Perspective: Advanced particle imaging

Abstract

Since the first ion imaging experiment [D. W. Chandler and P. L. Houston, J. Chem. Phys. 87, 1445–1447 (1987)], demonstrating the capability of collecting an image of the photofragments from a unimolecular dissociation event and analyzing that image to obtain the three-dimensional velocity distribution of the fragments, the efficacy and breadth of application of the ion imaging technique have continued to improve and grow. With the addition of velocity mapping, ion/electron centroiding, and slice imaging techniques, the versatility and velocity resolution have been unmatched. Recent improvements in molecular beam, laser, sensor, and computer technology are allowing even more advanced particle imaging experiments, and eventually we can expect multi-mass imaging with co-variance and full coincidence capability on a single shot basis with repetition rates in the kilohertz range. This progress should further enable “complete” experiments—the holy grail of molecular dynamics—where all quantum numbers of reactants and products of a bimolecular scattering event are fully determined and even under our control.

INTRODUCTION

Gravity waves, the game of billiards, and the Large Hadron Collider (LHC) have one feature in common: they are all examples of scattering phenomena. In each case, two objects collide and “products” emanate from the collision center. In the case of the billiards, we usually can see how the scattering collision happens. However, the collisions of black holes are too far away to see, and the collisions of protons are too small, but in each case the products—gravity waves or Higgs bosons—allow us to determine the nature of the interaction.

In chemical physics, as well, scattering experiments have played an important role in uncovering the nature of molecular interactions. The 1986 Nobel Prize in Chemistry to Dudley Herschbach, Yuan Lee, and John Polanyi was, in part, awarded for such experiments. As in the case of gravity waves and the LHC, new detector designs have been paramount in advancing the scattering technique for chemical physics. In this Perspective, we consider advanced techniques for the detection of the correlated speed, angular, and internal energy distributions of products of molecular scattering experiments. Relevant work to this special issue is cited, but in no way are the references complete. A full review of particle imaging techniques with recent applications can be found in Ref. 1.

The need for correlated information stems from our desire to understand the scattering process in as much detail as possible; it is often the case that the properties of one product and those of another are intimately intertwined. For example, one product may have large rotational energy and low vibrational or translational energy, while the other product may have high vibrational energy and low rotational or translational energy. The correlation between properties is often the information we need to narrow the range of possible interaction mechanisms. Until fifty or so years ago, it was difficult to determine the range of even a single property of a single product; but with the advent of lasers and new detection techniques, it is now possible not only to determine property ranges but also often to correlate multiple properties for multiple products. Here we briefly trace the history of how these capabilities became available, beginning with the first ion imaging experiment in 1987,² as well as the wide range of applications that are now being used and developed.

The properties on which we will focus are the speed, direction, and alignment of the products; the division between vibrational, rotational, and translational energies in the products; and the orientation or alignment of any vector quantities of the products, for example, the relationship between the rotational vector of one product molecule and its velocity or the relationship between the rotational vector of one product molecule and that of another as well as the correlation between the relative velocity vector of the collision, the outgoing velocity vector of the product, and the angular momentum vector of the product. Many of the newly developed techniques were first applied to photodissociation experiments. In these experiments, the incoming particles are a photon and a molecule, while the outgoing products are typically two atomic or molecular fragments.

ANTECEDENTS TO PRODUCT IMAGING

Photodissociation experiments were early examples of scattering experiments, and they underwent a significant step in their development in 1969. Almost simultaneously, two groups announced the development of machines combining a molecular beam of a precursor with a pulsed, polarized laser beam of photons.^3,4 The product fragments were detected by a fixed mass spectrometer along the flight direction perpendicular to the mutually orthogonal molecular and laser beams. For example, one study⁵ concentrated on the photodissociation of IBr and I₂. The velocity of the iodine product was determined from the distance and flight time between the time the laser pulse crossed the molecular beam and the arrival at the mass spectrometer of the iodine fragment. The conservation of energy and linear momentum was then used to determine if the iodine atom was in its electronic ground state or its first excited state. Additionally, the signal intensity at the mass spectrometer was measured as a function of the angle between the polarization vector of the laser and the flight direction. As first discussed by Zare and Herschbach,^6,7 the polarization of the light field aligns the absorbing molecule, so that a prompt dissociation (with respect to rotation) leads to an anisotropic product distribution in the laboratory frame. In the case of I₂ dissociation at 531 nm, iodine was found to dissociate predominantly along the direction of the polarization, indicating that the absorption process was due to a parallel dipole transition.

A second advance in detection came in the late 1970s when Phillips et al.⁸ first used Doppler spectroscopy to measure the angular scattering of Ar from Na(³P_1/2) to yield Ar + Na(³P_3/2). Light emitted from or absorbed by a moving atom or molecule is shifted slightly either to the blue or red depending on whether it is traveling toward or away from the observed. By resolving the frequency shift of the light, the authors were able to determine the angle between the direction of the scattered particle and the relative velocity vector. A similar technique was soon after used by Schmiedl et al.⁹ to study the photodissociation of HI. Like the photodissociation of I₂, the photodissociation of HI produces iodine atoms in either the excited or ground electronic state, in this case with different angular distributions with respect to the polarization direction of the dissociating light. By resolving the absorption frequency shift of the laser-induced fluorescence spectrum of the H atom product, the authors were able to determine the angular distributions of the H atoms that were produced in coincidence with either the ground or excited state of the I atom. The Doppler profile is essentially a one-dimensional projection of the three-dimensional velocity distribution. Application of the Doppler technique to molecular products followed rapidly, and led to the discovery of vector correlations between, for example, the direction of a diatomic product’s rotational vector and the relative recoil velocity vector.^10,11 A further development was made by Hall et al.¹² and Loo et al.,¹³ whose studies demonstrated that a Wiley-McLaren time-of-flight spectrometer could be used to create a one-dimensional projection of the three-dimensional velocity distribution of an ionized photodissociation product.

The stage was now nearly set for the introduction of the technique on which this Perspective is focused, but there are a few more inspirations to mention. In 1967, Solomon and co-workers showed that it was possible to visualize the direction of photodissociation products using a hemispherical bulb coated on the interior with tellurium.^14,15 When bromine or iodine at low enough pressure was dissociated with polarized light, the resulting atoms reached the walls of the bulb without collision and then etched the tellurium, resulting in an anisotropic pattern. This was the first time that researchers had “seen with their own eyes” the anisotropic nature of the photodissociation. In an experiment by Yates and his colleagues in 1974, ion desorption due to electron collisions with a molecules in a surface were imaged and visualized. The ions were visualized by acceleration of the ions through a hemispherical grid onto a microchannel plate amplifier backed by a phosphor screen and the image photographed. The resulting image was called an electron stimulated ion desorption ion angular distributions or ESDIAD.¹⁶ At low surface temperature, peaks in the angular distribution of deuterium ions from methyl groups in methyl acetate on the surface were observed.

PRODUCT IMAGING

In 1986, a Faraday Discussion on the dynamics of photodissociation was held in Bristol, where Paul Houston spoke on his group’s use of laser multiphoton ionization and time-of-flight ion detection to record the equivalent of Doppler profiles, essentially one-dimensional projections of the angular distribution of photodissociation products.¹² During the talk, it occurred to Dave Chandler that the combination of time-of-flight mass spectrometry and detection of charged products using microchannel plates and an fluorescent screen could allow one to “see” the products; he even drew a rough sketch of an apparatus during the talk. Further discussions ensued over the liquid refreshment of the sort for which the British are justly famous. While the Houston group had thought briefly about “two-dimensional” projections, they did not have the resources to pursue them. Nor was it immediately clear to them that the two-dimensional projection could be unraveled to provide more information than the one-dimensional projection. Nevertheless, the prospect of “seeing” the products motivated further work. The Chandler group had access to an under-used ESDIAD apparatus, and the two researchers agreed to spend time together at Sandia National Laboratories to attempt an experiment.

The first experiment, photodissociation of methyl iodide, worked extremely well after only a few days’ time. Figure 1 shows a photograph of the very first data.¹ A molecular beam of methyl iodide is dissociated with the pulsed, quadrupled output of a Nd:YAG laser at 266 nm. Methyl radicals in their ground vibrational level are ionized using a multiphoton process at 330 nm, a wavelength produced by a pulsed, double dye laser. Following ionization, the methyl fragments are accelerated toward the microchannel plates by an electric field; but during the flight time, they continue to expand away from their moving center of production. By the time they reach the microchannel plates, the circumference of their velocity sphere is of macroscopic dimension. At the microchannel plates, each impacting ion is converted to electrons, which are amplified by a cascade process and then accelerated to a fluorescent screen. In the first experiments, the image could clearly be seen by eye, and the screen was simply photographed with a polaroid film (see Figure 1 below). A later improvement used a CCD camera to record and average the results from each laser shot. The polarization of the dissociation laser is in the vertical direction relative to the photograph shown in Fig. 1, and it is clear that the methyl fragments fly predominantly to the north or south poles, parallel to the polarization as expected from the dominant transition moment.

FIG. 1.

Schematic drawings and first images obtained using particle imaging. On the top row, left, is the drawing of the expected three-dimensional distribution from an instantaneous dissociation for a parallel transition to a repulsive state, as is the case for the methyl iodide, CH₃I, dissociation. Top row middle is a drawing of the expected two-dimensional projection of the three-dimensional velocity distribution. Top row right is an artist’s rendition of the inverse Abel transformed distribution obtained from analyzing the 2-dimensional projection. The bottom row shows the black and white raw data (left), the digitized and false colored raw data (middle), and the inverse Abel transform of the image (right) of the CH₃I dissociation at 266 nm with the detection of CH₃ by laser ionization at 330 nm.

Advantages of this technique soon became apparent. First, the image is the two dimensional projection of the full, three-dimensional velocity distribution, correlating both speed and angular information about the ionized fragment. Second, the multiphoton ionization technique, when applied to molecules, can typically be used to ionize a single vibrational/rotational state of the product. Assuming that the energy of the photon and the internal energy of the parent molecule (here, methyl iodide) are known, conservation of energy and linear momentum can be used to calculate the translational energy of the unobserved product (here, the iodine atom). In the methyl iodide case, it was clear from the data that the iodine atoms were nearly all produced in their first excited (spin-orbit) state. Finally, it turned out that those who study image analysis had already demonstrated that, when the three-dimensional velocity distribution is cylindrically symmetric about an axis perpendicular to the projection direction (as it is here for methyl iodide), the full three-dimensional distribution can be constructed from a single two-dimension projection by applying an inverse Abel transform.^17,18 Several methods for performing this transform have since been developed, as reviewed in the introductions to two recent papers proposing alternative methods for extracting the three dimensional velocity distribution from the image without resorting to an inverse Abel transform.^19,20

Further application of product imaging to photodissociation quickly followed, mostly from the Chandler group and others associated with it. Many of these were reviewed in 1995.²¹ The first applications were to further study the dissociation of methyl iodide and its deuterated derivative^22–25 as well as H₂S,²² but further work was performed on methyl bromide,²⁶ methane,²⁷ acetylene,²⁸ and van der Waals molecules.²⁹ At the same time, product imaging was used for the first time to study bimolecular reactions, for example, H + HI³⁰ and H + D₂,³¹ and to analyze crossed-beam studies of an energy transfer process, Ar + NO(v = 0, J = 0) → Ar + NO(v = 0, J).³²

Another important extension of the imaging technique was to detect the speed and angular distribution of electrons. Helm et al. were among the first to image electrons.³³ By using a similar imaging system but reversing the polarity, they imaged photoelectrons produced in the multiphoton ionization of xenon. Blondel et al. soon followed by imaging the photodetachment of Br⁻³⁴ and O⁻.³⁵

LIMITATIONS OF PRODUCT IMAGING

Although at this stage it was clear that the product imaging technique had found a permanent home in chemical physics, its limitations were also becoming evident. First, the inverse Abel transform was unstable and introduced noise along the axis of symmetry. Might there be a way to conduct the experiment so that the full, three-dimensional velocity distribution of the state-selected product could be obtained without the transform? Second, even when a CCD camera was used to record and average the image, it was clear that a single event would “light up” more than one pixel. This was partly due to the limitations of channel plate amplifiers, but the loss of resolution was real. Might it be possible to avoid this loss? A third limitation was the limit in resolution caused by the finite volume of overlap between, for example, the molecular beam and the laser beam. Because the image was simply a velocity dependent spatial expansion of this overlap region, the spatial resolution could never be greater than the ratio of the size of the laser-molecular beam overlap to the size of the final image. In the decade following the mid-1990s, these and other limitations were confronted and reduced. Gebhardt et al.³⁶ introduced a “sliced imaging” technique that allowed determination of the three dimensional velocity distribution without the need for an inverse Abel transform. By allowing the Newton sphere to expand as it traveled down the time of flight detector and gating the detector to only see a part of the Newton sphere, one was able to record a slice of the Newton sphere instead of a crush of the Newton sphere. This offers some clear advantages in some experiments especially when the data do not contain cylindrical symmetry. Townsend et al.³⁷ and Lin et al.³⁸ created variations on the slice imaging technique and increased its resolution and application. Chang et al.³⁹ introduced pixel centroiding of each event for each laser shot, greatly improving the resolution. But most important was the technique of “velocity map imaging,” described in the section titled Velocity and spatial map imaging—Technical aspects.

VELOCITY AND SPATIAL MAP IMAGING—TECHNICAL ASPECTS

Velocity Map Imaging,⁴⁰ VMI, is a high resolution variant of ion imaging born in Nijmegen in 1996 out of frustration with the fine metal grids used in time-of-flight mass spectrometry to create homogenous electric fields. Ion images of heavy metal atom photoproducts (specifically, Hg from (CH₃)₂Hg) were found to be strongly “diced” by such grids. In an act of desperation, the grids were removed, leaving open annular rings for the extractor and ground electrodes to transmit the ions, with the hope that an image could still be formed. Serendipity also played a role because at the time laser ionization studies of Rydberg states of O₂ were taking place. While the Hg⁺ images obtained using this open lens approach seemed to improve slightly, the results for O₂⁴¹ were spectacular. The O₂⁺ image from O₂ “parent” molecules in the molecular beam collapsed at the correct or VMI extractor/repeller ratio to almost a single pixel on the detector. At this VMI ratio, O⁺ images from O₂ photoionization/dissociation came into sharp focus, revealing up to 30 concentric rings. Tuning the voltages on the extractor or the repeller plates of the ion lens, thereby moving the voltage ratio above or below the VMI ratio, showed a stripe clearly related to the spatial height and width of the O₂⁺ ions being formed. The so-called “spatial map image” of the O₂ molecular beam profile with a magnification factor of ∼10 was created by displacing the ionization laser stepwise in space. Nowadays, spatial map imaging (SMI) is extremely useful in ensuring the proper alignment of the laser and molecular beams along the imaging TOF axis, and Stei et al.⁴² have investigated several three-dimensional and two-dimensional variants of SMI with applications in characterizing the profile of the focused laser beam with 2 μm accuracy. Analytical applications of SMI for biomedical imaging and mass spectrometry have also been reported, for example, by the Oxford group,^43,44 and Lee et al.⁴⁵ described an alternative approach to SMI in this issue.

The great impact of velocity mapped ion imaging in molecular dynamics research is due to its capacity to map the velocity coordinates of points in phase space onto the imaging detector while suppressing the influence of the spatial coordinates. The shaped field lines provided by the electrostatic immersion lens used in VMI help to strongly reduce the blurring of ion or electron images that occurs in ion imaging experiments using homogeneous fields due to the finite source volume. In practice, the resolution improvements were convincing enough that velocity mapped ion imaging became widely adapted across the world, with close to 2000 papers now citing the original 1997 publication. With VMI, all of the parallel improvements in imaging could be put to good use, for example; delay line detectors with high spatial resolution and ultra-fast electronics can now routinely capture full three-dimensional Newton spheres for multiple masses, and new “active pixel” CMOS sensors⁴⁶ should soon show similar abilities. In a wider perspective, imaging directly reveals the differential state-to-state scattering cross section, i.e., the “Newton sphere” which is the basis for describing the physical and chemical change on the microscopic level. This powerful tool has changed the way chemical physicists visualize the molecular structure and dynamics processes.

No tool is perfect, however, and VMI has its own restrictions. In the first VMI publication,⁴¹ for example, parent O₂ molecules were seeded in a supersonic expansion, which reduces the transverse velocity spread down to only a few K. Even though the O₂⁺ ions are formed in a 3 mm long stripe through the molecular beam, they should velocity map to almost a single point on the detector. “Almost” is an important qualification built into the concept of ion imaging: the nascent velocity distribution of a neutral species must be converted without change to an ionic velocity distribution, typically by using state-specific laser resonance enhanced multi-photon ionization (REMPI). Along with the velocity spread due to the angular divergence of the skimmed molecular beam (typically up to 30 m/s), any extra velocity imparted to the ion by excess energy—the so-called “ion recoil”⁴⁷ (up to 200 m/s)—in the resonance enhanced multiphoton ionization (REMPI) will also be velocity mapped onto the nascent particle image.

In order to achieve the best possible velocity resolution⁴⁸ in any VMI experiment, it is thus necessary to limit the angular spread of the parent molecular beam(s), to perform state-specific threshold ionization to limit ion recoil, and furthermore, to avoid distortion due to Coulomb repulsion by limiting the number of ions formed in the source volume. The last restriction can be very annoying—for the first VMI experiment on O⁺ from O₂ mentioned above, a properly filled-in image requires >10 × 10⁶ events (i.e., O⁺ ions), which at one ion (any ion, not just O⁺) per laser shot at 10 Hz requires 277 h of averaging. This highlights one of the big advantages to VMI using a two-dimensional detector and a camera over using a time and position sensitive particle detector that can only detect one or a few particle hits per laser shot. With the camera system, many tens of ions or electrons originating from a single laser shot can be detected without inducing significant distortion to the velocity distribution due to charge repulsion, and the O⁺ image can be recorded in 15 minutes.

If the above requirements are met, imaging resolution can be further improved by using specially shaped electrodes⁴⁹ combined with “slicing”^36–38 or with optical slicing⁵⁰ using the simple 3-plate lens used in the original VMI setup. Since the inception of VMI, electrostatic lens designs have evolved to approach the space and velocity focusing capabilities inherent in the classic Wiley-McLaren optics for TOF-MS, to velocity map very high energy particles with kinetic energy release (KER) > 10 eV or to properly map and analyze low velocity products⁵¹ such as inelastically scattered cold molecules. Extraction of the three-dimensional information by inversion of the acquired 2D image has also received the necessary attention, with the BASEX method from Dbrinsiki et al.⁵² and its polar variant from Garcia et al.⁵³ now being the most widely adapted. New techniques for extracting the three-dimensional velocity distribution from the two dimensional projection without the use of a transform but with an optimized fitting procedure are also being developed and adopted.¹⁹

STATE-OF-THE-ART AND ULTIMATE IMAGING EXPERIMENTS

The ultimate unimolecular photodissociation experiment measures the quantum state, velocity, and alignment of the product to be imaged in a manner that provides correlated information about the product fragment that is not being ionized and observed. If one knows exactly the amount of energy deposited in the original molecule then the conservation of energy and momentum will provide information about the quantum states populated in correlation with the particle that was imaged. With fast lasers, it is even possible to intercept the molecule as it is falling apart, and to learn about the path atoms take on their way to dissociation. In addition, there are many instances where extra information is obtained by co-incidence measurements where multiple (typically 2 but not necessarily) fragments are detected from a single molecular dissociation event. For instance, by measuring the electron and the cation in a dissociative ionization event, one can determine the ejection direction of the electron relative to the dissociation axis thereby obtaining the molecular frame photoelectron image.⁵⁴ This capability of coincidence imaging is particularly important for experiments where femtosecond lasers are used to dissociate a molecule and ionize the products. This is because the conservation of energy and single quantum state resonance enhancement in the ionization are generally not possible due to the frequency bandwidth of the laser pulses.

To fully understand molecular scattering processes, VMI can indeed reveal the correlated speed, angular, internal energy distributions, and alignment and orientation of nascent products. However, the initial conditions of the scattering experiment must also be fully controlled. Ideally, this means producing a sample of molecules in a single (and variable) internal energy quantum state—including the direction of angular momentum and/or molecular axis in space—moving at a single (and variable) velocity, and interacting with an equally well-characterized scattering partner. Considerable progress is being made in this area by several groups using the hexapole state selection to select a single quantum state that can be oriented to the collision,^55,56 using a Stark decelerator that is also a state selector,^57,58 or using the laser preparation of a molecule in an excited state that is aligned by the laser to the relative velocity vector of the collision.^59–63

The applications of velocity map ion/electron imaging are wide ranging and often at the cutting edge of science. The latter claim is supported by the thousands of refereed papers that use VMI, with dozens published in the journals Science or Nature. The wide scope of VMI experiments includes studies ranging from high energy to ultra-cold molecule collisions, from slow to ultra-fast electron dynamics, using radiation from THz to extreme ultra-violet, with resolved product energy release from tens of eV down to meV and lower. For purposes of organization and to demonstrate the breadth of experiments presently being performed using these advanced imaging techniques, we group measurements that involve forming an image, usually of a Newton sphere, by projecting charged particles onto a two-dimensional (x,y) or three-dimensional (x,y,t) detector using VMI optics. This approach has found applications in a very wide range of topics concerning molecular dynamics and structure, and analytical chemistry. To put the special issue contributions into perspective, we can divide the field at present into two camps. In general, those using nanosecond lasers can utilize single-quantum-state preparation and detection and thereby utilize high resolution imaging and conservation of energy and momentum for the analysis of their process. In general, those using fast laser pulses give up this single-state preparation and detection in order to measure dynamics and the evolution of quantum wave packets. The ultimate example of giving up any sort of state selection in the detection of products is the growing number of coulomb-explosion imaging studies where high intensity lasers strip a molecule of many electrons and the recoil of the resulting ions is used to reconstruct the structure of the molecule at the moment of the explosion.^64–66

A perfect understanding of the bimolecular processes or photo-initiated processes is slowly progressing from simple diatomic molecules towards more complex systems. Imaging experiments on photodissociation, photoionization, and scattering are also slowly growing in the step-wise complexity from the generic diatomic molecule AB, to 1:1 van der Waals clusters of AB-X, where X is often a rare gas atom, to large clusters such as (AB)_n, and further to molecules on or in liquid helium droplets, LHeD-AB, and, most recently, to molecules adsorbed onto or scattered from surfaces such as AB(solid).

QUANTUM STATE RESOLVED IMAGING EXPERIMENTS

Consider first bimolecular collision or photo-initiated half-collision processes involving molecule AB prepared in a single initial state such as the generic AB + hν → A + B, where the product of interest, A or B, is a neutral atom or molecule. It will need to be ionized to be projected and detected, typically by state-selective REMPI: A+ nhν → A⁺ + e⁻. REMPI requires a narrow bandwidth and high intensity, meaning low repetition rate pulsed lasers that can deliver exquisitely the detailed information on the quantum state level for final state products. Examples of processes that produce neutral molecules that require resonance enhancement ionization for detection include the following:

• 1.Photodissociation: AB + hν → A + B, A+ nhν → A⁺ + e⁻.
• 2.Energy transfer: AB(v, J) + X → AB(v′, J′) + X, AB + nhν → AB⁺ + e⁻.
• 3.Chemical reactions: AB + C → AC + B, AC + nhν → AC⁺ + e⁻.

By utilizing not only the frequency of the laser but the polarization information, one can learn about the alignment or orientation of the product atom or molecule that is being detected.^67–71 It should be noted that it is not always necessary to perform resonance enhanced ionization to learn about many of these processes, and in some cases, it is actually preferable to not have resonance enhancement. For example, when one is studying reactions such as the Cl atom with ethane to form HCl and the ethyl radical, a single photon ionization scheme that ionizes all the ethyl radicals without fragmentation would be preferable as the spectroscopy of a hot ethyl radical would make it quite difficult to perform single-quantum-state-resolution detection. In some cases, the information one generally wants to know is best obtained from the angular distribution of all of the products. Often, a 157-nm excimer laser has been used for such studies^72,73 or single-photon ionization utilizing synchrotrons.

There are several types of experiments that require high frequency resolution light for the selective generation of a charged particle (as opposed to turning the neutral product into an ion after the fact) where the velocity of that charged particle provides information about the process. Examples are anion photo-detachment and cation photodissociation. When electrons are imaged from photo-detachment processes and the light energy is near the ion threshold energy slow electrons are generated and these have extremely high velocity resolution. This technique is termed Slow Electron Velocity Imaging or SEVI.

• 4.Photo-detachment, SEVI: AB⁻ + hν → AB + e⁻.
• 5.Multi-photon ionization: AB + nhν → AB⁺ + e⁻.
• 6.Dissociative ionization: AB + nhν → A⁺ + B + e⁻.
• 7.Combinations of the above: AB⁻ + nhν → A⁺ + B + 2e⁻.

In particular, in processes where one sees dissociation ionization, the measurement of the velocity can, in many experiments, distinguish between ionization followed by dissociation and dissociation followed by ionization to obtain the same fragment ion. For example, both processes were observed and quantified in the H⁺ formation from the 532-nm irradiation of H₂.⁷⁴

FEMTOSECOND TIME SCALE IMAGING EXPERIMENTS

When utilizing femtosecond lasers for the preparation of a molecule and the ionization of the products, one generally creates a time varying wavepacket by exciting many quantum states simultaneously and detection is not in general quantum state specific as the bandwidth of the laser precludes this. However much can be learned by utilizing the time dependence of the formation of the charged particles. For instance, many groups now combine the femtosecond excitation of an excited state of a molecule with time-resolved photoelectron spectroscopy (TRPES). By using TRPES one can observe directly non-adiabatic dynamics in excited electronic states of large molecules. As the energy flows from a singlet state to a triplet state the dynamics is reflected in the energy of the observed photoelectrons as the ionization spectra is strongly dependent on the nature of the electronic states in the neutral molecule.^75–78 This is one example of time-resolved, pump-probe (TRPP) experiments where the time-delay photodissociation or photoionization of an electronically excited state is used to follow reactive dynamics.

When the product of interest is already charged, for example, by photoionization, AB + hν → AB⁺ + e⁻, the final state information for AB⁺ is coded directly in the electron image. Because both partners are charged, coincidence imaging is possible to eliminate the signal from background or secondary processes. In this case, low count rates mean that high-repetition rate lasers are necessary. Photoelectron Circular Dichroism (PECD)^79–81 is a beautiful and powerful example of the advantages of coincidence imaging.

• 8.Coincidence photoionization: AB + hν → AB⁺ + e⁻.
• 9.TRPES photoionization: AB + hν → AB^* …AB^* + hν → AB⁺ + e⁻.
• 10.Anion photodissociation: AB⁻ + hν → A⁻ + B …A⁻ + nhν → A + e⁻ + nhν → A⁺ + e⁻.

Once again, for processes such as anion photodissociation,^82,83 one needs to distinguish between photo-detachment followed by ion dissociation or ionization and dissociation followed by photo-detachment. The measurement of coincidence fragments is needed to resolve the many possible channels.

IMAGING EXPERIMENTS WITHOUT LASERS

A long-term goal of the imaging field has been to apply the advantages of VMI and SMI to analytical chemistry, where expensive lasers would be replaced by inexpensive electron or metastable species from, e.g., discharge sources. Much can be learned from these non-state selective ionization processes. As mentioned earlier, single photon ionization is a very desirable process to use for detection when one is studying reactions that produce molecules whose spectroscopy is difficult. Several groups have adapted the VMI apparatus for the study of electron scattering, electron ionization, and dissociative electron attachment. A dissociative electron attachment in particular has shown to produce the interesting velocity distribution of product anions as the electrons only attach to the molecules if they are aligned appropriately to the electron beam and they dissociate faster than the molecule can vibrate or rotate producing images that provide information about the electron attachment process.^84,85 Penning ionization⁸⁶ is another soft ionization process that has recently been used along with the imaging, and the imaging of ion molecule reactions have had spectacular success in confirming basic chemical concepts such as SN₂ reactions.⁸⁷

• 11.Electron impact ionization: AB + e⁻ → AB⁺ + 2e⁻.
• 12.Dissociative electron attachment: AB + e⁻ → A + B⁺ + e⁻.
• 13.Penning Ionization: Rg* + AB → AB⁺ + e⁻ + Rg.
• 14.Ion molecule reactions: AB + C⁻ → AC + B⁻.

Many of the processes highlighted above are represented in this special issue of JCP.

CONCLUSION

The authors of this Perspective hope to convey the broad impact and cutting edge nature of the applications that imaging has brought to the fields of chemistry and physics. It all started with a conversation at a scientific conference and a few beers in Bristol where ion imaging was conceived, it was made possible in 1987 by combining the right tools at Sandia, and it took off in 1997 with the discovery of velocity mapping in Nijmegen, but it is the open sharing of ideas and techniques by the many imaging scientists that has brought the field so far. These days, critical developments in lasers, electronics, molecular beams, and sensors are still rapidly progressing, which promises increasingly more beautiful images and broader, deeper insights into physical and chemical changes in the coming years. This Perspective was not meant to be a review of the field but to highlight the sort of experiments that can and are being done using advanced particle imaging techniques.