPhysChem 2010, 11, 3896−3908. (16) Profeta, L. T. M.; Sams, R. L.; Johnson, T. J. Quantitative Infrared Intensity Studies of Vapor-Phase Glyoxal, Methylglyoxal, and 2,3-Butanedione (Biacetyl) with Vibrational Assignments. J. Phys. Chem. A 2011, 115, 9886−9900. (17) NIST Standard Reference Database 79: Quantitative Infrared Database; Data compiled by Chu, P. M.; Guenther, F. R.; Rhoderick, G. C.; Lafferty, W. J. webbook.nist.gov (accessed Oct 19, 2012). (18) Tadic, J.; Juranic, I.; Moortgat, G. K. Pressure Dependence of the Photooxidation of Selected Carbonyl Compounds in Air: nButanal and n-Pentanal. J. Photochem. Photobiol. A 2001, 143, 169− 179. (19) Atkinson, R.; et al. Evaluated Kinetic and Photochemical Data for Atmospheric Chemistry: Volume II  Gas Phase Reactions of Organic Species. Atmos. Chem. Phys. 2006, 6, 3625−4055.

■

NOTE ADDED AFTER ASAP PUBLICATION This paper was published on the Web on November 12, 2012, with incorrect TOC and Abstract graphics. The corrected version was reposted on November 26, 2012.

3526

dx.doi.org/10.1021/jz301701x | J. Phys. Chem. Lett. 2012, 3, 3522−3526