Femtosecond gas-phase mega-electron-volt ultrafast electron diffraction

Abstract

The development of ultrafast gas electron diffraction with nonrelativistic electrons has enabled the determination of molecular structures with atomic spatial resolution. It has, however, been challenging to break the picosecond temporal resolution barrier and achieve the goal that has long been envisioned-making space- and-time resolved molecular movies of chemical reaction in the gas-phase. Recently, an ultrafast electron diffraction (UED) apparatus using mega-electron-volt (MeV) electrons was developed at the SLAC National Accelerator Laboratory for imaging ultrafast structural dynamics of molecules in the gas phase. The SLAC gas-phase MeV UED has achieved 65 fs root mean square temporal resolution, 0.63 Å spatial resolution, and 0.22 Å^-1 reciprocal-space resolution. Such high spatial-temporal resolution has enabled the capturing of real-time molecular movies of fundamental photochemical mechanisms, such as chemical bond breaking, ring opening, and a nuclear wave packet crossing a conical intersection. In this paper, the design that enables the high spatial-temporal resolution of the SLAC gas phase MeV UED is presented. The compact design of the differential pump section of the SLAC gas phase MeV UED realized five orders-of-magnitude vacuum isolation between the electron source and gas sample chamber. The spatial resolution, temporal resolution, and long-term stability of the apparatus are systematically characterized.

FIG. 1.

Schematic of the gas-phase MeV UED at SLAC.

FIG. 2.

Cross-sectional view of the gas chamber (top left), the invacuum electron detector module (top right), and the differential pumping section (bottom right).

FIG. 3.

Photographs and computer-aided design (CAD) drawings of the incoupling mirror, capillary assembly, interaction point, and cold-trap.

FIG. 4.

Comparison of the simulated and measured electron beam spot size, σ_FWHM, and the bunch length, τ_FWHM. The rectangles mark the positions of different apparatus components.

FIG. 5.

(a) Schematic of the interaction region in the gas-phase MeV UED sample chamber. Gas-phase CF₃I molecules from the gas nozzle form pulsed gas jet samples (green). An ultrafast 266 nm pump laser (purple) excites the CF₃I molecules, while a 3.7 MeV electron pulse (blue) probes the ultrafast structural dynamics from the excited CF₃I molecules. Diffraction patterns as a function of the time delay between the pump laser and the probe electron pulses are captured on the electron detector 3.2 m downstream of the interaction point. (b) A cartoon of the CF₃I photodissociation process.

FIG. 6.

Temporal resolution characterization by the time-resolved ultrafast CF₃I photodissociation dynamics. (a) The modified scattering intensity as a function of momentum transfer $s$ and pump-probe time delay $t$. (b) The red squares show the integrated intensities in the region between the two dashed lines (between s = 1.5 Å⁻¹ and s = 2.2 Å⁻¹) in panel (a), while the black solid curve shows a best fit of error function to the red squares. The FWHM of $143\pm 36$ fs gives the upper limit of the temporal resolution.

FIG. 7.

Spatial resolution characterization. (a) Experimental diffraction pattern of CF3I without laser excitation. (b) Modified scattering intensity (blue) extracted from (a) and from simulation (red). (c) The corresponding pair distribution functions from experiment (blue) and simulation (red). The inset shows a cartoon of the molecular structure for CF3I without laser excitation. The peaks corresponding to different bond lengths are labeled.

FIG. 8.

Reciprocal-space resolution characterization. (a) A typical electron diffraction pattern from a single crystal gold sample. (b) A zoom-in view of the (200) Bragg reflection with Gaussian fitted FWHM of 0.22 Å⁻¹ as the upper limit of the reciprocal-space resolution.

FIG. 9.

Diffraction pattern centroid jitter in a typical UGED experimental run over 6 h. (a) Time trace of the horizontal (red) and vertical (blue) centroid jitter. (b) Corresponding histograms for the horizontal (red) and vertical (blue) centroid jitter. An rms position pointing jitter of 0.42 pixel corresponding to 14 μm is demonstrated in both directions. The pointing jitter at the interaction points is expected to be better than 14 μm, as it is located 3.2 m upstream of the detector.

FIG. 10.

Characterization of pump-probe time-zero stability by the plasma lensing effect. (a) Top graph shows the electron beam profile on the electron detector before time-zero, while the bottom graph shows that at pump-probe time delay t = 600 fs. (b) Intensities of electrons diffracted by the plasma field (red circles) as a function of the pump-probe time delay. An error function (black curve) is fitted to the raw data to determine the onset of the process as a measurement of time-zero. (c) A trace of time-zero changeover one hour monitored by the plasma lensing technique. The error bars reflect the fitting uncertainty in the estimated time zero change.