MeV electron diffraction from 200 nm Titanium foil
MeV electron diffraction from 200 nm Aluminum foil. The diffraction rings are transformed into ellipses by magnetic quadrupole focusing.
Bunch length vs charge for different initial laser spot sizes.
Real time resolution of atomic motion is one of the great open challenges in science today and is manifested in many fields of research. Any significant progress in this direction will undoubtedly leave a profound impact on how we view and understand the most basic processes in the study of molecules, materials, and biological systems. This broad research field is particularly new and active, as only in the last twenty years the progress in laser technology has enabled the generation of optical pulses of time duration comparable to the time-scale of atomic and molecular motion (less than 100 fs).
So far most of the structural changes on atomic length scales have only been inferred indirectly from the analysis of optical spectra (pump-and-probe spectroscopy). In order to directly observe atomic and molecular structures it is necessary to use waves of sufficiently short wavelength or, from an equivalent point of view, particles (either photons or electrons) of sufficiently high energy as probe projectiles. With the relevant spatial lengths of the order of the atomic dimensions, only x-rays or high energy electrons can provide the sufficient spatial resolution to directly resolve structural changes.
The time resolution of such investigations is set by the length of the electromagnetic or particle wave packet –i.e. the x-ray or electron pulse length--, and the challenge becomes then to produce suitable very short probe beams. In the last ten years, various ways of generating ultrashort x-ray pulses have been demonstrated and thus time resolved x-ray diffraction has been the primary choice in ultrafast diffraction studies. Due to the different nature of how the charged particles couple to matter, the use of electrons to directly probe matter is in many ways complementary to x-ray diffraction.
The X-ray photons are scattered by the electrons in the material under study. Since typically the electronic density is localized around the atomic positions, x-ray diffraction yields information on the atomic and molecular structure. The interaction strength can be quantified with the classical Thomson cross-section σT = (8π/3) r02 = 6.6 10-25cm2. Charged particles on the other hand, are predominantly scattered by the nuclei (with the immediate advantage of carrying direct information on the atomic locations) via the well known Rutherford cross section which, integrated with the proper screening factor, is typically larger than 10-20cm2.
The enormous difference (five orders of magnitude) in cross-section for the two physical processes is what makes electron diffraction attractive in many cases, like the study of surface phenomena, the bulk structure of thin foils or molecular structure of gas phase where for x-rays either the density or the interaction length are too small, and the number of scatter points integrated along the interaction distance (the thickness of the sample) does not provide a sufficient signal to noise ratio. Furthermore the De Broglie electron wavelength λ = h/p ~ 0.3 pm (for electrons @ 5 MeV energy) is dramatically shorter than the wavelength of typical x-ray source (1.5 Å @ 8 KeV) giving access to very different regions of the scattering momentum space.
We intend to develop a novel ultrafast electron diffraction system based on the use of rf photoinjector in order to study materials with atomic resolution at the 100 fs time scale. Currently the limit in time-resolution for conventional electron diffraction systems is determined by how short an electron pulse can be made. These systems use beam energies in the range of tens of KeV; at these energies space charge effects broaden the pulse during propagation. Researchers have been able to reduce the time resolution to sub-ps level only by dramatically reducing the number of electrons per pulse with the compromise of integrating over multiple pulses to collect a single diffraction image. On the other hand in order to capture transient structures with single 100 fs electron pulses, the highest possible beam intensity must be employed. The most promising solution is to increase the electron energy to the MeV level where relativistic effects significantly reduce the space charge forces. Rf photoinjectors can in principle deliver up to 107-108 electrons packed in bunches of 100 fs length. Using such beams for electron diffraction could have a revolutionary impact, allowing unprecedented time resolution and enabling the study of irreversible phenomena with single shot diffraction patterns.
Conventional UED
Relativistic UED
| ||
Energy |
20-300 KeV |
3-5 MeV |
Accelerating gradient at cathode |
2-5 MV/m |
80-100 MV/m |
Number of particles per bunch |
104-105 |
107-108 |
Pulse length |
~1 ps |
<100 fs |
Typical Bragg angle (d =2 Å) |
10 mrad |
0.5 mrad |
Elastic mean free path in Al |
20nm |
200 nm |
Another important difference due to the relativistic velocity of the probe electrons from an rf photoinjector is a better matching with the velocity of the pump (typically photons). In transmission mode, the velocity mismatch can be minimized for studies of solid phase samples by employing ultra-thin targets. On the other hand, for gas-phase targets where the probed sample thickness is on the order of a mm, the time resolution for non relativistic electrons is inherently limited to few ps. For 5 MeV electrons we have b = v/c ~ 0.995 and this broadening becomes negligible (<50 fs).
A first experimental demonstration was obtained at SLAC in 2005[Hastings et al. PRL]. In that experiment a 2 ps long laser beam was used to generated the beam which was then reduced down to 560 fs rms operating the photoinjector at a compressing phase of the rf accelerating wave. Recently we at UCLA have made significant progress in optimizing the operation of an rf photoinjector for UED replicating those first simple pioneering experiments obtaining diffraction images from thin metallic targets, with both i) a shorter electron bunch and ii) improvements on the quality of the spatial pattern[Musumeci et al. Ultramicroscopy].
The Pegasus laboratory is in fact at this time the only relativistic electron diffraction source operating in the country (and in the world) and thus in a prime position to bring out of its early stages this novel extremely promising technique and achieve the ultimate goal of a time resolved study of a physical process using relativistic electrons from an rf photoinjector.
By using the RF deflector, we measured the bunch length as a function of charge for different cathode spot sizes as measured by the x-band RF deflector. In these measurrements, we demonstrated that the Pegasus photoinjector can generated electron beams shorter than 100 fs with 2 pC charge. The 2D nature of the information obtained from the rf deflector can be exploited in an interesting way in electron diffraction experiments. What is typically intended in the literature (as well as in this proposal at least until this point) for single-shot time-resolved studies of dynamical changes is that only a single electron pulse is required to collect a diffraction pattern at a different position of the delay stage between the pump and the probe pulse. A full temporally resolved study still consists in a collection of data points obtained scanning the relative time of arrival between the pump and the probe (and for true irreversible processes moving the sample to a fresh spot). In principle, one could operate the RF deflecting cavity as a streak camera to time-resolve a relatively long (tens of ps) electron beam after its interaction with the diffraction sample. This different configuration has the fascinating potential to i) free UED by the limitation due to the length of the electron beam, ii) improve significantly the temporal resolution of the technique and iii) yield true single-shot structural change studies revolutionizing the approach of the conventional pump-probe experimental procedure. With a theoretical time-resolution of less than 20 fs one could in principle resolve the full dynamic of a lattice vibration or even irreversible reaction in a true single shot (see Fig.5 showing the diffraction images obtained by simulations of this setup). This possibility is predicated on first obtaining high quality diffraction images with large signal to noise ratio and good contrast, and then by turning the rf streaking voltage on to visualize the time dependence on the structure. We have already performed few preliminary tests of this idea showing that since the electron beam out of the gun is chirped (larger momentum particles are in the front), there is a small time-dependence on the scattering angle (which is inversely proportional to the electron momentum). This promising configuration could very well constitute another fundamental breakthrough in time resolved ultrafast techniques. Furthermore, the use of rf deflectors to extract temporal information imprinted in relatively long bunches can find application even in a conventional low energy (non relativistic) electron diffraction setup.