| People involved: |
| Former members: | Jens Uhlig, Wilfred Fullagar, Sophie Canton, Niklas Gador |
| Involved facilities: | Synchrotron facilities |
This technique has the following projects (and possibly other techniques) related to it:
The most prominent techniques for detailed molecular structural analysis at the present time are based on short wavelength scattering (including various forms of crystallography, diffuse scattering, reflectometry and EXAFS) and nuclear magnetic resonance (NMR). It hardly needs to be said that classical synthetic and physical chemical approaches had also established many details of numerous rather complex structures long before any of these techniques had been thought of. Of these, single crystal crystallography and NMR are the techniques best able to obtain extensive and detailed structural information from highly complex molecules. NMR is not adaptable to ultrafast measurements because of relatively long interaction times associated with the corresponding radiofrequency resonances. Diffraction approaches, however, are capable of ultrafast measurements, depending on the velocity of the radiation in relation to the sample depth needed for effective scattering. Thus thermal neutron scattering is not easily adaptable to ultrafast measurements, while electron and X-ray scattering can be.
X-rays are well known for their penetrating power. A single atom, for example, is simply unable to scatter X-ray photons to an experimentally significant extent. Partly for this reason, atomic line gratings would be of limited practical use at X-ray energies, even if other complications could be overcome. On the other hand, there are enough atoms even in small crystals that X-ray scattering becomes practically efficient. Here, the relevant "grating" does not consist of lines, rather parallel planes. X-rays can only be diffracted (specularly) by these planes if the Bragg relationship is fulfilled (nl = dsinq). Given a particular d-spacing, the wavelength and crystal orientation must mutually satisfy Bragg's formula for diffraction to occur. If the orientation is slightly wrong, a monochromatic source simply will not be diffracted. Thus, when observing reflections from crystals in the context of monochromatic structure refinements, it is necessary to mechanically rotate the crystal during the exposure. However, in ultrafast measurements it is impossible to rotate the crystal fast enough!
The Ewald sphere construction (described elsewhere, eg Wikipedia) gives a pictorial account of the situation. Some understanding of the reciprocal lattice is also assumed in what follows (again, Wikipedia, etc).
To see a reasonable number of diffraction spots in one orientation of the crystal, the spacing of the reciprocal lattice should be moderately small compared to the nest of Ewald spheres, though not so small that detected reflections overlap. The X-ray beam's energy range and divergence, as well as crystal mosaicity and detector resolution are also considerations. The detector should subtend large solid angles around the crystal. There is generally a rather good chance of seeing at least one of whatever symmetry-equivalent high order reflections are present (favoured by high crystal symmetry). It will also be noticed in the sketch above that low order reflections have a rather low chance of being seen. In practice this is alleviated slightly by some focusing (non-collimation) of the beam, however the relative invisibility of low order reflections (which contain much of the gross structure) can sometimes present difficulties.
In the ground state structure, once the intensity of a sufficient number of reflections has been determined a phase must be assigned to each reflection such that the relative positions of different atoms can be deduced. By knowing the ground state structure and observing changes, the excited state structure can be inferred. This makes it possible to observe molecular structure changes in crystallographic detail, on timescales limited by the duration of the X-ray pulse. To date it has only been possible to achieve suitable pulse characteristics at a handful of synchrotrons, where the pulse duration is on the order of ~100 ps.
This time resolution is four orders of magnitude slower than the collimated, broadband betatron X-rayradiation pulses that can be produced by laser-based electron acceleration schemes in jitter-free arrangements, and three orders of magnitude slower than the broadband collimated pulses expected from the short pulse X-ray source at the proposed MAX-IV facility. The likelihood that both these approaches can be pursued locally accounts for our particular interest in these sources.
Needless to say, pump-probe Laue crystallography is not trivial. The need to generate high levels of excitation in a crystal means that a given crystal will not last many laser shots. Stroboscopic experiments are therefore impossible, meaning that each X-ray pulse needs to be suitably intense. A reasonable sampling of reciprocal space is desirable for each crystal before it "dies". From a sample perspective, the concentration of the chromophore must be fairly low, since energy is largely deposited as heat in non-chromophoric parts of the crystal (in the case of fluorescent molecules, the bluer and more efficient the fluorescence, the better). Of course it needs to be a crystal that can be reproducibly grown, at a size appropriate for the available experimental infrastructure. A reasonably high symmetry can be desirable, to reduce the volume of reciprocal space that needs to be interrogated. The crystal has to intercept the X-ray beam, meaning that the X-ray beam typically requires some focusing. (Focusing is a nontrivial exercise for X-rays, since critical reflection angles are very shallow, typically requiring long and potentially very expensive grazing incidence mirrors. Various compact broadband options such as lobster eye lenses and various mono- and poly-capillary devices are also worth consideration.) A good ground state structure for the crystal is of course mandatory, though this is often known and otherwise may not present much difficulty. The source should not provide more than one octave of energy, since higher order reflections increasingly lead to diffraction at the same angles. Spots on the (image) detector must not overlap, which sets limits on the relative size of the wavelength vs the lattice dimensions, as well as on the source bandwidth and divergence, and potentially also the detector's spatial resolution. Several software suites exist for the analysis of Laue data, both at ultrafast X-ray and neutron scattering facilities. Such experimental matters need to be borne in mind as collimated polychromatic ultrafast sources begin to enter the scenario.
Despite any such complications, the setup required by a Laue crystallography experiment at a laser source is conceptually extremely simple, illustrated below. Pulse isolation choppers, heat load choppers and jitter management procedures are completely unnecessary on such laser-based sources, where avaliable laser powers are such that sample excitation is largely a matter of sample preservation. Here a relatively simple betatron radiation laser-generated beam is suggested, though more sophisticated X-ray generation procedures are possible (given the production of a high energy electron beam, laser-based means of accelerating it to yet higher energies, and many further mechanisms for X-ray generation alluded to in our Time Resolved Measurements pages). The electron beam is magnetically deflected aside, and associated colinear laser beam reflected aside with thin foils, so as not to interfere with the diffraction experiment. The extreme laser intensities up to the point of X-ray generation require vacuum systems, though once generated, the X-rays could perhaps be more easily dealt with after transmission into a helium chamber through a suitable X-ray window. The wraparound detector could be image plates in the first instance, though readout of these is rather cumbersome in practice and more convenient alternatives exist.
To our knowledge, nobody has successfully performed such a laser-based Laue experiment to date (we haven't, either!). Certain aspects of the engineering may be very laborious, but the rewards are incredible too: the full knowledge of photoinduced chemical reactions in real-time, often referred as "molecular movies". Differential maps of the electron density would allow charge transfer and excitation migration over the molecule to be directly seen, if data was sufficiently well resolved. Such knowledge would likely assist in the design of molecules for future applications.
A monochromatic analogue of the Laue approach might be to use a single crystal, but direct X-rays at it from a range of different angles simultaneously. Certain angles would then match the Bragg condition, producing diffraction in specific directions and amenable to analysis. Although this approach might have the advantage of substantially dealing with crystallographic "blind" zones, it appears that no suitable ultrafast X-ray source exists.
In crystalline powders all orientations are represented simultaneously. This corresponds to rotating the lattice around all possible axes passing through the origin such that they form a set of concentric spheres of different reflection intensity. These reflection rings now intersect the monochromatic Ewald sphere in a set of concentric rings, corresponding to the well known powder diffraction rings. The intensity of the corresponding single crystal reflections can be deduced. Since many reflections are monitored in parallel (given an appropriate detector), the need to rotate a single crystal in ultrafast measurements is avoided. Uniform and suitably intense stimulation of a sufficiently large number of necessarily small crystallites is not easy, however. A related approach (essentially involving a two dimensional powder) involves molecules at interfaces. These can have an in-plane crystallinity that allows in-plane scattering of grazing incidence monochromatic beams.
Another approach is to not use a crystal at all, but instead attend to the interatomic distance correlations in molecular liquids. The distribution of distances between atoms (pair distribution function) gives rise to diffuse scattering rings from a collimated, monochromatic beam. The scattering may be dominated by particular pairs of (heavy) atoms, making it possible to observe changes in their associated distances. It would also be possible to use a collimated polychromatic beam in this experiment if the energy of photons could be established afterwards. This may be possible by in-house approaches using the low flux but high repetition rate of highly collimated few-keV broadband X-ray beams generated by high harmonic generation procedures. Scattered radiation would be detected one photon at a time in an energy sensitive array detector. By knowing the energy of individual X-rays and their position, the necessary momentum transfers could be established. Counting statistics would build up at a rate essentially governed by the capabilities of the detector, assisted by the diffuseness of the spread of photons. Our detector pages give some allusion to this possibility. Given the diffuseness of scatter, energy resolution of individual photons would not need to be particularly great, meaning that the use of ordinary direct detection CCDs may be quite adequate.
General overview and motivation for X-ray based measurements
X-ray absorption spectroscopy
X-ray Diffraction and the argumentation for broad bandwidth (This page)
Overview and motivation for ultrafast X-ray measurements
Developments done on ultrafast X-ray sources in Lund
Developments on X-ray Detectors
Measurements done on ultrafast synchrotron user facilities
Steady state measurements done on synchrotrons