Stereodynamics of the Photodissociation of Nitromethane at 193 nm

Article pubs.acs.org/JPCA

Stereodynamics of the Photodissociation of Nitromethane at 193 nm: Unravelling the Dissociation Mechanism J. D. Rodríguez,† M. G. González,† L. Rubio-Lago,*,† L. Bañares,*,† P. C. Samartzis,‡ and T. N. Kitsopoulos‡,¶ †

Departamento de Quı ́mica Fı ́sica I, Facultad de Ciencias Quı ́micas, Universidad Complutense de Madrid, 28040 Madrid, Spain Institute of Electronic Structure and Laser, Foundation for Research and Technology Hellas, P.O. Box 1527, 71110 Heraklion, Greece ¶ Department of Chemistry, University of Crete, Voutes, 71003 Heraklion, Greece ‡

ABSTRACT: The photodissociation of nitromethane at 193 nm is reviewed in terms of new stereodynamical information provided by the measurement of the first four Dixon’s bipolar moments, β20(20), β00(22), β20(02), and β20(22), using slice imaging. The measured speed-dependent β20(20) (directly related with the spatial anisotropy parameter β) indicates that after one-photon absorption to the S3(2 1A″) state by an allowed perpendicular transition, two reaction pathways can compete with similar probability, a direct dissociation process yielding ground-state CH3 and NO2(1 2A2) radicals and a indirect dissociation through conical intersections in which NO2 radicals are formed in lower-lying electronic states. A particularly important result from our measurements is that the low recoil energy part of the methyl fragment translational energy distribution presents a contribution with parallel character, irrespective of the experimental conditions employed, that we attribute to parent cluster dissociation. Moreover, the positive values found for the β00(22) bipolar moment indicates some propensity for the fragment’s recoil velocity and angular momentum vectors to be parallel.

■

INTRODUCTION Nitromethane (CH3NO2), the simplest nitroalkane, is a prototype energetic material whose physical and chemical properties have attracted considerable interest from scientists of different fields such as combustion (fuels, propellants), ignition (explosives), and atmospheric pollution. The photochemistry of nitromethane has been studied extensively, but yet, its photodissociation dynamics are unclear due to the ravelled electronic structure. The ultraviolet absorption spectrum of CH3NO2 is composed of two bands centered at 198 and 270 nm.1 The strong band at 198 nm was first observed by Nagakura and was assigned to a π → π* transition localized on the NO2 moiety;2 the much weaker band at 270 nm was assigned to a n → π* transition from a nonbonding electron of the O atom by Bayliss and McRae.3 The study of the photodissociation dynamics after excitation to the strong band has been possible since the development of excimer lasers. In 1983, Butler and co-workers4 using photofragment translational spectroscopy and fluorescence techniques determined that the primary photofragmentation process upon absorption at 193 nm is the C−N bond cleavage, yielding CH3 and NO2 radicals. The measured translational energy distributions (TEDs) for the CH3 fragment showed a bimodal character suggesting two © 2013 American Chemical Society

competing dissociation pathways. A decade later, Houston and co-workers5 using resonance enhanced multiphoton ionization (REMPI) and time-of-fight (TOF) mass spectrometry confirmed the concurrence of two dissociation channels, a minor channel characterized by a sharp, low-recoil TED of the CH3 fragment in correlation with highly electronically excited NO2 and a major channel with a significantly broader CH3 TED correlating with NO2 in a lower excited state. In both experiments, the internal energy content of the CH3 fragment was found to be rather modest, and there was clear evidence of the production of ground-state NO(X2Π) and electronically excited-state NO(A2Σ+) fragments. The observed experimental results were explained in terms of the following mechanisms CH3NO2 + hν → CH3 + NO2 (1 2 B2) → CH3 + NO(X2Π) + O(3P)

(1)

Special Issue: Stereodynamics Symposium Received: April 2, 2013 Revised: May 28, 2013 Published: May 28, 2013 8175

dx.doi.org/10.1021/jp403272x | J. Phys. Chem. A 2013, 117, 8175−8183

The Journal of Physical Chemistry A

Article

formation of highly excited NO2 radicals in the main dissociation event, which justifies the observed broad and structureless TEDs. In order to unravel the intricate dissociation mechanism of CH3NO2, in the present work, we have used the slice imaging technique to combine the measurement of TEDs and speeddependent anisotropy parameters. We have followed the recent work by Grubb et al.9,10 to determine speed-dependent Dixon’s bipolar moments11 from the phenomenological anisotropy parameters extracted from sliced images of the CH3 fragment. Four bipolar moments have been determined, β20(20), β00(22), β20(02), and β20(22), which account, respectively, for the μ−v, v−J, μ−J, and μ−J−v correlations, where μ is the transition dipole moment, v the fragment recoil direction, and J the total angular momentum of the detected fragment. The interpretation of the four bipolar moments in terms of the CH3 speed distribution provides information about the excited-state symmetry, couplings, and dissociation time scales and details about forces and torques between the separating fragments.

2 CH3NO2 + hν → CH3 + NO** 2 (2 B2 ) + hν

→ CH3 + NO(A2Σ+) + O(3P)

(2)

where eqs 1 and 2 correspond to the major and minor channels, respectively. The secondary dissociation products of the major channel showed a distribution of internal and translational energy consistent with a dissociation on a repulsive surface, and the electronically excited NO(A2Σ+) was assigned to a secondary product of the minor channel, after absorption of a second 193 nm photon. Relatively little translational energy was found in the fragments corresponding to the second decomposition step of the minor channel. More recently, Arenas and co-workers6−8 have devoted a series of theoretical works to study the CH3NO2 photodissociation at 193 nm using high-level ab initio (MS-CASPT2) methods, with emphasis on the role of surface crossings in the photochemistry of this molecule.7 In view of the detected stationary points and conical intersections (CIs) for the accessible electronic states of the molecule, these authors have suggested7 possible dissociation pathways for CH3NO2 at 193 nm, which are depicted schematically in Figure 1. After

■

EXPERIMENTAL SECTION The setup employed in the slice imaging experiments has been described in detail previously,12−14 and only a brief description is presented here. A molecular beam is created by expanding a gas mixture of CH3NO2 in He (10%, 1 atm backing pressure) into vacuum using a piezoelectric pulsed valve. The beam is skimmed and collimated to a diameter of 2 mm prior to intersecting the focused laser beams at right angles. CH3NO2 excitation is carried out at 193 nm using an excimer laser (Lambda Physik COMPEX, operating with ArF) and the detection of vibrationless CH3(ν=0) products at 333.45 nm (the Q branch of the 3pz transition for a two-photon process) by a MOPO laser (Spectra-Physics 730D10) pumped by a Nd:YAG laser (Spectra-Physics Pro Series 450) delayed 10 ns with respect to the excitation pulse. The generated CH+3 ions are projected onto the position-sensitive detector system, two microchannel plates (MCPs) coupled to a phosphor screen. Slice images of CH+3 are recorded using a 500 ns extraction delay applied to the repeller plate and an effective 10 ns detector gate on the front MCP. Ion images are recorded using four laser polarization configurations, X(pump)X(probe), XZ, ZX, and ZZ, where X is the direction perpendicular to the laser propagation axis (Y) and Z is parallel to the molecular beam. The recorded slice images are quadrant-symmetrized prior to extracting the kinetic energy and angular distributions. The slice images are calibrated using the TEDs of CH3(ν=0) fragments from the well-known CH3I photodissociation at 193 and 333.45 nm. The strength of the slice imaging technique is usually limited by the ratio between the experimental detection gate and the speed of the fragments. For continuous speed distributions, like those measured in this work, a possible partial slicing would have the effect of overlapping signals from different fragment velocities, distorting the photofragment speed and angular distributions. We have quantified the magnitude of this effect by direct comparison of the results obtained with two different gate widths, 10 and 20 ns, and with pure unsliced images. An additional set of results is obtained from Abel inversion of the unsliced XX images, that is, the usual velocity map imaging (VMI) procedure. As expected, the unsliced image speed distribution differs substantially from the real VMI distribution, but the difference is sensibly reduced for the 20 ns gate, becoming inappreciable for the 10 ns gate. The VMI technique

Figure 1. Sketch of the potential energy curves, including TSs and CIs involved in the photodissociation of CH3NO2 at 193 nm based on the ab initio calculations by Arenas and co-workers.6,7 The main routes for dissociation are indicated with arrows. The dissociation channels yielding CH3 + NO2(2 2A2) and CH3 + NO2(2 2B2) are energetically closed at this excitation wavelength.

one-photon absorption at 193 nm, the CH3NO2 molecule is promoted to the 2 1A″ state (S3) through an allowed perpendicular transition, with enough excess energy to overcome the transition state (TSS3) connecting adiabatically with ground-state CH3 and NO2(1 2A2). However, considering the properties of the S3 surface, the authors suggest that the most plausible process would be deactivation through the S3/S2 CI, where several other possibilities open up, direct dissociation on S2 leading to CH3NO(1 1A″) and O(1D), further deactivation through the S2/S1 CI and dissociation on S1 leading to CH3 and NO2(1 2B2), and dissociation in the ground S0 state after deactivation through the S2/S1 CI from where dissociation yields CH3 and highly excited NO2(1 2A1). An important point to be considered in the analysis of the photodissociation dynamics of nitromethane, in which all of the experimental works reported coincide, is the production of NO radicals as a secondary photodissociation product. In a onephoton absorption process, the production of NO implies 8176



Article

terms of the expansion appear. The sum extends to 2n + 2 terms, where n is the number of photons involved in the resonant step of the detection scheme employed. For a (2+m) REMPI detection scheme, three coefficients, β2, β4, and β6, are needed to fit the experimental data; when a (1+m) REMPI scheme is used, the parameters are reduced to β2 and β4. The connection between the βi and any other set of parameters must be viewed as a nonlinear equation system where either the βkq(k1k2) bipolar moments or the akq(p) parameters represent the unknowns. Rakitzis has devoted some effort to connect the pure phenomenological βi coefficients with the akq(p) parameters for the case of Abel-invertible ion images;18 the connection between the βkq(k1k2) and akq(p) parameters up to the second order (k = 2) has been established as well.19 The connection between the βkq(k1k2) and βi parameters for sliced ion images and (1+1′) REMPI photofragment detection are the central result of North and co-workers’ work.9,10 In their work, the four βi parameters obtained from ion images acquired at three different geometries, XX, XZ, and ZX, can be related to the five second-order bipolar moments. In the semiclassical (high-J) limit, the linear independent bipolar moments are reduced to four, and therefore, they can be uniquely determined. In the experiments presented in this work, a (2+1) REMPI photofragment detection scheme is used instead. The additional photon enlarges the set of parameters involved. In eq 3, the sum extends up to β6; in the bipolar harmonic expansion, the four forth-order bipolar moments must be considered and, therefore, the four k = 4 multipoles in the akq(p) formalism. Because the XZ ion image contributes with a single parameter in both REMPI schemes, the number of equations is increased XX only by two (corresponding to the βZX 4 and β4 parameters), while the number of unknowns is increased by four. Clearly, without further considerations, the problem is algebraically unsolvable. As will be shown in the following section, in all of the cases studied in the present work, the β4ZX and β6XX parameters obtained from the fit of the experimental data to eq 3 were found to be nearly zero. Although this fact does not make the problem algebraically solvable, as we show in the following lines, under the experimental conditions employed in this work, it effectively reduces the k = 4 case to a k = 2 case. According to Rakitzis,18 for Abel-invertible ion images, for which the cylindrical symmetry is preserved, the β2, β4, and β6 coefficients are expressed in terms of the akq(p) parameters in the following way

does not allow one to obtain the bipolar moments because the XZ and ZX ion images are not Abel-invertible. The comparison between the three sets of measurements, unsliced and 20 and 10 ns gate images, allows an estimation of the effect of partial slicing in the reported bipolar moments. In agreement with the analysis of North and co-workers,9 the anisotropic character of all of the measured bipolar moments increases as the detection gate is decreased. In summary, the measurements reported in this work, taken with a 10 ns gate, are minimally affected by a possible partial slicing effect.

■

METHODOLOGY Analysis of the (2+1) REMPI Angular Distributions. In the past decades, different photofragment angular momentum formalisms have been developed, with the aim of providing operational tools for the analysis of the experimental signals arising from one-photon dissociation of diatomic and polyatomic molecules.15 In the mid 1980s, Dixon expressed the Doppler-broadened line profiles obtained from laserinduced fluorescence (LIF) of single quantum states of the photofragments in a photodissociation process as an expansion of bipolar harmonics, which form a complete set in the molecular frame.11 The expansion coefficients, the βKQ(k1k2) bipolar moments, provide a complete characterization of the correlations between the three vectors involved, the photofragment recoil direction and total angular momentum (v and J, respectively), and the photolysis laser polarization (μ). The power of Dixon’s description is revealed when the system under study can be treated semiclassically and the bipolar moments acquire a clear physical meaning. Any semiclassical approximation involves a large amount of quanta at a certain magnitude, so that it can be considered continuous. In molecular polarization, the characteristic magnitude is the total angular momentum, J, and in the semiclassical approximation, the so-called high-J limit is reached. Outside the high-J limit, the validity of the bipolar harmonic expansion remains unaltered, although the bipolar moments lose their physical meaning. The phenomenological formalism presented by Rakitzis and Zare in 1999 in terms of the akq(p) moleculeframe polarization parameters16 constituted an effective alternative to the complex full quantum approaches.17 The akq(p) parameters possess distinct physical meaning in the quantum mechanical frame in the axial recoil limit.16 Despite the efforts made, none of the mentioned formalisms have been chosen by the majority of the experimentalists, who prefer a pure phenomenological approach, lacking of any physical meaning but effortlessly implemented. For one-photon photodissociation and REMPI or LIF detection, the photofragment angular distribution can be expressed as an expansion of Legendre polynomials 2n + 2

I (θ ) =

∑ i=1

σ [1 + βi Pi cos θ ] 4π

⎧ 11 β⎞ 2 10 ⎛ ⎜1 − ⎟a ( ⊥ ) β2XX N = s2⎨ (1 + β)a02(∥) − ⎝ ⎠ 0 21 21 2 ⎩ ⎪

(3)

⎪

where θ is the angle between the photofragment recoil direction and the photolysis polarization direction, σ is the absorption cross section (because the experimental setup has not been calibrated for total intensities, the quotient is treated in general as a normalization fitting parameter), and Pi are the Legendre polynomials of the ith order. The βi expansion coefficients comprise the information related to the dissociation dynamics and the photofragment polarization. When both photolysis and probe lasers are linearly polarized, only the even

+

8177

1 7

8 2 Re[a12(∥, ⊥)] − 3 7

⎫ β⎞ 2 32 ⎛ ⎜1 − ⎟a ( ⊥ ) ⎬ 2 3⎝ 2⎠ ⎭

+ s4

⎛ β⎞ 4 ⎧ ⎨(1 + β)a04(∥) − ⎜1 − ⎟a04( ⊥) ⎝ 21 ⎩ 2⎠

+

5 Re[a14(∥, ⊥)] −

⎫ ⎛ β⎞ 10 ⎜1 − ⎟a 24( ⊥)⎬ ⎝ ⎠ ⎭ 2

⎪

⎪

(4)


The Journal of Physical Chemistry A β4XX N −

Article

⎛ β⎞ 12 ⎧ = s2 ⎨(1 + β)a02(∥) − ⎜1 − ⎟a02( ⊥) ⎝ 35 ⎩ 2⎠

geometry, with identical outcome, that is, a β4ZX ≈ 0 experimental result implies that the k = 4 terms vanish. The above analysis justifies the use of the methodology developed by North and co-workers9,10 for the experimental results presented in this work. To extract the information comprised in the k = 4 terms, more sophisticated strategies should be employed.21 Detection Sensitivity Factors. In the analysis of the experimental data presented in the following section, the (2+1) REMPI sensitivity factor s2 has been used. In ref 20, the sensitivity factors sk=1−4 are explicitly given for ΔJ = 0, ±1, and ±2 transitions as

⎪

⎪

8 Re[a12(∥, ⊥)] + 3

⎫ β⎞ 2 32 ⎛ ⎜1 − ⎟a ( ⊥ ) ⎬ 2 3⎝ 2⎠ ⎭ ⎪

⎪

⎧ 117 β⎞ 4 114 ⎛ ⎜1 − ⎟a ( ⊥ ) (1 + β)a04(∥) − + s4 ⎨ 0 ⎝ ⎩ 231 231 2⎠ +

⎫ ⎛ β⎞ 4 5 Re[a14(∥, ⊥)] − 24 10 ⎜1 − ⎟a 24( ⊥ )⎬ ⎝ ⎭ 77 2⎠ (5)

⎧ ⎛ β⎞ β6XX N = s4 ⎨10(1 + β)a04(∥) − 10⎜1 − ⎟a04( ⊥) ⎝ ⎩ 2⎠ ⎫ ⎛ β⎞ − 4 5 Re[a14(∥, ⊥)] + 2 10 ⎜1 − ⎟a 24( ⊥)⎬ ⎝ ⎠ ⎭ 2

sk = Pk

where N is a normalization factor18 and sk are the detection sensitivity factors for a (2+1) REMPI detection scheme.20 The XX configuration is not sensitive to the q ≠ 0 terms arising from coherent (q = 1) or incoherent (q = 2) interferences, and hence, eqs 4−6 reduce to ⎧ 11 ⎫ β⎞ 2 10 ⎛ ⎜1 − ⎟a ( ⊥ ) ⎬ β2XX N = s2⎨ (1 + β)a02(∥) − 0 ⎭ ⎩ 21 21 ⎝ 2⎠ ⎧4 ⎫ β⎞ 4 4 ⎛ ⎜1 − ⎟a ( ⊥ ) ⎬ + s4 ⎨ (1 + β)a04(∥) − 0 ⎝ ⎠ ⎩ 21 ⎭ 21 2 (7)

(8)

⎧ ⎫ ⎛ β⎞ β6XX N = s4 ⎨10(1 + β)a04(∥) − 10⎜1 − ⎟a04( ⊥)⎬ ⎝ ⎭ ⎩ 2⎠

(9)

β4XX N = s2

The β6 ≈ 0 experimental result can be viewed as a constraint, implying that ⎛ β⎞ (1 + β)a04( ) ≈ ⎜1 − ⎟a04( ⊥) ⎝ 2⎠

c(k)⟨J∥J (k)∥J ⟩

(11)

where the mathematical form of the reduced matrix elements ⟨J∥J(k)∥J⟩, the c(k) coefficients, and the Pk line strength factors are given in the literature.20,22−24 However, the expressions reported in ref 20 for sk are not valid in the case of ΔJ = 0 because they incorporate Pk line strength factors incompatible with the concomitance of linearly polarized light and ΔJ = 0 transitions.24 The problem lies in the fact that, in order to calculate the Pk factors for a ΔJ = 0 transition detected with linearly polarized light, the branching ratio between the different intermediate states of the two-photon processes must be known.24,25 Such constrain is not necessary for ΔJ ≠ 0 transitions or when circularly polarized light is employed. In this work, the detection of the methyl radical is carried out through the Q branch, which, fortunately, might be one of the pairs radical−transition more studied in the literature. According to Hudgens and co-workers,26 the ratio between the perpendicular Σ ← Π ← Σ and parallel Σ ← Σ ← Σ pathways can be set to ∼0.56. Using the eq 11 of ref 25 for the Pk line strength factors, we have obtained a value for s2 = 1.88. High-J Limit. The formalism developed by North and coworkers9,10 considers both the high-J limit and the general case. Outside of the high-J limit, the problem is algebraically unsolvable unless additional independent information is considered. Experimental approaches to measure alignmentfree anisotropy parameters (β = 2β20(20)) have been proposed and used.27,28 The rotational alignment parameter β20(02) can be extracted from the intensity of unsliced images9

(6)

⎫ ⎛ β⎞ 12 ⎧ ⎨(1 + β)a02(∥) − ⎜1 − ⎟a02( ⊥)⎬ ⎝ ⎭ 35 ⎩ 2⎠ ⎧ 117 ⎫ β⎞ 4 114 ⎛ ⎜1 − ⎟a ( ⊥ )⎬ (1 + β)a04(∥) − + s4 ⎨ 0 ⎝ ⎠ ⎭ ⎩ 231 231 2

[J(J + 1)]k /2 2k + 1 2J + 1

(10)

For absorption transitions involving a single excited state, the transition possesses a single character, either parallel or perpendicular. In other words, the condition imposed by eq 10 is fulfilled only if a40(∥) ≈ a40(⊥) ≈ 0. This result reduces to XX zero the k = 4 contributions in both βXX 2 and β2 . Such a conclusion is not unexpected because the XX geometry is in general not very sensitive to the k = 4 terms. The angular distributions of the XZ and ZX images, as stated by North and co-workers,9,10 do no produce any β4 term because only the polarization of one of the two lasers is along the image plane. However, for a (2+1) REMPI scheme, this assertion stands only for the XZ geometry. The second photon involved in the detection through the resonant state of the photofragment gives rise again to a β4 term. In the experiments carried out in the present work, the βZX 4 parameter has been found consistently to be zero. A similar scrutiny to the one carried out for the XX geometry has been carried out for the ZX

β02(02) =

5 ⎛ I − I⊥ ⎞ ⎟⎟ ⎜⎜ 2s2 ⎝ I + 2I⊥ ⎠

(12)

where I⊥ and I|| represent the ion image intensity for perpendicular and parallel pump and probe polarizations, respectively. In the present work, we have followed the later approach. We have measured the β20(02) parameter from unsliced images and compared the outcomes of eqs 5 and 8 of refs 9 and 10. Taking into consideration that the β20(02) parameter determined through eq 12 is speed-independent, the results produced by both equations are undistinguishable within the experimental error, indicating that the methyl fragments are produced with estimable rotational excitation, and thus, we are in the high-J limit. This conclusion is in agreement with the experiment of Houston and co-workers5 who estimated a rotational temperature of ∼200 K for the methyl fragment. 8178



Article

Figure 2. Sliced images corresponding to CH3(ν=0) fragments produced in the photodissociation of CH3NO2 at 193 nm, taken at the different pump−probe polarization configurations XX, XZ, ZX and ZZ, for a pump laser energy of 20 μJ per pulse and a delay time between the laser and molecular beam pulses of t0 + 10 μs (see the text for details). At the top right side of each image are the pump and probe linear polarizations (an arrow means perpendicular and a circle with dot means parallel to the Z axis, which is parallel to the molecular beam).

■

RESULTS AND DISCUSSION Figure 2 shows a series of sliced images corresponding to CH 3 (ν=0) recorded at the four different polarization configurations of the pump and probe laser beams employed in this work. Strikingly, the observed main feature, a highly anisotropic disk, changes dramatically in shape from one image to the other. A close inspection of the images reveals that the central part varies in a different way than the outer annulus depending on the polarization geometries, constituting by itself a differentiate low-recoil contribution. The identification and characterization of the underlying dynamics leading to such features has been carried out by angular and radial integration of the images. The speed distributions of the produced CH3 fragments are obtained by direct angular integration of the XX image and transformed into the TED shown in Figure 3 using the calibration procedure described above. It must be noticed that the XX image provides the true shape of the photofragment TED because it comprises all of the information of the dissociation and detection processes. The XZ distribution neglects the alterations induced by the detection process on the fragment speed, and it should be used when the information relative to the dissociation kinematics is sought, as in the present case. The observed features in the images appear as a broad and unstructured distribution peaking at around 0.4 eV with a tail extending up to 1.5 eV and a shoulder at low

Figure 3. Center-of-mass (CM) CH3 TED obtained by angular integration of the XX image shown in Figure 2. The vertical bars are located at the available energies for the three proposed channels. The arrow indicates the contribution in the low-recoil part of the distribution.

translational energies that accounts for the low-recoil contribution. The relative intensity of both contributions shows a strong dependence on the laser polarizations. Specifically, the low-recoil part seems to disappear when the pump laser polarization is set parallel to the molecular beam (Z). Such behavior is not unfamiliar to the slicing technique 8179



Article

Figure 4. The four vector correlations, β20(20), β00(22), β20(02), and β20(22), as a function of the CH3(ν=0) recoil speed in ∼60 ms−1 bins, obtained via the equations provided in refs 9 and 10. The speed distribution corresponding to the XX image of Figure 2 is shown as the solid line, with an arbitrary intensity scale. The vertical lines are as those in Figure 3.

and denotes a cos2 θ distribution of fragments, where θ is the angle between the pump laser polarization vector and the recoil fragment direction. When the pump laser polarization is set parallel to the molecular beam (Z), the plane that separates the two lobes of the distribution is rotated and coincides with the central slice of the distribution, which averages a region of zero fragment density, and no signal is detected. This result will be discussed later on, based on the anisotropy analysis. The CH3 fragment center-of-mass (CM) translational energies for the different dissociation channels are calculated according to the energy balance m NO2 [hν − D0 + Ei(CH3NO2 ) Etrans(CH3) = mCH3NO2 − Ee(NO2 ) − Ei(CH3) − Ei(NO2 )]

rovibrational excitation in any of the dissociation channels involved, confirming previous works.5 In fact, NO2 must be produced with energy above the dissociation threshold because ground-state NO radicals have been observed as dissociation products.5 Quantitatively, such an amount of extra energy would shift the maximum translational energy bars of Figure 3 to ∼0.5 eV, that is, around the center of the distribution. The stereodynamics of the photodissociation process has been studied using the formalism described recently by North and co-workers,9,10 as explained in the previous section. The angular distributions measured in the ZZ configuration are used as a reference to avoid systematic errors, such as detector inhomogeneities. The images shown in Figure 2 were radially integrated pixel-by-pixel using homemade software, and the resulting angular distributions fitted to eq 3 in the commonly used version for linearly polarized one-photon dissociation and (2+1) REMPI detection processes11,16,18,22

(13)

where ν is the frequency of the photolysis laser, D0 = 2.64 eV is the dissociation energy of the C−N bond,29 Ei refers to the vibrational and rotational energies of the respective species, and the term Ee (NO2) denotes the NO2 electronic state energy. Six different NO2 electronic states, related with six different dissociation channels correlating with ground-state CH3, are considered (in order of increasing energy): 1 2A1, 1 2B2, 1 2B1, 1 2 A2, 2 2A2, and 2 2B2. In Figure 3, the vertical lines correspond to the maximum translational energy (i.e., when the internal energy of both CH3 and NO2 fragments and the parent molecule are considered zero) of three of these channels. The available energy corresponding to formation of NO2 in the ground 1 2A1 state lies at 2.85 eV and cannot account for the experimental results; similarly, the 2 2A2 and 2 2B2 excited states are not energetically allowed with excitation at 193 nm and can be ignored (see Figure 1). The thresholds of the other three channels are located well beyond the distribution edge, indicating that NO2 must be produced with considerable

I (θ ) =

σ [1 + β2P2 cos θ + β4 P4 cos θ + β6P6 cos θ ] 4π (14)

The fit produces a set of speed-dependent βi parameters for each of the images measured. As commented on above, a single parameter (β2) was needed to fit the XZ and ZX images. In all cases, the calculated β6 parameters were found to be close to zero. Thus, the formalism developed by North and co-workers for probing the fragments by (1+1) REMPI is valid for the present case where (2+1) REMPI is used to probe the CH3 fragments. The ZZ image cannot contain any dynamical information, and nonzero βi parameters in this configuration are taken as an instrumental function. The calculated speeddependent βi parameters are introduced in eq 7 of refs 9 and 10 to calculate the speed-dependent βKQ(k1k2) bipolar moments. In the present case, we are probing CH3(ν=0) fragments by (2+1) 8180



Article

REMPI through the Q branch of the 000 transition, which is spectrally congested, and we cannot resolve for individual N states. Thus, the polarization dependence of the measured images permits only the determination of N,K-averaged bipolar moments. The results depicted in Figure 4 confirm that the two contributions observed in the images and TED show a clearly different anisotropy. The β20(20) bipolar moment, which is directly related to the spatial anisotropy parameter β (2β20(20) = β), takes negative values (indicating a perpendicular transition) for the high-speed part of the distribution and positive values (indicating a parallel transition) for the lowspeed part of the distribution. The rest of the bipolar moments provide additional information about the stereodynamics of the dissociation process. For the low-speed part of the distribution, all of the remaining bipolar moments are around zero; however, for the high-speed part of the distribution, some correlations appear. In particular, the β20(20) parameter, associated with fragment’s rotational alingment, that is, the v−J correlation, shows a positive, constant, and non-negligible value. The speed dependence of the β20(20) bipolar moment depicted in Figure 4 suggests three different mechanisms underlying the production of CH3 radicals. In the high-speed region (from ∼2200 to 3300 m/s), β02 (20) takes an approximate constant value of −0.37; below 2200 m/s, β20(20) drops continuously, reaching a value of zero at around ∼1100 m/s; in the low-speed region (from ∼1100 m/s to zero speed), β20(20) takes small positive values. Considering that the limit for a pure perpendicular transition is β20(20) = −0.5, corresponding to β = −1, a value of −0.37 not only corroborates the predicted perpendicular nature of the transition6 but implies dissociation on a repulsive surface. From the possible dissociation channels quoted above, direct dissociation above the TSS3 saddle point yielding CH3 + NO2(1 2 A2) is the only possibility that matches our experimental results. The observed behavior in the intermediate region, the decrease of β20(20) as the fragment speed diminishes, is in agreement with indirect dissociation involving CIs, where there is some loss of anisotropy, and according to Arenas et al.,7 the possible indirect dissociation channels (through the S3/S2 CI) would be CH3 + NO2(1 2B2) and CH3 + NO2(1 2A1). Moreover, according to the energy balance, the CH3 TED spans the vibrational and rotational content of both dissociation products. The loss of anisotropy must be related to the time that the wavepacket spends in the CIs, sampling different structural configurations, which increases with the internal energy of excited molecule.30 The small positive β20(20) values observed in the low-speed region (which correspond to β ≈ +0.19) indicate that the absorption step corresponds to a parallel transition, which, according to the potential energy surface structure of nitromethane, reduces to absorption to the S2 surface. The negligible oscillator strength of this transition at 193 nm,6 however, indicates that a different source is responsible for the signal observed at low speeds or translational energies. Interestingly, the β00(22) bipolar moment presents a constant value of ∼+0.1 through the intermediate and high-speed regions, indicating a mild tendency of the CH3 total angular momentum J to lie parallel to the recoil velocity vector. Such rotational alignment has been observed in similar CH3containing systems31,32 and is produced when no rotational torques are applied to the dissociating bond during the dissociation step.

The TED corresponding to the XX measurement resembles the TEDs measured in the past, where a clear sharp peaked contribution appeared in the low-recoil part of the distribution and was attributed to a dissociation channel yielding groundstate CH3 and NO2(2 2B2),4,5 although this channel is not energetically allowed (see above). In order to investigate further this contribution, we have carried out experiments at different pump laser intensities and at different delay times between the gas and pump laser pulses to scan different regions of the molecular beam pulse. Nitromethane is known to produce small clusters easily, which can yield methyl fragments after cluster multiphoton absorption.33 The onset of the molecular beam pulse is characterized by a certain t0 reference delay time corresponding to a region where the vibrational content of the sample molecules has not been significantly transferred to translational energy in the expansion (hot region where vibrational relaxation is poor). Because vibrational energy in polyatomic molecules usually overcomes van der Waals binding energies, such a region is supposed to be clusterfree. In the lower panel of Figure 5, the CH3 distributions

Figure 5. Upper panel: CM CH3 TEDs obtained at different pump laser intensities and a delay time of t0 + 10 μs. The contribution at low recoil increases with the pump energy. The distribution obtained by Houston et al.5 is also shown for comparison. Lower panel: CM CH3 TEDs obtained scanning different regions of the molecular beam at a pump energy per pulse of 20 μJ.

obtained from CH3NO2 dissociation in different regions of the molecular beam (referenced to t0) and for a pump laser energy of 20 μJ per pulse are shown. At the hottest region of the molecular beam (t0), the low-recoil contribution to the TED appears as a shoulder of the broad and major contribution, but this low-recoil contribution increases as soon as colder regions of the molecular beam pulse (sampled by increasing the delay time) are interrogated. The observed dependence with the molecular beam temperature suggests that the low-recoil part of the distribution can be assigned to parent cluster dissociation, present even at t0. After the initial enhancement, the increase of the low-recoil relative intensity reaches a maximum, suggesting that the cluster contribution must be restricted to small aggregates, such dimers and trimers.33 In the upper panel of Figure 5, CH3 TEDs obtained for different pump laser intensities and for the delay time t0 + 10 μs 8181



Article

(3) Bayliss, N. S.; McRae, E. G. Solvent Effects in the Spectra of Acetone, Crotonaldehyde, Nitromethane and Nitrobenzene. J. Phys. Chem. 1954, 58, 1006−1011. (4) Butler, L. J.; Krajnovich, D.; Lee, Y. T.; Ondrey, G.; Bersohn, R. The Photo-dissociation of Nitromethane at 193 nm. J. Chem. Phys. 1983, 79, 1708−1722. (5) Moss, D. B.; Trentelman, K. A.; Houston, P. L. 193 nm Photodissociation Dynamics of Nitromethane. J. Chem. Phys. 1992, 96, 237−247. (6) Arenas, J. F.; Otero, J. C.; Peláez, D.; Soto, J. The Ground and Excited State Potential Energy Surfaces of Nitromethane Related to Its Dissociation Dynamics after Excitation at 193 nm. J. Chem. Phys. 2003, 119, 7814−7823. (7) Arenas, J. F.; Otero, J. C.; Peláez, D.; Soto, J. Role of Surface Crossings in the Photochemistry of Nitromethane. J. Chem. Phys. 2005, 122, 084324. (8) Soto, J.; Arenas, J. F.; Otero, J. C.; Peláez, D. Effect of an S1/S0 Conical Intersection on the Chemistry of Nitramide in Its Ground State. A Comparative CASPT2 Study of the Nitro−Nitrite Isomerization Reactions in Nitramide and Nitromethane. J. Phys. Chem. A 2006, 110, 8221−8226. (9) Grubb, M. P.; Warter, M. L.; Freeman, C. D.; West, N. A.; Usakoski, K. M.; Johnson, K. M.; Bartz, J. A.; North, S. W. A Method for the Determination of Speed-Dependent Semi-classical Vector Correlations from Sliced Image Anisotropies. J. Chem. Phys. 2011, 135, 094201. (10) Grubb, M. P.; Warter, M. L.; Freeman, C. D.; West, N. A.; Usakoski, K. M.; Johnson, K. M.; Bartz, J. A.; North, S. W. Erratum: “A Method for the Determination of Speed-Dependent Semi-Classical Vector Correlations from Sliced Image Anisotropies” [J. Chem. Phys. 135, 094201 (2011)]. J. Chem. Phys. 2012, 136, 219901. (11) Dixon, R. N. The Determination of the Vector Correlation between Photofragment Rotational and Translational Motions from the Analysis of Doppler-Broadened Spectral-Line Profiles. J. Chem. Phys. 1986, 85, 1866−1879. (12) Rubio-Lago, L.; Amaral, G. A.; Arregui, A.; Izquierdo, J. G.; Wang, F.; Zaouris, D.; Kitsopoulos, T. N.; Bañares, L. Slice Imaging of the Photodissociation of Acetaldehyde at 248 nm. Evidence of a Roaming Mechanism. Phys. Chem. Chem. Phys. 2007, 9, 6123−6127. (13) Gebhardt, C. R.; Rakitzis, T. P.; Samartzis, P. C.; Ladopoulos, V.; Kitsopoulos, T. N. Slice Imaging: A New Approach to Ion Imaging and Velocity Mapping. Rev. Sci. Instrum. 2001, 72, 3848−3853. (14) Papadakis, V.; Kitsopoulos, T. N. Slice Imaging and Velocity Mapping Using a Single Field. Rev. Sci. Instrum. 2006, 77, 083101. (15) The aim of the present work is to use the formalism developed by North and co-workers9,10 for the photodissociation of nitromethane at 193 nm. The treatment of North and co-workers is based on the work by Dixon11 and Rakitzis and Zare,16 and therefore, only those works and the connections between them are briefly reviewed here. References to alternative photofragment angular polarization treatments can be found in the works of Dixon and Rakitzis and Zare. (16) Rakitzis, T. P.; Zare, R. N. Photofragment Angular Momentum Distributions in the Molecular Frame: Determination and Interpretation. J. Chem. Phys. 1999, 110, 3341−3350. (17) Siebbeles, L. D. A.; Glass-Maujean, M.; Vasyutinskii, O. S.; Beswick, J. A.; Roncero, O. Vector Properties in Photodissociation  Quantum Treatment of the Correlation between the Spatial Anisotropy and the Angular-Momentum Polarization of the Fragments. J. Chem. Phys. 1994, 100, 3610−3623. (18) Rakitzis, T. P. Direct Measurement of Photofragment Alignment from Unnormalized Abel-Invertible Images. Chem. Phys. Lett. 2001, 342, 121−126. (19) Rakitzis, T. P.; Hall, G. E.; Costen, M. L.; Zare, R. N. Relationship Between Bipolar Moments and Molecule-Frame Polarization Parameters in Doppler Photofragment Spectroscopy. J. Chem. Phys. 1999, 111, 8751−8754. (20) Rakitzis, T. P.; Alexander, A. J. Photofragment Angular Momentum Distributions in the Molecular Frame. II. Single State Dissociation, Multiple State Interference, and Nonaxial Recoil in

are shown and compared with the result of Houston and coworkers.5 As can be seen, the low-recoil contribution increases significantly as the pump laser energy is increased. The dependence of this low-recoil contribution with the pump laser intensity indicates multiphoton processes. The most likely candidates are nitromethane dimers and trimers, easily formed in the expansion and that, after two- (dimers) or three(trimers) photon absorption, dissociate into CH3 and NO2 fragments. It must be pointed out that the agreement between our work (at the highest pump intensity studied (40 μJ) and for the delay time t0 + 10 μs) and the results of Houston and coworkers is remarkable, taking into account the different techniques employed. Such evidence suggests that the source of the low-recoil contribution in previous works is dissociation of small parent clusters.

■

CONCLUSIONS In this work, we have employed the slice imaging technique in combination with REMPI detection of the methyl products to determine the TED and stereodynamics of the photodissociation of nitromethane at 193 nm. The TED of the CH3 radical presents two contributions, as reported in previous experimental works, but the assignment of the dissociation channels, made on the basis of the measured bipolar moments, is in disagreement with the literature. The sharp and lowrecoiled contribution is proved to come from small parent cluster dissociation. The broad distribution agrees with three possible different dissociation channels yielding ground-state CH3 and NO2 in different electronic states, involving direct dissociation and deactivation through different CIs.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (L.R.-L.); [email protected] (L.B.). Phone:+34913944228. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Andreas Kartakoullis and Pavle Glodic for their help in setting up the experiments. J.D.R. acknowledges financial support from MICINN (Spain) through a FPI fellowship. M.G.G. and L.R.-L. thank MICINN for predoctoral and postdoctoral contracts, respectively, through the Consolider program “Science and Applications of Ultrafast Ultraintense Lasers” Grant No. CSD2007-00013. M.G.G. also acknowledges financial support from the MICINN program “Personal Técnico de Apoyo”. This work has been financed by the Spanish MICINN through Grants CTQ2008-02578, CTQ2012-37404-C02-01, and CSD2007-00013 and by EU FP7 programs LASERLAB-EUROPE II (Grant No. 228334) and ICONIC (Grant No. PITN-GA-2009-238671). P.C.S. further acknowledges EU FP7 support under the Marie Curie Reintegration Grant GPSDI (GA No. PIRG07-GA-2010268305).

■

REFERENCES

(1) Loos, K. R.; Wild, U. P.; Gunthard, H. H. Ultraviolet Spectra and Electronic Structure of Nitroethylene. Spectrochim. Acta, Part A 1969, 25, 275−281. (2) Nagakura, S. Ultra-Violet Absorption Spectra and π-Electron Structure of Nitromethane and the Nitromethyl Anion. Mol. Phys. 1960, 3, 152−162. 8182



Article

Photodissociation of Polyatomic Molecules. J. Chem. Phys. 2010, 132, 224310. (21) Rakitzis, T. P.; Samartzis, P. C.; Kitsopoulos, T. N. Complete Measurement of S(1D2) Photofragment Alignment from AbelInvertible Ion Images. Phys. Rev. Lett. 2001, 87, 123001. (22) Zare, R. N. Angular Momentum: Understanding Spatial Aspects in Chemistry and Physics; Wiley: New York, 1998. (23) Orr-Ewing, A. J.; Zare, R. N. Orientation and Alignment of Reaction−Products. Annu. Rev. Phys. Chem. 1994, 45, 315−366. (24) Mo, Y. X.; Suzuki, T. Geometrical Factors of Two-Photon Absorption for the Determination of Alignment and Orientation. J. Chem. Phys. 1998, 109, 4691−4692. (25) Mo, Y. X.; Suzuki, T. Vector Correlation in Molecular Photodissociation: Quantum Mechanical Expression and Comparison with the Formal Expansion Formula. J. Chem. Phys. 2000, 112, 3463− 3473. (26) Hudgens, J. W.; DiGuiseppe, T. G.; Lin, M. C. Two Photon Resonance Enhanced Multiphoton Ionization Spectroscopy and State Assignments of the Methyl Radical. J. Chem. Phys. 1983, 79, 571−582. (27) Wouters, E. R.; Ahmed, M.; Peterska, D. S.; Bracker, A. S.; Suits, A. G.; Vasyutinskii, O. S. In Imaging in Chemical Dynamics; Suits, A. G., Continetti, R. E., Eds.; American Chemical Society: Washington, DC, 2001; p 238. (28) Smolin, A. G.; Vasyutinskii, O. S.; Wouters, E. R.; Suits, A. G. Orbital Alignment in N2O Photodissociation. I. Determination of All Even Rank Anisotropy Parameters. J. Chem. Phys. 2004, 121, 6759− 6770. (29) Luo, Y. R. Handbook of Bond Dissociation Energies in Organic Compounds; CRC Press: Boca Raton, FL, 2003. (30) Zener, C. Non-Adiabatic Crossing of Energy Levels. Proc. R. Soc. London, Ser. A 1932, 137, 696−702. (31) Cruse, H. A.; Softley, T. P. Velocity-Map Imaging Study of the Photodissociation of Acetaldehyde. J. Chem. Phys. 2005, 122, 124303. (32) Pisano, P. J.; Cline, J. I. Determination of μ-v-j Vector Correlations in Photodissociation Experiments Using 2+n ResonanceEnhanced Multiphoton Ionization with Time-of-Flight Mass Spectrometer Detection. J. Chem. Phys. 2000, 112, 6190−6200. (33) Kwok, H. S.; He, G. Z.; Sparks, R. K.; Lee, Y. T. On the PhotoDissociation of Nitromethane at 266 nm. Int. J. Chem. Kinet. 1981, 13, 1125−1131.

8183


Stereodynamics of the Photodissociation of Nitromethane at 193 nm

Recommend Documents