Balerna and Mobilio 2015 Introduction To Synchrotron Radiation
Balerna and Mobilio 2015 Introduction To Synchrotron Radiation
Balerna and Mobilio 2015 Introduction To Synchrotron Radiation
1.1 Introduction
When electrons or other charged particles moving at relativistic speeds are forced by
magnetic fields to follow curved trajectories they emit electromagnetic radiation in
the direction of their motion, known as synchrotron radiation. Synchrotron radiation
is emitted, for example, by cosmic sources. The blue light visible in Fig. 1.1, that
shows a mosaic image of the Crab Nebula, the remnant of a supernova explosion seen
on earth by many astronomers in 1054, comes from high energy electrons whirling
A. Balerna (B)
Laboratori Nazionali di Frascati, INFN, via E. Fermi 40, 00044 Frascati, RM, Italy
e-mail: [email protected]
S. Mobilio
Dipartimento di Scienze, Universita’ Roma Tre, via della Vasca Navale 84, 00146 Rome, Italy
e-mail: [email protected]
around magnetic field lines due to the presence of a pulsar (pulsating radio source). In
this lecture we present the properties and characteristics of the synchrotron radiation
produced in particle accelerators by electrons or positrons; such radiation is extremely
intense and extends over a broad energy range from the infrared through the visible
and ultraviolet, into the soft and hard x-ray regions of the electromagnetic spectrum.
It was observed for the first time in 1947 at the General Electric synchrotron in
the USA; for a long time it was considered only as a problem for particle physics
since represents the major source of energy loss in high energy particle accelerators.
Only in the late sixties it was realized that synchrotron radiation was very useful for
condensed matter research. Since then, there was an explosive growth in its use and
dedicated synchrotron radiation facilities were built, making this radiation a unique
tool in many research fields. Nowadays synchrotron radiation is being used to study
many aspects of the structure of matter at the atomic and molecular scale, from
surface properties of solids to the structure of protein molecules.
Electrons traveling at a speed close to c, the speed of light, and forced to change
the direction of their motion under the effect of magnetic fields (perpendicular to
the direction of their motion), emit light, with peculiar characteristics, known as
synchrotron radiation. The ‘natural’ man-made sources of this radiation [1] are high
energy, electron or positron circular accelerators, like storage rings. They consist
1 Introduction to Synchrotron Radiation 5
of circular evacuated pipes where the electrons are forced to follow circular paths
under the action of magnets placed along the circumference (bending magnets). The
electrons enter the storage ring only after they have been accelerated by a linear
accelerator or ‘linac’ until their energy reaches several millions of electron volts
(MeV) and then by a booster ring that gives them a boost in energy from millions
to billions or giga electron volts (GeV); at that point they are transferred to the final
circular accelerator as shown in Fig. 1.2.
Here the electrons, if needed, may be further accelerated to higher energies by the
radio frequency (RF) electric fields. When the electrons reach the expected energy
they are in a quasi-stationary situation; forced to follow circular paths by the mag-
netic field of the bending magnets, they loose, during each turn, part of their energy,
emitting synchrotron radiation. The energy lost in this way is fully regained in pass-
ing through the RF cavities. More precisely, storage rings consist of an array of
magnets for focusing and bending the beam, connected by straight linear sections
(see Fig. 1.3). In one or more of these linear sections, RF cavities are installed in
order to accelerate the particles.
After its discovery in 1947, during the 1960s and early 70s pioneering use of
the light was made in the so called first generation synchrotron radiation facilities;
these machines were not dedicated to synchrotron radiation studies, but were used
in a parasitic way, during high energy physics experiments. In the mid-70s facilities
totally dedicated to synchrotron light, were built. These early-dedicated facilities, in
which synchrotron light was mainly produced by bending magnets, are called second
generation sources; also the use of high magnetic field devices, known as wigglers,
started in this period. Afterwards, the optimization of magnetic structures, like wig-
glers and undulators, placed in straight sections made possible the realization of new
synchrotron radiation sources the so called third generation sources like the European
Synchrotron Radiation Facility (E.S.R.F. Grenoble France) and many other [2].
6 A. Balerna and S. Mobilio
Fig. 1.3 Schematic view of a storage ring where some main elements like bending magnets,
focusing and de-focusing magnets (quadrupoles), insertion devices (undulator, wigglers) and the
RF (radio frequency) cavity are visible; the injection system is not shown (courtesy of S. Tomassini
INFN-LNF)
Due to the advanced designs and beam control of third generation light sources,
source size and divergence have been greatly improved. Nowadays for many exper-
iments, like the ones involving imaging or microscopy, the beam has to be focused
down to very small spots and so a relevant synchrotron radiation property becomes
its spectral brightness1 or flux per unit source area and per unit solid angle [3].
Undulators are excellent sources of high brightness radiation. This is shown
in Fig. 1.4, where the gain in brightness achieved by the new synchrotron radi-
ation sources, as a function of time, is shown; its increase of many orders of
magnitude gives an idea of the spectacular progress obtained in beam quality by
third generation sources.
While improvements in third generation synchrotron radiation sources are still
possible, fourth generation sources are being developed, based on free electron lasers
(FELs) able to produce very short coherent light pulses of very high peak intensity and
brightness (FEL sources characteristics and applications are described in Chaps. 2
and 30).
1 Brightness or brilliance? Several authors give different definitions while other authors use both
terms to define the same quantity. The best thing to do, is to look to units to be sure of what you are
dealing with. In any case in this lecture we will use the conclusions of [3] and define the spectral
brightness as photons per second, per unit source size and divergence in a given bandwidth.
1 Introduction to Synchrotron Radiation 7
Photons
Brightness = 106
s · mm2 · mr 2 · (0.1 % BW)
• Sun
Photons
Brightness = 1010
s · mm2 · mr 2 · (0.1 % BW)
Photons
Brightness = 1016
s · mm2 · mr 2 · (0.1 % BW)
8 A. Balerna and S. Mobilio
where p, e and v are respectively the particle momentum, charge and velocity and E
and B are the electric and magnetic fields.
The power radiated [1] by a relativistic electron forced to move along a circular
orbit, with a radius of curvature, R, is given by the Schwinger’s formula [5]:
2 e2 c E 4
Pe = P(λ, ψ)dλdψ = (1.1)
3 R2 mc2
In order to replenish the energy lost with the emission of radiation and keep the
electrons at a constant energy, radio frequency (RF) cavities are used. In a RF cavity a
longitudinal electric field accelerate the electrons. The RF fields have an accelerating
effect only during one half of their period and a decelerating action during the other
half; so the RF is effective in restoring the electron energy only for one half of the
time. Additional considerations have to be performed, regarding the stability of the
electron orbit. Let us suppose that at the time t 0 an electron passing through the RF
finds exactly the electric field he needs to fully restore the energy lost during a turn.
We will call this electron a “synchronous” electron (Fig. 1.6).
Synchronous electrons continue their motion along the ring returning in the RF
again in time to regain the exact amount of energy lost along the circular path. So
10 A. Balerna and S. Mobilio
the synchronous electrons are in a stable condition. Let us now consider electrons
arriving in the RF a little bit later than the synchronous electrons. They are slower
than the first ones, i.e. they have a lower energy. In order to restore the energy,
they have to find an electric field higher than that found by the synchronous electrons
otherwise they will continue to lose energy with respect to them. In the next turn they
will arrive later, and after some turns they will enter the RF during the decelerating
semi-period and will be lost. Vice versa electrons arriving before the synchronous
ones must find a lower electric field, otherwise their energy will increase with respect
to the synchronous electrons. These considerations show that only one half of the
accelerating semi-period (i.e. one fourth of the period) is effective in maintaining the
electrons on the orbit (Fig. 1.6).
The stability condition is effectively more strict, and only 5–10 % of the RF period
is effective in restoring the electron energy. All the electrons, passing through the RF,
not in phase with this 5–10 % effective time, do not follow the ideal circular orbit of
the ring and therefore are lost. As a consequence the electrons in the storage ring are
grouped in bunches with time lengths that are typically 5–10 % of the RF period. Also
the radiation appears in pulses with the same time duration and separation (Fig. 1.7).
Along the storage ring many bunches can be distributed. The time interval between
them is an integer multiple of the RF period (called harmonic number of the ring).
The maximum separation between two pulses is obtained in the single bunch mode,
i.e. when only one bunch in the full ring is present. In this case the time interval is
equal to the period of revolution, typically of the order of microseconds. When more
bunches are present the time interval is lower; the minimum possible time interval
between bunches is equal to the RF period. The filling of bunches in a machine is a
parameter that can be completely controlled; it is possible to choose how many and
which bunches have in the ring. This flexibility is often used to relate to the photon
1 Introduction to Synchrotron Radiation 11
Fig. 1.7 Electron bunches moving in the storage ring produce radiation peaks having the same
bunch length (ps) and separation (ns)
source different time dependence characteristics. The total current depends on the
number of filled bunches. The current is lower when few bunches are filled, because
the total amount of current that can be stored in a single bunch is limited.
Application to a third generation electron storage ring like the European Syn-
chrotron Radiation Facillity (ESRF, Grenoble -FR).
• The energy of the electrons at ESRF is E = 6 GeV, it has a circumference
(C) of 844 m, the frequency of the RF cavity is νRF = 355 MHz (8 MV RF
accelerating voltage) and I max = 200 mA is the maximum current.
T0 = C/c = 2.81 µs
where TRF represents the minimum time separation between bunches achiev-
able only in uniform bunch mode (corresponding to a separation of about
86 cm between them). The maximum number of bunches that can be stored
depends on νRF and on ν0 = (1/T0 ):
In order to understand the angular and spectral distribution of the emitted radiation
let us first remind the emission from a classical electron moving at a speed, v, much
lower than the speed of light, c (v c)(see Fig. 1.8).
In this case the emitted pattern is similar to that of an oscillating dipole with its
maximum of intensity in the direction perpendicular to the acceleration and does not
depend on the electron speed.
For a relativistic effect, when the speed of the emitting electrons increases to
relativistic values (v ≈ c) the radiation pattern is compressed into a narrow cone in
the direction of motion, resulting into an emission tangential to the particle orbit.
The vertical half-opening angle, ψ, is given by:
ψ ≈ mc2 /E ≈ γ −1 .
γ = 1957E(GeV)
so for a storage ring of energy E = 1 GeV it follows that ψ ≈ 0.5 mrad ≈ 0.029◦ :
synchrotron radiation is highly collimated. This characteristic provides extremely
high fluxes on very small areas also at distances of tens of meters from the storage ring.
In a bending magnet the horizontal collimation is lost because the electrons move
along a circular orbit emitting the radiation along the tangent. The radiation is
1 Introduction to Synchrotron Radiation 13
Fig. 1.8 Qualitative radiation patterns related to charged particles moving in a circular orbit. The
dipole pattern achieved for slow particles (left) (β = v/c << 1) is distorted into a narrow cone
when β ≈ 1 (right) (the figure was redrawn: it is based on Fig. 1.1 of [6])
Fig. 1.10 Universal curve function of the spectral distribution of BM synchrotron radiation drawn
as a function of λc /λ
that depends only on the machine parameters γ and R. The critical wavelength, λc ,
represented by the discontinuous line in Fig. 1.10), divides the spectrum into two
parts of equal radiated power: 50 % of the total power is radiated at wavelengths
shorter than λc and 50 % at wavelengths longer than λc .
The physical origin of such broad distribution can be qualitatively understood.
Let us consider to have a single electron moving in the storage ring and to record the
emitted radiation using a point detector which looks tangentially at a specific point
of the orbit (see Fig. 1.11). The detector will receive a short pulse of radiation every
time the electron passes through that point of the orbit, i.e. at a frequency equal to
the frequency of the period of motion ω = v/2π R.
In the frequency domain the spectral distribution will be composed of the
fundamental frequency and of its harmonics. At high frequency a cutoff will be
present because the detector will receive the radiation emitted by the electron along
the arc 2/ γ and this originates a pulse of non zero duration, Δτ , given by the dif-
ference between the time for the electron to travel along the arc and the time for the
light to travel along the chord subtended by this arc (see Fig. 1.11):
R 1 1 R
Δτ = − 2sin ∼ 3.
c γβ 2γ cγ
Fig. 1.11 Graphical view of the origin of the duration of the light pulse emitted by a bending
magnet. The detector records the radiation emitted by the electron along the path from A to B (the
figure was redrawn: it is based on Fig. II.2 of [7])
ωcutoff ≈ Δτ −1 = cγ 3 /R
31/2 e2 Δλ
N(λ) = γ G1 Δθ (1.2)
2π hR λ
where all quantities are expressed in CGS units. G1 , known as the universal
synchrotron radiation function is the function plotted in Fig. 1.10.
In order to calculate the emission from a storage ring, it is necessary to multiply
(1.2) by Ne the number of electrons in the storage ring given by:
Ne = I (2π R)/ec
where I is the electron current. This gives for N (λ) the following equation:
16 A. Balerna and S. Mobilio
Fig. 1.12 Spectral distribution of synchrotron radiation as a function of the value of critical energy,
εc , of the storage ring
e Δλ
N(λ) = 31/2 γ IG1 Δθ
hc λ
that in practical units [8], assuming Δλ/λ = 10−3 and Δθ = 1 mrad can be written as:
photons
N(λ) = 2.457 × 1013 E[GeV ]I[A]G1 . (1.3)
s · 0.1 % BW · mrad θ
The spectral distribution of the emitted photons, given by (1.3), is that shown in
Fig. 1.10. For λ λc the spectral distribution falls off exponentially while for λ λc
it decreases slowly, it is practically independent of the energy of the electrons and it
is mainly determined by the current of the machine.
The photon flux is frequently expressed as a function of energy (Fig. 1.12); in
order to do this, y = λc /λ must be replaced by y = ε/εc where εc = hc/λc represents
the critical energy given by:
εc = 3hcγ 3 /(4π R)
or in practical units
As shown in Fig. 1.12, the maximum photon energy achievable with a storage ring
is few times its critical energy, εc .
1 Introduction to Synchrotron Radiation 17
1.6 Polarization
I// − I⊥
PLinear = .
I// + I⊥
In the orbital plane, (ψ = 0), P is equal to 1, i.e. the light is 100 % linearly polarized.
When the observation is off axis the linear polarisation decreases as a function of γ ψ.
The integration over all wavelengths gives:
Fig. 1.14 Intensity distribution of the parallel and perpendicular components of the radiation as a
function of ψ at three different photon energies, for a storage ring with E = 3.5 GeV, R = 12.12 m;
the critical energy is 7.8 keV and 1/γ is 0.146. The behaviour of the linear and circular degrees of
polarisation are also shown (the figure was redrawn: it is based on Fig. 1.8 of [1])
IR − IL ±2 (I// I⊥ )
PC = =
IR + IL I// + I⊥
where I R and I L are the intensities of the right and left polarized light and the sign, ±,
corresponds to positive or negative values of ψ, is shown in Fig. 1.14, and compared
with the degree of linear polarisation.
The vertical angular dependence of the SR beam on the vertical angle ψ can be
well approximated by a Gaussian function 1/(2π σR2 )0.5 exp(−ψ 2 /2σR2 ) [9]; σ R , in
the range 0.2 < λ/λc ) < 100, is approximately given by:
0.565 λ 0.425
σR =
γ λc
0.682 λ 0.425
Δψ(FWHM) = 2.35 · σ R =
E(GeV ) λc
An useful quantity for experiments which do not use focussing optics, is the photon
flux per solid angle, because from it the flux on the sample can be easily evaluated
using simple geometrical considerations. This quantity, defined as the number of
photons emitted per second, in a spectral bandwidth ΔE/E = 0.1 %, into a unit solid
angle, is given by:
second, in a spectral bandwidth ΔE/E = 0.1 % in an unit source area and per unit of
solid angle. As well known due to the Liouville’s theorem, focussing preserves the
brightness, i.e. the brightness of the source is equal to the brightness of the beam when
focussed on the sample. The brightness is determined by the size of the source, that
is given by the size of the electron beam and by the angular spread of the radiation,
given by the convolution of the angular distribution of synchrotron radiation, Δψ,
with the angular divergence of the electron beam. Therefore the characteristics of
the electron beam source are essential in order to determine the brightness of the
photon source. In a storage ring the product of the electron beam transverse size and
angular divergence is a constant along the ring and is called emittance. Although
the electron beam transverse size and angular divergence vary around the ring, their
product or emittance is a constant. There is a horizontal and a vertical emittance.
The horizontal emittance is measured in nanometer-radians (nm-rad).The vertical
emittance is normally a few percent of the horizontal one. In order to get high
brightness, small emittance machines must be used.
Brightness is the main figure of merit of synchrotron radiation sources and its huge
increase, shown in Fig. 1.4, was obtained thanks to the designing of low emittance
machines, like the European Synchrotron Radiation Facility (ESRF), that minimize
the source size and the beam divergence.
Insertion devices (ID) are periodic magnetic structures installed in the straight
sections of storage rings. Passing through such alternating magnetic field structures,
electrons oscillate perpendicularly to the direction of their motion and therefore emit
synchrotron radiation during each individual wiggle. The primary effects of the IDs
are:
(1) the shift of the critical energy to higher values due to the smaller bending radius
with respect to the bending magnets;
1 Introduction to Synchrotron Radiation 21
Fig. 1.17 Artistic view of the radiation beam emission from a multipole wiggler magnetic structure
(courtesy of V. Tullio INFN-LNF)
Fig. 1.18 Schematic view of the ‘wiggler’ regime, where λu represents the period of the oscillations
(2) the increase of the intensity of the radiation by a factor related to the number of
wiggles induced by the many poles of the magnetic structure;
(3) the relevant increase of the spectral brightness.
Insertion devices are of two kinds: wigglers and undulators. Inside both these
devices the electron beam is periodically deflected but outside no deflection or dis-
placement of the electron beam occurs (Fig. 1.17).
A wiggler is a multipole magnet made up of a periodic series of magnets
(N periods of length λu , the overall length being L = Nλu ), whose magnetic field
forces the electrons to wiggle around the straight path (Fig. 1.18).
The alternating magnetic field is normally applied in the vertical direction so
the sinusoidal trajectory of the electron beam lies in the horizontal plane. Electrons
follow in this way a curved trajectory with a smaller local radius of curvature with
respect to the one of the dipole-bending magnet, because in a wiggler, magnetic
fields, higher than in a bending magnet can be used.
Using the value of the instantaneous radius of curvature of the electron trajectory,
R(m):
R(m) = 3.34E(GeV)/B(T)
22 A. Balerna and S. Mobilio
where B is the magnetic field strength given in Tesla and using (1.4) the critical
energy, εc , can be expressed as:
The use of higher magnetic fields increases the critical energy with respect to the
values achievable with bending magnets and extends the spectral range of a storage
ring towards higher energies (Fig. 1.19).
The radiation observed is the incoherent sum of the radiation emitted by each
individual pole. Therefore the overall characteristics of the beam are the same as those
of a bending magnet with the same magnetic field but with an intensity enhanced by
the factor N, the number of poles.
The total power generated by a wiggler is given, in practical units, by the equation:
where B and L are the maximum magnetic field and length of the device.
In order to introduce undulators and clarify their difference with respect to
wigglers, we make use of the dimensionless parameter K. It is given by the ratio
between the wiggling angle of the trajectory, α, and the natural angular aperture of
synchrotron radiation, 1/γ , (Fig. 1.18) i.e.
K = αγ .
Fig. 1.20 Artistic view of the radiation beam emission from a multipole undulator magnetic struc-
ture (courtesy of V. Tullio INFN-LNF)
where, λu , is the period of the device. In a wiggler the transverse oscillations of the
electrons are very large and the angular deviations, α, (Fig. 1.18) much wider than
the natural opening angle ψ = γ −1 , therefore K >> 1. In these large K devices, the
interference effects between the emission from the different poles can be neglected
and the overall intensity is obtained by summing the contribution of the individual
poles.
An undulator is very similar to a wiggler, but its K value is less than 1, that means,
that the wiggling angle α is smaller than, or close to, the photon natural emission
angle γ −1 (Fig. 1.20).
In this case interference occurs between the radiation emitted by electrons at
different points along the trajectory. Considering the phase differences between the
photons emitted at different points along the sinusoidal orbit, it is easy to show
that observing the radiation in a direction forming an angle θ with the axis of the
undulator, constructive interference occurs at the wavelength, λ:
λu K2
λ= 1+ +γ θ .
2 2
(1.5)
2γ 2 2
Fig. 1.21 Calculated flux spectrum of the APS (Advanced Photon Source (7 GeV)- Argonne
(Illinois)) undulator A, a planar device (L = 2.4 m, N = 72) optimized to generate x-ray photons in
a wide energy range, using the first, third and fifth radiation harmonics (sharp peaks), with I = 100
mA, K = 2.6 giving E1 = 3.21 keV and no slits (the figure was redrawn: it is based on data of [11])
Fig. 1.22 Left panel Looking head on into the upper and lower rows of magnets (jaws) in an undu-
lator. The adjustable vertical distance (gap) between the jaws determines the emitted wavelength.
A laser was used to simulate the burst of light produced (courtesy of Advanced Light Source). Right
panel Lower row of magnets of one of the SPARC undulators (courtesy of SPARC_ LAB group
INFN-LNF)
gap distance in an undulator has an inverse effect on the amplitude of the magnetic
field and is routinely used to change λ.
In an undulator, the amplitudes of the fields radiated by each individual period of
the undulator add up coherently, so the intensity increases with N 2 while it increases
only as 2N in a wiggler.
Each harmonic has a limited wavelength bandwidth approximately given by:
1 Introduction to Synchrotron Radiation 25
Δλ/λ = 1/(nN)
Values of 10−2 can be easily achieved on the fundamental. Note that the bandwidth
decreases with the number, N, of periods of the undulator and with the harmonic
number, n.
The angular distribution, of the n-th harmonic is concentrated in a narrow cone in
both the horizontal and vertical directions (Fig. 1.20) whose half width is given by:
3 1 + (K 2 /2) 1 1
σ = ≈ √
4π γ nN2 γ nN
so it is always lower than the natural emission cone, 1/γ of a bending magnet and
decreases as the square root of the number of poles and of the harmonic number.
This very narrow angular distribution together with the N 2 dependence of the
intensity radiated in the ‘undulator’ regime explain why the spectral brightness
achievable with undulators exceeds by several order of magnitude that of bending
magnets and of wigglers (Fig. 1.23).
We underline here that the observed properties of the photon beam are determined
also by the electron beam emittance; while first and second-generation rings have
an horizontal emittance of the order of 100–200 nm-rad, third generation ones are
26 A. Balerna and S. Mobilio
in the 3–25 nm-rad range. Such low emittance values made possible to achieve the
incredible brightness of the third generation synchrotron radiation facilities. However
the increase in photon beam brightness as a function of the emittance decrease has a
limit given by diffraction effects. The diffraction limited emittance or lowest useful
emittance to produce high brightness photons of a given wavelength is λ/4π : for
a photon energy of 1 keV the diffraction limited electron beam emittance is of the
order of 0.1 nm-rad.
IDs are normally classified also as a function of the different technologies used
to generate the magnetic fields, that include the use of electromagnets at room tem-
perature (conventional iron and coil magnets) or superconducting (superconducting
coils with or without iron) and the use of permanent magnets (NdFeB and SmCo).
At room temperature if high photon energies are required, the highest magnetic field
at the shortest possible period is necessary and for this reason normally permanent
magnets are used (Bmax = 3–4 T): electromagnets are an economical choice but pro-
duce smaller magnetic fields (factor 2–5). To overcome the problem related to the
maximum achievable magnetic field, superconducting electromagnets can be used
(Bmax = 10–12 T). Operating at the liquid He temperature (4 K) superconducting IDs
are more expensive to build and operate than permanent magnets but represent the
most important and cheap solution to achieve high critical energies with a high mag-
netic field rather than increasing the whole storage ring energy. Nowadays most of the
IDs, operating with a periodic vertical field, are realized using permanent magnets.
In order to perform very specific studies like the ones needing circular polarization,
complex IDs, like helical undulators can be also designed and built. Helical undula-
tors generate helical magnetic fields that force electrons to move along spirals and
emit circularly polarized radiation. In practice the use of the insertion devices in the
third generation facilities has opened also the possibility to perform a quite large
number of experiments that were quite impossible or difficult to realize in the past.
Acknowledgments The authors are very grateful to Mr. Vinicio Tullio for his help in the realization
of many figures of this lecture.
1 Introduction to Synchrotron Radiation 27
or in practical units:
ψ ≈ mc2 /E ≈ γ −1
γ = 1957 E (GeV).
• The critical wavelength, λc :
λc = (4/3)π Rγ −3
or in practical units:
• Bending radius, R:
References
1. E.E. Koch, D.E. Eastman, Y. Farges, Synchrotron radiation—a powerful tool in science, in
Handbook of Synchrotron Radiation, ed. by E.E. Koch (North-Holland Publishing Company,
Amsterdam, 1983), pp. 1–63
2. Lightsources of the World. http://www.lightsources.org/regions
3. D.M. Mills et al., Report of the Working Group on Synchrotron Radiation Nomenclature
brightness, spectral brightness or brilliance? J. Synchrotron Rad. 12, 385 (2005)
4. H. Winick, Synchrotron radiation sources—present capabilities and future directions. J. Syn-
chrotron Rad. 5, 168–175 (1998)
5. J. Schwinger, On the classical radiation of accelerated electrons. Phys. Rev. 75, 1912–1925
(1949)
6. D.H. Tomboulian, P.L. Hartman, Spectral and angular distribution of ultraviolet radiation from
the 300-mev cornell synchrotron. Phys. Rev. 102, 1423–1447 (1956)
7. D. Raoux, in Introduction to Synchrotron Radiation and to the Physics of Storage Rings, ed.
by J. Baruchel et al. Neutron and Synchrotron Radiation for Condensed Matter Studies (Les
Editions de Physique, Springer, 1993), pp. 37–78
8. P.J. Duke, Synchrotron Radiation Sources, ed. by A.G. Michette, C.J. Buckley. X- Ray Science
and Technology (Institute of Physics Publishing, London, 1993), pp. 64–109
9. G.K. Green, Spectra and Optics of Synchrotron Radiation, Report BNL 50522: Particle Accel-
erators and High-Voltage Machine-TID 4500 (Brookhaven National Lab, Upton, 1976), pp.
1–63
10. G. Materlik, in Properties of Synchrotron Radiation, ed. by H.B. Stuhrmann. Uses of Syn-
chrotron Radiation in Biology (Academic Press Inc., London, 1982), pp. 1–21
11. W.-K. Lee, P. Fernandez, D.M. Mills, in X-Ray Crystal Optics, ed. by H. Onuki, P. Elleaume.
Undulators, Wigglers and Their Applications (Taylor & Francis, London, 2003), pp. 369–420
12. M. Boge et al., The Swiss Light Source Accelerator Complex: An Overview. http://accelconf.
web.cern.ch/Accelconf/e98/PAPERS/MOP28G.PDF
http://www.springer.com/978-3-642-55314-1