Conformational Switching of HOCO Radical: Selective Vibrational

Jun 18, 2017 - Conformers of carboxyl radical (HOCO) have been studied by IR spectroscopy in argon and nitrogen matrices. In an argon matrix, only the...
7 downloads 19 Views 1MB Size
Download PDF
Article pubs.acs.org/JACS

Conformational Switching of HOCO Radical: Selective Vibrational Excitation and Hydrogen-Atom Tunneling Sergey V. Ryazantsev,†,‡ Vladimir I. Feldman,† and Leonid Khriachtchev*,‡ †

Department of Chemistry, Lomonosov Moscow State University, Moscow 119991, Russia Department of Chemistry, University of Helsinki, P.O. Box 55, Helsinki FIN-00014, Finland

‡

ABSTRACT: Conformers of carboxyl radical (HOCO) have been studied by IR spectroscopy in argon and nitrogen matrices. In an argon matrix, only the lower-energy conformer transHOCO is observed, whereas both cis and trans conformers are found for deuterated carboxyl radical DOCO. In a nitrogen matrix, both conformers of HOCO and DOCO isotopologues can be prepared, indicating strong stabilization of the higher-energy cis conformer by a nitrogen matrix. Selective vibrational excitation promotes the trans-to-cis and cis-to-trans conversions of DOCO in an argon matrix and HOCO and DOCO in a nitrogen matrix, which is the first conformational photoswitching of an open-shell species. In a nitrogen matrix, the cis-to-trans and trans-to-cis conversions of HOCO is also found upon broadband IR light of the spectrometer, and the ratio of the quantum yields of these processes is about 3.3. The photoswitching peculiarities are in agreement with the available theoretical energy barriers. The higher-energy cis conformer decays to the lower-energy trans conformer via hydrogenatom tunneling through the torsional barrier, which is also a unique observation for an open-shell species. The tunneling mechanism of the cis-to-trans switching is supported by the low-temperature limit of the reaction rate and by the H/D kinetic isotope effect. Our results suggest a large difference in the H/D kinetic isotope effects in nitrogen and argon matrices (∼5 and >100, respectively). The stabilizing effect on cis-DOCO by a nitrogen matrix (by 2 orders of magnitude versus an argon matrix) is much smaller than that on cis-HOCO (estimated to be >104).

■

INTRODUCTION

conformation may strongly affect their reactivity; thus, the conformational switching in radicals is important in this respect. Carboxyl (or hydrocarboxyl) radical (HOCO) is one of the simplest open-shell species that is particularly significant for atmospheric chemistry, astrochemistry, and combustion.30−33 Similarly to formic (HCOOH) and nitrous (HONO) acids, it can accept the trans or cis conformations, the former one being computationally more stable by ∼600 cm−1, and the trans−cis barrier is ∼2900 cm−1 (with zero-point vibrational energy correction).32,34 Both conformers were observed in the gas phase35−37 and in nitrogen and carbon monoxide matrices;38,39 however, no conformational switching was reported. Recently, we have demonstrated a photochemical method for the efficient preparation of carboxyl radicals in noble-gas (Ng) matrices.40 The approach is based on a two-step photolysis of a HCOOH/Ng matrix: first with UV light to produce the CO···H2O complexes and then with VUV light to generate carboxyl radicals (in general, together with the OH···CO complexes; see Scheme 1). Only trans-HOCO was observed in those experiments, similarly to the results obtained for X-ray irradiation of matrix-isolated formic acid.41 The failure to observe cis-HOCO in those experiments is possibly due to its low stability against hydrogen-atom tunneling. Indeed, the stabilization barrier of cis-HOCO (∼2300 cm−1)32,34 is lower than that of formic and acetic acids,42 which suggests its fast

Conformational change is an important concept in various fields of chemistry and beyond.1−8 Matrix-isolation technique provides a powerful tool for direct probing of the conformational changes in small molecules in the condensed phase.9−21 Switching between different conformers isolated in a lowtemperature matrix can be achieved by their selective vibrational excitation. In some cases, the lifetime of the higher-energy conformers is limited by hydrogen-atom tunneling through the stabilization (torsional) barrier. Solvation in solid matrices can affect the tunneling rate.22 For example, the higher-energy cis conformers of formic and acetic acids are orders of magnitude more stable in a nitrogen matrix than in an argon matrix.23 In addition, the higher-energy conformers can be stabilized against hydrogen-atom tunneling by the deuteration due to the mass effect.24 To the best of our knowledge, the conformational switching has been previously reported only for closed-shell systems. The main motivation of the present work is to examine the conformational change promoted by selective vibrational excitation and hydrogen-atom tunneling for an open-shell species (radical) in a cryogenic matrix. Free radicals are the key intermediates in a wide variety of different processes, ranging from interstellar and atmospheric chemistry to biological reactions.25−29 Conformational changes in free radicals are significant for understanding the detailed mechanisms of these processes. These species are typically very reactive and © 2017 American Chemical Society

Received: March 15, 2017 Published: June 18, 2017 9551

DOI: 10.1021/jacs.7b02605 J. Am. Chem. Soc. 2017, 139, 9551−9557

Article

Journal of the American Chemical Society

trans- and cis-HOCO in a nitrogen matrix do not coincide, in contrast to the suggestion in ref 38. In fact, the ν4 band of transHOCO is very weak and appears at 1084.8 cm−1. Conformational Photoswitching. Figure 2 shows the effects of narrow-band IR pumping of carboxyl radical in a nitrogen matrix. Excitation of the OH stretching mode of transHOCO decreases the amount of this species and increases the amount of cis-HOCO (spectrum a). In turn, excitation of the OH stretching mode of cis-HOCO leads to the cis-to-trans conversion (spectrum b). Thus, the conformers of HOCO can be interconverted by selective vibrational excitation of the OH stretching mode. The efficiency of photoswitching of the HOCO forms is comparable with that of formic acid.50 To the best of our knowledge, the observed IR-induced trans−cis interconversion of HOCO is the first example of conformational photoswitching of an open-shell species. To recall, the calculated trans−cis and cis−trans barriers for HOCO are ∼2900 and ∼2300 cm−1, respectively;32,34 thus, the energy of the OH stretching modes (3400−3600 cm−1) is sufficient to promote these conformational changes. It should be emphasized that in an argon matrix, cis-HOCO was not observed even after excitation of the OH stretching mode of trans-HOCO at 3603 cm−1, which suggests its very short lifetime. For DOCO, the trans-to-cis conversion was not achieved by excitation of the OD stretching fundamental (see trace c in Figure 2 for a nitrogen matrix). However, this process occurs upon selective excitation of the OD stretching overtone (trace d in Figure 2 for a nitrogen matrix). It is remarkable that the opposite cis-to-trans conversion was observed after excitation of the OD stretching fundamental of cis-DOCO (see trace e in Figure 2 for a nitrogen matrix). (The same processes can also be promoted in an argon matrix.) These observations are in full agreement with the calculated barriers.32,34 Indeed, the energy of the OD stretching modes (2500−2700 cm−1) is sufficient for the cis-to-trans conversion (calculated barrier ∼2300 cm−1), but it is not sufficient for the trans-to-cis process (calculated barrier ∼2900 cm−1). Nevertheless, it should be noted here that the calculated barriers refer to the system in a vacuum, whereas the experiments are performed in a solid matrix that can change the transition barriers.22,42 In addition, conformational change can be obtained at the energies, which are somewhat below the barriers.10 These facts complicate the quantitative comparison of experiment and theory. As another note, the OD stretching overtones of trans-DOCO were not observed directly in the FTIR spectra because of their weakness. The corresponding frequencies were found by maximizing the conversion effect, which represents an example of reactive vibrational excitation spectroscopy.50 Broadband Conformational Conversion. Broadband light of the FTIR spectrometer also produces conformational changes of HOCO in a nitrogen matrix. After the cis-to-trans conversion by narrow-band IR pumping, the bands of transHOCO tend to decrease in intensity upon broadband IR excitation and the bands of cis-HOCO increase. The opposite effect is observed after conversion of trans-HOCO to cisHOCO, i.e., the cis-HOCO bands decrease and the transHOCO bands increase under broadband IR light of the spectrometer. The results of the kinetic measurements of the normalized concentrations of trans-HOCO to cis-HOCO are shown in Figure 3. The kinetic curves indicate the existence of a photostationary state reached in about 1 h. This means that the broadband IR light simultaneously promotes both trans-to-cis and cis-to-trans conversions and these processes are in

Scheme 1. Photochemical Production of HOCO in Cryogenic Matrices40

decay to the trans form. In the present work, we intend to stabilize the higher-energy conformer of carboxyl radical against hydrogen-atom tunneling by solvation in a nitrogen matrix and by deuteration. As a result, we can achieve IR-induced conformational photoswitching of carboxyl radicals and directly observe the hydrogen-atom tunneling decay of the cis form.

■

RESULTS AND DISCUSSIONS Spectral Assignments. Two key observations are made on the photoproduction of carboxyl radicals. First, in a nitrogen matrix, both trans and cis conformers of HOCO are observed after two-step UV/VUV photolysis of HCOOH (Figure 1). It is

Figure 1. Formation of HOCO radicals. The upper and lower traces are difference FTIR spectra showing the results of two-step UV/VUV photolysis of formic acid in argon and nitrogen matrices. The peak marked with an asterisk originates from the OH···CO complex.40 The peak marked with a cross is the (2ν2 + ν3) absorption of CO2 (this peak is masked by the band of trans-HOCO in the argon matrix). The negative peaks are from formic acid.

in contrast with argon and other Ng matrices, where only transHOCO appears under these conditions.40 Second, for deuterated carboxyl radicals (DOCO), both trans and cis conformers are observed not only in a nitrogen matrix but also in an argon matrix. These observations show that the higherenergy form is stabilized by solvation in a nitrogen matrix and by deuteration, similarly to small carboxylic acids23,24 and parasubstituted derivatives of benzoic acid.16 Table 1 shows the vibrational frequencies of the HOCO and DOCO conformers in nitrogen and argon matrices. The assignments are supported by the IR-pumping and kinetic experiments described below. The experimental spectra agree well with the available calculations34−36,43−45 and experimental data.35,38−41,46−49 The difference between the gas-phase and matrix data are due to the matrix effect on vibrational frequencies, and smaller differences from the gas-phase values are observed for transHOCO in a neon matrix.40 The present work adds some new spectroscopic information on this species (new assignments are marked in bold in Table 1). In particular, selective conformational photoswitching (see below) allows detecting relatively weak absorptions. As a result, it is found that the ν4 bands of 9552

DOI: 10.1021/jacs.7b02605 J. Am. Chem. Soc. 2017, 139, 9551−9557

Article

Journal of the American Chemical Society Table 1. Vibrational Frequencies (in cm−1) of the Conformers of Carboxyl Radicala trans-HOCO N2 matrix

mode ν1 (O−H str.) 2ν4 ν2 (CO str.) ν3 (H−O−C bend) ν4 (C−O str.) ν5 (O−C−O bend) ν6 (torsion)

cis-HOCO

Ar matrix

N2 matrix

Ar matrix

gas

3411.1 3406 sh 2108.7 1802.3 1800 sh 1797 sh −

− − − −

− − − 1290f

3452.3 2073.9 1824.1 1280.2

− −

1040f 605f

1042.4 (151.9) 601.2 (28.4)

calcb,c

gas d

3571.0 3567 sh 2125 1837.9 1836 sh 1833 sh 1241.4

3604.8 3602.6 3600 sh 2108.7 1843.5 1839.8 1825.4 1211.1 1208.8 1205.6

3635.7 − 1852.6e 1194f

3641.0 2089.0 1862.0 1212.7

1084.8 −

1066 sh 1064.5 −

1048f 629f

1052.0 (79.9) 616.0 (5.8)

1063.0 1059 sh 610.4g

542.2

−

−

475.4 (84.5)

573.2

(80.4) (15.9) (203.6) (221.6)

trans-DOCO mode ν1 ν3 ν2 ν3 ν4

(O−D str.) + ν4 (CO str.) (D−O−C bend) (C−O str.)

− − cis-DOCO

calcb,c (15.0) (15.5) (276.8) (0.9)

540.2 (102.4)

N2 matrix

Ar matrix

gas

calcb,c

N2 matrix

Ar matrix

gas

calcb,c

2638.0 − 1828.3 1825 shg 1109.2 926.0g

2662.0 2660.2 1987.9 1841.8 1838 1092.6 907.2

2684.1h − − − −

2685.1 (48.0) 1977.6 (12.6) 1859.8 (214.6) 1086.4 (229.9) 902.6 (10.7)

2522.3 − 1804.3 1140.7 972.9

2531.2 2073.0 1809.4 1805 1127.7 957.0

− − − − −

2551.6 (14.8) 2073.4 (9.4) 1827.5 (177.7) 1123.1 (84.4) 960.9 (75.7)

a

New assignments are in bold (including the available matrix works);38,39 bands of the thermally unstable matrix site are in italics (these bands disappear after annealing at 20 and 25 K for nitrogen and argon matrices, respectively; see Experimental Methods for more details). bAnharmonic frequencies obtained by the CcCRE QFF method at the CCSD(T)-aug-cc-pVXZ (X = T, D, and 5) level of theory are from ref 43 (intensities in km mol−1 are in parentheses). cAdditional calculations on vibrational spectra of carboxyl radical can be found, for example, in refs 35, 36, 44, and 45. d Ref 48a. eRef 49. fRef 35. gTentative assignment. hRef 48b.

part of the spectrum. In these experiments, we used a long-pass IR filter that cuts light above 1850 cm−1. This filter allows monitoring several relatively intense bands of HOCO and DOCO (in particular, ν2). With the filter, the amount of transHOCO in a nitrogen matrix does not decrease after the cis-totrans conversion, i.e., the lower-energy part of the spectrometer light does not promote the conformational switching. Nevertheless, the reversed cis-to-trans-HOCO conversion occurs with the filter and in the dark in a time scale of ∼20 h (Figure 4a). In general, the cis-to-trans decay is improbable over the barrier of ∼2300 cm−1 at such low temperatures,22 and quantum tunneling is presumably involved. The cis-to-trans decay is also observed for DOCO in argon and nitrogen matrices (Figure 4a). The tunneling decay of cis-DOCO is expected to be very slow and the OD fundamental is relatively close to the filter edge (compared to the OH fundamental). We wanted to exclude completely the effect of the spectrometer source on the tunneling decay and limited the number of measurements in a nitrogen matrix. The IR light of the spectrometer was blocked between the measurements. The standard error of the fit for the lifetime (at the level of 1/e) is 2%, which is accurate enough for our purposes. To the best of our knowledge, it is the first example of tunneling-induced conformational switching in an open-shell species. Tunneling-induced valence isomerization of radicals through a hydrogen atom shift have been repeatedly reported.52−55 In addition, there have been many EPR and theoretical studies on the quantum rotational dynamics in small radicals and radical cations involving hydrogen atom tunneling between equivalent structures (see, for example, refs 56, 57). However, the case reported here is essentially different since it involves nondegenerate conformers. It is also worth mentioning that the cis-HOCO → H + CO2 tunneling reaction was observed in the anion photodetachment experiments58 and simulated theoretically.59−61 In the present work, no formation of CO2 is found upon decay of cis-HOCO which indicates the very low rate of the cis-HOCO → H + CO2

equilibrium. The equilibrium trans/cis concentration ratio is independent of the initial concentration ratio. The total concentration of the trans and cis forms does not change in time. A similar behavior was observed, for example, for the HONO conformers in a krypton matrix.51 The kinetic data can be fitted to the exponential functions as described in ref 51 (Figure 3). This procedure yields the ratio of the conversion rate constants kcis−trans/ktrans−cis ≈ 1.87. Taking into account the experimental intensities of the OH stretching modes (Itrans/Icis ≈ 1.78), we obtain a ratio of the quantum yields of the IR-induced processes φcis−trans/φtrans−cis ≈ 3.3. These estimates are based on the realistic assumptions: (i) the conformational change mainly occurs by excitation of the OH stretching fundamentals, (ii) the intensity of IR light of the spectrometer is similar at these positions (3411 and 3571 cm−1), and (iii) the tunneling decay of cis-HOCO is relatively slow (see below). For HONO in a krypton matrix, similar estimates gave φcis−trans/φtrans−cis ≈ 14.51 The explanation of these results is a challenge for computational chemistry, and the matrix can affect the isomerization efficiency. For DOCO in a nitrogen matrix, the IR light of the spectrometer significantly accelerates the decay of the cis form (by the factor of ∼25); however, no photoequilibrium is achieved. This different behavior (from HOCO) is fully consistent with the conclusion derived from the narrow-band IR pumping that only the cis-to-trans conversion occurs upon excitation of the OD stretching fundamentals. This experiment also supports our assumption (i) that excitation of the OH stretching fundamentals plays a dominant role in the broadband photoswitching, and the effect of the overtones and combinations is negligible. This conclusion is reasonable because the overtones and combinations are much weaker than the fundamentals and the intensity of the spectrometer source decreases toward the near-IR region. Cis-to-Trans Tunneling Decay. The interconversion of the HOCO conformers promoted by broadband light of the spectrometer can be suppressed by filtering the high-frequency 9553

DOI: 10.1021/jacs.7b02605 J. Am. Chem. Soc. 2017, 139, 9551−9557

Article

Journal of the American Chemical Society

Figure 3. Conformational changes of HOCO in a nitrogen matrix under broadband IR light of the spectrometer (T = 4.5 K). At the beginning of these two measurements, the matrices mainly contained trans-HOCO (open symbols) or cis-HOCO (solid symbols). All concentrations are normalized by these initial amounts. The relative concentrations are obtained by integration of the OH stretching bands taking into account the experimental ratio of the trans-HOCO and cisHOCO absorption intensities (1.78). The trans-HOCO and cisHOCO concentrations are shown by red circles and blue squares, respectively, and the sum concentration is shown by gray triangles. Experimental points are fitted using the kinetic scheme described elsewhere.51

Figure 2. Conformational photoswitching of carboxyl radicals in a nitrogen matrix. All traces are difference FTIR spectra showing the results of IR pumping of (a) trans-HOCO at the OH stretching fundamental; (b) cis-HOCO at the OH stretching fundamental; (c) trans-DOCO at the OD stretching fundamental; (d) trans-DOCO at the OD stretching overtone; (e) cis-DOCO at the OD stretching fundamental. The increase of the noise on the left side of the spectra is due to the long-pass IR filter used in these experiments. The spectra are measured immediately after the IR pumping. It should be noted that practically complete conversion is achieved in these experiments by pumping at the fundamental frequencies for several minutes, and the difference in band intensities is due to varying amounts of carboxyl radical in different experimental days. For pumping of trans-DOCO at the OD stretching overtone, the efficiency is more than an order of magnitude lower, which is due to the low intensity of this mode. The pumping frequencies are shown in the figure.

tunneling reaction in the ground state as compared to the cisHOCO → trans-HOCO switching. The cis-to-trans tunneling decay should be considered in future work on this fundamental system. In addition, no formation of CO2 occurs upon excitation of trans- and cis-HOCO by IR light, suggesting a low probability of tunneling to H + CO2 from the low vibrational states. These results are in agreement with the calculations.60 Quantum tunneling has two characteristic features: kinetic isotope effect and low-temperature limit of the reaction rate.62 It has been also found for carboxylic acids that the cis-to-trans tunneling rate depends on the matrix material due to the change of the tunneling barrier.22 Figure 4a shows the cis-totrans conversion kinetics measured at 4.5 K. The decay of cisHOCO and cis-DOCO in a nitrogen matrix shows the H/D kinetic isotope effect kH/kD ≈ 5 at 4.5 K. It is worth noting that this value is significantly smaller than that found for carboxylic acids in an argon matrix (for formic acid, kH/kD ≈ 103).24 However, this modest kinetic isotope effect may be specific for a nitrogen matrix that strongly stabilizes cis-HOCO. (To the best of our knowledge, the kinetic isotope effect has not been reported for conformational change in a nitrogen matrix. For example, Amiri et al. could prepare the higher-energy form only for O-deuterated para-substituted derivatives of benzoic acid but failed with the normal (nondeuterated) species.16) It is

Figure 4. (a) Tunneling decay of cis-HOCO in a nitrogen matrix and cis-DOCO in argon and nitrogen matrices at 4.5 K. The experimental points are fitted with single exponential functions. (b) Arrhenius plots of the decay rate of cis-HOCO in a nitrogen matrix and cis-DOCO in an argon matrix. The lines guide the eye. The decay rate constants are measured in s−1.

possible that the tunneling is affected by the medium reorganization, which is an isotopically independent factor and partially masks the kinetic isotope effect.63 In other words, the true reaction coordinate is not exclusively the hydrogen 9554

DOI: 10.1021/jacs.7b02605 J. Am. Chem. Soc. 2017, 139, 9551−9557

Article

Journal of the American Chemical Society

DOCO, both trans and cis conformers are observed in both matrices, which means the deuteration-induced stabilization of the cis conformer. The spectroscopic information on the HOCO and DOCO conformers has been improved (Table 1). The first conformational photoswitching of an open-shell species has been obtained by selective vibrational excitation (Figure 2). In particular, excitation of the OH stretching transition of trans-HOCO in a nitrogen matrix produces cisHOCO and vice versa. For DOCO, only the cis-to-trans process is achieved upon excitation of the OD stretching fundamental, whereas the opposite conversion can be promoted by excitation of the OD stretching overtone. This remarkable fact agrees with the calculated trans−cis and cis− trans barriers (∼2900 and 2300 cm−1, respectively).32,34 The trans-to-cis and cis-to-trans conversions of HOCO in a nitrogen matrix also occur upon broadband IR light of the spectrometer (Figure 3). The trans/cis concentration ratio in photoequilibrium estimates the ratio of the quantum yields of the IR-induced conformational changes φcis−trans/φtrans−cis ≈ 3.3 for excitation of the OH stretching fundamentals. The higher-energy conformer of carboxyl radical decays in the dark to the lower-energy form via hydrogen-atom tunneling through the torsional barrier, which is a unique observation for open-shell species. The tunneling mechanism is confirmed by the kinetic isotope effect and by the characteristic lowtemperature limit of the reaction rate (Figure 4). The kinetic isotope effect on cis-HOCO in a nitrogen matrix (∼5) appears to be significantly smaller than that on formic and acetic acids in an argon matrix. However, in an argon matrix, cis-DOCO is rather stable, whereas cis-HOCO is not observed even after selective IR pumping, presumably due to the short lifetime. This fact allows us to estimate the lower limit of the kinetic isotope effect of the cis-HOCO decay in an argon matrix at the level of 102−103, which supports the tunneling mechanism. The strong effect of the matrix material on the kinetic isotope effect is a remarkable observation that needs theoretical interpretation. We speculate that this difference is due to the medium reorganization in the tunneling process, which is an isotopically independent factor. Our results indicate the very low rate of the cis-HOCO → H + CO2 tunneling reaction in the ground state as compared to the cis-HOCO → trans-HOCO switching. The cis-to-trans tunneling decay of HOCO is worth considering in future work on this fundamental system. The higher-energy form of HOCO is strongly stabilized by a nitrogen matrix compared to Ng matrices, where cis-HOCO has not been observed.40,41,47 Our estimates give a huge stabilization factor for cis-HOCO in a nitrogen matrix as compared with an argon matrix (kAr/kN2 > 104). This stabilization effect on cis-DOCO is smaller (kAr/kN2 ≈ 94) and comparable to that on cis-HCOOH and cis-CH3COOH.23 It is highly remarkable that cis-HOCO in a nitrogen matrix is more stable than cis-DOCO in an argon matrix (Figure 4). In other words, the nitrogen matrix effect on the stability of the cis form of carboxyl radical is stronger than the effect of deuteration! The opposite has been found for formic and acetic acids.23,24 Such a strong effect of a nitrogen matrix on the stability of cis-HOCO means a significant increase of the cis− trans barrier (with respect to an argon matrix), which is the energy difference between the cis form and the transition state. This result is not well understood and poses a challenge for computational chemistry. For example, it cannot be fully explained by the formation of the OH···N2 hydrogen bond. Indeed, the 1:1 complexes with nitrogen can provide some

motion but rather the collective coordinate involving the reorganization of matrix molecules. To remind, cis-HOCO is not observed experimentally in an argon matrix even after selective IR pumping, which presumably originates from its short lifetime. The cis form of acetic acid has a lifetime of ∼50 s in an argon matrix, and it is easily observable,11 as well as cis-formic acid in a neon matrix with a lifetime of ∼5 s.64 We propose that the lifetime of cisHOCO in an argon matrix is much shorter than 1 min. Because the lifetime of cis-DOCO in an argon matrix is ca. 1 h, this estimate means the H/D kinetic isotope effect of >100. In general, the H atom tunneling should be very fast through the barrier of ∼2300 cm−1 predicted for cis-HOCO.32,34 The computational barriers for the cis-to-trans conversion of formic and acetic acids are 3204 and 2860 cm−1, respectively (in a vacuum),42 and the corresponding lifetimes are 7 min and 50 s in an argon matrix.11,22 The extrapolation of these lifetimes to the 2300 cm−1 barrier gives a reasonable value of 3 s, which means for cis-HOCO in an argon matrix kH/kD ∼ 103. This comparison of HOCO with carboxylic acids is based on similar structures of the HOC groups and hence similar barrier widths. In this respect, it is worth mentioning the HOCH-to-H2CO tunneling process that occurs in an argon matrix with a halftime of ∼2 h through a very high energy barrier of >10 000 cm−1.65 Hydrogen-atom tunneling in that case is only possible due to a relatively narrow barrier. The effect of barrier width on the tunneling control of chemical reactions is discussed elsewhere.66 The presented data clearly demonstrate the strong matrix effect on the tunneling rate. The decay of cis-DOCO is ∼94 times slower in a nitrogen matrix than in an argon matrix, which is similar to the corresponding values for cis-HCOOH and cis-CH3COOH (55 and 600).23 On the basis of our estimates of the cis-HOCO lifetime in an argon matrix, the stabilization of this species by a nitrogen matrix seems to be very strong. The lifetime of cis-HOCO in a nitrogen matrix is ∼18 h, which means kAr/kN2 > 104. The mechanism of this strong stabilization is not fully understood. In addition to the specific interactions with the nitrogen matrix molecules,23 it may be contributed by the medium reorganization required for tunneling. We have also measured the tunneling rate at elevated temperatures. The Arrhenius plots for HOCO in a nitrogen matrix and DOCO in an argon matrix are shown in Figure 4b. Both plots have a low-temperature limit, which strongly support the tunneling mechanism. It should be reminded that the increase of the decay rate at higher temperatures is not due to an overbarrier reaction but explained in terms of reorganization energy.67 This behavior is characteristic for tunneling in the solid phase, particularly for carboxylic acids.22,23 The decay of cis-DOCO in a nitrogen matrix is too slow to measure its temperature dependence.

■

CONCLUSIONS Carboxyl radical is important for atmospheric chemistry, astrochemistry, and combustion. In the present work, the conformers of this species have been studied by IR spectroscopy in argon and nitrogen matrices. For HOCO in an argon matrix, only the lower-energy conformer trans-HOCO is observed after the photochemical preparation from formic acid. In contrast, both trans- and cis-HOCO appear in a nitrogen matrix, which suggests the stabilization of the higherenergy cis conformer by a nitrogen matrix (Figure 1). For 9555

DOI: 10.1021/jacs.7b02605 J. Am. Chem. Soc. 2017, 139, 9551−9557

Article

Journal of the American Chemical Society stabilization of the higher energy conformer as found for the cisHCOOH···N2 complex in an argon matrix and explained by the stronger OH···N2 interaction compared to the OH···Ar one.68 However, the effect of the nitrogen matrix is much stronger and the lifetime of cis-HCOOH in a nitrogen matrix is 12 times longer than that of the cis-HCOOH···N2 complex in an argon matrix.23,68 For formic acid in noble-gas matrices, the polarizable continuum model provides reasonable results with respect to the relative tunneling rates (ref 22) but it completely fails (by orders of magnitude) for a nitrogen matrix (ref 42).

■

EXPERIMENTAL METHODS

■

AUTHOR INFORMATION

Leonid Khriachtchev: 0000-0002-1146-5212 Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This work was supported by the Academy of Finland (Projects No. 1277993 and No. 1288889). REFERENCES

(1) Seeman, J. I. Chem. Rev. 1983, 83, 83−134. (2) Ito, Y.; Ochiai, Y.; Park, Y. S.; Imanishi, Y. J. Am. Chem. Soc. 1997, 119, 1619−1623. (3) Donhauser, Z. J.; Mantooth, B. A.; Kelly, K. F.; Bumm, L. A.; Monnell, J. D.; Stapleton, J. J.; Price, D. W., Jr.; Rawlett, A. M.; Allara, D. L.; Tour, J. M.; Weiss, P. S. Science 2001, 292, 2303−2307. (4) Moresco, F.; Meyer, G.; Rieder, K.-H.; Tang, H.; Gourdon, A.; Joachim, C. Phys. Rev. Lett. 2001, 86, 672−675. (5) Roach, P.; Farrar, D.; Perry, C. C. J. Am. Chem. Soc. 2005, 127, 8168−8173. (6) Zwier, T. S. J. Phys. Chem. A 2006, 110, 4133−4150. (7) Shen, X.; Viney, C.; Johnson, E. R.; Wang, C.; Lu, J. Q. Nat. Chem. 2013, 5, 1035−1041. (8) Secundo, F. Chem. Soc. Rev. 2013, 42, 6250−6261. (9) Pettersson, M.; Lundell, J.; Khriachtchev, L.; Räsänen, M. J. Am. Chem. Soc. 1997, 119, 11715−11716. (10) Pettersson, M.; Maçôas, E. M. S.; Khriachtchev, L.; Fausto, R.; Räsänen, M. J. Am. Chem. Soc. 2003, 125, 4058−4059. (11) Maçôas, E. M. S.; Khriachtchev, L.; Pettersson, M.; Fausto, R.; Räsänen, M. J. Am. Chem. Soc. 2003, 125, 16188−16189. (12) Marushkevich, K.; Khriachtchev, L.; Lundell, J.; Räsänen, M. J. Am. Chem. Soc. 2006, 128, 12060−12061. (13) Fausto, R.; Khriachtchev, L.; Hamm, P. In Physics and Chemistry at Low Temperatures; Khriachtchev, L., Ed.; Pan Stanford: Singapore, 2011; pp 51−84. (14) Lapinski, L.; Nowak, M. J.; Reva, I.; Rostkowska, H.; Fausto, R. Phys. Chem. Chem. Phys. 2010, 12, 9615−9618. (15) Lopes Jesus, A. J.; Reva, I.; Araujo-Andrade, C.; Fausto, R. J. Am. Chem. Soc. 2015, 137, 14240−14243. (16) Amiri, S.; Reisenauer, H. P.; Schreiner, P. R. J. Am. Chem. Soc. 2010, 132, 15902−15904. (17) Wagner, J. P.; Reisenauer, H. P.; Hirvonen, V.; Wu, C.-H.; Tyberg, J. L.; Allen, W. D.; Schreiner, P. R. Chem. Commun. 2016, 52, 7858−7861. (18) Schreiner, P. R.; Wagner, J. P.; Reisenauer, H. P.; Gerbig, D.; Ley, D.; Sarka, J.; Császár, A. G.; Vaughn, A.; Allen, W. D. J. Am. Chem. Soc. 2015, 137, 7828−7834. (19) Mardyukov, A.; Quanz, H.; Schreiner, P. R. Nat. Chem. 2016, 9, 71−76. (20) Bazsó, G.; Najbauer, E. E.; Magyarfalvi, G.; Tarczay, G. J. Phys. Chem. A 2013, 117, 1952−1962. (21) Nanbu, S.; Sekine, M.; Nakata, M. J. Phys. Chem. A 2011, 115, 9911−9918. (22) Pettersson, M.; Maçôas, E. M. S.; Khriachtchev, L.; Lundell, J.; Fausto, R.; Räsänen, M. J. Chem. Phys. 2002, 117, 9095−9098. (23) Lopes, S.; Domanskaya, A. V.; Fausto, R.; Räsänen, M.; Khriachtchev, L. J. Chem. Phys. 2010, 133, 144507. (24) Maçôas, E. M. S.; Khriachtchev, L.; Pettersson, M.; Fausto, R.; Räsänen, M. Phys. Chem. Chem. Phys. 2005, 7, 743−749. (25) Encyclopedia of Radicals in Chemistry, Biology and Materials; Chatgilialoglu, C., Studer, A., Eds.; John Wiley & Sons: New York, 2012. (26) Halliwell, B.; Gutteridge, J. M. C. Free Radicals in Biology and Medicine, 5th ed.; Oxford University Press: Oxford, 2015. (27) Herbst, E. Annu. Rev. Phys. Chem. 1995, 46, 27−54. (28) Sorrell, W. H. Astrophys. J. 2001, 555, L129−L132. (29) Anderson, J. G. Annu. Rev. Phys. Chem. 1987, 38, 489−520. (30) Francisco, J. S.; Muckerman, J. T.; Yu, H.-G. Acc. Chem. Res. 2010, 43, 1519−1526.

HCOOH (>98%, Merck) and DCOOH (99%, Icon Isotopes) were degassed by several freeze−pump−thaw cycles. Nitrogen (>99.9999%, Linde) and argon (>99.9999%, Linde) were used as purchased. The gas FA/M (FA = HCOOH and DCOOH; M = N2 and Ar) mixtures were made in a glass bulb by using standard manometric techniques. Because FA is easily adsorbed on glass surfaces, the bulb was passivated with FA vapors by several fill−keep−evacuate cycles prior to mixture preparation. The FA/M (1/1000−1/2000) matrices were deposited onto a CsI substrate held at 10 and 15 K for nitrogen and argon, respectively, in a closed-cycle helium cryostat (RDK-408D2, SHI). The IR absorption spectra in the 4000−540 cm−1 range were measured at 4.5 K using an FTIR spectrometer (Vertex 80, Bruker) equipped with an MCT-B detector with 1 cm−1 resolution and 500 scans (or 100 scans in some kinetics measurement). In order to prepare carboxyl radicals, the deposited FA/M matrices were first irradiated by UV light at 193 nm (∼3000 pulses) or 250 nm (∼65 000 pulses) and then with VUV light, as described elsewhere.40 It should be mentioned that a number of other species are also formed in this preparation procedure.40 In particular, the OH···CO complex is observed in an argon matrix, but it is not found in a nitrogen matrix. In addition, the H2O···CO complex, H2O and CO monomers, and CO2 (possibly complexed with H2) are observed after photolysis. 193 nm photolysis was performed using an ArF excimer laser (MSX-250, MPB, 1 Hz, ∼5 mJ cm−2). An optical parametric oscillator (OPO, Sunlite, Continuum, 10 Hz, ∼5 mJ cm−2) provided photolysis at 250 nm. VUV-photolysis was performed with a Kr lamp (130−170 nm, Opthos). Selective vibrational excitation was made by narrow-band light (fwhm ∼0.1 cm−1, 10 Hz, ∼10 mJ cm−2) of another OPO (LaserVision). Some matrix-site splitting of the HOCO and DOCO bands is observed after preparation from formic acid (see Table 1). In this case, the tunneling decay of the cis form follows a biexponential function. The faster component of this function has a smaller amplitude and it disappears after annealing of the matrices together with the weaker matrix-site components (at 20 and 25 K for nitrogen and argon matrices). Thus, the faster component originates from the species in an unstable (unrelaxed) matrix site. For cis-HOCO in a nitrogen matrix at 4.5 K, the faster component has a weight of ∼25% and a lifetime of ∼2 h, whereas the slower component has a weight of ∼75% and a lifetime of ∼18 h. After annealing, the tunneling decay follow a single-exponential function with a lifetime of ∼18 h. The tunneling decay measurements (Figure 4) were performed on annealed matrices, which increases the accuracy of the results especially for measurements at elevated temperatures. The concentration of the cis form was maximized before these measurements by selective vibrational excitation of the trans form. To suppress the effect of broad-band IR light of the spectrometer on the tunneling process, a long-pass IR filter (Spectrogon, transmission range below 1850 cm−1) was used in these measurements.

Corresponding Author

*leonid.khriachtchev@helsinki.fi ORCID

Sergey V. Ryazantsev: 0000-0001-8662-580X 9556

DOI: 10.1021/jacs.7b02605 J. Am. Chem. Soc. 2017, 139, 9551−9557

Article

Journal of the American Chemical Society

(67) Trakhtenberg, L. I. In Atom Tunneling Phenomena in Physics, Chemistry and Biology; Miyazaki, T., Ed.; Springer: Berlin, 2004; pp 33−58. (68) Marushkevich, K.; Räsänen, M.; Khriachtchev, L. J. Phys. Chem. A 2010, 114, 10584−10589.

(31) Boxe, C. S.; Francisco, J. S.; Shia, R.-L.; Yung, Y. L.; Nair, H.; Liang, M.-C.; Saiz-Lopez, A. Icarus 2014, 242, 97−104. (32) Johnson, C. J.; Otto, R.; Continetti, R. E. Phys. Chem. Chem. Phys. 2014, 16, 19091−19105. (33) Nilsson, E. J. K.; Konnov, A. A. Energy Fuels 2016, 30, 2443− 2457. (34) Wang, Y.; Carter, S.; Bowman, J. M. J. Phys. Chem. A 2013, 117, 9343−9352. (35) Johnson, C. J.; Harding, M. E.; Poad, B. L. J.; Stanton, J. F.; Continetti, R. E. J. Am. Chem. Soc. 2011, 133, 19606−19609. (36) Oyama, T.; Funato, W.; Sumiyoshi, Y.; Endo, Y. J. Chem. Phys. 2011, 134, 174303. (37) McCarthy, M. C.; Martinez, O.; McGuire, B. A.; Crabtree, K. N.; Martin-Drumel, M.-A.; Stanton, J. F. J. Chem. Phys. 2016, 144, 124304. (38) Mielke, Z.; Olbert-Majkut, A.; Tokhadze, K. G. J. Chem. Phys. 2003, 118, 1364−1377. (39) Milligan, D. E.; Jacox, M. E. J. Chem. Phys. 1971, 54, 927−942. (40) Ryazantsev, S. V.; Duarte, L.; Feldman, V. I.; Khriachtchev, L. Phys. Chem. Chem. Phys. 2017, 19, 356−365. (41) Ryazantsev, S. V.; Feldman, V. I. Phys. Chem. Chem. Phys. 2015, 17, 30648−30658. (42) Tsuge, M.; Khriachtchev, L. J. Phys. Chem. A 2015, 119, 2628− 2635. (43) Huang, X.; Fortenberry, R. C.; Wang, Y.; Francisco, J. S.; Crawford, T. D.; Bowman, J. M.; Lee, T. J. J. Phys. Chem. A 2013, 117, 6932−6939. (44) Mladenovich, M. J. Chem. Phys. 2012, 137, 014306. (45) Wang, J.; Li, J.; Ma, J.; Guo, H. J. Chem. Phys. 2014, 140, 184314. (46) Jacox, M. E. J. Chem. Phys. 1988, 88, 4598−4607. (47) Ryazantsev, S. V.; Feldman, V. I. J. Phys. Chem. A 2015, 119, 2578−2586. (48) (a) Petty, J. T.; Moore, C. B. J. Mol. Spectrosc. 1993, 161, 149− 156. (b) Petty, J. T.; Moore, C. B. J. Chem. Phys. 1993, 99, 47−55. (49) Sears, T. J.; Fawzy, W. M.; Johnson, P. M. J. Chem. Phys. 1992, 97, 3996−4007. (50) Maçôas, E. M. S.; Khriachtchev, L.; Pettersson, M.; Juselius, J.; Fausto, R.; Räsänen, M. J. Chem. Phys. 2003, 119, 11765−11772. (51) Khriachtchev, L.; Lundell, J.; Isoniemi, E.; Räsänen, M. J. Chem. Phys. 2000, 113, 4265−4273. (52) Brunton, G.; Griller, D.; Barclay, L. R. C.; Ingold, K. U. J. Am. Chem. Soc. 1976, 98, 6803−6811. (53) Brunton, G.; Gray, J. A.; Griller, D.; Barclay, L. R. C.; Ingold, K. U. J. Am. Chem. Soc. 1978, 100, 4197−4200. (54) Sirjean, B.; Dames, E.; Wang, H.; Tsang, W. J. Phys. Chem. A 2012, 116, 319−332. (55) Cao, Q.; Berski, S.; Latajka, Z.; Räsänen, M.; Khriachtchev, L. Phys. Chem. Chem. Phys. 2014, 16, 5993−6001. (56) Knight, L. B., Jr.; King, G. M.; Petty, J. T.; Matsushita, M.; Momose, T.; Shida, T. J. Chem. Phys. 1995, 103, 3377−3385. (57) Yamada, T.; Komaguchi, K.; Shiotani, M.; Benetis, N. P.; Sørnes, A. R. J. Phys. Chem. A 1999, 103, 4823−4829. (58) Johnson, C. J.; Continetti, R. E. J. Phys. Chem. Lett. 2010, 1, 1895−1899. (59) Ma, J.; Li, J.; Guo, H. Phys. Rev. Lett. 2012, 109, 063202. (60) Wang, X.; Bowman, J. M. J. Phys. Chem. A 2014, 118, 684−689. (61) Wang, J.; Li, J.; Guo, H. J. Chem. Phys. 2014, 140, 184314. (62) Benderskii, V. A.; Goldanskii, V. I. Int. Rev. Phys. Chem. 1992, 11, 1−70. (63) Kohen, A. In Isotope Effects In Chemistry and Biology; Kohen, A., Limbach, H.-H., Eds.; CRC Press: Boca Raton, 2006; pp 743−764. (64) Marushkevich, K.; Khriachtchev, L.; Räsänen, M. J. Chem. Phys. 2007, 126, 241102. (65) Schreiner, P. R.; Reisenauer, H. P.; Pickard, F. C.; Simmonett, A. C.; Allen, W. D.; Mátyus, E.; Császár, A. G. Nature 2008, 453, 906− 909. (66) Schreiner, P. R.; Reisenauer, H. P.; Ley, D.; Gerbig, D.; Wu, C. H.; Allen, W. D. Science 2011, 332, 1300−1303. 9557

DOI: 10.1021/jacs.7b02605 J. Am. Chem. Soc. 2017, 139, 9551−9557