Steric Control of the Reaction of CH Stretch–Excited CHD3 with Chlorine Atom

See allHide authors and affiliations

Science  18 Feb 2011:
Vol. 331, Issue 6019, pp. 900-903
DOI: 10.1126/science.1199771


Exciting the CH-stretching mode of CHD3 (where D is deuterium) is known to promote the C-H bond’s reactivity toward chlorine (Cl) atom. Conventional wisdom ascribes the vibrational-rate enhancement to a widening of the cone of acceptance (i.e., the collective Cl approach trajectories that lead to reaction). A previous study of this reaction indicated an intriguing alignment effect by infrared laser–excited reagents, which on intuitive grounds is not fully compatible with the above interpretation. We report here an in-depth experimental study of reagent alignment effects in this reaction. Pronounced impacts are evident not only in total reactivity but also in product state and angular distributions. By contrasting the data with previously reported stereodynamics in reactions of unpolarized, excited CHD3 with fluorine (F) and O(3P), we elucidate the decisive role of long-range anisotropic interactions in steric control of this chemical reaction.

The directional nature of chemical bonds generally renders intermolecular interactions and the Born-Oppenheimer potential energy surfaces (PESs) that describe such interactions anisotropic. In the course of reactions that form and break those bonds such anisotropy will tend to dynamically steer the scattering trajectories either toward or away from the transition state (a critical configuration along the reaction path, serving as a bottleneck to reaction), thereby promoting or suppressing reactivity. This steering or reorientation effect has long been recognized by chemists, and mechanistic insights at the molecular level have been garnered for a number of elementary reactions (18). Particularly relevant to this work are two recent studies on the reactions of CH stretch–excited, unpolarized CHD3(v1 = 1) with F atom (9) and O atom (10).

The reaction of F + CHD3 is highly exothermic (change in enthalpy ΔH0 = −31.3 kcal/mol) with a small early barrier, for which the transition-state structure is reactant-like. Our group found that one quantum excitation of the CH stretch (v1 = 1) in CHD3 inhibits C-H bond cleavage, resulting in a deceleration of the overall reaction rate (9). This unexpected finding was conjectured, and later calculated theoretically (11), to be a result of steering or deflection of the approaching F atom away from the targeted H atom. By contrast, an opposite effect was discovered in the O + CHD3 reaction (10), which is slightly endothermic (ΔH0 = 2.1 kcal/mol) with a barrier height of ~9.6 kcal/mol. The location of the barrier is nearly at the midpoint of the pathway; namely, the transition-state structure is neither reactant-like nor product-like. CH-stretching excitation led to a substantial rate promotion at fixed collision energy (Ec). Moreover, the product angular distribution broadened markedly from backward peaking to sideways dominant, suggesting that the vibrational enhancement operates by extending the range of impact parameters, thus opening up the cone of acceptance to reaction; this mechanism has been theoretically predicted to be particularly prominent in thermoneutral atom + diatom reactions (6). The driving force behind this vibrationally induced steric mechanism was further attributed to long-range anisotropic interactions, which pull or focus the trajectories toward the transition state (10, 12).

A deeper implication of these two studies is that in a reaction with strong anisotropic interactions, prealigning the reagents will have little impact on reactivity because of the reorientation effects en route to the barrier. That is, only reactions with weak steering interactions are viable candidates for steric control (active influence over the outcome by prealignment). In this regard, a previous report on the Cl + CHD3(v1 = 1) reaction, in which a reagent alignment effect on reactivity was observed (13), raises puzzling questions. That study probed a single rovibrational product state, for which Cl atoms approaching perpendicularly to the stretch-excited C-H bond preferentially yielded forward-scattered HCl(v = 1, j = 1) product. Here, v refers to vibrational excitation and j refers to rotational excitation. The Cl + CHD3 reaction is slightly endothermic (ΔH0 = 1.8 kcal/mol) with a moderate barrier height of ~4 kcal/mol. The transition-state structure is more product-like (14, 15) than in the O(3P) reaction. A detailed study of this reaction indeed demonstrated (16, 17) an appreciable rate promotion upon CH-stretching excitation, as observed in the O(3P) + CHD3(v1 = 1) case. Does the vibrationally induced steric mechanism also operate in the Cl + CHD3(v1 = 1) system to promote reactivity? If so, how can the CHD3(v1 = 1) remain asymptotically aligned, as is necessary to explain the reagent alignment effect (13), given the anisotropic steering force that underlies the vibrationally induced steric mechanism (6)?

We report here an in-depth study of the stereodynamics of Cl + CHD3(v1 = 1), aimed at clarifying this apparent paradox and shedding more light on the general tug-of-war between the reorientation effect exerted by the interaction potential and the stereoselectivity induced by prealigning the reagents. We performed a crossed-beam scattering experiment under single-collision conditions. The details of the apparatus have been described previously (1619), except for the preparation of aligned CHD3(v1 = 1) reagents, which we performed in this work at the molecular beam crossing region [rather than in the source chamber (16, 17)]. Figure 1 depicts the experimental setup and time-sliced, velocity-map imaging detection scheme. Two molecular beams were rotated so that the relative velocity vector lay parallel to the ultraviolet (UV) probe laser direction. The infrared (IR) excitation laser, tuned to the R(0) transition of the CHD3(v1 = 1 ← 0) band at 2999.21 cm−1 (20), was directed perpendicularly to the UV laser. Thus, a single rovibrationally excited reactant CHD3(v1 = 1, |JKM ≥ |100〉) was prepared, where the quantum numbers J, K, and M denote the rotational angular momentum, the projection of J onto the top axis, and the projection of J onto the IR-polarization axis, respectively. A variable waveplate was used to control the linear polarization axis of the IR laser. The // and symbols refer, respectively, to parallel and perpendicular orientation of the IR-polarization axis relative to the initial relative velocity vector of the collision system. The UV probe laser was fired 3.5 μs after the IR excitation pulse to allow product buildup and to maximize the differential signals of the // and ⊥ configurations. To unravel the stereodynamics, we acquired four product images: IR-off, IR-//, IR-⊥, and IR-αM [i.e., an IR polarization at the magic angle, 54.7°, with respect to the relative velocity vector axis (21)]. Experimentally, imaging data at each configuration were accumulated for 5 min in sequence, and this cycle was then repeated, typically >20 times, to generate the final set of product images.

Fig. 1

Experimental setup for state-selected, aligned molecular beam scattering experiments. A linearly polarized IR laser directed perpendicularly to the relative velocity vector of the two molecular beams prepares the vibrationally excited CHD3 at the scattering center. Reagent alignment is controlled by IR laser polarization direction: “//” refers to an end-on attack and “⊥” to a side-on approach. Time-sliced velocity-map imaging reveals the alignment effects on product pair–correlated distributions.

Figure 2 presents raw images of the probed CD3(v = 0) products at Ec = 3.84 kcal/mol. Consistent with a previous report (17), no IR-generated signals for CHD2 products could be detected at this energy (fig. S1). The IR-off and IR-αM images are almost identical to those reported previously (16, 22). The IR-off image features two back-scattered structures, labeled as (00, 0)g and (00, 0)b (see Fig. 2 legend), respectively, whereas the IR-αM image exhibits distinct, additional ringlike structures reflecting the impact of C-H stretch excitation on the reaction dynamics. On energetic grounds, the sharp forward peak in the inner ring can be assigned to the (00, 1)s product pair, whereas the back-scattered outer ring is ascribed to the (00, 0)s pair. Compared to the IR-αM image, even casual inspection reveals that the relative intensities of the sharp forward peak and the outer-ring feature show opposite propensities as the IR-polarization axis flips from // to ⊥.

Fig. 2

Four CD3(v = 0) product images with a superimposed axis indicating the scattering directions; the 0° angle refers to the initial CHD3 beam direction in the center-of-mass frame. The pair-correlated labeling is defined as follows: The numbers in the parentheses denote the quanta of vibrational excitation in the modes of CD3 (left) and HCl (right) products; the outer subscript indicates the reactant state (“g” for ground-state CHD3, “b” for bend-excited CHD3, and “s” for stretch-excited CHD3). αM indicates magic angle polarization.

To map out the polarization-angle dependence of the alignment effect, we acquired images at many intermediate polarization angles. The results are summarized in Fig. 3 (top). For the (00, 0)s pair the total intensity (I) was included, whereas for (00, 1)s, only the forward signals within a 70° cone (i.e., excluding the IR-off feature) were counted. Notably, the two product channels display distinct, out-of-phase modulations in signals. The polarization-angle dependences can be fitted (solid lines) with I(α) = Iiso [1 + βP2(cosα)] (Eq. 1), analogous to the form that describes the one-photon excited, aligned CHD3 reactant (13, 23): Iiso is proportional to the product-state–specific reaction cross section from an unpolarized reagent, β characterizes the stereoselective reactivity with values ranging from +2 to −1, and P2(cosα) = (3cos2α − 1)/2 is the second-order Legendre polynomial with α denoting the IR polarization angle with respect to the initial relative velocity. After full image analysis accounting for the density-to-flux transformation (18) and the hyperfine depolarization (see below), the resulting anisotropy parameters β could be extracted (table S1). For R(0) rotational transition pathways, these were 1.38 ± 0.24 for HCl(v = 0) and −0.38 ± 0.05 for the forward-scattered HCl(v = 1). A negative β value (signifying that the ⊥-polarization is favored) for forward-HCl(v = 1) corroborates the previous finding (13).

Fig. 3

(Top) Polarization-angle dependence of the probed signals from the stretch-excited reaction. The v = 0 data correspond to the outer-ring feature labeled (00,0)s in Fig. 2; for the v = 1 data [the (00, 1)s pair], only the forward signals within a 70° cone are counted. Two sets of data are normalized by I(αM) = 1 for display. Solid lines are best fits based on Eq. 1. (Bottom) Temporal profiles of the state-specific signals in the top panel were monitored to determine the degree of hyperfine depolarization. Imin (Imax) denotes I// (I) for v = 1, and I (I//) for v = 0. The results for each of the two product states agree within 10%; averaging with another set of v = 0,1 measurements yields C/IM) = 0.32 ± 0.02 at Δt= 3.5 μs, which is shown by the shaded area in the top figure. By subtracting the depolarization component C from the data shown in the top panel, the genuine β parameters can then be determined.

As alluded to earlier, the products were probed 3.5 μs after the IR excitation to build up product yield. During this time, recoupling of the randomly distributed nuclear spins in CHD3 with the initially aligned rotational angular momentum leads to a dephasing of coherently excited hyperfine components (13, 24), which scrambles (or depolarizes) the initial alignment and contaminates the observed product signals. That is, if there were no hyperfine depolarization, the observed alignment effects would have been more pronounced. To determine the degree of such depolarization, we measured the temporal profiles of I// and I with three IR laser polarizations (0°, 54.7°, and 90°) at different time delays (Δt) between two lasers. As shown in Fig. 3 (bottom), a linear dependence of Imin/Imax on Δt is evident. By reexpressing the polarization-angle dependence as I = Iiso [1 + βP2(cosα)] + C, one can deduce the depolarization component C at Δt using the extrapolated Imin/Imax value at Δt = 0. The results of four independent determinations yielded C/IM) = 0.32 ± 0.02 at Δt = 3.5 μs, which is sketched as the shaded area in Fig. 3 (top) and is to be subtracted to derive the genuine β values.

The image analysis followed the previous procedures (1619); the resulting pair-correlated product angular and speed distributions are presented in figs. S2 and S3. With the depolarized contributions subtracted out from the measured // and ⊥ distributions, the final product angular distributions [Fig. 4 (top)] and speed distributions (fig. S4) then represent the genuine stereodynamics of the reaction from an optically aligned CHD3(v1 = 1, |100〉) reactant. From the product speed distribution (fig. S4), the polarization-dependent cross sections are obtained and summarized in Fig. 4 (bottom) (see table S2 for the numerical data).

Fig. 4

(Top) Polarization-dependent angular distributions of HCl(v = 0) and (v = 1) states formed concomitantly with the probed CD3(v = 0) products. A large disparity is evident for different IR-polarizations. (Middle) Polarized integral cross sections presented graphically (see table S2 for the numerical data). All data are normalized to Embedded Imageat αM, and the error represents two SDs. Previous study of Cl reaction with an unpolarized CHD3(v1 = 1) demonstrated a vibrational enhancement factor of 10 relative to the CHD3(v = 0) cross section at Ec = 3.84 kcal/mol (16). (Bottom) Molecular cartoons illustrating the stereodynamics of the scattering directions of the HCl(v = 0) product channel.

All polarization-dependent cross sections presented in Fig. 4 are normalized to σtotalM). The //-polarized CHD3(v1 = 1) yields a higher total reactivity, σ// = 2.7, and this preference for //-polarization originates solely from the HCl(v = 0) product channel (σ//= 7.71); HCl(v = 1) shows an opposite propensity slightly in favor of ⊥-polarization, σ//= 0.87. In terms of the correlated HCl vibrational branching ratio, σ(v = 1)/σ(v = 0) changes from a cold distribution of 0.31 for //- polarization to an inverted distribution, 2.71, for ⊥-polarization—a marked manifestation of stereoselectivity on product state populations.

Although the state-specific steric effects are pronounced, their mechanistic origin is less obvious. An optically prepared CHD3 reagent can yield alignments from both the rotational angular momentum and the vibrational transition dipole moment (13). The former arises from unequal population of (2J + 1)-fold degenerate mJ sublevels induced by optical excitation, whereas the latter, relying on a parallel transition for v1 = 1 ← 0, dictates that the C-H bond axis lie along the polarization direction of the IR radiation. These two types of alignment can be distinguished by means of different rotational-branch excitations (13): The alignment of the C-H bond will always peak along the laser polarization axis for all branches, whereas the rotational alignment reverses when switching from R- to Q-branch excitation. Hence, to unravel the origin of the observed alignment effects, we also determined the β values via the Q(1) and P(1) excitations. Quantitative comparisons of the resulting β values (table S1) lead us to conclude that the observed polarization dependence for the (00, 0)s product pair is due primarily to the C-H bond axis alignment, whereas the forward-scattered (00, 1)s pair is subject to both the bond axis and rotational angular momentum alignments. The need to consider both alignment origins complicates physical interpretations. More sophisticated probes will be required to disentangle the two alignment effects.

With this caveat in mind, we examine the polarization-dependent angular distributions. The trajectories of the (00, 0)s pair from an unpolarized CHD3(v1 = 1) at αM are mainly back-scattered, suggestive of a direct rebound reaction mechanism at this Ec (16, 25). Moreover, all polarized angular distributions for HCl(v = 0) products span the same angular range as that of the ground-state reaction (fig. S5). We surmise that neither the vibrational enhancement nor alignment preference can be attributed to enlarging the range of reactive impact parameters; rather, they are ascribed to the increase of reaction probabilities at fixed impact parameters. This conclusion is in sharp contrast to the findings in studying the O(3P) + CHD3(v1 = 1) reaction, for which a greater range of impact parameters contribute to reaction upon vibrational excitation, yielding a much broader product angular distribution than the ground-state reaction (10). A closer inspection of the HCl(v = 0) angular distributions for the //- and ⊥-configurations reveals that whereas the former remains backward-peaked, the latter, with lower reactivity, appears sideways-dominant. Qualitatively, the result for (00, 0)s is in keeping with a direct reaction with collinear transition state, in which a //-aligned CHD3 favors small–impact parameter, end-on collisions (26), whereas the ⊥-aligned CH bond prefers larger–impact parameter, side-on attack of Cl atoms (27). This intuitive classical picture implicitly assumes little reorientation during the approach motion to the barrier.

The reactive fluxes into the (00, 1)s pair show a slight preference for ⊥-polarized CHD3(v1 = 1). Both polarization-dependent angular distributions cover the full angular range. The observed propensity is, however, evident only in a very narrow range of forward-scattering angles; little differential reactivity is discernible at larger angles. The origin of this channel’s dynamics is more subtle owing to the presence of both bond-axis and rotational angular momentum alignment effects, as mentioned above. Mechanistically, the formation of HCl(v = 1) is likely mediated by a reactive resonance pathway (16, 17, 28), rather than a direct reaction, further complicating the issue.

Irrespective of the underlying mechanism, the marked state-specific stereoselectivity reported here is a testament to the conservation of reactant prealignment and corresponding absence of substantial steering effects in this reaction—in sharp contrast to the analogous reactions of CHD3(v1 = 1) with F and O(3P) atoms (9, 10). In retrospect, this implication may not be too surprising. It has been demonstrated (22, 29) that in the ground-state Cl + CHD3 reaction, the product angular distribution essentially mirrors the opacity function (the reaction probability as a function of impact parameters) in a one-to-one correspondent manner [see figures 2 and 5 of (22)]. In accounting for enforcement of such a mirror-like correspondence, a rather weak anisotropic PES in the entrance valley could have been inferred. The structure of the transition state is product-like, and thus the reaction barrier is recessed in the exit valley. Upon vibrational excitation of CHD3(v1 = 1), the elongation of the C-H bond can enlarge the range of attack angles at the reaction barrier, thereby increasing the reaction probability at fixed impact parameters, but the resulting anisotropic interactions may not extend into the entrance valley far enough to appreciably steer the prealigned reagents.

Supporting Online Material

Figs. S1 to S5

Tables S1 and S2


References and Notes

  1. The contrasting behaviors of O and F can also be qualitatively rationalized from the transition-state theory perspective. The transition state in F + CHD3 possesses a reactant-like structure, and thus a stretched or compressed C-H bond of CHD3 deviates from the transition-state structure, disfavoring the reaction. By contrast, ab initio calculations predicted that in O(3P) + CHD3, both the breaking C-H bond and forming O-H bond are elongated. Excitation of the C-H bond of CHD3 therefore helps attain the transition-state structure.
  2. Magic angle imaging serves two purposes. First, at this polarization angle, the approaching Cl atom encounters a practically unpolarized CHD3(v1 = 1, j = 1) reagent. Previously, when we prepared CHD3(v1 = 1) in the source chamber (16, 17), the excited reagents traveled for >100 μs before reacting, which we presumed was long enough to depolarize initially aligned CHD3(v1 = 1) molecules by hyperfine interactions. Indeed, the results from the acquired αM image agree broadly with the previous findings. Second, the signals from the stretch-excited reagents under the three polarization angles are related by (I// + 2I) = 3IαM, which provides a stringent check (within ±1% in this work) of the consistency of the data.
  3. In an optically aligned (not oriented) CHD3 molecule, the H atom of the aligned C-H bond can point either toward or away from the approaching Cl atom; thus, no distinction of the head-versus-tail dynamics can be made.
  4. In a direct reaction, the forward-scattered product is normally associated with large–impact parameter collisions, for which there is a purely kinematic smearing of the alignment effect even for a perfectly aligned reagent (30).
  5. We are indebted to S. Yan for earlier attempts of this project and to J. Lam for help with the experiment. This work was supported by National Science Council (NSC-99-2113-M-011-016), Academia Sinica, and the Air Force Office of Scientific Research (AOARD-10-4034).

Stay Connected to Science

Navigate This Article