Berkeley Lab’s ALS generates femtosecond synchrotron radiation

27 June 2000

A team at Berkeley’s Advanced Light Source has shown how a laser time-slicing technique provides a path to experiments with ultrafast time resolution.


A Lawrence Berkeley National Laboratory team has succeeded in generating 300 fs pulses of synchrotron radiation at the ALS synchrotron radiation machine. The team’s members come from the Materials Sciences Division (MSD), the Center for Beam Physics in the Accelerator and Fusion Research Division and the Advanced Light Source (ALS).

Although this proof-of-principle experiment made use of visible light on a borrowed beamline, the laser “time-slicing” technique at the heart of the demonstration will soon be applied in a new bend-magnet beamline that was designed specially for the production of femtosecond pulses of X-rays to study long-range and local order in condensed matter with ultrafast time resolution. An undulator beamline based on the same technique has been proposed that will dramatically increase the flux and brightness.

The use of X-rays to study the course of solid-state phase transitions, the kinetic pathways of chemical reactions and the efficiency and function of biological processes on the fundamental timescale of a molecular vibration (about 100 fs) is an emerging field of research.

Ahmed H Zewail of Caltech was awarded the 1999 Nobel Prize for Chemistry for demonstrating how rapid laser techniques can reveal how atoms move during chemical reactions. Pump-probe methods in which a pump pulse stimulates the process followed by a probe pulse to examine it at intervals thereafter constitute a common way of following the dynamics of ultrafast processes with infrared and visible lasers. However, there is a dearth of ultrafast X-ray sources to provide structural data on this timescale. The pulse length of synchrotron radiation, for example, is limited by the bunch length of the electron beam – about 30 ps at the ALS.

Ultrashort pulses

A solution to the bunch-length problem was described four years ago by Alexander Zholents and Max Zolotorev of the Center for Beam Physics. In short, a high-power femtosecond laser synchronized with the electron bunches passes collinearly with an electron bunch through an insertion device (undulator or wiggler) as in a free electron laser. The high electric field of the shorter laser pulse modulates a portion of the longer electron bunch, with some electrons gaining energy and some losing energy.

The condition for optimum energy modulation occurs when the laser wavelength matches the wavelength of the fundamental emission from the insertion device. Subsequently, when the energy-modulated electron bunch reaches a section of the storage ring with a non-zero dispersion, a transverse separation occurs, resulting in slices of the bunch roughly as long as the laser pulse. A collimator or aperture selects the synchrotron radiation from the displaced bunch slices.

Femtosecond time structure

The team led by MSD’s Robert Schoenlein implemented the time-slicing scheme by using a high-power titanium sapphire laser to modulate the electron beam in a 16 cm period wiggler already in straight section 5 of the 12-fold symmetric storage ring. Bend magnets between the wiggler and the beamline provide horizontal dispersion and the synchrotron radiation, and a test chamber on an existing bend-magnet beamline in the curved sector after straight section 6 records the femtosecond pulses (figure 1).

Schoenlein’s group verified the femtosecond time structure by imaging visible light from the beamline onto a nonlinear optical crystal along with a delayed 50 fs cross-correlation pulse from the laser system and then counting photons at the sum frequency as a function of delay between the modulating and the cross-correlation laser pulses. An adjustable knife edge located in the beamline at an intermediate image plane provided a means of selecting radiation from different transverse regions of the electron beam. In this way the team measured a dark 300 fs hole in the central cone of the synchrotron radiation and a bright 300 fs peak in the wing of the synchrotron radiation (figure 2).

This success was the result of a synergistic collaboration between two complementary groups at Berkeley working at the ultrafast science frontier – the Center for Beam Physics, headed by Swapan Chattopadhyay and the Femtosecond Spectroscopy Group, led by Berkeley lab director Charles Shank. As part of a growing femtosecond X-ray science programme at the ALS, new beamlines are under construction and proposed under the leadership of Schoenlein and Roger Falcone of the University of California, Berkeley. A bend-magnet beamline, with an anticipated completion date of June 2000, has a performance goal of 100 fs pulses at a repetition rate of 5 kHz with a flux of about 105 photons/s/0.1% bandwidth and a brightness of about 10photons/s/mm2/mrad2/0.1% bandwidth for photon energies up to 10 keV. A proposed undulator beamline would increase the flux and brightness by factors of about 100 and 10 000 respectively. An in-vacuum device, the planned undulator has a 5 mm gap, almost a factor of three smaller than the current smallest magnetic gap (14 mm) and nearly a factor of two smaller than the narrowest vacuum chamber (9 mm) in the ring. A vertical rather than horizontal dispersion would also be used. A complete mini beta lattice with large vertical dispersion bumps is being designed to accommodate these features.

Further reading

R W Schoenlein et al.2000 Science287 2237.

bright-rec iop pub iop-science physcis connect