Radiation-induced transformations of acetaldehyde molecules at cryogenic temperatures: a matrix isolation study

Abstract

Acetaldehyde is one of the key small organic molecules involved in astrochemical and atmospheric processes occurring under the action of ionizing and UV radiation. While the UV photochemistry of acetaldehyde is well studied, little is known about the mechanism of processes induced by high-energy radiation. This paper reports the first systematic study on the chemical transformations of CH₃CHO molecules resulting from X-ray irradiation under the conditions of matrix isolation in different solid noble gases (Ne, Ar, Kr, and Xe) at 5 K. CO, CH₄, H₂CCO, H₂CCO–H₂, C₂H₂⋯H₂O, CH₂CHOH, CH₃CO˙, CH₃˙, HCCO˙, and CCO were identified as the main radiolysis products. The dominant pathway of acetaldehyde degradation involves C–C bond cleavage leading to the formation of carbon monoxide and methane. The second important channel is dehydrogenation resulting in the formation of ketene, a potentially highly reactive species. It was found that the matrix significantly affected both the decomposition efficiency and distribution of the reaction channels. Based on these observations, it was suggested that the formation of the methyl radical as well as vinyl alcohol and the C₂H₂⋯H₂O complex presumably included a significant contribution of ionic pathways. The decomposition of acetyl radicals under photolysis with visible light leading to the CH₃˙–CO radical-molecule pair was observed in all matrices, while the recovery of CH₃CO˙ in the dark at 5 K was found only in Xe. This finding represents a prominent example of matrix-dependent chemical dynamics (probably, involving tunnelling), which deserves further theoretical studies. Probable mechanisms of acetaldehyde radiolysis and their implications for astrochemistry, atmospheric chemistry and low-temperature chemistry are discussed.

---

1. Introduction

Acetaldehyde (CH₃CHO) is known to be a crucially important molecule for astrochemistry and atmospheric chemistry. Considering astrochemistry, one may note that CH₃CHO was detected in different interstellar media (e.g. ref. 1–6) and, due to the presence of the carbonyl group (C=O), it could be involved in various chemical reactions in the interstellar medium (ISM) to yield complex organic molecules (COMs), including those of biological relevance. For example, acetaldehyde may play an important role in the formation of carbohydrates,⁷ acrolein (crucial intermediate in the synthesis of different amino acids),⁸ and deoxyribonucleosides (the major component of DNA).⁹ Actually, this small and chemically reactive molecule presumably originating from the simplest highly abundant precursors (such as H₂O, CO, CO₂, NH₃, and CH₄ (see ref. 10 and references therein)) can be considered as one of the key intermediate molecular species in organic and prebiotic astrochemistry. On the other hand, the phototransformations of acetaldehyde play a meaningful role in the chemistry of polluted atmospheres and therefore constitute a matter of concern to atmospheric chemistry.^11,12 For this reason, investigations of the possible reactions involving acetaldehyde under the conditions relevant to the space and Earth atmosphere radiation environment (including UV-photolysis, gamma-rays, low-energy electrons, and protons) were a subject of extensive theoretical and experimental studies (see ref. 13–28 and references therein).

The photochemistry of CH₃CHO in the gas phase was investigated experimentally for several decades. It is well established that the UVB (280–315 nm) photolysis leads to a population of the S₁ electronic state, which decays generally via three possible reaction channels: decomposition to CH₃˙ and HCO˙, CH₄ and CO, and CH₃CO˙ and H˙ though the impact of the third channel is quite small in the UVB region (at 300 nm ϕ₃ ∼ 0.025 and decreases to zero at 320 nm).^{15–17,19,21} The first channel strongly predominates in the UVB region (e.g. at 280 nm ϕ₁ = 0.58 and ϕ₂ = 0.06)¹⁹ and proceeds almost exclusively through the intersystem crossing to T₁ state.^20,21 Meanwhile, the internal conversion or sequential intersystem crossing (S₁ → T₁ → S₀) may result in a population of highly vibrational excited S₀ state of acetaldehyde where all three channels compete.^20,21 The contribution of the second channel (decomposition to molecular products) increases at a shorter wavelength and it becomes dominant at λ ≤ 266 nm.¹⁹ Additionally, vinyl alcohol was found to be the primary, collision-induced photoproduct following photolysis of neat CH₃CHO with 295–330 nm light.¹⁶ Formation of H₂CCO and H₂ under 305.6 nm photolysis of CH₃CHO was also observed but the quantum yield of this channel was determined to be rather small (at 305 nm ϕ ∼ 0.0075).²¹ It should be noted that the decay to three fragments (CH₃˙ + CO + H˙) becomes thermodynamically feasible at λ < 294 nm.^17,20 The occurrence of this channel under photolysis of CH₃CHO with a 248 nm light was deduced from analysis and comparison of primary quantum yields of CH₃˙, HCO˙, and H˙ species¹⁷ and experimentally observed by time-resolved infrared spectroscopy.¹⁸ Recently, Yang et al.^22,23 tracked the time-evolution of the pair-correlated products of CH₃CHO photolysis with 267 nm light and made the detailed roadmap of acetaldehyde photodissociation which shed light on the competing pathways that lead to the CH₃˙ + HCO˙ and CH₄ + CO products. Quite different photochemistry of gaseous CH₃CHO was observed under VUV photolysis at 157.6 nm.¹⁵ Absorption of a photon with such energy by an acetaldehyde molecule results in a population of the Rydberg states of CH₃CHO (n → 3p transition), mainly followed by conversion to T₁ and S₀ states but some acetaldehyde molecules isomerize to vinyl alcohol followed by multichannel decomposition. In addition to CH₃CO˙, H₂CCO, HCO˙, CO, CH₃˙, H₂, and H˙ species several new photofragments (namely, H₂CO, C₂H₃˙, C₂H₂, H₂O, OH˙, and CH₂) were detected using time-of-flight mass spectrometry. It is worth noting that the formation of C₂H₃˙ and OH˙ and C₂H₂ and H₂O photofragments observed in these experiments indicates that isomerization of acetaldehyde to vinyl alcohol occurs before fragmentation.¹⁵ However, the precise nature of the processes induced by high-energy photons (close to the ionization limit) is not well understood.

In contrast to photochemistry, the radiation chemistry of acetaldehyde is poorly studied. CO, CH₄, and H₂ were identified as the main products of the CH₃CHO radiolysis with gamma-rays in the gas phase.²⁴ According to the early EPR results,²⁵ the radiolysis of frozen neat acetaldehyde at 77 K produces CH₃CO˙ radicals and alkyl-type radicals (the structure of the latter species cannot be deduced unequivocally from the available data). Irradiation of CH₃CHO ices with low-energy electrons (4–15 eV) at 35 K was reported to result in the formation of CO, CH₄, CO₂, C₂H₄, CH₃CH₂OH, CH₃CH₂CHO, and CH₃CH(OH)CH₃ as measured by thermal desorption spectrometry (TDS).²⁶ CO, CH₄, and CH₃CO˙ were detected in the FTIR spectra of the low-temperature (5 K) acetaldehyde ices irradiated with 5 keV electrons (additionally, ion signals of H₂CCO, CH₂CH₂OH, CH₃COCH₃, (CH₃CHO)₂H⁺, and (C₄H₇HO₂)H⁺ were observed in photoionization reflectron time-of-flight mass spectrometry (PI-ReToF-MS) detections).²⁷ The proton radiolysis of pure solid acetaldehyde at 18 K produces mostly CO, CH₄, H₂CCO, CO₂, CH₃CO˙, and CH₂CHO˙.²⁸ We may note that the mechanism of radiolysis of CH₃CHO both in the gas and in condensed phases remains unclear.

The lack of mechanistic information on the radiation-induced transformations of acetaldehyde at the molecular level presents a considerable gap in a basic understanding of the chemical evolution of this ubiquitous molecule in space (particularly, in the ISM) and upper layers of the atmosphere. Indeed, under such conditions, CH₃CHO (like most other organic molecules) occurs mainly in the form of small admixture and does not form phases. Generally speaking, the matrix isolation technique provides a powerful solution for this problem, particularly useful for basic simulation of the radiation-induced transformations of small molecules in inert ices.^29,30 It opens up opportunities for step-by-step simulation of the evolution of the focus isolated molecules at different stages of radiolysis (up to very high conversions), which gives a deeper insight into the mechanisms of both primary processes and secondary transformations and makes it a valuable tool for radiation chemistry, astrochemistry, and atmospheric chemistry.³⁰ Recently this method was applied by our group to comprehensively investigate the radiation-induced chemistry of several small organic molecules of astrochemical interest, including hydrocarbons,^31–33 methanol,³⁴ formic acid,³⁵ and simple nitriles^36–38 using FTIR spectroscopy as the most versatile and convenient spectroscopic method. A few studies on acetaldehyde photochemistry using matrix isolation FTIR spectroscopy were reported previously. Védova and Sala³⁹ observed the formation of CH₄ and CO after photolysis of CH₃CHO/Ar and CH₃CHO/N₂ samples with a medium-pressure mercury lamp, whereas Schriver et al.⁴⁰ have found CH₄, CO, C₂H₂⋯H₂O complex, and H₂CCO species in the CH₃CHO/N₂ matrix irradiated with microwave discharge hydrogen flow lamp. Thompson et al.⁴¹ have carried out a sort of different experiment using in situ electron beam processing of the CH₃CHO/Ar gaseous mixture before the sample deposition on a low-temperature wafer and reported the formation of CO, CH₄, H₂CCO, HCCO˙, HCO˙, CH₃CO˙, and vinyl alcohol trapped in a deposited matrix. It is worth mentioning that matrix isolation was has been applied for characterization of the CH₃CHO˙⁺ radical cation and radical products of its fragmentation and deprotonation observed after irradiation of the CH₃CHO/Ng (Ng = Ne, Ar, or Xe) solid matrices in the presence of electron scavengers using EPR spectroscopy.^42–44 These works provided some important clues on possible ionic pathways. However, EPR spectroscopy using alone does not provide any information on diamagnetic products, which makes it impossible to deduce unequivocal mechanistic conclusions.

In the present work, we report the first systematic study on the X-ray radiation-induced transformations of acetaldehyde at 5 K in various solid noble gas hosts (Ne, Ar, Kr, or Xe) over a wide range of absorbed dose using FTIR matrix isolation spectroscopy and discuss possible mechanisms and their implications for astrochemistry, atmospheric chemistry, and solid-state radiation chemistry. In addition, we present new data on reversible photoinduced and thermal transformations of the closely related acetyl radical.

2. Experimental details

Acetaldehyde (CH₃CHO, 99.5%, Acros Organics; CD₃CDO, ≥98% purity, ≥99% atom D, Sigma Aldrich) was additionally purified by several freeze–pump–thaw cycles before use in mixture preparation. Neon (99.996%, Akela-N), argon (99.9995%, Voessen), krypton (99.9998%, Akela-N), and xenon (99.9994%, Medxenon) were used as received. Gaseous mixtures of acetaldehyde with noble gas (typical ratio – 1 : 1000, noble gas = Ne, Ar, Kr, Xe), which were used in the main set of experiments, were prepared by standard manometric technique. Complementary experiments were carried out with CH₃CHO/CFCl₃/Ng (Ng = Ar, Xe) 1 : 2 : 1000 mixtures (CFCl₃, 99.8%, Khimprom, Volgograd, Russia).

The experiments were performed using a closed-cycle helium cryostat based on the SHI RDK-101E cryocooler (a detailed description of the experimental setup can be found elsewhere).⁴⁵ The pressure inside the cryostat chamber was less than 10⁻² Pa before the deposition procedure. Prepared mixtures were slowly (ca. 2 mmol h⁻¹) deposited onto a cooled KBr substrate of the cryostat and the deposition temperatures were typically about 8, 18, 25 and 33 K for Ne, Ar, Kr and Xe matrices, respectively. The sample temperature was controlled using a Lakeshore 325 temperature controller connected to the Cernox-type sensor, or a t-STAT 310 cm device (RTI Cryomagnetic Systems, Russia) connected to the calibrated Cu:Cu(Fe) thermocouple. The composition and thickness of the depositing samples were monitored by the FTIR spectra recorded during the deposition procedure. Typical thicknesses of deposited samples were ca. 120–140 μm for a Ne matrix and ca. 70–100 μm for Ar, Kr and Xe matrices.

The deposited matrices were cooled down to 5 K (minimal temperature attainable at the cryostat) and then irradiated with X-rays through an aluminium foil (thickness of the foil is 45 μm) window mounted in the cryostat. Irradiation was performed using a 5BKhV-6(W) tube with a tungsten anode (45 kV_p, anode current 80 mA, effective X-ray energy ca. 20 keV). The absorbed dose rate was estimated in our previous work;⁴⁶ the values are ca. 15.4, 38.6, 72.9, and 55.0 Gy s⁻¹ (1 Gy = 1 J kg⁻¹, 1 kGy = 10³ Gy) for Ne, Ar, Kr, and Xe, respectively (in the case of 80 μm thick noble gas matrix which is a typical thickness in our experiment while in Ne the absorbed dose is virtually independent on sample thickness).⁴⁶ The irradiation time varied from 1 to 120 min, depending on the matrix material (the absorbed dose varied from ca. 4 to 111 kGy). Photolysis of the irradiated samples with visible light was performed using a set of ARPL-STAR-3W LEDs (λ_D = 620 nm and λ_D = 520 nm).

The FTIR spectra of the samples were recorded at 5 K in the 6500–420 cm⁻¹ range using a Bruker Tensor II spectrometer equipped with a cooled MCT detector (resolution 1 cm⁻¹, averaging 144–500 scans).

3. Results

The FTIR spectra of the deposited CH₃CHO/Ng 1 : 1000 mixtures (Ng = Ne, Ar, Kr, or Xe) demonstrate characteristic absorption bands of the isolated acetaldehyde molecules in good agreement with previous studies^39,47–49 (the list of main absorption bands is given in Table S1 of ESI†). It should be noted that the assignments in Ne, Kr, and Xe matrices were made based on previous reports on the corresponding features in an Ar matrix^39,48 taking into account reasonable matrix shifts. In addition to the isolated CH₃CHO, traces of carbon dioxide, water and water associates, which are common impurities in matrix samples, were also detected in the deposited matrices.⁵⁰

Radiolysis of the deposited matrix samples with X-rays results in the efficient decomposition of acetaldehyde (Fig. 1). The radiation-chemical yield of the CH₃CHO degradation estimated from initial slopes of the recorded kinetics profiles are ca. 0.18, 0.27, 0.27, and 0.08 μmol J⁻¹ for Ne, Ar, Kr, and Xe, respectively, i.e. the efficiency of acetaldehyde decomposition increases in a row Xe < Ne < Kr ∼ Ar.

Fig. 1 Kinetics of the radiation-induced decay of acetaldehyde molecules in the CH₃CHO/Ng 1 : 1000 samples (Ng = Ne, Ar, Kr, or Xe) irradiated with X-rays. The relative concentration of acetaldehyde in the samples irradiated to different doses was calculated by integration of the C–C_str absorption band. The initial molar mass concentrations of CH₃CHO were determined based on a molar mixture ratio of 1 : 1000.

Several new absorption bands appear in the FTIR spectra of the irradiated CH₃CHO/Ng (Ng = Ne, Ar, Kr, or Xe) samples (Fig. 2). These features should be ascribed to the radiolysis products. The main new species observed in the irradiated samples are CO, CH₄, H₂CCO, H₂CCO–H₂, C₂H₂⋯H₂O, CH₂CHOH, CH₃CO˙, CH₃˙, HCCO˙, and CCO, which were assigned (see Table 1) based on literature data.^49,51–63 Traces of HCO˙^64–66 (1866.2; 1863.0; 1862.3, 1859.9; and 1858.6, 1856.9 for Ne, Ar, Kr, and Xe, respectively) were also found in the irradiated CH₃CHO/Ng samples. Additionally, the signals of positive Ng₂H⁺ ions were also observed in Ar, Kr, and Xe irradiated matrices (1139.5 and 903.4 cm⁻¹ for Ar₂H⁺; 1159.8, 1007.7 and 852.6 cm⁻¹ for Kr₂H⁺; 952.5, 842.4 and 730.5 cm⁻¹ for Xe₂H⁺, respectively).⁶⁷

Fig. 2 Difference FTIR spectra illustrating the effect of X-ray irradiation on the CH₃CHO/Ng 1 : 1000 samples recorded at 5 K: Ng = Ne (a), Ar (b), Kr (c), and Xe (d). In the case of Xe, absorption band of unperturbed H₂CCO is masked by H₂CCO–H₂ one. Absorption bands of the CO_rock-opla ketene band, which could include both perturbed and unperturbed ketene signals, are marked with asterisks. Note that both unperturbed (H₂CCO) and perturbed (H₂CCO–H₂) ketene forms may contribute to the demonstrated absorption band of ketene in a Ne matrix which is designed as H₂CCO–H₂ (see text for details). The unassigned absorption band with the maximum at 1288.3 cm⁻¹ in a Ne matrix is marked with a question sign.

Table 1 Absorption maxima (cm⁻¹) of the species appear under radiolysis of the CH₃CHO/Ng 1 : 1000 (Ng = Ne, Ar, Kr, or Xe) samples. Tentative assignments are in italic

Species	Assignment	Ne^a	Ar	Kr	Xe	Ref.
sh – shoulder, OPLA – out of plane.a In a Ne matrix both H2CCO/H2CCO–H2 (D2CCO/D2CCO–D2) may contribute to all observed absorption bands of ketene except CH2 s-str (CD2 s-str), see text for details. Owing to this reason we provide the same values of absorption maxima both for H2CCO and H2CCO–H2 (D2CCO/D2CCO–D2).b Fermi resonance with an overtone or a combination band.^59
CH3CHO
CO	COstr	2143.6sh	2137.9	2135.6	2133.0	51 and 52
		2141.3			2131.0sh
CH4	CH3 d-str	3036.3	3032.0	3025.0	3020.1	53
	CH3 d-deform	1309.0	1305.6	1306.4	1309.4
				1303.0	1303.7
					1300.6
H2CCO	CH2 s-str	3069.4	3063.0	3054.9	3052.5	52,54 and 55
	COstr	3076.7	2142.2	2139.7	2136.7
	CH2 scis	2151.3
	CH2 rock-opla	1386.5
	COrock-opla	592.2
		528.3
H2CCO–H2	CH2 s-str	3079.3	3071.7	3066.5	3056.2	56
		3076.7		3063.9
	COstr	2151.3	2148.0	2145.0	2139.8
	CH2 scis	1386.5	1380.9	1378.2	1374.8
	CH2 rock-opla	592.2		590.7	590.8
	COrock-opla	528.3	525.5	522.4
C2H2⋯H2O	C–Hstr	3252.2	3240.0	3234.1	—	56 and 57
		3248.2
	C–Hbend		785.8	783.1
CH2CHOH	Mixed mode	1077.7	1078.8	1082.1	1078.9	49 and 58
				1077.2
	OPLA	823.6	816.8	816.4	812.8
		818.0	813.7	811.1
CH3CO˙	COstr^b	1884.3	1874.9	1886.2	1880.7	49,52 and 59
			1837.7	1849.7	1844.4
	CH3 deform	1326.0	1321.2	1322.7	1318.6
CH3˙	OPLA	610.0sh	607.1	609.5	602.9	52 and 60
		607.6		604.2
HCCO˙	CCOdeform	2024.3	2030.8	2026.2	2018.9	61 and 62
		2019.5sh	2023.2	2019.7	2016.2
		2015.8	2019.4		2013.7
CCO	COstr	1972.9	1973.5	1970.5	1965.9	63
CD3CDO
CO	COstr	2143.6sh	2137.9	2135.7	2132.9	51 and 52
		2141.2			2131.0sh
CD4	CD3 d-deform	997.5	994.1	997.7	997.3	68
				995.0sh	992.9sh
				992.2sh	990.0sh
D2CCO	CD2 s-str	2266.1	2260.0	2252.1	2250.4sh	52,54 and 55
	ν 3 + ν4	2158.0	2149.3	2146.0
	COstr	2121.4	2112.7	2109.9	2106.8
	CD2 scis	1229.1
	CD2 rock-opla	543.1
	COrock-opla	434.8		431.2
D2CCO–D2	CD2 s-str	2271.1	2264.2	2262.5	2255.7	—
	ν 3 + ν4	2158.0	2155.2	2151.7	2146.4
	COstr	2121.4	2119.2	2115.9	2110.7
	CD2 scis	1229.1	1229.2	1228.1	1225.7
	CD2 rock-opla	543.1	542.0	538.7	537.2
	COrock-opla	434.8	431.6	434.6	429.9
C2D2⋯D2O	C–Dstr	2419.3	2411.5	2405.8	2405.2	57
		2416.6			2402.9
		2414.3
	C–Dbend		585.1	584.5
				581.5
CD2CDOD	Mixed mode	923.4	925.2	923.6	921.1	58
			921.5	919.9	916.2
	OPLA	650.5	650.0sh	648.1	645.5
			648.4	646.6
CD3CO˙	COstr	—	1866.7sh	1864.3	1861.2	52 and 59
			1859.5
			1856.0
	CD2 scis		1024.6	1026.4	1021.7
	CCDdeform			896.1	893.1
				891.8
CD3˙	OPLA	457.5	454.9	454.0	453.8	52 and 60
DCCO˙	CCOdeform	1995.1	1994.1	1990.5	1989.2sh	61 and 62
		1992.0	1990.0	1987.4sh	1986.4
CCO	COstr	1973.1	1973.9	1969.8	1966.0	63

---

To verify the assignment of the observed absorption bands to the above-mentioned radiolysis product, we have performed a set of complementary experiments with deuterated acetaldehyde. The FTIR spectra of the deposited CD₃CDO/Ng (Ng = Ne, Ar, Kr, Xe) samples demonstrate absorption bands of CD₃CDO (presented in Table S1 of ESI†)^47,48 as well as traces of CO₂, water, and water associates.⁵⁰ The effect of X-ray irradiation on the deposited CD₃CDO/Ng samples is shown in Fig. 3. Absorption bands of the main radiolysis products of deuterated acetaldehyde (CO, CD₄, D₂CCO, D₂CCO–D₂, C₂D₂⋯D₂O, CD₂CDOD, CD₃CO˙, CD₃˙, DCCO˙, and CCO) observed in the spectra of the irradiated samples are summarized in Table 1 (the assignment was made based on previously published data; see ref. 51, 52, 54–63 and 68). Very weak signals of DCO˙ 1925.7; 1917.7; 1910.4, and 1907.4 cm⁻¹ for Ar, Kr, and Xe, respectively)⁶³ were also detected in the irradiated samples. Additionally, the signals of positive Ng₂D⁺ ions were found in Ar, Kr, and Xe matrices (643.0 cm⁻¹ for Ar; 770.4 and 605.9 cm⁻¹ for Kr; 750.7, 634.2, and 516.3 cm⁻¹ for Xe).⁶⁷ We have also detected several relatively weak absorption bands in the irradiated CH₃CHO/Ng and CD₃CDO/Ng matrices, which were not identified (the list of these bands is provided in Table S2 of ESI†).

Fig. 3 Difference FTIR spectra illustrating the effect of irradiation on the CD₃CDO/Ng 1 : 1000 (Ng = Ne, Ar, Kr, or Xe) samples with X-rays. Each spectrum represents the data obtained at the maximum absorbed dose achieved in the corresponding experiment. In the case of Ar, the signal of CD₃CO˙ is masked by the absorption band of CD₃CDO. Note that both unperturbed (D₂CCO) and perturbed (D₂CCO–D₂) ketene forms may contribute to the observed absorption bands of ketene in a Ne matrix, which are designated as D₂CCO–D₂ (see text for details).

Analysis of the FTIR spectra shown in Fig. 2 and 3 allows one to make some preliminary conclusions about the medium effect on the radiation-induced transformations of acetaldehyde. First, we may note the matrix effect on the peculiarities of stabilization of ketene and ketenyl radical. In fact, two forms of ketene (H₂CCO) result from the radiation-induced transformations of acetaldehyde, as manifested by a characteristic splitting pattern for the corresponding absorption bands in Ar, Kr and Xe matrices. These forms were previously identified as “unperturbed” and “perturbed” ketene. The former term corresponds to the isolated H₂CCO molecules, while the latter one stands for a H₂CCO–H₂ caged molecular pair.⁵⁶ Intermolecular interactions in this pair were found to be rather small,⁵⁶ therefore, we refrain from the term “intermolecular complex” in this case. The proportion of the unperturbed ketene species to the perturbed one strongly increases from Xe to Ar. In a Ne matrix, however, most of the H₂CCO absorption bands did not demonstrate such a pattern (Fig. 2) as well as infrared absorption bands of D₂CCO in the irradiated CD₃CDO/Ne samples (Fig. 3). Nevertheless, characteristic splitting was observed for the CH_2 s-str vibrational mode of ketene in a Ne matrix, which implies that both unperturbed and perturbed ketene may be formed in a Ne matrix as well. Thus, we may tentatively assume that both of these ketene forms can contribute to the intensity of all other absorption bands of ketene species isolated in Ne, but the splitting is too small to be observed in this matrix. The nature of this effect may be of certain interest for molecular spectroscopy, but it is out of the scope of the present study. Meanwhile, we may conclude that the proportion of unperturbed and perturbed ketene extracted from the absorption intensity related to the CH_2 s-str mode of these species in Ne is roughly comparable to that in an Ar matrix. Similarly, the absorption band of ketenyl radical (HCCO˙) also splits into two components, which may be explained by the formation of perturbed and unperturbed HCCO˙ from the corresponding ketene forms, respectively. Second, the efficiency of the vinyl alcohol (CH₂CHOH) and C₂H₂⋯H₂O complex production is strongly matrix dependent and markedly decreases from Ar to Xe. In Ne and Ar matrices the yields of C₂H₂⋯H₂O complex are roughly comparable, while the relative yield of CH₂CHOH tends to decrease from Ar to Ne. Third, rather strong signals of methyl radical (CH₃˙) were observed in Ne and Ar matrices while that in Kr is significantly weaker and there is virtually no sign of methyl radical in Xe matrix. Fourth, the absorption bands of acetyl radical (CH₃CO˙) dramatically increases from Ne to Xe. The possible origin of the matrix effect will be considered in the Discussion section.

Essential information on the mechanisms of the acetaldehyde radiation-induced transformations may be derived from the analysis of the accumulation profiles of its radiolysis products. These profiles are presented in Fig. 4. One should notice that the invariant coordinates (normalized concentration vs. conversion degree of acetaldehyde) were used to compare the kinetic profiles in different matrices. The infrared intensities used in this work are provided in Table S3 of ESI.† We should note that the infrared intensities were mainly taken from the computational works completely neglecting the matrix effect (see Table S3, ESI† for the references), which brings considerable uncertainty. Nevertheless, we believe that such representation is suitable to identify the dominant channels of the acetaldehyde radiation-induced transformations.

Fig. 4 Build-up profiles of the major CH₃CHO radiolysis products in (a) Ne, (b) Ar, (c) Kr and (d) Xe matrices. Note that the H₂CCO curves represent the total integrated intensity of perturbed and unperturbed ketene signals. The integrated intensities of the absorption bands were normalized to an appropriate absorption coefficient (see Table S3 of ESI†), multiplied by the absorption coefficient of CO, and divided by the maximum integrated intensity of the CO band found in a given experiment.

Analysis of the kinetic curves shown in Fig. 4 reveals that carbon monoxide (CO) and methane (CH₄) are the principal radiolysis products of isolated CH₃CHO in all matrices. These products appear in considerable amounts already at low conversion of acetaldehyde, which indicates their predominant formation as primary radiolysis products. Meanwhile, we suppose that some deviations in their accumulation from linearity (increasing formation rate at higher absorbed doses) may be caused by the contribution of secondary reactions in their formation (see Discussion section). One may notice that the CO yield is higher than that of CH₄ and the ratio of CO to CH₄ is highest in Kr and smallest in Ar (Fig. 4). It should be taken into account that the same value of CO/CH₄ molar absorption coefficients ratio was taken for all matrices, which may result in significant systematic error. However, the discrepancy between CO and CH₄ amounts (particularly, in a krypton matrix) seems to be too large to be explained solely by this uncertainty. Thus, one should consider some additional reaction channels, which lead to the formation of CO without the production of CH₄ (see Discussion section). Another important pathway of the CH₃CHO radiolysis is the formation of ketene species in different forms, which also exhibit a roughly linear accumulation profile. Regarding other radiolysis products, one may notice that the amounts of C₂H₂⋯H₂O and CH₃˙ increase linearly with the CH₃CHO conversion, whereas the production of CH₂CHOH and CH₃CO˙ tend to saturate at higher conversion degrees of the parent acetaldehyde molecules. Finally, the formation rate of HCCO˙ and CCO species increases with increasing acetaldehyde conversion (see Fig. 4).

We have examined the effect of electron scavenger Freon-11 (CFCl₃) on the formation of CH₃CHO radiolysis products in Ar and Xe matrices to get a deeper insight into the mechanisms of its radiation-induced transformations. The decomposition rate of CH₃CHO decreases in the presence of CFCl₃: for instance, the conversion degree of CH₃CHO was found to be ca. 37% in the CH₃CHO/Ar and ca. 17% in the CH₃CHO/CFCl₃/Ar mixture at the same irradiation dose (ca. 23 kGy). A similar-type effect was observed in a Xe matrix, where the addition of CFCl₃ led to a significant decrease in the acetaldehyde decomposition rate (39 and 9% conversion of CH₃CHO was reached at the absorbed dose of ca. 40 kGy for the samples containing only acetaldehyde and acetaldehyde and Freon-11, respectively). This effect is likely to result from competition between CFCl₃ for positive holes and excitons produced in a matrix (in other words, a considerable part of absorbed energy is used for Freon decomposition). Indeed, ca. 18% and ca. 41% of CFCl₃ molecules were decomposed under these conditions in the cases of Ar and Xe matrices, respectively (we also observed the absorption bands of CFCl₃ radiolysis products).⁶⁹ The effect of electron scavenger on the CH₃CHO radiolysis is illustrated by Fig. 5 (build-up profiles of the major CH₃CHO radiolysis products in Ar and Xe matrices doped with an electron scavenger are provided in Fig. S1 of ESI†). One can see a noticeable decrease of the CO, H₂CCO, H₂CCO–H₂, C₂H₂⋯H₂O, CH₃˙, CH₃CO˙, HCCO˙, and CCO yields in the presence of CFCl₃, whereas the yields of CH₄ and Ar₂H⁺ are almost unaffected. Doping of a Xe matrix with an electron scavenger have no prominent effect on CH₄ and CO yields and leads to moderate (by ca. 1.4 times) decrease of H₂CCO, H₂CCO–H₂, HCCO˙, and CCO signals, while the CH₃CO˙ yield decreases dramatically. In contrast, the yield of Xe₂H⁺ noticeably increases. It is worth noting that absorption bands of CH₂CHOH and CH₃˙ both in Ar and Xe matrices overlaps with CFCl₃, so no definite conclusion can be made on the effect of a scavenger on the formation of these species.

Fig. 5 Difference FTIR spectra of the irradiated CH₃CHO/CFCl₃/Ng (Ng = Ar, Xe; 1 : 0–2 : 1000) illustrating the effect of an electron scavenger (CFCl₃) addition on the CH₃CHO radiation-induced transformations. Conversion degree of acetaldehyde is ca. 22, 17, 29, and 29 per cent for the CH₃CHO/Ar, CH₃CHO/CFCl₃/Ar, CH₃CHO/Xe, and CH₃CHO/CFCl₃/Xe samples, respectively. Note that in the case of Xe matrix absorption band of unperturbed H₂CCO is masked by H₂CCO–H₂ one. Absorption bands of CFCl₃ and CFCl₂⁺ are marked with diamonds and bullets, respectively.

Photolysis of irradiated CH₃CHO/Ng and CD₃CDO/Ng (Ng = Ne, Ar, Kr, Xe) 1 : 1000 samples (λ_D = 620 nm) results in bleaching of acetyl and formyl radicals and formation of their photolysis products. Growth of signals of methyl radicals and carbon monoxide molecules was observed in the FTIR spectra of the photolysed samples.^43,59,70 Hydrogen atoms should be also produced in the process of HCO˙ photolysis, but they cannot be observed via FTIR spectroscopy. It is worth noting that the efficiency of acetyl radical photolysis depends on matrix material. It was found that acetyl radicals completely decompose after 15 min photolysis with λ_D = 620 nm light in Ne and Ar matrices, while in Kr and Xe matrices only partial decay of CH₃CO˙ was observed. It may be explained by more effective quenching of excited states in more polarizable matrices (Kr and Xe) which should reduce the efficiency of photoinduced decomposition.

Photolysis of the X-ray irradiated samples with the shorter wavelength light (λ_D = 520 nm) for 30 min results in complete bleaching of the remaining acetyl radicals in a Kr matrix while in the case of a xenon a major part (more than a half) of these radicals persists even after prolonged (>60 min) photolysis. Furthermore, it was observed that after about 40 minutes of photolysis further action of λ_D = 520 nm light had virtually no effect on remaining acetyl radical in a Xe matrix. It is worth noting that new absorption bands appear in the CH₃CHO/Ng and CD₃CDO/Ng samples after photolysis (see Table 2; an example of the CH₃CHO/Xe sample is provided in Fig. 6, panel a). It is worth noting that some of these absorption bands are slightly shifted from bands of CO and CH₃˙ (CD₃˙) monomeric species (see Table 2 and Fig. S2 of ESI†). These absorption bands could be attributed to the CH₃ and CO radical-molecular pair (CH₃–CO) trapped within the same matrix cage (to note, the formation of the CH₃–CO pairs was previously identified in a solid p-H₂ matrix).⁵²

Table 2 Absorption maxima (cm⁻¹) of the species appear after λ_D = 520 nm photolysis of the irradiated CH₃CHO/Ng and CD₃CDO/Ng 1 : 1000 (Ng = Ne, Ar, Kr, or Xe) samples (ph) and absorption maxima of the CO, CH₃˙, and CD₃˙ species, appear after irradiation of the deposited CH₃CHO/Ng and CD₃CDO/Ng samples (irr). Shifts were calculated as ν_ph − ν_irr

Species	Assignment	Sample	Ne	Ar	Kr	Xe
OPLA – out of plane.
CH3CHO
CO	COstr	ph	2141.3	2137.6	2135.6	2131.4
		irr	2141.3	2137.9	2135.6	2133.0
		shift	0.0	−0.3	0.0	−1.6
CH3˙	OPLA	ph	607.6	607.4	609.3	604.8
					604.6
		irr	607.6	607.1	609.5	602.9
					604.2
		shift	0.0	+0.3	−0.2	+1.9
					+0.4
CD3CDO
CO	COstr	ph	2141.1	2137.5	2135.6	2131.5
		irr	2141.2	2137.9	2135.7	2132.9
		shift	−0.1	−0.4	−0.1	−1.4
CD3˙	OPLA	ph	457.6	454.8	453.7	453.1
		irr	457.5	454.9	454.0	453.8
		shift	+0.1	−0.1	−0.3	−0.7

---

Fig. 6 (a) Difference FTIR spectra of the irradiated CH₃CHO/Xe 1 : 1000 sample demonstrating the effect of photolysis (10 min with λ_D = 620 nm light and 40 min with λ_D = 520 nm light; top spectrum) and subsequent standing in the dark at 5 K for 40 min (bottom spectrum). Absorptions maxima of isolated monomeric CO and CH₃˙ species observed in the experiment are illustrated with solid lines. Changes in the amount of H₂CCO–H₂ species are negligible as compared to that of CH₃˙–CO radical-molecular pairs taking into account the absorption intensities of carbon monoxide and ketene (see Table S3 of ESI†). (b) Kinetic data on the CH₃˙–CO → CH₃CO˙ reaction in a Xe matrix at 5 K.

Remarkably, the recovery of acetyl radicals in the dark at 5 K (minimum temperature attainable at the cryostat) was found to occur after photolysis of the X-ray irradiated CH₃CHO/Xe samples (Fig. 6, panel a), whereas this process was not observed in other noble gas matrices. This reverse reaction probably accounts for the observed “photoresistance” of acetyl radicals in a Xe matrix. We have measured the kinetics of this reaction (Fig. 6, panel b) and evaluated its rate constant (k = (2.5 ± 0.5) × 10⁻³ s⁻¹; τ_1/e = 420 ± 80 s) assuming that the transformation of CH₃˙–CO to CH₃CO˙ is a first-order reaction. A rough estimation of the isotopic effect yields k_H/k_D ∼ 2 for the recovery reaction based on recovery kinetics of CH₃CO˙ and CD₃CO˙. It is worth noting that photolysed acetyl radicals can be partially (Ar, Kr) or completely (Xe) restored by annealing the samples up to 30 K (we did not perform annealing in a Ne matrix due to very low range of temperature stability of solid Ne).

4. Discussion

4.1. Mechanisms of acetaldehyde radiation-induced transformations

Considering the mechanistic issues, one should first bear in mind that in matrix isolation experiments the X-ray radiation is primarily absorbed by a matrix because of a very large excess of matrix substance and higher X-ray absorption coefficients of noble gas atoms as compared to the guest organic molecules. Thus, the isolated molecules are destroyed due to efficient positive charge (hole) and/or excitation transfer, which leads to the formation of radical cations (both in the ground and excited states) or a population of various excited states leading to further chemical transformations. Electrons produced in the ionization processes may either recombine with holes resulting in the formation of neutral isolated molecules in excited states (both singlet and triplet) or fall into different physical and chemical traps (structural defects, ions, radicals or molecules with positive electron affinity). We have to stress out that this kind of “indirect” activation of isolated species (generally applicable to all dilute icy media) implies a strong effect of a matrix medium on the efficiency and channel distribution of the radiation-induced transformations of isolated molecules because of strong variations of the excess energy and efficiency of its dissipation determined by matrix electronic characteristics.^45,71

Taking into account this basic consideration, we may suggest that the relatively small yield of the CH₃CHO radiation-induced decomposition in Xe may be related to the rather efficient quenching of acetaldehyde excited states in this matrix.^45,71 The reason for the relatively low yield of the CH₃CHO decomposition in a neon matrix is not fully clear. It may be associated either with the effect of matrix on the fate of primary radical cations (see below) or with the rigidity of matrix cage reducing the probability of fragment separation.

It was found out that decomposition to CH₄ and CO (Fig. 4) represents the dominant channel of acetaldehyde radiolysis in all matrices. Generally, these products appear in significant amounts at low absorbed doses, which imply that their formation should occur through a single-step (or quasi-single-step) radiation-induced decomposition of acetaldehyde molecule (1):

CH₃CHO [long arrow, wavy then straight] CH₄ + CO (1)

Reaction (1) may actually occur through the formation of CH₃˙⋯HCO˙ radical pair immediately reacting within the matrix cage, although we have no direct evidence for the formation of such caged intermediates. In any case, considering this pathway, one must expect that CO and CH₄ yields should be formed in equal amounts. In fact, as mentioned above, the observed excess of CO over CH₄ looks rather surprising and it deserves some explanation. First, one could suggest that this seeming disbalance might result from the inaccuracy of used absorption intensities (see explanation in the Results section). Nevertheless, in view of the different behaviour of these key radiolysis products in different matrices (especially, at high doses) we should also consider possible sources of additional carbon monoxide formed upon radiolysis. In this connection, we should note that the addition of the electron scavenger considerably changes the balance in favour of CH₄ in an argon matrix (the production of CO is suppressed). Meanwhile, the addition of the same scavenger has little effect on the CH₄/CO balance in the case of xenon (Fig. 5). We may suppose that the electron scavenger has a minor influence on the direct production of CH₄ and CO but drastically diminishes the formation of CO in other reactions implying a significant contribution of ionic channels in the production of additional carbon monoxide. The ionic channels are generally more probable in argon because of higher excess energy in the positive hole transfer occurring in this matrix.^45,71

Another important channel of acetaldehyde radiolysis in matrices is the production of ketene species which also shows almost linear accumulation with the acetaldehyde conversion (Fig. 4). As mentioned above, two types of spectroscopically distinguishable ketene molecules were found in our experiments. The formation of perturbed and unperturbed ketene may be presented as follows:

CH₃CHO [long arrow, wavy then straight] [H₂CCO–H₂] and/or H₂CCO + H₂/2H˙ (2)

Note that H˙ atoms produced by radiolysis of matrix isolated organic molecules normally escape the matrix cage due to excess kinetic energy, which was directly proved by the estimation of distance between trapped radicals and H atoms in a xenon matrix.⁷² However, this is not the case for hydrogen molecules, which may be trapped within the same matrix cage forming the so-called perturbed ketene (see the Results section). The proportion of the perturbed to unperturbed ketene increases from Ar to Xe. The probable explanation of this effect implies that a meaningful excess of energy is needed for escaping hydrogen molecule from the matrix cage. The probability of this process should increase from Xe to Ar, because the matrix polarizability (and, therefore, the efficiency of the excited states quenching) increases from Ar to Xe.^45,71 Decreasing of the ketene yield in the presence of an electron scavenger in an argon matrix may indicate a significant role of triplet excited states (T_n), which are populated predominately by the ion-electron recombination:

CH₃CHO˙⁺ + e⁻ → CH₃CHO* (S_n,T_n) (3)

CH₃CHO* (T_n) → H₂CCO + H₂/2H˙ (4)

Indeed, the proportion of the states with different spin multiplicity should approach the statistical limit (S_n/T_n = 1/3) for recombination of completely non-correlated ion–electron pairs in the bulk of matrix and this could be the case because of the large distance of hole and electron migration in the noble gas matrices. On the other hand, one can expect that direct excitation yields mainly singlet states, which degrade through other channels (see Introduction). This scheme is in accord with the low yield of ketene under the gas-phase photolysis of acetaldehyde.²¹ Probable contribution of triplet states into this process is also supported by the consideration of matrix effect. In contrast to argon, the addition of an electron scavenger in a xenon matrix only slightly affects the relative yield of ketene. As mentioned above, the role of ionic channels in xenon is less significant, while the population of triplet excited states in this matrix may occur due to more efficient intersystem crossing (external heavy atom effect):

CH₃CHO* (S_n) → CH₃CHO* (T_n) → H₂CCO + H₂/2H˙ (5)

This mechanism is apparently unaffected by an electron scavenger.

As shown in our previous study, the matrix isolated ketene (both in perturbed and unperturbed form) is unstable under prolonged irradiation with X-rays.⁵⁶ The first pathway of its transformation is dehydrogenation⁵⁶ leading to HCCO˙ and CCO (channels (6) and (7)), which is in accord with the observation that accumulation of both of these species demonstrates a prominent induction period (Fig. 4). It is worth noting that perturbed ketene (caged with hydrogen) may transform into corresponding perturbed species, i.e., [HCCO˙–H₂] and, probably, [CCO–H₂] (see the Results section). Because HCCO˙ and CCO are predominately formed from ketene, the addition of an electron scavenger also results in a decrease in the yield of these species in argon and xenon matrices. The second possible pathway of the radiation-induced degradation of ketene is concerned with the C=C bond cleavage resulting in the formation of CO and CH₂(8). As mentioned above, a clear spectroscopic sign of the CH₂ species isolated in solid noble gas matrices by IR spectroscopy was not obtained so far.⁷³ However, in the case of perturbed ketene, one may expect the reaction of CH₂ with H₂ trapped within the same matrix cage leading to the synthesis of CH₄(9).⁵⁶ Interestingly, this reaction sequence actually represents a pathway of delayed formation of CH₄ and CO from CH₃CHO under radiolysis and it may account for the acceleration of the formation of these products with the increasing conversion of acetaldehyde at higher absorbed doses (see Fig. 4).

H₂CCO [long arrow, wavy then straight] HCCO˙ + H˙ [long arrow, wavy then straight] CCO + H˙ (6)

H₂CCO [long arrow, wavy then straight] CCO + H₂/2H˙ (7)

H₂CCO [long arrow, wavy then straight] CO + CH₂ (8)

H₂CCO–H₂ [long arrow, wavy then straight] [CO + CH₂–H₂] → CO + CH₄ (9)

As mentioned above (see Introduction), the matrix isolated CH₃CHO can isomerize into CH₂CHOH under radiolysis. It is worth noting that the relative yield of CH₂CHOH in argon is significantly higher than those in krypton and, particularly, xenon (Fig. 2), which may imply a significant contribution of ionic channels into its formation. The concentration of CH₂CHOH tends to level off at a high conversion degree of acetaldehyde. Presumably, it can be explained by its radiation-decomposition onto ketene and hydrogen⁵⁸ at high absorbed doses. Another opportunity for the radiation-induced decomposition of vinyl alcohol is concerned with its decay to C₂H₂ and H₂O. Indeed, it was shown theoretically⁷⁴ that vinyl alcohol is an intermediate on the pathway leading to acetaldehyde decomposition to these fragments. It should be noted that the C₂H₂⋯H₂O complex was actually observed in the irradiated acetaldehyde-containing noble gas matrices and the matrix effect on the relative yield of this species shows the same trend as for CH₂CHOH (Fig. 2 and 4). However, the kinetic consideration (Fig. 4) indicates that the complex may be formed from CH₃CHO directly (i.e., without stabilization of intermediate CH₂CHOH). This result can be tentatively explained by the formation of “hot” CH₂CHOH, which can either relax or decay to fragments immediately (we may note that a similar explanation was suggested for a quasi-single-step CO formation through “hot” formaldehyde from methanol).³⁴ Absence of clear saturation on the accumulation profile of C₂H₂⋯H₂O complex (Fig. 4) implies that further radiation-induced transformation of this complex are not dominant in the achieved range of acetaldehyde conversion degrees (to note, the radiation-induced transformations of the matrix isolated C₂H₂⋯H₂O complex were examined in our previous work).⁵⁶

CH₃˙ radical was found in all the matrices. It accumulates approximately linearly with the acetaldehyde conversion in Ar and Ne matrices (Fig. 4). The relative yield of this radical becomes lower when going to Kr, and, particularly, Xe (Fig. 2), which may imply its formation through ionic channels (in particular, fragmentation of “hot” acetaldehyde radical cations (10).^44,75 Nevertheless, the yield of methyl radicals is reduced in the presence of an electron scavenger in solid argon indicating the contribution of chemical transformations origination from excited states populated by a charge recombination process (for example, a triple decomposition channel (11)).⁷⁵

CH₃CHO [long arrow, wavy then straight] (CH₃CHO˙⁺)* → CH₃˙ + HCO⁺ (10)

CH₃CHO [long arrow, wavy then straight] CH₃˙ + CO + H˙ (11)

The second common radical product of the CH₃CHO radiolysis in matrices is CH₃CO˙ radical. It was argued previously⁴³ that acetyl radical in solid Xe may be produced by deprotonation or rearrangement of CH₃CHO˙⁺ radical cations. This explanation generally agrees with the increase of the relative CH₃CO˙ yield when going from Ar to Xe observed in this work (Fig. 2) because of polarizability and proton affinity of a matrix increases in the same row.^45,71 The deprotonation of acetaldehyde radical cations is also evidenced by a formation of Ng₂D⁺ (Ng = Ar, Kr, or Xe) in the irradiated CD₃CDO/Ng samples. However, somewhat surprisingly, we did not observe a clear signal of the CH₃CO˙ radical in the irradiated Ar and Xe matrices containing an electron scavenger (Fig. 5). This is in apparent contradiction with previous EPR results, where a distinct signal of CH₃CO˙ was observed in the EPR spectra of the irradiated CH₃CHO/CFCl₃/Xe samples, while the yield of hydrogen atoms was found to be negligibly small.⁴³ Following the suggestion made previously,^40,68 we may assume that the conversion of acetaldehyde radical cation to CH₃CO˙ radical (or to closely related species) may occur not via direct deprotonation to a matrix (12), but also via intramolecular H˙ transfer to form a distonic radical cation (13).

CH₃CHO˙⁺ + 2Ng → CH₃CO˙ + Ng₂H⁺ (12)

CH₃CHO˙⁺ (Ng) → (CH₃CO)H˙⁺ (13)

It should be noted that the distonic radical cation is actually a protonated form of CH₃CO˙ and these species are very similar in terms of spin density distribution, so, most probably, they cannot be distinguished in the EPR spectra in the solid phase.⁴⁰ On the other hand, the IR spectroscopic features of neutral and protonated species (CH₃CO˙ and (CH₃CO)H˙⁺, respectively) may differ substantially because of a strong effect of protonation on both frequency and bandwidth (see e.g. HCOO⁻/HCOOH).^76,77 If it is the case, one can suggest that the CH₃CO˙ radicals observed by EPR in the presence of electron scavenger in xenon are actually protonated species resulting from reaction (13) and they are “invisible” in the IR spectra due to a rather strong shift from normal (neutral) CH₃CO˙ radicals and absorption band broadening. Indeed, previous experimental and theoretical results demonstrate that (CH₃CO)H˙⁺ is a stable C₂H₄O˙⁺ isomer.^78,79 In fact, according to the theoretical calculations it is only ca. 2.8 kcal mol⁻¹ less stable than CH₃CHO˙⁺ radical cation in the gas phase⁷⁹ and may become thermodynamically favourable in condensed phases. The kinetic obstacle for reaction (13) is a rather large barrier height for the intramolecular H˙ transfer; however, it may be dramatically reduced due to the effect of polarizable environment, which is expected to be particularly strong for xenon (the so-called “matrix catalysis”).^43,71,80

Such an explanation implies that the neutral CH₃CO˙ radicals observed in the IR spectra in all matrices in the absence of an electron scavenger mainly originate from some other precursor rather than from deprotonation of the primary radical cation. This idea is indirectly supported by the observation of a weak signal of CH₃CO˙ in a Ne matrix, where deprotonation is impossible. Presumably, this precursor is a certain triplet excited state of CH₃CHO (14) populated either by charge recombination or due to intersystem crossing (the latter process is particularly probable in a heavy-atom xenon matrix, in line with the observed matrix effect).

CH₃CHO* (T_n) → CH₃CO˙ + H˙ (14)

As it was mentioned in the Introduction, mercury lamp photolysis of inert matrix isolated CH₃CHO produces CO and CH₄³⁹ while a wider range of photoproducts (CH₄, CO, C₂H₂⋯H₂O, and H₂CCO species) are formed during microwave discharge hydrogen flow lamp irradiation of CH₃CHO/N₂ samples.⁴⁰ It is worth noting that CO and CH₄ are also the main products of radiolysis of the CH₃CHO/Ng (Ng = Ne, Ar, Kr, or Xe) samples but, in contrast to photolysis, radiolysis leads to a variety of new species (CH₂CHOH, CH₃CO˙, CH₃˙, HCCO˙, and CCO). It may be generally explained by a greater range of excited states available for population and involvement of ionic channels in the case of radiolysis. It also should be noted that absorption of ionizing radiation is non-specific in contrast to light which leads to simultaneous chemical transformations of acetaldehyde and its radiolysis products with the formation of secondary radiolysis products (HCCO˙ and CCO) under processing of samples with ionizing radiation.^45,71

4.2. Reversible photoinduced and thermal transformations of the CH₃CO˙ radical

The matrix-dependent reversible transformations of acetyl radicals found in this work may be of specific interest for a better understanding of the chemical dynamics of this important radical in the solid phase. While the photodissociation of CH₃CO˙ under the action of visible light (λ_D = 520 nm) is well known (see ref. 43 and 59), its dark recovery occurring at extremely low temperatures after photolysis is a new observation, which looks rather surprising. As follows from our study, the photoinduced decomposition of CH₃CO˙ was observed in all investigated matrices, whereas subsequent dark recovery at the lowest available temperature was found to occur only in a Xe matrix. According to the previous work of Das and Lee,⁵² the CH₃CO˙ does not result from the CH₃˙–CO pair trapped in a solid p-H₂ matrix at 3.2 K, which was explained by the significant reaction barrier (to note, the calculated barrier height is about 27 kJ mol⁻¹).⁸¹ To explain the observed effect in Xe, one should suggest either a strong matrix effect on the barrier height or a meaningful contribution of tunnelling results in the recovery of acetyl radical in solid xenon. Considering the former suggestion, we could roughly estimate the upper limit of the energy barrier value for this process in a xenon matrix within the framework of the classical theory (assuming the pre-exponential factor of 10¹⁴ s⁻¹) as ca. 1.6 kJ mol⁻¹. In principle, the reduction of the barrier height in this case in comparison with other noble gas media can be expected because a relatively loose Xe matrix cage should not strongly hinder the recovery of CH₃CO˙ radicals. In other examined matrices (Ne, Ar, and Kr) a more tight matrix cage may strongly hinder the rearrangement of the caged CH₃˙–CO pair to the CH₃CO˙ radical. It should prevent the process at low temperatures in these matrices and may result in the incomplete recovery of the acetyl radical during matrix annealing (taking into account dispersion of the cage geometry and the corresponding rotational barriers, typical for the solid-state kinetics). Nevertheless, in any case, the experimental estimation of barrier value for classical reaction looks very low in comparison with the calculated value actually related to vacuum (no hindrance). Thus, most probably, the reaction in xenon occurs via tunnelling mechanism, assuming that in this matrix CH₃˙ and CO occupy favourable positions for the tunnelling process. We could speculate that the observed relatively small (but still discernible) isotopic effect (k_H/k_D ∼ 2, see Results section) is in qualitative agreement with this mechanism since the tunnelling occurs through a soft intermolecular C⋯C mode, which is only slightly affected by the H/D isotopic substitution. In this connection, one should recall another example of an unexpected matrix effect on the radical-molecule reaction occurring within the OH˙⋯CO complex, which yields the trans-HOCO˙ radical. It was shown^82,83 that this process was realized as a classical “through-the-barrier” reaction upon thermal annealing in Ar, Kr, and Xe matrices, while in the case of neon it slowly occurs even at 4.5 K in the dark indicating some involvement of the tunnelling process. Detailed investigation of the effects of this kind deserves special consideration, which would require demanding dynamic calculations explicitly taking into account the cage environment.

5. Summary and conclusions

The results obtained in this work clearly demonstrate that the C–C bond cleavage is the dominating process under the radiolysis of isolated acetaldehyde molecules in inert solid matrices. Similar to the case of UV-photolysis, the most abundant products are CO and CH₄, which are chemically quite stable and relatively resistant to further radiation-induced degradation; the third most abundant product is H₂CCO (ketene). Furthermore, we have identified a number of molecular and radical species containing C₂ moieties (HCCO˙, CCO, and CH₃CO˙ as well as C₂H₂⋯H₂O complex) and the methyl radical. It was demonstrated that the matrix noticeably affected both the efficiency and channel distribution of the acetaldehyde radiation-induced transformations, which suggests the involvement of different mechanisms, dependent on the matrix electronic properties (such as IE and polarizability). Generally, we conclude that a significant difference between the radiation chemistry and photochemistry of CH₃CO˙ results from the involvement of ionic channels and a higher population of optically unattainable states (especially, triplets) in the former case.

Regarding the potential implications to ice astrochemistry, it is important to note that ketene and radicals produced from the radiolysis of acetaldehyde can be reactive even at very low temperatures. In particular, further evolution of these products may lead to the synthesis of COMs. For example, ketene potentially can be converted into more complex molecules such as acetic acid (by reaction with water) or acetamide (by reaction with ammonia).⁸⁴ Additionally, it can also serve as a source of methylene, a highly reactive carbene, which readily attaches even to relatively inert molecules leading to the sophistication of the molecular structure.⁷³ The methyl and acetyl radicals may be involved in radical chemistry also leading to the formation of various COMs.

In a broader context of low-temperature chemistry and chemical physics, we have to pay attention to partially reversible dynamics of the CH₃CO˙ radical produced by acetaldehyde radiolysis. These species undergo photodissociation rather easily in all the studied matrices, while the observation of its matrix-dependent recovery in dark at very low temperatures is rather unusual (among a few known examples). We tentatively suggest that this back-reaction occurring at 5 K in a Xe matrix (but not in Ar or Kr matrices) involves the contribution of tunnelling. Further studies of the dynamics of this process is a challenge for the modern theory of quantum effects in chemical reactions and it may improve our better understanding of the low-temperature chemistry in a solid phase.

Author contributions

Pavel V. Zasimov: conceptualization, methodology, investigation, visualization, data curation, writing – original draft; Elizaveta V. Sanochkina: conceptualization, methodology, investigation, visualization, data curation, writing – original draft; Vladimir I. Feldman: conceptualization, methodology, writing – review & editing, supervision.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the Russian Science Foundation (project no. 21-13-00195). We are thankful to D. A. Tyurin and I. V. Tyulpina for their contribution to the experimental procedure.

References

1. B. E. Turner, R. Terzieva and E. Herbst, Astrophys. J., 1999, 518(2), 699–732.
2. M. Ikeda, M. Ohishi, A. Nummelin, J. E. Dickens, P. Bergman, Å. Hjalmarson and W. M. Irvine, Astrophys. J., 2001, 560(2), 792–805.
3. S. B. Charnley, Adv. Space Res., 2004, 33(1), 23–30.
4. F. Goesmann, et al. , Science, 2015, 349(6247), aab0689.
5. S. Scibelli and Y. Shirley, Astrophys. J., 2020, 891, 73.
6. M. De Simone, et al. , Astron. Astrophys., 2020, 640, A75.
7. A. Córdova, I. Ibrahem, J. Casas, H. Sundén, M. Engqvist and E. Reyes, Chem. – Eur. J., 2005, 11, 4772–4784.
8. H. J. Cleaves II, Monatsh. Chem., 2003, 134(4), 585–593.
9. J. S. Teichert, F. M. Kruse and O. Trapp, Angew. Chem., Int. Ed., 2019, 58, 9944–9947.
10. E. Herbst, Int. Rev. Phys. Chem., 2017, 36, 287–331.
11. D. B. Millet, et al. , Atmos. Chem. Phys., 2010, 10, 3405–3425.
12. M. F. Shaw, B. Sztáray, L. K. Whalley, D. E. Heard, D. B. Millet, M. J. Jordan, D. L. Osborn and S. H. Kable, Nat. Commun., 2018, 9, 1–7.
13. F. E. Blacet, J. Phys. Chem., 1948, 52(3), 534–545.
14. E. K. Lee and R. S. Lewis, Adv. Photochem., 1980, 12, 1–97.
15. S. H. Lee, J. Chem. Phys., 2009, 131(17), 174312.
16. A. E. Clubb, M. J. Jordan, S. H. Kable and D. L. Osborn, J. Phys. Chem. Lett., 2012, 3(23), 3522–3526.
17. P. Morajkar, A. Bossolasco, C. Schoemaecker and C. Fittschen, J. Chem. Phys., 2014, 140(21), 214308.
18. K. C. Hung, P. Y. Tsai, H. K. Li and K. C. Lin, J. Chem. Phys., 2014, 140(6), 064313.
19. J. B. Burkholder, S. P. Sander, J. Abbatt, J. R. Barker, R. E. Huie, C. E. Kolb, M. J. Kurylo, V. L. Orkin, D. M. Wilmouth and P. H. Wine, Chemical Kinetics and Photochemical Data for Use in Atmospheric Studies, Evaluation No. 18, JPL Publication 15–10, Jet Propulsion Laboratory, Pasadena, 2015, http://jpldataeval.jpl.nasa.gov.
20. B. W. Toulson, K. M. Kapnas, D. A. Fishman and C. Murray, Phys. Chem. Chem. Phys., 2017, 19(22), 14276–14288.
21. A. W. Harrison, A. Kharazmi, M. F. Shaw, M. S. Quinn, K. K. Lee, K. Nauta, K. N. Rowell, M. J. T. Jordan and S. H. Kable, Phys. Chem. Chem. Phys., 2019, 21, 14284–14295.
22. C. H. Yang, S. Bhattacharyya, L. Liu, W. H. Fang and K. Liu, Chem. Sci., 2020, 11, 6423–6430.
23. C. H. Yang, S. Bhattacharyya and K. Liu, J. Phys. Chem. A, 2021, 125(29), 6450–6460.
24. T. Dohmaru and S. Taniguchi, Annu. Rep. Radiat. Cent. Osaka Prefect., 1972, 12, 86–90.
25. S. Y. Pshezhetskii et al. , EPR of Free Radicals in Radiation Chemistry, Wiley, 1974, p. 446.
26. E. Burean and P. Swiderek, J. Phys. Chem. C, 2008, 112, 19456–19464.
27. N. F. Kleimeier and R. I. Kaiser, Chem. Phys. Chem., 2021, 22(12), 1229–1236.
28. R. L. Hudson and R. F. Ferrante, Mon. Not. R. Astron. Soc., 2020, 492(1), 283–293.
29. V. I. Feldman, S. V. Ryazantsev, E. V. Saenko, S. V. Kameneva and E. S. Shiryaeva, Radiat. Phys. Chem., 2016, 124, 7–13.
30. V. I. Feldman, S. V. Ryazantsev and S. V. Kameneva, Russ. Chem. Rev., 2021, 90(9), 1142–1165.
31. S. V. Ryazantsev, P. V. Zasimov and V. I. Feldman, Radiat. Phys. Chem., 2018, 151, 253–260.
32. M. A. Lukianova, E. V. Sanochkina and V. I. Feldman, J. Phys. Chem. A, 2019, 123(25), 5199–5205.
33. M. A. Lukianova, I. S. Sosulin, D. A. Tyurin and V. I. Feldman, Radiat. Phys. Chem., 2020, 176, 109022.
34. E. V. Saenko and V. I. Feldman, Phys. Chem. Chem. Phys., 2016, 18(47), 32503–32513.
35. S. V. Ryazantsev and V. I. Feldman, Phys. Chem. Chem. Phys., 2015, 17(45), 30648–30658.
36. S. V. Kameneva, D. A. Tyurin and V. I. Feldman, Radiat. Phys. Chem., 2016, 124, 30–37.
37. S. V. Kameneva, A. D. Volosatova and V. I. Feldman, Radiat. Phys. Chem., 2017, 141, 363–368.
38. A. D. Volosatova, S. V. Kameneva and V. I. Feldman, Phys. Chem. Chem. Phys., 2019, 21, 13014–13021.
39. C. D. Védova and O. Sala, J. Raman Spectrosc., 1991, 22(9), 505–507.
40. A. Schriver, J. M. Coanga, L. Schriver-Mazzuoli and P. Ehrenfreund, Chem. Phys., 2004, 303, 13–25.
41. M. G. Thompson, M. R. White, B. D. Linford, K. A. King, M. M. Robinson and J. M. Parnis, J. Mass Spectrom., 2011, 46, 1071–1078.
42. L. B. Knight, B. W. Gregory, S. T. Cobranchi, F. Williams and X. Z. Qin, J. Am. Chem. Soc., 1988, 110(2), 327–336.
43. V. I. Feldman, F. F. Sukhov, A. Y. Orlov and N. A. Shmakova, High Energy Chem., 2001, 35(5), 319–327.
44. V. Feldman, F. Sukhov, A. Orlov and I. Tyulpina, Phys. Chem. Chem. Phys., 2003, 5(9), 1769–1774.
45. V. I. Feldman, in Application of EPR in Radiation Research, ed. A. Lund and M. Shiotani, Springer, 2014, pp. 151–188.
46. P. V. Zasimov, A. V. Belousov, I. V. Baranova and V. I. Feldman, Radiat. Phys. Chem., 2020, 177, 109084.
47. H. Hollenstein and H. H. Günthard, Spectrochim. Acta, Part A, 1971, 27(10), 2027–2060.
48. K. B. Wiberg, Y. Thiel, L. Goodman and J. Leszczynski, J. Phys. Chem., 1995, 99(38), 13850–13864.
49. M. L. Estep, W. J. Morgan, A. T. Winkles, A. S. Abbott, N. Villegas-Escobar, J. W. Mullinax, W. E. Turner, X. Wang, J. M. Turney and H. F. Schaefer III, Phys. Chem. Chem. Phys., 2017, 19(40), 27275–27287.
50. S. V. Ryazantsev and V. I. Feldman, J. Phys. Chem. A, 2015, 119(11), 2578–2586.
51. H. Dubost, Chem. Phys., 1976, 12(2), 139–151.
52. P. Das and Y. P. Lee, J. Chem. Phys., 2014, 140(24), 244303.
53. M. G. Govender and T. A. Ford, J. Mol. Struct., 2000, 550, 445–454.
54. C. B. Moore and G. C. Pimentel, J. Chem. Phys., 1963, 38(12), 2816–2829.
55. I. Haller and G. C. Pimentel, J. Am. Chem. Soc., 1962, 84(15), 2855–2857.
56. P. V. Zasimov, S. V. Ryazantsev, D. A. Tyurin and V. I. Feldman, Mon. Not. R. Astron. Soc., 2020, 491(4), 5140–5150.
57. A. Engdahl and B. Nelander, Chem. Phys. Lett., 1983, 100(2), 129–132.
58. M. Hawkins and L. Andrews, J. Am. Chem. Soc., 1983, 105(9), 2523–2530.
59. M. E. Jacox, Chem. Phys., 1982, 69(3), 407–422.
60. D. E. Milligan and M. E. Jacox, J. Chem. Phys., 1967, 47(12), 5146–5156.
61. M. E. Jacox and W. B. Olson, J. Chem. Phys., 1987, 86(6), 3134–3142.
62. D. Forney, M. E. Jacox and W. E. Thompson, J. Mol. Spectrosc., 1995, 170(1), 178–214.
63. M. E. Jacox, D. E. Milligan, N. G. Moll and W. E. Thompson, J. Chem. Phys., 1965, 43(10), 3734–3746.
64. D. E. Milligan and M. E. Jacox, J. Chem. Phys., 1969, 51(1), 277–288.
65. G. Maier and C. Lautz, Eur. J. Org. Chem., 1998, 769–776.
66. M. Pettersson, L. Khriachtchev, S. Jolkkonen and M. Räsänen, J. Phys. Chem. A, 1999, 103(45), 9154–9162.
67. H. M. Kunttu and J. A. Seetula, Chem. Phys., 1994, 189(2), 273–292.
68. L. H. Jones and S. A. Ekberg, J. Chem. Phys., 1987, 87(8), 4368–4370.
69. E. S. Shiryaeva, D. A. Tyurin and V. I. Feldman, J. Phys. Chem. A, 2016, 120(40), 7847–7858.
70. S. V. Ryazantsev, D. A. Tyurin and V. I. Feldman, Spectrochim. Acta, Part A, 2017, 187, 39–42.
71. V. I. Feldman, Radiat. Phys. Chem., 1999, 55, 565–571.
72. V. I. Feldman, F. F. Sukhov and A. Y. Orlov, J. Chem. Phys., 2008, 128, 214511.
73. W. Sander, G. Bucher and S. Wierlacher, Chem. Rev., 1993, 93(4), 1583–1621.
74. Y. C. Han, B. C. Shepler and J. M. Bowman, J. Phys. Chem. Lett., 2011, 2, 1715–1719.
75. K. M. Kapnas, L. M. McCaslin and C. Murray, Phys. Chem. Chem. Phys., 2019, 21, 14214–14225.
76. D. Forney, M. E. Jacox and W. E. Thompson, J. Chem. Phys., 2003, 119(20), 10814–10823.
77. R. L. Redington, J. Mol. Spectrosc., 1977, 65(2), 171–189.
78. J. K. Terlouw, J. Wezenberg, P. C. Burgers and J. L. Holmes, J. Chem. Soc., Chem. Commun., 1983, 20, 1121–1123.
79. Y. Apeloig, M. Karni, B. Ciommer, G. Depke, G. Frenking, S. Meyn, J. Schmidt and H. Schwarz, J. Chem. Soc., Chem. Commun., 1983, 24, 1497–1499.
80. T. D. Fridgen and J. M. Parnis, Int. J. Mass Spectrom., 1999, 190, 181–194.
81. C. J. Bennett, C. S. Jamieson, Y. Osamura and R. I. Kaiser, Astrophys. J., 2005, 624(2), 1097–1115.
82. S. V. Ryazantsev, L. Duarte, V. I. Feldman and L. Khriachtchev, Phys. Chem. Chem. Phys., 2017, 19(1), 356–365.
83. S. V. Ryazantsev, P. V. Zasimov and V. I. Feldman, Chem. Phys. Lett., 2020, 753, 137540.
84. R. L. Hudson and M. J. Loeffler, Astrophys. J., 2013, 773, 109.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1cp03999g

---