Berkeley plans CIRCE to fill the terahertz gap

22 August 2005

A machine has been proposed at Berkeley that would provide radiation at terahertz frequencies, a valuable source for research.

A team at the Advanced Light Source (ALS) of the Lawrence Berkeley National Laboratory (LBNL) has proposed the construction of a ring-based photon source optimized for generating coherent synchrotron radiation (CSR) at terahertz frequencies. The Coherent Infrared Center (CIRCE) will exploit all the CSR production mechanisms currently available for achieving top-level performance, including a photon flux exceeding by more than nine orders of magnitude that of existing conventional broadband terahertz sources.


Interest in the scientific use of radiation at terahertz frequencies is rapidly increasing: the fields that would benefit range from solid-state physics (semiconductors, metals, superconductors, strongly correlated materials, etc) through chemistry and biology to applications in medical science and security. However, a major problem is that generating radiation of significant intensity in this frequency range, which lies between microwaves and infrared, is not straightforward. Owing to the lack of sources, this region is often referred to as the “terahertz gap”, but storage-ring-based CSR sources are very promising candidates for addressing this situation.


CSR occurs when the synchrotron emission from the relativistic electrons in a beam bunch is in phase. This happens when the length of an electron bunch is comparable to, or shorter than, the wavelength of the radiation being emitted. At 1 THz, this is about 300 μm. In the coherent regime, the radiation intensity is proportional to the square of the number of particles per bunch, in contrast with the linear dependence of conventional incoherent synchrotron radiation. Considering that the number of electrons per bunch in a storage ring is typically very large (106-1011), the potential intensity gain for a CSR source is huge. However, achievable bunch lengths and the shielding effect of the conductive vacuum chamber in storage rings mean CSR can only be generated in the terahertz frequency range (from about 100 μm to a few millimetres).


Although CSR was predicted to occur in high-energy storage rings over half a century ago, it has only been observed in the past few years. Intense bursts of CSR with a stochastic character have been measured in the terahertz frequency range in storage rings at several synchrotron light sources. Work carried out by groups at the Stanford Linear Accelerator Center (SLAC), LBNL and the Berliner Elektronenspeicherring-Gesellschaft für Synchrotron Strahlung (BESSY) showed that this bursting emission of CSR is associated with a single bunch instability (G Shipakov et al. 2002, M Venturini et al. 2002, J M Byrd et al. 2002, M Abo-Bakr et al. 2003a). This “microbunching instability” (MBI) is driven by the fields of the synchrotron radiation emitted by the bunch itself. Although interesting in terms of accelerator physics, these bursts of CSR are not very useful as a terahertz source, because they are intrinsically unstable and stochastic.

However, CSR emission with remarkably different characteristics was observed at BESSY when the storage ring was tuned to a special mode for short bunches (M Abo-Bakr et al. 2002 and 2003b). The emitted radiation was not the quasi-random bursting previously observed but a powerful and stable flux of broadband CSR in the terahertz range – exactly what is required for a source that is useful for scientific experiments. The LBNL, SLAC and BESSY groups together drew up a model that reproduces the observations and can be used for designing a ring-based source optimized for generating stable terahertz CSR (F Sannibale et al. 2004a and 2004b).

Terahertz CSR in storage rings

An interesting feature of the CSR spectra measured at BESSY is that they extend to significantly shorter wavelengths than those expected from a Gaussian longitudinal distribution of the bunch. The model developed showed that the synchrotron radiation fields can potentially produce a stable distortion of the bunch distribution from Gaussian towards a sawtooth-like shape with a sharp leading edge. This was ultimately responsible for the observed extension of the CSR spectra towards shorter wavelengths in BESSY. We will refer to this configuration as the “ultra-stable” mode of operation.

Another development in CSR in storage rings, first demonstrated at the ALS and more recently at BESSY, was obtained by exploiting parasitically the “femtoslicing” technique used for producing femtosecond X-ray pulses. In the femtoslicing scheme, the co-propagation in a wiggler of a femtosecond optical laser pulse with a much longer electron bunch generates a modulation of the electron energy in a femtosecond slice of the bunch. When the bunch propagates in a dispersive region, the energy-modulated particles are transversely displaced. Properly masking the synchrotron radiation can remove the part emitted by the core of the bunch while allowing the transmission of the part emitted by the displaced electrons. In this way, femtosecond X-ray pulses are obtained.

At the same time, because of the longitudinal dispersion in the ring, the modulation in energy induces a density variation in the longitudinal distribution as the bunch propagates along the ring. The characteristic length of these longitudinal structures starts from tens of micrometres (a few tens of femtoseconds duration) immediately after the laser-beam interaction region in the wiggler. It quickly increases to the order of a millimetre, before finally disappearing in a few ring turns. These structures radiate intense CSR in the terahertz range with appealing characteristics: very short CSR pulses (of the same order as the laser pulse length), which extend the CSR spectrum towards shorter wavelengths (to about 10 μm or about 30 THz) than those in the ultra-stable mode; high energies per terahertz pulse (tens of micro-joules); and terahertz CSR pulses intrinsically synchronous with the femtosecond laser and X-ray pulses (allowing for a variety of pump-probe experiments and/or electro-optic sampling techniques). The main limitation is the relatively low repetition rate (a few kilohertz), which is imposed by present laser technology.

Designing CIRCE

In designing the CIRCE ring, the team has provided for optimized versions of all the techniques for generating terahertz CSR as described. Figure 1 shows a 3D layout of the ring inside the ALS facility. The ring, 66 m in circumference and operating at 600 MeV, is designed to be located on top of the ALS Booster Ring shielding and will share the injector with the ALS Storage Ring.

Figure 2 shows the impressive flux of CIRCE, calculated for three settings of the ultra-stable mode of operation. The gain of many orders of magnitude in the terahertz frequency range over the existing conventional source is clearly visible. Figure 3 shows how the femtoslicing mode complements the ultra-stable mode of operation in CIRCE. The calculated spectra for the two modes together cover the entire terahertz range from wavelengths of about 10 μm (30 THz) to about 10 mm (0.03 THz). The energy per terahertz pulse in the example used for the femtoslicing case is about 8.5 μJ, which when focused onto a sample would provide an electric field of about 106 V/cm. Current laser technology should allow repetition rates as high as 10-100 kHz.

The vacuum chambers in the dipole magnets and the first in-vacuum mirror have been designed for the efficient collection of terahertz synchrotron radiation. The design calls for three ports with 100 mrad horizontal by 140 mrad vertical acceptance for each of the 12 dipole magnets, giving a potential total of 36 dipole beam lines in CIRCE. The layout of the ring also includes six 3.5 m straight sections that can be used for insertion devices for possible future sources (as for the case of the wiggler in the femtoslicing scheme).

The CIRCE team has completed a detailed feasibility study that includes electron-beam linear and nonlinear dynamics studies, the design of all the magnets, the design of the special high-acceptance dipole vacuum chamber, and evaluating the compatibility of CIRCE with the ALS facility. Also, the team has experimentally investigated resonating modes that could be excited by the electron beam in the high-acceptance dipole vacuum chamber.

These modes, potentially dangerous for the electron-beam stability, have been measured and characterized by means of radio-frequency measurements in a prototype dipole chamber. No “show-stoppers” have been identified and CIRCE is part of the current five-year strategic plan for the ALS.

Further reading

M Abo-Bakr et al. 2002 Phys. Rev. Lett. 88 254801.
M Abo-Bakr et al. 2003a 2003 IEEE Particle Accelerator Conference 3023.
M Abo-Bakr et al. 2003b Phys. Rev. Lett. 90 094801.
J M Byrd et al. 2002 Phys. Rev. Lett. 89 224801.
J B Murphy 2004 ICFA Beam Dynamics Newsletter 35 20.
F Sannibale et al. 2004a ICFA Beam Dynamics Newsletter 35 27.
F Sannibale et al. 2004b Phys. Rev. Lett. 93 094801.
G Shipakov et al. 2002 Phys. Rev. ST AB 4 5 054402.
M Venturini et al. 2002 Phys. Rev. Lett. 89 224802.

bright-rec iop pub iop-science physcis connect