Infrared Spectroscopy and 193 nm Photochemistry of Methylamine

Article pubs.acs.org/JPCA

Infrared Spectroscopy and 193 nm Photochemistry of Methylamine Isolated in Solid Parahydrogen Fredrick M. Mutunga and David T. Anderson* Department of Chemistry, University of Wyoming, Laramie, Wyoming 82071, United States S Supporting Information *

ABSTRACT: The in situ UV photolysis of a precursor molecule trapped in a parahydrogen (pH2) matrix is a simple method used to generate isolated radical photofragments that are well suited for infrared spectroscopic studies. However, for molecules that can dissociate via multiple pathways, little is known about how the pH2 matrix influences the branching among these open pathways. We report FTIR spectroscopic studies of the 193 nm photodecomposition of methylamine (MA, CH3NH2) isolated in pH2 quantum matrixes at 1.8 K. We observe single exponential decay of the MA precursor upon irradiation and the quantum yield for MA photodissociation is measured to be Φ = 0.26(2) consistent with a weak pH2 cage effect. By comparing to gas-phase results, we show the in situ photolysis results in greater production of molecular products (CH2NH + H2) compared to radical products (CH3NH + H) consistent with the idea of partial caging of the H atom photofragments. The information gained in this work can be used to guide future photolysis studies in pH2 matrixes.

1. INTRODUCTION The photodissociation of a small molecule trapped in a noble gas matrix has long been studied as a prototype for a chemical reaction in the presence of a solvent.1 Photoexciation of a precursor molecule above the dissociation threshold initiates nuclear dynamics along one of the open reaction channels and typically results in a strong collision of the nascent photofragments with the surrounding matrix cage. This perturbation of the solutes photodissociation dynamics caused by collisions with the matrix cage is termed the cage effect. The cage effect can dominate the condensed-phase photochemistry of a molecule that produces heavy fragments with sizes comparable to the matrix atoms. Under these conditions the first collision of the outgoing photofragments with the matrix cage can very efficiently remove the translational energy of the fragments and force geminate recombination. Indeed, photoexcitation of I2 in an Ar matrix results in rapid dissipation of the kinetic energy of the I atom fragments and subsequent recombination caused by the cage effect and has been measured directly in the time domain using femtosecond pump−probe experiments.2 In contrast, the in situ photochemistry of precursor molecules that produce H atom photofragments can show cage exit and permanent dissociation.3 Due to its small size, the H atom can escape directly through openings in the matrix cage. In addition, the small mass of the H atom means the energy loss per collision with the matrix is small such that the H atom can collide multiple times with the cage and still keep sufficient kinetic energy to exit the cage.3 For these reasons the in situ photodissociation of hydrogen halides in noble gas matrixes have been used as a way to produce H atoms for subsequent reaction studies4 and as a way to produce noble gas molecules such as HArF.5,6 Our group is investigating the in situ photochemistry of a variety of precursor molecules trapped in frozen parahydrogen © XXXX American Chemical Society

(pH2) quantum matrixes to gain a better understanding of how the pH2 solvent cage alters the photochemistry.7−14 Compared to noble gas matrixes, pH2 matrixes represent the other extreme with respect to cage exit due to the small mass of H2. The caging ability of solid pH2 is very weak for heavy photofragments that easily push their way out of the pH2 solvent cage.15−17 Indeed, this negligible cage effect is frequently touted as one of the advantages of pH2 matrixes for the preparation of well isolated photofragments. Upon UV photoexcitation of a precursor molecule, the primary photofragments can readily escape the solvent cage to form permanently dissociated products at high quantum yields. For example, our group showed the 355 nm in situ photodissociation of Br2 in pH2 and orthodeuterium (oD2) matrixes can be used to produce high yields of isolated Br atoms.7−9 It is therefore of interest to explore the other extreme in pH2 matrixes; precursor molecules that produce H atom photofragments. In this case the pH2 matrix may show a substantially larger cage effect because the mass of the H atom is less than a single pH2 molecule. We can follow the in situ photokinetics of a small molecule by irradiating the sample with UV light from a high repetition rate excimer laser and measuring the growth and decay of various species using rapid scan FTIR spectroscopy. By choosing a precursor molecule whose gas-phase photochemistry has been well studied, we can compare our results with the gas-phase results to infer the effect of the pH2 matrix. One lesson learned in studies of the 193 nm photodecomposition of N-methylformamide (NMF) in solid pH2 is the branching between open dissociation pathways can be quite Special Issue: Markku Rasanen Festschrift Received: August 21, 2014 Revised: November 5, 2014

A

dx.doi.org/10.1021/jp508476j | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

flows onto a BaF2 optical substrate cooled to approximately 2.5 K in a sample-in-vacuum liquid helium bath cryostat (Janis SSVT-100). The substrate temperature is measured with a silicon diode temperature sensor mounted to the bottom of the optical substrate holder. Room temperature normal hydrogen (nH2) gas that contains approximately 75% orthohydrogen (oH2) and 25% pH2 is converted during deposition to 99.97% pH2 gas by use of an ortho/para converter. The ortho/para converter consists of a 1/8 in. outer diameter by 2 m long copper tube packed with granular hydrous ferric oxide (Fe(OH)3) catalyst particles of size ranging from 10 to 100 μm, wound with thermally conducting epoxy to a copper bobbin. To reach these enrichment levels, the ortho/para converter is operated near T = 14.0 K by attaching the converter bobbin to the cold tip of a separate closed-cycle helium cryostat (Advanced Research Systems CSW-204). Sample thickness (path length) is determined from the integrated intensity of the Q1(0) + S0(0) and S1(0) + S0(0) pH2 crystal “double” infrared transitions using the calibration method developed by Fajardo.25,26 The dopant, anhydrous MA (Matheson Tri-Gas INC, 100%), was used as received without further purification. The dopant and pH2 gas flows are adjusted to achieve MA concentrations in the range of 30−70 ppm. The concentration of different species within the pH2 crystal is determined using a Beer’s Law expression,

different from the gas phase.12 One obvious reason for different branching ratios occurs when a primary photolysis product reacts with the pH2 host. Another difference can arise if one of the primary photolysis products can undergo efficient secondary photolysis; secondary photolysis can occur in solid pH2 experiments that are repeatedly performed on the same sample but typically make minor contributions in gas-phase studies because the initial sample is continuously replenished. However, even when these obvious differences are identified and characterized, photochemistry in solid pH2 typically favors “molecular” over “radical” channels that may signal effects of what we term “partial caging”. For example, in recent studies of the 193 nm photodissociation of NMF in solid pH2 we measured significant branching (>50%) into the CO + CH3NH2 molecular channel that can arise from partial caging of the major gas-phase radical channel that produces HCO + NHCH3. The basic requirement for partial caging is the production of two radical cofragments that due to the cage effect are forced to react with each other and then are ejected from the solvent cage. We decided to study the in situ photochemistry of methylamine (MA, CH3NH2) trapped in solid pH2 for two primary reasons: (1) similar to NMF there are multiple open radical channels and (2) methylamine is one of the primary photolysis products identified in the NMF photochemical studies and we want to check the IR assignment of this species with a genuine sample. Previous gas-phase studies of the photodissociation of MA have reported four primary dissociation pathways.18−22 CH3NH 2 + hν → CH3NH + H

(75%)

→ CH 2NH 2 + H → CH3 + NH 2

(< 10%) (5%)

→ CH 2NH + H 2

(10%)

[X] =

2.303 ∫ log10(I /I0) dν ̃ ε (cm mol−1) d (cm)

(23.16 cm 3 mol−1)(1 × 106) (2)

(1a)

where [X] is the concentration of species X in parts per million (ppm), ε is the gas-phase integrated absorption coefficient, d is the IR path length through the crystal, and V0 = 23.16 cm3 mol−1 is the molar volume27 of solid pH2. The MA concentration is determined by using the v9 and v11 peaks and the integrated absorption coefficients reported in Table S1 in the Supporting Information. Using six different vibrational modes for MA, all not involved in Fermi resonances, we get a standard deviation in concentration of 50%. By comparing the average value with the value determined from the dopant and pH2 gas flow rates, we settled on measuring the MA concentration using the average of the ν9 and ν11 vibrational modes. Sample temperatures in the 1.8−4.3 K range are achieved by changing the He gas pressure in the He cryostat using a dedicated vacuum pump. IR absorption spectra at maximum resolutions of 0.03 cm−1 (nominal with boxcar apodization) from 700 to 1800 cm−1 are obtained using an FTIR spectrometer (Bruker IFS-120HR) equipped with a glowbar source, a Ge-coated KBr beamsplitter, a 5500 cm−1 long pass optical filter, and a liquid nitrogen cooled HgCdTe detector. Additional IR spectra are recorded from 1800 to 5000 cm−1 using the FTIR spectrometer with a liquid nitrogen cooled InSb detector. The entire optical path outside the cryostat and the spectrometer is purged using dry air to reduce atmospheric (CO2 and H2O) absorptions. The ultraviolet (UV) photolysis is achieved using the unfocused 193 nm output of an ArF excimer laser (Gam Laser EX5) configured to pass through the sample at 45° with respect to the surface normal of the BaF2 substrate. The IR probe beam is focused (8 in. off-axis parabolic mirrors) at normal incidence through the sample to permit FTIR spectra to be recorded either during or immediately after 193 nm irradiation. The typical experimental procedure is to photolyze

(1b) (1c) (1d)

In 1963 Michael and Noyes 18 completed the first comprehensive study of the gas-phase photodissociation of MA using a broad band UV light source (194−244 nm). As a result, they reported that reaction 1a, N−H bond fission, accounted for 75% of the primary photolysis products. These gas-phase branching ratios have to a large extent been supported by more recent H atom photofragment translational studies of the 222 nm photolysis of MA.20−22 One important conclusion from these studies is the photochemistry of MA is more complex than the isoelectronic species CH3OH where the photochemistry is dominated by the O−H bond fission channel.20−22 We were therefore interested to see how the branching between the three radical channels (1a−1c) and the one molecular channel (1d) is altered for the 193 nm in situ photolysis of MA in solid pH2. The challenge in the present work therefore is to spectroscopically characterize the photodissociation products, measure the photokinetics to determine the primary photolysis products, and finally determine the branching ratios.

2. EXPERIMENTAL METHODS The experimental approach has already been described in detail.12,13 Briefly, MA doped pH2 crystals are synthesized using the “rapid vapor deposition” method developed by Tam and Fajardo.23,24 The chemically doped pH2 crystals are prepared by codeposition of independent MA and precooled pH2 gas B

dx.doi.org/10.1021/jp508476j | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

a sample with short 193 nm exposures (5−10 min) and record rapid scan FTIR spectra before, during, and after each exposure. In these experiments, the average ArF laser fluence is approximately 40 μJ cm−2 pulse−1 and were carried out at 250 Hz repetition rates. The laser fluence is measured through an aperture with a power meter placed in the UV beam path just outside a CaF2 window on the cryostat.

3. EXPERIMENTAL RESULTS AND ANALYSIS A. Infrared Spectroscopy of Methylamine Trapped in Solid pH2. For a relatively small molecule, MA has a complicated gas-phase infrared spectrum due primarily to the presence of two different large amplitude vibrations (LAVs).28−32 The LAVs are related to the umbrella-type inversion motion of the hydrogen atoms of the amine (NH2) group (ν9 wagging motion) and the internal rotation of the methyl (CH3) group (ν15 torsional motion). These two LAVs can impact the high-resolution FTIR spectroscopy of the other vibrations by introducing tunneling splittings. Another aspect of the infrared spectroscopy of MA is the strong CH3 bend− stretch Fermi resonance (2:1) that breaks down the normal mode picture of the vibrational dynamics in the C−H stretching region.29,33 Although a detailed description of the high-resolution FTIR spectroscopy of MA isolated in solid pH2 is desirable, it is outside the scope of this work. Accordingly, only the vibrational assignment of the MA infrared absorption features will be presented to characterize the vibrational modes used to monitor the MA concentration in the photolysis studies and to test the infrared assignment presented in an earlier photolysis study of NMF.12 Shown in Figure 1 is the FTIR spectrum recorded at 1.8 K of an as-deposited sample of MA isolated in solid pH2 displaying many of the assigned MA features. The MA peak positions are presented in Table 1. The MA peak assignments are made by comparing the measured peak frequencies with gas-phase values.29 However, in the C−H stretching region we instead compare our measured values to full anharmonic ab initio calculations of the Fermi resonances.33 Inspection of Figure 1 shows the v8 and v9 absorptions both show additional fine structure not observed for the other vibrational modes. The v8 peak shows fine structure with three resolved peaks at 1043.9, 1044.2, and 1044.5 cm−1 in the as-deposited sample and the v9 peak shows evidence for four resolved peaks at 786.8, 789.6, 790.6, and 791.9 cm−1. Excitation of v9 (NH2 wag) greatly increases the tunneling splitting (37-fold) due to the LAV umbrella-type inversion mode28,31 of the NH2 moiety and therefore the fine structure observed for v9 is likely due to resolvable tunneling splittings. Further, the v8 (C−N stretch) mode is a small amplitude vibration but also may show resolvable tunneling splittings because in the excited state the CH3 torsional tunneling splitting28 increases by 50%. However, assigning this fine structure to increased tunneling splittings can be tricky. The basic splitting pattern for both absorptions remains intact with annealing (not shown here) except additional peaks are produced on the low- and high-energy wings of the v8 and v9 peaks, respectively. Thus, it is not clear if the resolved fine structure can be assigned solely to tunneling splittings. Further complicating the assignment, Ruzi reported12 a broad doublet for the v9 mode and two closely spaced peaks for v8. Thus, it appears the detailed line shape also depends on the MA concentration because in the study by Ruzi the MA concentration was considerably less (∼6 ppm). A more detailed assignment of this fine structure is not possible at this time and

Figure 1. Infrared absorption spectrum of an as-deposited MA/pH2 sample recorded at 1.8 K displaying most of the peaks assigned to MA. The A′ and A″ symmetries of the modes are indicated with blue and red vibrational labels, respectively. Note the absorption intensity in the region from 3315 to 3450 cm−1 has been multiplied by a factor of 5 to better show the peaks. This sample contains 47 ppm of MA and is 0.25(1) cm thick.

Table 1. Peak Positions (Uncertainties) in cm−1 for CH3NH2 Isolated in Solid pH2 mode

gas29,31,33

pH2

matrix shift

A′ ν1 NH2 d-stretch ν2 CH3 d-stretch

3361 2980.24

−8.57 −18.98

2ν5

2922.75

ν5 + ν6 ν3 CH3 s-stretch 2ν6

2907.09 2875.94 2813.64

ν4 NH2 scissor ν5 CH3 d-deform ν6 CH3 s-deform ν7 CH3 rock ν8 CN stretch ν8 CN stretch ν8 CN stretch ν9 NH2 wag A″ ν10 NH2 a-stretch ν11 CH3 d-stretch ν12 CH3 d-deform

1623 1473 1430 1130 1044.813 1044.813 1044.813 779.61

3352.43(6) 2961.26(5) 2955.30(5) 2932.21(14) 2912.12(12) 2902.49(14) 2895.50(1) 2874.46(1) 2815.94(2) 2788.47(8) 2770.28(11) 1631.1(2) 1463.80(6) 1427.38(11) 1130 1044.512(6) 1044.215(6) 1043.936(3) 786.84(2)

3427 2985 1485

3415.20(8) 2981.96(4) 1480.37(4)

−11.80 −3.04 −4.63

−10.63 −11.59 −1.48 +2.30

+8.1 −9.20 −2.62 0 −0.301 −0.598 −0.877 +7.23

will require more experiments under a greater variety of conditions. C

dx.doi.org/10.1021/jp508476j | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

There are also strong Fermi resonances29,33 that complicate the MA spectrum in the C−H stretching region from approximately 2700 to 3100 cm−1. These Fermi resonances result from mixing of vibrational states of the same symmetry when they are close in energy. In this case two quanta in CH3 bending modes (ν5 or ν6) are close in energy and can interact with one quanta of CH3 stretch (ν2 or ν3). Mixing caused by the Fermi resonance results in both changes in the frequencies and intensities of the MA absorptions in the C−H stretching region. We see evidence of this phenomenon in the spectral region from 2750 to 3025 cm−1 in Figure 1. In the absence of Fermi resonances we should only observe three peaks (ν3, ν2 and ν11 absorptions) in this region. However, instead we observe multiple peaks, which we can assign on the basis of the anharmonic calculations of Levi et al.33 For example, the intensity of the 2v6 overtone and the v5 + v6 combination band, which are generally expected to be much weaker than any of the fundamental vibrational absorptions, are greatly amplified in this region. This amplification results from intensity “borrowing” of the weak overtone (dark) transition from the fundamental (bright) transition by mixing in fundamental character to the overtone transition. Levi et al. performed detailed calculations on the Fermi resonances in MA and reported that all the A′ states in this region are strongly mixed except for the ν11 peak, which has A″ symmetry and therefore is not perturbed.33 Accordingly, we use the ν9 and ν11 peaks to monitor the concentration of MA for the reasons discussed in the Experimental Methods. We point out that uncertainties in the absolute concentration of MA constitute the largest errors in the determined branching ratios and therefore the reported branching ratios are only semiquantitative. As discussed at the beginning of this section, another motivation for the present work is to check the spectral assignment reported by Ruzi and Anderson that MA is produced in the 193 nm in situ photolysis of NMF.12 By comparing the as-deposited spectra recorded in this work for a sample that contains only MA with the spectrum recorded after the partial photolysis of NMF in solid pH2, we can very quickly identify which peaks are due to MA. Comparison of the two different sets of data shows that whereas the earlier spectral data are for lower MA concentrations and complicated by infrared absorptions due to other photoproducts and by the NMF precursor, it is clear the assignment is correct and MA is one of the primary photoproducts of the 193 nm in situ photolysis of NMF. A comparison of the spectra in the C−H stretching region for the two different samples is provided in Figure S1 in the Supporting Information. A detailed comparison shows, however, the ν2 peak assignment was incorrect in the earlier work.12 B. 193 nm in Situ Photolysis of CH3NH2. We can identify the photoproducts of the 193 nm photolysis of MA by using IR difference spectra constructed from spectra recorded before and after irradiation (e.g., after−before). All the positive peaks in the IR difference spectrum are species produced during irradiation (photoproducts) whereas the negative peaks correspond to the MA precursor. Shown in Figure 2 is a difference spectrum in the 900−1400 cm−1 region constructed from spectra recorded before and after 5 min of 193 nm irradiation of 71 ppm MA doped pH2 sample at 1.7 K. The one negative peak in this region is due to the ν8 absorption of MA (Figure 1) and the many positive peaks are due to NH3, CH2NH, CH4, and CH3NH. The CH3NH assignment is only tentative because the IR spectrum of this species has not

Figure 2. Infrared difference spectrum obtained by subtracting the spectrum recorded before and after the 193 nm photolysis of MA in solid pH2. The positive peaks are assigned to the NH3, CH3NH, CH2NH, and CH4 photoproducts, and the one negative peak in this region is due to the ν8 mode of MA. This difference spectrum was generated from spectra recorded at 1.7 K for a sample that contains 71 ppm of MA before photolysis.

been reported in other matrixes or under gas-phase conditions. This assignment is based on DFT calculations presented in the earlier NMF photolysis study.12 Two peaks observed previously for CH3NH at 1025.2 and 1365.2 cm−1 are again observed in this work (Figure 2). The main reason the CH3NH assignment is still tentative is because the calculated frequencies and intensities only qualitatively match the measured values.12 Part of this discrepancy may be due to LAVs in the CH3NH radical that are not accounted for in harmonic vibrational calculations. A detailed comparison of the CH3NH absorption features measured in this study and the previous NMF photolysis study are presented in Figure S2 in the Supporting Information. Three of the assigned features match very well, but the line shape of the peak assigned to the N−H stretch is different in the two studies. The fact that the same peaks are observed in both studies suggests the carrier is the same, but still more work is required for a definitive assignment. To address this issue, we plan to conduct similar photolysis experiments on 13CH3NH2 to see if we observe peaks for 13CH3NH and use the DFT predicted isotopic shifts to assign this species. We can gain more insight into the mechanism of the MA photodissociation by looking at the photokinetics of all the various species. Shown in Figure 3 is representative data for the photolysis of a MA doped pH2 sample at 1.8 K showing the concentrations of MA, CH2NH, CH4, and NH3 as a function of time. The tan bars represent the timing and duration of the 193 nm photolysis. Repeated FTIR scans before photolysis show the sample consists of 44(1) ppm of MA and approximately 5(1) ppm of NH3 present as an impurity in the MA sample. During photolysis FTIR spectra are recorded with 67.5 s acquisition times (4 scans at 0.03 cm−1 resolution) to map out the photokinetics during the 10 min photolysis of the sample. Note during irradiation the MA concentration decays exponentially to near zero values and then is relatively constant after photolysis is stopped. In contrast, the methanimine (CH2NH) concentration is zero before photolysis and exponentially increases to a value of 24(1) ppm. As we will show, the rate constants for the decay of MA and the increase in CH2NH are approximately equal and thus this simple analysis already shows that channel 1d has a branching ratio of 56(4)%. The rotational resolution in the infrared spectroscopy of CH4 and NH3 trapped in solid pH2 allow the growth kinetics of D

dx.doi.org/10.1021/jp508476j | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

whereas in our case we photolyze both MA and CH3 at 193 nm simultaneously. Therefore, the fact that we measure a branching ratio into the I = 2 CH4 state that is closer to the statistical limit is not surprising. We can test this interpretation further by looking for CH3 peaks in only partially photolyzed MA/pH2 samples. Once produced in the ground electronic state, the CH3 radical is remarkably stable in solid pH2 because the tunneling reaction of CH3 with the pH2 host is extremely slow.36 Shown in Figure 4 are spectra in the ν4 and ν3 regions of CH3 that

Figure 3. Kinetic data for the 193 nm photolysis of MA in solid pH2 showing the decay of MA and the growth of CH2NH, CH4, and NH3. The solid tan bar represents the timing and duration of the 10 min 193 nm photolysis (40 μJ cm−2 pulse−1, 250 Hz). Note the concentration of para-NH3 has been multiplied by a factor of 5. See text for details.

Figure 4. High-resolution (0.03 cm−1) spectra of CH3 features measured in the ν4 and ν3 regions after partial photolysis of a MA/pH2 sample. The spectra are offset vertically for comparison, and the intensity in the ν3 region is multiplied by 2. The measured spectra are shown as black lines, and the results of least-squares fits of the data to the sum of three Lorentzian line shapes are shown as red lines. The lower spectra are recorded for a photolyzed sample before annealing, and the upper spectra are recorded after annealing.

these two species to be studied in greater detail. We suspect that CH4 is produced from the secondary photolysis of CH3 after MA photodissociates via channel 1c. Shown in Figure 3 is the concentration of CH4 in the I = 2 and I = 1 nuclear spin states. Previous research34 has shown there is fast equilibration between the I = 0 and I = 1 CH4 nuclear spin manifolds in solid pH2, but much slower (τ = 1/k = 437(6) min at 2.5 K) nuclear spin conversion between I = 2 and I = 1. Therefore, we can temporally resolve the branching between these two nuclear spin states using rapid scan FTIR spectroscopy. Before photolysis there is no detectable CH4 present in the sample and during irradiation the concentrations of CH4 in the I = 1 and I = 2 nuclear spin states increase linearly, but with a greater slope for I = 1. Momose and co-workers have shown the 193 nm in situ photolysis of CH3 likely follows the following twostep reaction mechanism,35 CH3 + hν → CH 2 + H

(3)

CH 2 + pH2 → CH4

(4)

were recorded after partial photolysis of a MA doped pH2 sample. Although the peaks are very weak because the 193 nm light efficiently photodissociates the CH3 radicals as well as MA, they are clearly present. We can estimate the CH 3 concentration by fitting the two absorption features to a sum of three Lorentzian line shapes and the results of the fit are shown as red lines in Figure 4. On the basis of this analysis, we estimate the CH3 concentration is 4(2) ppm for this sample that originally contained about 71(1) ppm of MA. It is difficult to produce higher CH3 concentrations due to efficient secondary photolysis of the CH3 at 193 nm. However, the fact that we can detect the CH3 intermediate provides circumstantial evidence that (1) CH3 is produced by the 193 nm in situ photolysis of MA and (2) CH4 is produced via reaction steps (3) and (4) as postulated. Nonetheless, we cannot rule out that some minor amounts of CH4 are produced directly in the photolysis of MA. This analysis also implies the measured CH4 concentration after photolysis can be used as a proxy for the amount of branching into channel (1c). Another interesting observation is the CH3 spectra shown in Figure 4 reproduce the reported peak frequencies to within experimental uncertainties but show evidence in both the ν4 and ν3 regions for a third peak not reported previously.37 That is, the CH3 radical is known to freely rotate in solid pH2 and the peaks in both spectral regions correspond to the N,K = 1,1 ← 0,0 transition of ortho-CH3 (I = 3/2). The observed fine structure therefore is due to a crystal field splitting of the upper N = 1, |K| = 1 rotational level into |M| = 1 and 0 states where the M quantum number represents the space-fixed projection of N on the c-axis of the hexagonal close packed (hcp) pH2 crystal.37 The peaks at the low- and high-energy extremes in

with the result that a 0.2 mole fraction of CH4 is produced in the I = 2 nuclear spin state (and 0.8 in the I = 1). This experimentally measured value falls between the theoretical limit of 0.00, if nuclear spin is conserved in reactions 3 and 4 and the reaction starts with pure I = 3/2 CH3, and 0.31 for the high temperature limit. Momose and co-workers interpreted the measured 0.2 value as evidence that either nuclear spin is not conserved or reaction 4 proceeds through a stepwise abstraction mechanism.35 In our studies we measure an I = 2 mole fraction of 0.32(1) just after photolysis, which is very close to the high temperature limit. We ascribe the greater I = 2 mole fraction measured in our work to photolysis conditions, not to a different photolysis mechanism. In the previous studies of Momose and co-workers,35 first CH3I is photolyzed at 253.7 nm to generate the CH3 radical and then in a second step the CH3 is photolyzed at 193 nm. When the photolysis is conducted in a stepwise manner, all the CH3 is in the lower energy I = 3/2 nuclear spin state prior to 193 nm photolysis, E

dx.doi.org/10.1021/jp508476j | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

produced during photolysis react with the pH2 host to form defect-NH3 faster than the temporal resolution of the FTIR scans (∼6 min between spectra after photolysis). This initial jump in the defect-NH3 concentration after the laser is turned off occurs whenever the photoproduction rate of NH2 exceeds the reaction rate with the pH2 host.13 In contrast, the para-NH3 concentration does not show this initial jump implying that reactions of NH2 with the pH2 host do not produce para-NH3. The para-NH3 is produced only during photolysis, presumably by a different reaction mechanism. The mechanism by which NH3 is produced shows a strong preference for ortho-NH3. The defect-NH3 is known to be an ortho-NH3 molecule in a metastable solvation site and thus the defect-NH3 concentration is proportional to the ortho-NH2 concentration produced in the photolysis of MA. As can be seen, the defect-NH3 concentration rises to a significantly greater value during photolysis than para-NH3. In fact, the fraction of ortho-NH3 produced during photolysis is calculated to be 0.95(1) using the defect-NH3 and para-NH3 concentrations, which is significantly larger than the 0.66 value predicted from nuclear spin conservation for the ortho-NH2(I = 1) + pH2(I = 0) reaction. We determined the 0.66 value on the basis of the ratio’s reported for the related CH2(I = 1) + pH2(I = 0) → CH3(I = 3/2,1/2) + H reaction.35 In this case orthoNH3 is associated with the lowest VIR state and therefore the 0.95(1) value for ortho-NH3 could result if some of the paraNH3 is rapidly converted to ortho-NH3 during photolysis at 1.8 K. These discrepancies between the expected nuclear spin fraction of CH4 and NH3 based on the assumed reaction mechanism and nuclear spin selection rules deserve further study. The other important point to make is the NH3 concentration measured during photolysis cannot be used to determine the branching into channel 1c because depending on the 193 nm fluence used in the experiment, some of the NH3 will not be accounted for because it is tied up as ortho-NH2. We can again check our proposed mechanism that NH3 is produced by reactions of NH2 with the pH2 matrix by looking for the orthoNH2 intermediate during photolysis. This is harder than identifying the CH3 intermediate because as we have shown elsewhere ortho-NH2 rapidly reacts with the pH2 host and therefore can only be observed during photolysis when its steady-state concentration is significant.13 For the data shown in Figure 3 we were not able to detect the ortho-NH2 intermediate during photolysis, but in similar studies of a more concentrated MA sample ([MA] = 71 ppm) we could detect the ortho-NH2 ν2 111 ← 000 rovibrational transition near 1527.85 cm−1 and representative spectra are shown in the Supporting Information. The only other photoproducts detected in this study are HCN and HNC. However, both of these species are clearly identified as secondary photoproducts. C. 193 nm in Situ Photokinetics of CH3NH2. Having identified the photoproducts and some of the possible side reactions, we now examine the 193 nm in situ photolysis of MA in more detail. We can recast the kinetic data shown in Figure 3 to evaluate the photokinetics of MA by plotting the data as a function of photolysis time. The recast kinetic data from the experiment depicted in Figure 3 are plotted in Figure 5. The concentration of each species has also been transformed into a mole fraction by dividing the measured concentration of each species by the concentration of MA before photolysis. Inspection of Figure 5 shows the decay of MA is well represented by a single exponential decay expression and that

both spectral regions were reported previously and assigned to transitions to the |M| = 1 and 0 states, respectively.37 The new peak at intermediate wavenumbers (1402.40 and 3171.59 cm−1) in both the ν4 and ν3 regions likely arises from CH3 molecules in single substitution sites of locally face centered cubic (fcc) crystal structure. As shown in Figure 4, we observe a small decrease in the intensity of this new intermediate peak between the as-deposited and annealed pH2 samples. It is wellknown that rapid vapor deposited pH2 samples contain mixed fcc and hcp crystal structures for the as-deposited sample and that by annealing the sample at 4.3 K the fcc sites can be irreversibly converted to hcp sites consistent with the observed intensity changes.9,38 Such an assignment is also consistent with the fact that the previous studies have been conducted on samples in enclosed optical cells which are known to produce pH2 crystals with only the lower energy hcp crystal structure.37 We plan to study the spectroscopy of CH3 further using a different precursor to see if we can achieve more intense CH3 absorptions. We also examine the NH3 photokinetics in greater detail on the basis of our previous studies of the nuclear spin conversion39 and 193 nm in situ photolysis of NH3 in solid pH2 matrixes.13 The NH3 molecule has I = 1/2 and 3/2 nuclear spin states that are commonly referred to as para and ortho, respectively. The nuclear spin conversion of NH3 trapped in solid pH2 is quite rapid (τ = 1/k = 530(41) s) such that a room temperature sample of NH3 (approximately a 50:50 mixture of para and ortho) that is rapidly deposited into a pH2 matrix relaxes to pure ortho-NH3 in the matter of minutes.39 Therefore, as shown in Figure 3, only ortho-NH3 is present prior to photolysis (NH3 is an impurity in the MA sample). Then, as the sample is photolyzed, we observe changes in both the ortho-NH3 and para-NH3 concentrations. The other NH3 species labeled defect-NH3 in Figure 3 corresponds to orthoNH3 molecules in defect sites. As discussed previously,13 if NH2 is produced photolytically in solid pH2 it rapidly reacts with the pH2 host, NH 2 + pH2 → defect‐NH3 + H

(5)

where defect-NH3 represents ortho-NH3 molecules that are produced in single substitution sites next to a vacancy or an H atom due to the chemical reaction with the pH2 host. We can spectrally resolve the peaks of defect-NH3 from the other ortho-NH3 peaks because the angular anisotropy of the crystal field for an NH3 molecule in a defect site splits the aR(0,0) vibration-inversion-rotation (VIR) transition into two peaks we term satellite peaks. The defect-NH3 molecules then irreversible start to convert to ortho-NH3 in the normal solvation site after photolysis is stopped. This behavior is displayed in Figure 3 where we propose NH2 is produced in the 193 nm photodissociation of MA via channel 1c and then reacts rapidly with the pH2 host to form defect-NH3. After photolysis these defect-NH3 molecules decay to ortho-NH3 molecules as studied previously.13 However, during photolysis all the NH3 (defect, ortho, and para) can also undergo secondary photolysis. This can be recognized by inspection of Figure 3; the ortho-NH3 concentration initially remains constant and then grows slightly with photolysis exposure. In contrast, the defect-NH3 and para-NH3 concentrations rapidly photoequilibrate (grow exponentially) as the rate of photodestruction of these species ramps up with concentration (initially zero for both these species). After the photolysis laser is stopped, the defect-NH3 concentration rapidly increases as NH2 molecules F

dx.doi.org/10.1021/jp508476j | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

MA decay constant is determined to be k = 4.74(8) × 10−3 s−1 and the constant for CH2NH growth is k = 3.69(9) × 10−3 s−1. For the two cases where the CH2NH rate constant is >90% the MA rate constant, we use the fitted mole fraction as a measure of the branching into this channel. This gives a value of 0.59(2). The growth curves of both CH4 and HCN identify these species as not being primary photolysis products. We know CH4 is produced by the CH3 photofragments, and therefore, we can use the concentration of this species at infinite photolysis times as a measure of the branching into channel 1c. The value is determined to be 0.33(1). Therefore, the corresponding sum of branching ratios for channels 1c and 1d is 0.92(2), and therefore, these two channels represent more than 90% of the photochemistry. We speculate the remaining 8% proceeds through channel 1a, thereby producing CH3NH + H on the basis of the tentative assignment of the CH3NH radical presented in this work. We also did not observe any peaks that could be assigned to CH2NH2, which suggests that channel 1b is not significant. However, the greatest weakness in this analysis is the lack of quantifying the amount of CH3NH produced and thus the reported branching ratios can only be viewed as semiquantitative. Even though the branching ratios are only semiquantitative and more work is needed, one of the major findings of this work is the branching into channel 1d is significantly enhanced compared to the gas phase. This is an interesting observation because in the gas phase the dominant channel is N−H bond rupture (75%) and the channel that produces CH2NH is less than 10%. The photokinetics of CH2NH is consistent with this species being a primary photoproduct. We speculate the enhanced photoproduction of CH2NH may stem from partial caging of the CH3NH + H photoproducts that leads to further reaction of these two species and production of the CH2NH + H2 products. Although the pH2 solid has a small cage effect for heavy photofragments, the caging of H atoms might be anomalous because the H atom has a smaller mass than a single pH2 molecule. The other photofragments all have masses greater than pH2 such that even at moderate translational energies they have sufficient momentum to exit the pH2 solvent cage. This may not be true of nascent H atoms, which can back-scatter off the pH2 cage and undergo further collisions with the other photofragment inducing chemistry. However, this interpretation of the data is preliminary and we intend to perform additional work using a different isotopomer of the MA to more definitively assign the IR spectroscopy of the CH3NH fragment, which clearly would help with testing this hypothesis. The concept of partial caging could also be tested in a more straightforward fashion by investigating the in situ photochemistry of a simpler precursor molecule such as HCl or HCN.

Figure 5. Full photokinetic data for the experiment depicted in Figure 3. The mole fractions of MA, CH2NH, CH4, and HCN are plotted as a function of the photolysis time (40 μJ cm−2 pulse−1 at 250 Hz). The filled circles are the data and the solid lines are the results of leastsquares fits of the data (see text for details).

MA can be photodissociated to near zero concentrations. This demonstrates that under these conditions there are no photofragments that limit or complicate the MA photochemistry. We can therefore fit the MA data to a simple firstorder decay expression, X(t ) = X(t = 0) exp( −σI193Φt )

(6)

where t is the photolysis time, X(t) is the mole fraction of MA at photolysis time t, σ is the gas-phase cross section40 for MA at 193 nm (σ = 1.84 × 10−18 cm2), I193 is the laser fluence (9.54 × 1015 photons cm−2 s−1) used in the experiment, and Φ is the quantum yield. Based on three separate experiments with different laser fluences and initial MA concentrations, the quantum yield for the 193 nm photolysis of MA is Φ = 0.26(2). The relatively large quantum yield is consistent with a negligible cage effect in solid pH2 whereby the photofragments easily escape the solvent cage and form well separated product molecules. This quantum yield is large in comparison to the quantum yield for Cl2 photodissociation at 308 nm in an argon matrix which is estimated41 to be