26
P. R. Willmott
can be effectively achieved using multilayer monochromators or, in some instances,
refractive optics such as compound refractive lenses or prisms [16].
Improvements don’t stop here, however. A critical aspect of undulator design is
that the magnetic field must be exceedingly homogeneous in the x-direction (that
is, in the horizontal plane, perpendicular to the undulator axis) in the region where
the electron beam propagates. This has meant that, for third-generation facilities, the
magnet yokes need to be a few centimetres wide or more in the x-direction. The
approximately five to ten times smaller extent of electron beams passing through
undulators installed in DLSRs means that the yokes can be of the order of one
centimetre. This allows the design of compact magnetic ‘funnels’, concentrating
magnetic field lines and thus increasing the magnetic field strength, which, in turn,
allows for a more compact design and shorter undulator periods. Moreover, this
means that more periods (N ) can be fit into a given undulator length L.
The reduction of the horizontal electron emittance in DLSRs thus lends many
potential scientific opportunities and will drive exciting innovations along the full
technology chain, from the sources (in particular, undulators), through improved Xray optics (regarding minimization of optical imperfections and aberrations), detector
technology, to handling of large data volumes. The first DLSR, MAX-IV in Lund,
Sweden, came online in Summer 2016, while a second greenfield facility. Sirius in
Campinas, Brazil and the ESRF, the first facility to undergo an upgrade from third
generation to DLSR status, both began pilot experiments in early 2020.
Several orders of magnitude higher peak brilliance are provided by high-gain
X-ray free-electron lasers (XFELs). The machine science and technologies for this
paradigm shift in X-ray sources are based on fundamentally different principles, and
consequently, XFELs are discussed separately in the next section.
1.5 X-Ray Free-Electron Lasers
Radiation from fourth-generation DLSRs has some important common features with
laser radiation: it is very intense, collimated and, in the case of radiation from undulators, partially monochromatic. An important distinction, however, is that the degree
of transverse (spatial) coherence of synchrotron radiation, although much improved
at DLSRs compared to that produced by third-generation facilities, is still only of the
order of a few percent in the hard X-ray regime. This is because, although radiation
from any single electron travelling along an undulator is coherent, there is no spatial
(i.e. phase) correlation between different electrons, and hence their combined output
remains largely incoherent. In contrast, visible lasers are normally close to being
100% coherent.
Moreover, the shortest pulse duration of X-rays from synchrotrons is a few tens
of picoseconds and may be as large as a few hundred picoseconds in the case of
DLSRs. Lasers in the visible and near-visible regimes can have pulse lengths as small
as 70 attoseconds in some exceptional cases; more representatively, pulse lengths of
a few femtoseconds are routinely achieved. The production of femtosecond light
P. R. Willmott
can be effectively achieved using multilayer monochromators or, in some instances,
refractive optics such as compound refractive lenses or prisms [16].
Improvements don’t stop here, however. A critical aspect of undulator design is
that the magnetic field must be exceedingly homogeneous in the x-direction (that
is, in the horizontal plane, perpendicular to the undulator axis) in the region where
the electron beam propagates. This has meant that, for third-generation facilities, the
magnet yokes need to be a few centimetres wide or more in the x-direction. The
approximately five to ten times smaller extent of electron beams passing through
undulators installed in DLSRs means that the yokes can be of the order of one
centimetre. This allows the design of compact magnetic ‘funnels’, concentrating
magnetic field lines and thus increasing the magnetic field strength, which, in turn,
allows for a more compact design and shorter undulator periods. Moreover, this
means that more periods (N ) can be fit into a given undulator length L.
The reduction of the horizontal electron emittance in DLSRs thus lends many
potential scientific opportunities and will drive exciting innovations along the full
technology chain, from the sources (in particular, undulators), through improved Xray optics (regarding minimization of optical imperfections and aberrations), detector
technology, to handling of large data volumes. The first DLSR, MAX-IV in Lund,
Sweden, came online in Summer 2016, while a second greenfield facility. Sirius in
Campinas, Brazil and the ESRF, the first facility to undergo an upgrade from third
generation to DLSR status, both began pilot experiments in early 2020.
Several orders of magnitude higher peak brilliance are provided by high-gain
X-ray free-electron lasers (XFELs). The machine science and technologies for this
paradigm shift in X-ray sources are based on fundamentally different principles, and
consequently, XFELs are discussed separately in the next section.
1.5 X-Ray Free-Electron Lasers
Radiation from fourth-generation DLSRs has some important common features with
laser radiation: it is very intense, collimated and, in the case of radiation from undulators, partially monochromatic. An important distinction, however, is that the degree
of transverse (spatial) coherence of synchrotron radiation, although much improved
at DLSRs compared to that produced by third-generation facilities, is still only of the
order of a few percent in the hard X-ray regime. This is because, although radiation
from any single electron travelling along an undulator is coherent, there is no spatial
(i.e. phase) correlation between different electrons, and hence their combined output
remains largely incoherent. In contrast, visible lasers are normally close to being
100% coherent.
Moreover, the shortest pulse duration of X-rays from synchrotrons is a few tens
of picoseconds and may be as large as a few hundred picoseconds in the case of
DLSRs. Lasers in the visible and near-visible regimes can have pulse lengths as small
as 70 attoseconds in some exceptional cases; more representatively, pulse lengths of
a few femtoseconds are routinely achieved. The production of femtosecond light