X-ray probes

Although x-rays probe inner rather than valence electrons, in light elements the chemical state of the emitting atom may affect inner-shell energies enough to be detected at high resolution. Thus the K d lines of sulfur at 0.537 nm shift by 0.3 pm between the oxidation states and. ... 

This chapter contains articles on six techniques that provide structural information on surfaces, interfeces, and thin films. They use X rays (X-ray diffraction, XRD, and Extended X-ray Absorption Fine-Structure, EXAFS), electrons (Low-Energy Electron Diffraction, LEED, and Reflection High-Energy Electron Diffraction, RHEED), or X rays in and electrons out (Surfece Extended X-ray Absorption Fine Structure, SEXAFS, and X-ray Photoelectron Diffraction, XPD). In their usual form, XRD and EXAFS are bulk methods, since X rays probe many microns deep, whereas the other techniques are surfece sensitive. There are, however, ways to make XRD and EXAFS much more surfece sensitive. For EXAFS this converts the technique into SEXAFS, which can have submonolayer sensitivity. 

An important specific feature of the present experiment is worth noting. The X-ray photons have energies that are several orders of magnitude larger than those of optical photons. The pump and probe processes thus evolve on different time scales and can be treated separately. It is convenient to start with the X-ray probing processes, and treat them by Maxwellian electrodynamics. The pumping processes are studied next using statistical mechanics of nonlinear optical processes. The electron number density n(r,t), supposed to be known in the first step, is actually calculated in this second step. 

One important use of X-ray probes is in the study of local order and displacements, but this is not within the scope of the present book. The recent availability in intense synchrotron sources with selectable X-ray energies permits high-precision measurements of chemically specific atomic-pair correlations in solid solution alloys. A recent review of the technique is given by G.E. Ice and C.J. Sparks (Modern Resonant X-ray studies of alloys local order and displacement) in Annual Reviews of Materials Science 1999, 29, 25-52. 

We shall concern ourselves here with the use of an X-ray probe as a surface analysis technique in X-ray photoelectron spectroscopy (XPS) also known as Electron Spectroscopy for Chemical Analysis (ESCA). High energy photons constitute the XPS probe, which are less damaging than an electron probe, therefore XPS is the favoured technique for the analysis of the surface chemistry of radiation sensitive materials. The X-ray probe has the disadvantage that, unlike an electron beam, it cannot be focussed to permit high spatial resolution imaging of the surface. 

X-ray probes for surface analysis are used in X-ray photoelectron spectroscopy (XPS), and examples are given of a wide range of applications of this technique in materials science. 

Time-resolved X-ray diffraction (TRXRD), illustrated in Fig. 3.1, provides a powerful technique to probe directly the structural dynamics of crystals far from the equilibrium. It employs visible pump pulses from a laser, and laser-or accelerator-based X-ray probe pulses [1, 3]. As X-ray diffraction can in principle probe k 0 phonons, TRXRD has the potential to reveal the energy transfer dynamics, for example, from the zone-center to the zone-boundary phonons. 

Fig. 3.1. Left visible pump/X-ray probe scheme for femtosecond TRXRD experiments. Hard X-ray pulses are generated by shining intense femtosecond laser pulses on a metal target (laser plasma X-ray source). Right geometrical structure factor of bismuth as a function of inter-atomic distance for diffraction from (111) and (222) lattice planes. From [1] and [2]...

We describe beamline ID09B at the European Synchrotron Radiation Facility (ESRF), a laboratory for optical pump and x-ray probe experiments to 100-picosecond resolution. The x-ray source is a narrow-band undulator, which can produce up to 1 x 1010 photons in one pulse. The 3% bandwidth of the undulator is sufficiently monochromatic for most diffraction experiments in liquids. A Ti sapphire femtosecond laser is used for reaction initiation. The laser mns at 896 Hz and the wavelength is tunable between 290-1160 nm. The doubled (400 nm) and tripled wavelength (267 nm) are also available. The x-ray repetition frequency from the synchrotron is reduced to 896 Hz by a chopper. The time delay can be varied from 0 ps to 1 ms, which makes it possible to follow structural processes occurring in a wide range of time scales in one experiment. 

The high intensity of x-ray beams from modern synchrotrons has made it possible to conduct optical pump and x-ray probe experiments to a time resolution of 100 ps, i.e. the limit given by the x-ray pulse length of a synchrotron. In these machines, the brightest... 

Research based on time-resolved XAS in an optical pump-x-ray probe scheme has first been implemented at synchrotron radiation sources. Mills et al. [2] used a 20 Hz repetition rate Nd YAG laser to photolyse carbonmonomyoglobin (MbCO) and monitor the photolysis product with time-resolved XAS around the K-edge of the iron atom. Other studies were carried out on different types of photolyzed systems in liquids, by Thiel et al. [3], Clozza et al. [4], Chance et al. [5,6] and Chen et al. [7,8,9]. All these studies were limited to the nanosecond or longer time domain. We recently reported on time-resolved XANES studies of a Ruthenium complex in water solution reaching the picosecond time scale [10]. This work allows us to evaluate the feasibility of future time-resolved XAS experiments, which we present below together with our new results. 

In the experiment, the transmission intensities for the excited and the dark sample are determined by the number of x-ray photons (/t) recorded on the detector behind the sample, and we typically accumulate for several pump-probe shots. In the absence of external noise sources the accuracy of such a measurement is governed by the shot noise distribution, which is given by Poisson statistics of the transmitted pulse intensity. Indeed, we have demonstrated that we can suppress the majority of electronic noise in experiment, which validates this rather idealistic treatment [13,14]. Applying the error propagation formula to eq. (1) then delivers the experimental noise of the measurement, and we can thus calculate the signal-to-noise ratio S/N as a function of the input parameters. Most important is hereby the sample concentration nsam at the chosen sample thickness d. Via the occasionally very different absorption cross sections in the optical (pump) and the x-ray (probe) domains it will determine the fraction of excited state species as a function of laser fluence. 

Regardless of the choice of the sample thickness, the total amount of sample particles in the x-ray probe beam under optimized conditions is directly proportional to the x-ray spot size and inversely proportional to the x-ray absorption cross section, whose photoinduced (small) changes we want to measure [12]. Typical x-ray foci at synchrotrons are in the 0.1 - 0.3 mm range. For the examples treated below, this means that we have between 1014 and 1016 molecules in the probed volume. In order to achieve a reasonable photoinduced signal we should excite as many solute molecules as possible. Neglecting the optical absorption cross sections for photoexcitation for the moment, this requires on the order of 1015 laser photons per pulse, or ca. 0.25 mJ of pulse energy (e.g., at 800 nm). In other words, one should aim to... 

Figure 4b shows the measured transient difference absorption spectrum as a function of the x-ray probe energy E, recorded 50 ps after laser excitation (data points with error bars) for a sample containing 80 mmol/1 solution of [Ru"(bpy)3]2+ in H2O. This transient contains all the electronic changes from the reactant state absorption spectrum, R E), to the product state absorption spectrum, P(E,t), at the time t after photoexcitation. WithXO being the fraction of excited state species at time t, the transient absorption spectrum T(E,t) is given by... 

Fig. 5. Scans of the pump-probe time delay between the laser pump pulse and the x-ray probe of the transient X-ray absorption signal of [Run(bpy)3]2+ (the x-ray energy is set at the minimum of the signal in fig. 20.b) over several tens of nsec (a) and in the psec time domain (b). The latter represents the cross-correlation signal between the 200 fs laser pulse and the 70-80 ps long x-ray pulse (indicated by the Gaussian-shaped derivative of the fit curve to the data).

This work demonstrated a number of new results and opportunities for ultrafast XAS 1) it is possible to work with highly dilute solutions in transmission mode without dramatic loss of signal-to-noise ratio. This is very promising as one can envision the study of samples, for which large concentrations are impossible to reach. Biologically relevant samples are usually investigated in solutions with up to 1 mmol/1 concentration, and we therefore can envision such studies on the ultrafast time scales in the near future. 2) It is possible to scan the time delay between the laser pump pulse and the x-ray probe pulse, and therefore follow the evolution of the system from the start. 3) It also demonstrated the operation of an optical-x-ray cross-correlator (Fig. 6.b). The time resolution is not a limiting factor and the experiments are feasible with sources of shorter x-ray pulses, provided the flux is not too low. 

In the optical bands, thousands of individual knots have been observed to yield kinematic information, whereas the X-rays probe more dynamically important information since a much larger fraction of the ejecta mass is probed by X-rays... 

---