Velocity Map Imaging the Scattering Plane of Gas Surface Collisions

The ability of gas-surface dynamics studies to resolve the velocity distribution of the scattered species in the 2D sacattering plane has been limited by technical capabilities and only a few different approaches have been explored in recent years. In comparison, gas-phase scattering studies have been transformed by the near ubiquitous use of velocity map imaging. We describe an innovative means of introducing a surface within the electric field of a typical velocity map imaging experiment. The retention of optimum velocity mapping conditions was demonstrated by measurements of iodomethane-d3 photodissociation and SIMION calculations. To demonstrate the systems capabilities the velocity distributions of ammonia molecules scattered from a PTFE surface have been measured for multiple product rotational states.

velocity map imaging. We describe an innovative means of introducing a surface within the electric field of a typical velocity map imaging experiment. The retention of optimum velocity mapping conditions was demonstrated by measurements of iodomethane-d3 photodissociation and SIMION calculations. To demonstrate the system's capabilities the velocity distributions of ammonia molecules scattered from a PTFE surface have been measured for multiple product rotational states.
It has been noted by several authors 1-5 that the study of gas-surface scattering could be revolutionized by using a combination of resonance enhance multiphoton ionization (REMPI) 6 and velocity map ion imaging (VMI). 7 REMPI-VMI allows the velocity distribution of quantum state selected products to be recorded; this approach has already been widely adopted in the gas-phase scattering community 8 as an vast improvement on previous techniques, e.g. traditional rotatable mass spectrometer approach 9 , which do not detect the whole 2D scattering plane in a single measurement.
VMI uses the interpenetrating electric fields generated by annular electrodes to map ions with the same velocity, but created in different locations, onto the same spot on a position sensitive detector. 7,10 Even small perturbations to these carefully created electric fields result in the loss of 'velocity mapping' conditions, so approaches to imaging the scattering from surfaces have endeavored to minimize these effects by either mounting the surface onto an electrode, [2][3][4] or outside the electrodes altogether. 1 These approaches mean that it is no longer possible to image the whole 2D scattering plane, and so multiple measurements are still required to obtain the total scattering distribution.
Detecting the velocity distribution of the whole 2D scattering plane in a single measurement requires the scattering plane to be parallel to the position sensitive detector. If the surface is mounted onto one of the electrodes 3, 4 then the scattering plane, which must contain the surface normal, is necessarily perpendicular to the detector preventing its direct measurement. Thus, either complex modeling or multiple time slicing measurements are required to deconvolute the velocity profile. Furthermore, only grazing angles of incidence are possible, as gas molecules have to pass between the electrodes to strike the surface. 3,4 Experiments that mount the surface outside the electrodes 1 are limited to having a large surface to laser distance; that means that the ionization region has to be particularly large to avoid preselecting molecules and biasing the measured distribution to a small range of scattering angles. This requires very shallow electric gradients that limit such techniques to only spatially imaging the scattered products and all velocity information has be inferred from images collected at individual time steps. In this paper, we present a novel adaptation to a standard VMI apparatus that overcomes these issues and provides a mechanism for directly imaging the scattering plane irrespective of the incident angle and should allow novel REMPI-VMI studies to be performed on numerous gas (dielectric)surface scattering systems.
The Surface-Scattering Velocity Map Imaging (SS-VMI) experimental set-up consisted of two differentially pumped chambers: a source chamber, that contained the molecular beam source; and a scattering chamber, which housed the laser focal region, the VMI electrodes (ion optics), the surface, and time-of-flight (TOF) region. The laser, TOF, and the molecular beam axes are mutually perpendicular (labeled x, y and z, respectively, see Figs. 1 and 2) and meet at the center of the scattering chamber.
The molecular beam source consisted of a pulsed solenoid valve (General Valve Corporation, series 9) with a 200 µs opening time, that was used to form a supersonic expansion of ammonia (2.5% in 4 Bar He), or iodomethane-d3 (2.5% in 4 Bar Ar), with mean velocities of 1550 ms -1 and 540 ms -1 , respectively. The molecular beam was collimated by a 1.01 mm skimmer (Beam Dynamics, model 1) mounted at the intersection between the two chambers, 33 mm from the front face of the nozzle and 176 mm from the laser axis.  The ion optics (see Fig. 1) employed are based on those designed by Wrede et al. 11 , which have seen widespread use in photodissociation experiments. [12][13][14][15][16] The design implemented in these studies (shown in fig 4 of ref. 12) comprised five electrodes: a cup shaped repeller, a conical extractor, two additional annular lenses and ground (Labelled R, E, L1, L2 and G respectively in Fig. 1). Careful optimization of the ion optics' voltages and fast pulsing of the second MCP (40 ns temporal width) allowed for dc-slice imaging, where the ion packets were elongated and only the central "slice" was acquired by pulsing the detector. 17 This allowed the exclusive imaging of only those products in the scattering plane defined by the molecular beam and the surface (i.e. the xz plane).
The surface studied in this work was PTFE (Polytetrafluoroethylene) with an exposed top face of 1 mm (along the TOF y-axis) by 12.7 mm (along the laser x-axis) that was normal to the molecular beam direction (zaxis, with the positive z direction defined as away from the surface). The PTFE surface was mounted in a PEEK surface. These scalpel blades were employed as compensating electrodes to stabilize the electric field and maintain optimum velocity mapping conditions between the repeller and extractor. Scalpel blades were employed for these electrodes because of the dual benefits of their sharp edges: producing minimal electric field perturbation; and a narrow cross-section for scattering, which otherwise may mask scattering from the PTFE surface. The PEEK holder was mounted in the scattering chamber via an XYZ translator, which allowed the surface to be positioned relative to the TOF and laser axes, or removed from the electric field region altogether to perform comparative measurements using the "standard" VMI technique (referred to as VMI mode from hereon). All SS-VMI experiments were performed with the PTFE surface held 10 mm from, and aligned centrally to, the TOF and laser axes. Table I. shows the voltages applied to the ion optics and stabilizing electrodes in SS-VMI and VMI mode.  To compare the velocity mapping conditions of the experiment in VMI and SS-VMI modes one color photodissociation of CD3I and subsequent ionization of I( 2 P1/2) atoms was performed at 310.6 nm, using a (2 + 1) REMPI transition (via the 5p 4 ( 3 P0) 6p 1 2 [1]°3/2 state). 18 This acted as a useful benchmark for velocity calibration. nm. 19 The rotational levels studied were isolated spectroscopic transitions that represent low, medium and high JK cases in the scattered signal. Each NH3 scattering image contains two features: signal from the incident molecular beam (lower in Fig. 4) and signal from surface scattering (upper in Fig. 4). As scattered molecules must make a 20 mm round trip from the detection region to the surface and back again, scattering signal is expected to appear at a minimum time of 13 µs after the initial molecular beam appearance; the images in Fig. 4 are taken at a 90 µs delay.
It should be noted that the slight perturbation of the electric field, predicted by the SIMION calculations ( Fig. 2), is far more significant for ions whose lab frame velocity is directed towards the surface (negative z), while ions with a lab frame velocity away from the surface (positive z) will experience minimal perturbation of the electric field. For the CD3I images in Fig. 3b both the upper and lower crescents of the rings have negative z velocities in the lab frame, as the photodissociation induced velocity of the iodine is less than the molecular beam velocity. However, the NH3 scattered signal only possess positive z velocity taking these ions away from the surface to an area where the velocity mapping field is unperturbed by the presence of the surface.
Speed distributions of the scattered NH3 are obtained by radially integrating the signal in a 90° wedge centered on the positive z direction from zero lab frame velocity. Radial integration of 90° wedges perpendicular to the molecular beam (i.e. along the x-axis), from zero lab frame velocity, was used to obtain and subtract the background signal caused by NH3 scattering from the chamber walls and electrodes. This background is visible in Fig 4. close to zero lab frame velocity. The radial signal is then converted to speed by a calibration factor (17.92 ms -1 pixel -1 for NH3) determined from the CD3I images. Normalized speed profiles of the scattering signal (from Fig. 4) are shown in Fig. 5. These speed profiles show that there is a large exchange of translational energy during the collision between NH3 and surface; by fitting the speed profile to a Maxwell-Boltzmann distribution a mean NH3 speed 20 can be obtained (694 ms -1 for JK = 10) that is significantly slower than the incoming molecular beam velocity. We note that the loss in translational energy does not follow a trend dependent on J, as the mean speed increases from JK = 10 to 53 (821 ms -1 ) and then decreases again from 53 to 86/109 (799 ms -1 ).
The angular distribution of scattered signal about the zero lab frame velocity in Fig. 4 show a narrow range with all signal in an approximately 75° wedge centered around the z axis. The angular distribution of signal shows only a slight change across the three rotational levels studied, with JK = 10 showing a narrower spread by approximately 15°. It should be noted that the range of measurable scattering angles is limited by our detection geometry; this limit is close to an 80° region, centered on the positive z direction from the zero lab frame velocity, calculated using the Raleigh range of laser focus and the surface to laser distance.
In conclusion, we have demonstrated the viability of SS-VMI as a means of directly detecting the 2D velocity distribution in the scattering plane of a gas-surface collision, which can lead to more rigorous studies of scattering dynamics. The calculations of electric field contours and the experimental iodomethane-d3 photodissociation images have shown the feasibility of using stabilizing electrodes to maintain velocity mapping conditions while introducing a dielectric object (in this case the PTFE surface) inside the VMI ion optics.
Ammonia scattering results have highlighted the quality of the velocity distributions that can be obtained with these experiments. Although the range of detectable scattering angles currently limits the technique, it is expected that an alternative ion optics design will allow a reduction of the surface to laser distance (while maintaining velocity mapping conditions). Coupled with greater laser Rayleigh range, this will minimize the level of background signal, improve velocity resolution and allow a greater than 150° range of measurable angles.