An interesting web site with information about generating X rays was prepared by Grzegorz Jezierski
X Ray Generation
Table of Contents:
- Conventional Generators
- Other Sources
- Choice of Radiation
- Monochromatization and Collimation of X Rays
X-Ray photons are electromagnetic radiation with wavelengths typically in the range 0.1 - 100 Å. X Rays used in diffraction experiments have wavelengths of 0.5 - 1.8 Å. X Rays can be produced by conventional generators, by synchrotrons, and by plasma sources. Electromagnetic radiation from nuclear reactions, called γ radiation, can also occur at the same energies as X rays, but γ radiation is differentiated from X ray radiation by the fact that it originates from nuclear reactions.
X rays are sometimes called Röntgen rays after their discoverer, Wilhelm Conrad Röntgen.1 He called these new rays X rays after the unknown quantity X in mathematics. These new rays had no charge, and were much more penetrating than cathode rays discovered by Johann Hittorf in 1876. X Rays were able to pass through a variety of objects. X rays could expose film. R&oml;ntgen found that X rays could pass through the tissues of a living person and illustrate the bones and other tissues in the body. For this discovery he was awarded the Nobel Prize in physics in 1901.
Röntgen wanted to determine whether X rays were particles or waves. At the time it was known that waves were involved if a stream could be shown to exhibit reflection, refraction, or diffraction. Unfortunately, Röntgen was not able to verify any of these properties of X rays.
From slit measurements, the wavelength of X rays were calculated to be on the order of Angstroms. Diffraction can occur when radiation is scattered off of an object with a repeat spacing of approximately the same size as the wavelength of the radiation. Thus, Laue looked for an object with a repeat spacing on the order of Angstroms. Another quick calculation showed that crystals could have the needed lattice spacings. Along with Friedrich and Knipping, Laue showed that X rays could be diffracted by a crystal of zinc blende.
A great deal of information about the properties of X rays and X-ray generation is available at the X-Ray Data Book. Electromagnetic radiation is made up of waves of energy that contain electric and magnetic fields vibrating transversely and sinusoidally to each other and to the direction of propogation of the waves.
This graphic representation of an electromagnetic wave, showing its associated electric (E) and magnetic (H) fields, moving forwards at the speed of light. (Copied from the CSIC web site.)
X Rays are produced in labs by directing an energetic beam of particles or radiation, at a target material. The energetic beam can be electrons, protons, or other X rays. X Rays for crystallographic studies are typically generated by bombarding a metal target with an energetic beam of electrons. The electrons are usually produced by heating a metal filament which then emits electrons. The electrons coming from the filament are then accelerated towards the target by a large applied electrical potential between the filament and the target.
When the beam of electrons hits the target (or anode) a variety of events occur. This rapid deceleration of electrons causes a variety of events including the emission of X-ray radiation, photoelectrons, Auger electrons, and a large amount of heat. Actually two types of X rays are emitted in this process. A continuous band of white radiation is always emitted. If the energy of the electron beam is sufficient then a series of intense, discrete lines that are characteristic of the target material are also observed.
Figure 1. X-ray Tube Schematic.2
White Radiation - Bremsstrahlung
Some of the collisions between the thermionic electrons and the target result in the emission of a continuous spectrum of X rays called white radiation or Bremsstrahlung. White radiation is believed due to the collision of the accelerated electrons with the atomic nuclei of the target atoms. If all of the kinetic energy carried by an electron is converted into radiation, the energy of the X-ray photon would be given by
Emax = hνmax= eV
where h = Plank's constant, νmax = the largest frequency, e = charge of an electron, V = applied voltage. This maximum energy or minimum wavelength is called the Duane-Hunt limit.
hνmax = hc/λmin = eV
λmin = hc / eV = 12398. / V (volts)
Figure 2. White Radiation from an X-Ray Generator.2 The intensity of the beam is plotted as a function of the wavelength of the radiation.
The majority of collisions that produce white radiation do not completely dissipate the kinetic energy of the electron in a single collision. Typically, these colliding electrons hit electrons in the target material with a glancing blow dissipating some energy as emitted X-ray photons. Then these photoelectrons hit other electrons in the target material emitting lower energy X-ray photons or hit valence electrons producing heat.
Thus the white radiation spectrum does have a minimum wavelength or maximum energy related to the kinetic energy of the incident radiation beam, and continues to longer wavelengths or lower energies until all of the kinetic energy is absorbed. The highest intensity of emitted white radiation spectrum is obtained at a wavelength that is about 1.5 time the minimum wavelength. The white radiation intensity curve may be fit to an expression of the form:
Iw = A i Z Vn, n ~ 2
where i is the applied current, Z is the atomic number of the target, V is the applied voltage and A is a proportionality constant. The only type of diffraction experiment that uses white radiation is the Laue experiment.
When the energy of the electron beam is above a certain threshold value, called the excitation potential, an additional set of discrete peaks is observed superimposed on the white radiation curve. The energies of these peaks are characteristic of the type of target material.
These peaks are
generated by a two-stage process. First an electron from the filament
collides with and removes a core electron from an atom of the target. Then
an electron in a higher energy state
drops down to fill the lower energy,
vacant hole in the atom's structure, emitting an X-ray photon. These emitted
X-ray photons have energies that are equal to the difference between
the upper and lower energy levels of the electron that filled the core
hole. The excitation potential for a material is the minimum energy needed to
remove the core electron.
Figure 3. Characteristic radiation from an X-ray generator.2
The characteristic lines in an atom's emission spectra are called K, L, M, ... and correspond to the n = 1, 2, 3, ... quantum levels of the electron energy states, respectively. When the two atomic energy levels differ by only one quantum level then the transitions are described as α lines (n = 2 to n = 1, or n = 3 to n = 2). When the two levels are separated by one or more quantum levels, the transitions are known as β lines (n = 3 to n = 1 or n = 4 to n = 2).
Figure 4. Electronic energy levels of an atom of the anode.
Because all K lines (n = 1) arise from a loss of electrons in the n = 1 state, the Kα and Kβ lines always appear at the same time. The n = 2 and higher energy levels (L, M, N, O) are actually split into multiple energy levels causing the α and β transitions to split into a variety of closely spaced lines at high resolution. Thus, the observed Cu Kα line can be resolved at high scattering angle (high resolution) into Kα1 and Kα2 lines with separate wavelengths. The Kα1 line is about twice as intense as the Kα2 line. At low resolution (lower scattering angle) the Kα wavelength is considered as a weighted average of the Kα1 and Kα2 lines with λ(Kαave) = [2*(λ(Kα1)) + λ(Kα2)]/3. The Kα line is about 5 - 10 times as intense as the Kβ line.
The intensity of the Kα line can be approximately calculated by
Ik = B i (V - Vk)1.5
where i = applied current, Vk = excitation potential of the target material, V = applied voltage. It can be shown that the ratio Ik / Iw is a maximum if the accelerating voltage is chosen to be about 4 times the excitation potential of the anode.
The wavelengths of characteristic X-ray lines were found to be inversely related to the atomic number of the atoms of the target material. Moseley found that
√(f) = K1 [Z - σ]
where f is the frequency of the radiation, K1 is a proportionality constant, Z is the atomic number of the target atom type, and σ is the shielding constant that typically has a value of just less than 1. Today this formula is more typically recast as
1/λ = K2 [Z - σ]2
where λ is the wavelength of the radiation, K2 is a proportionality constant, Z is the atomic number of the target atoms, and σ is the shielding constant.
The notation for describing the
characteristic X-ray lines shown above was first presented by Siegbahn. In 1991,
the International Union of Pure and Applied Chemists (IUPAC) recommended
that X-ray lines be referred to by writing the initial and final levels separated
by a hyphen, e.g. Cu K- L3, rather than using the
Siegbahn notation, e.g. Cu Kα1,
which is based on the relative intensities of the lines.3 A table
of the correspondence between IUPAC and Siegbahn notations is given in the
International Tables for Crystallography, Vol. C.4 The
Siegbahn notation remains common in the chemical and crystallographic literature.
The shape of the incident beam depends on the focal projection of the filament onto and the anode material. X-Ray beams that are parallel with wide projection of the filament have a focal shape of a line. X-Ray beams that are parallel with the narrow projection of the filament have an approximate focal shape of a square, which is usually labeled as a spot. These two focal projections are necessarily about 90 ° apart in the plane normal to the filament-anode axis. The X-ray beams emitted from the anode travel in a variety of angular directions from the anode surface. As the angle from the anode surface is increased, the intensity of the beam increases, but the spot also becomes less focused. Thus take-off angles are typically selected in the 3 - 6 ° range.
Two cartoons of an X-ray tube. Drawing a) shows the line and spot focus patterns of a typical sealed tube. Drawing b) shows the take-off angle of a tube.
The generation of X rays is very inefficient. In addition to white radiation and characteristic lines, laboratory sources also produce Auger electrons and photo-electrons. However, the vast majority of the power used in generating X rays results in the collision of accelerated electrons with valence electrons of the target material producing heat. A small fraction of the energy applied to the tube actually produces the characteristic radiation used in diffraction experiments.
Sealed-tube X-ray generators use a stationary anode. These tubes are limited in the power that can be applied to the tube by the amount of heat that can be dissipated through water cooling. One way to increase the heat dissipating ability of the system, and thus increase the X-ray beam intensity, is to move or rotate the anode surface so that the beam of electrons continually hits a new region of the anode. These rotating-anode generators typically yield about 5 times the flux of X-rays as is routinely produced by sealed-tube generators with normal-focus X-ray tubes.
Because macromolecular crystallographers need the most intense beam available, they typically use rotating-anode X-ray generators. Rotating-anode generators require a considerable amount of maintenance to replace filaments, and repair or replace the anode bearings as well as vacuum and water seals. To keep from burning the filament, it must remain in a high vacuum. The anode with its constant flow of cooling water must be continuously rotating at speeds of 6000 rpm or more. Special ferro-fluidic seals are used to maintain the vacuum along the rotating shaft of the anode. Sealed-tube sources with their minimal maintenance requirements are generally quite adequate for most small molecule needs.
Another type of sealed-tube source that produces beam fluxes comparable to rotating-anode systems is a micro-focus generator. Because heat dissipates rather quickly in a metal block, manufacturers have found that when the focal size is reduced to 10-300 μm then the power can be increased to make the beam flux much higher than for normal- or even fine-focus sealed tube sources. One of the great advantages of a micro-focus radiation source is that the electrical power needs are in the range of 30-80 Watts not the 2-3 kWatts that are required of a typical sealed tube generator, or the 3-12 kWatts required by a rotating anode generator.
Synchrotron Radiation Sources
In 1943, Dmitri Ivanenko and Isaak Pomeranchuk predicted that electrons traveling at relativistic speeds when directed through a magnetic field would emit radiation. This prediction was observed in the lab in 1946 by scientists at GE. A synchrotron radiation source is a very intense, tunable source of radiation with wavelengths from hard X-rays through visible wave, to microwaves. Because it is so costly to maintain large quantities of electrons traveling at near the speed of light in a high vacuum storage ring, few of these radiation sources are built.
To make the best use of this type of radiation, synchrotrons are shaped roughly as rings with ports that emit the photons located at nearly each bend of the ring. The radiation from each of these ports is then directed to one or more experimental chambers. Synchrotron radiation is used in crystallography to collect data on biological macromolecules, on tiny small-molecule single crystals, and on various polycrystalline materials. In addition, synchrotron radiation in X-ray energies is used in a variety of scattering and absorption studies as well as a multitude of physics experiments. Aside from the very high flux, synchrotron radiation also has the added benefit of being tunable to a specific wavelength.
In most synchrotrons, electrons are generated by an electron gun, then accelerated first by a linear accelerator, linac, and then transferred to a booster ring where they are accelerated by resonating rf cavities until the energies are 3-6 Gev and the speeds of the electrons are near the speed of light. These electrons are then directed into the main storage ring. When radiation is emitted, the electrons loose energy. The electrons are reenergized by resonating rf cavities located in the straight parts of the storage ring.
Two types insertion devices, devices to send radiation beams towards an instrument, are used in the straight parts of the storage ring to boost the flux of radiation. Wigglers are a series of electromagnetic plates with opposite charge. Undulators are similar to wigglers with lower energy applied to the plates, but the plates are spaced to give optimum intensity to particular wavelengths and their harmonics.
The main problem limiting brightness in laboratory-based X-ray sources is the removal of heat. A new type of X ray source, that offers a novel solution to this problem, uses a liquid metal, gallium anode.5 The scientists that developed this source have already reported achieving beam brightnesses greater that modern rotating anodes, with the theoretical capability of increasing this flux by another 3 orders of magnitude. A Swedish company, Excillum, is currently producing these sources.
A new type of tube that utilizes carbon nanotubes as the cathode are most likely to be developed as portable and miniature X-ray sources.6 As of this writing, these sources are not commercially available.
Medical X Rays
X Rays for medical use are generally produced by one of two methods. Diagnostic X rays for examining bones and teeth are usually produced by sealed tube equipment with a tungsten target. X Rays for CT (computed tomography) scans and radiation therapy are produced by a linear accelerator, linac. Electrons from an electron gun are accelerated through the linac by a series of charge plates. These electrons then collide with a target giving off Bremsstrahlung. The medical X rays from sealed tube equipment have typical energies of 50-80 kev. The X rays from CT or tomography equipment typically have energies around 4-8 Mev.
Choice of Radiation
Most X-ray tubes used for diffraction studies have targets (anodes) made of copper or molybdenum metal. The characteristic wavelengths and excitation potentials for these materials are shown below. Copper radiation is preferred when the crystals are small or when the unit cells are large. Copper radiation (or softer) is required when the absolute configuration of a compound is needed and the compound only contains atoms with atomic numbers < 10. A copper source is preferred for most types of powder diffraction. Chromium anodes are sometimes used to enhance anomalous scattering effects for some macromolecular samples.
Molybdenum radiation is preferred for larger crystals of strongly absorbing materials and for very high resolution, sin (θ) / λ < 0.6 Å, data. The scintillation point detectors, often used in small molecule diffraction, have somewhat higher quantum efficiencies for molybdenum radiation than for copper radiation. Because the diffraction spots are closer together for molybdenum radiation than for copper radiation, molybdenum is the preferred radiation source when using area detectors to study small molecules. The solid angle coverage of most area detectors is such that with molybdenum radiation, it is usually possible to collect an entire data set with the detector sitting at a single position. However, because a brighter incident beam of X-rays is produced from a copper tube than from a molybdenum tube at the same power level, very small crystals of even strongly absorbing materials will often yield better diffracted intensities from copper radiation than from molybdenum radiation.
Occasionally, other types of target materials, e.g. Cr, Fe, W, or Ag, are chosen for specialized diffraction experiments. Sources with Cr or Fe targets are often chosen when protein crystals are very small or when anomalous differences need to be enhanced. When samples are very strongly absorbing or when extremely high resolution data are needed then X-ray tubes with sources such as W or Ag are usually selected.
|Excit. Pot. (kV)||5.99||7.11||8.98||20.0|
Monochromatization and Collimation of X Rays
Nearly all of the data collection experiments require that the energy of the X-ray radiation be limited to as narrow a band of energies (and hence wavelengths) as possible. Using a narrow wavelength band of X rays significantly reduces the fluorescent radiation given off by the sample and makes absorption corrections much simpler to perform. It has been noted that when the applied voltage for K excitation occurs, both the Kα and Kβ lines as well as the white radiation curve are observed. Usually the Kα band is selected for diffraction experiments because of its greater intensity.
Also, typical data collection methods require that the incident beam be a parallel beam of photons. To assure that the beam is as parallel as possible (lacking divergence), the incident beam path is collimated to produce an incident beam that is about 0.5 mm in diameter for normal focus sources and 0.1-0.3 mm for micro focus sources.
When the energy of a photon beam is just above the excitation potential or absorption edge of a material, that material strongly absorbs the given photon beam. If another substance can be found that has an absorption edge between the Kα and Kβ lines of the incident photon beam, this other substance can be used to significantly reduce the intensity of the Kβ line relative to the Kα line. The absorption edges of elements with ZFilter = ZTarget - 1 (or - 2) meet this requirement. The thickness of the filtering material is usually chosen to reduce the intensity of the Kβ line by a factor of 100 while reducing the intensity of the Kα line by a factor of 10 or less.
The absorption of X rays follows Beer's Law:
I / Io = exp(-μ × t)
where I = transmitted intensity, Io = incident intensity, t = thickness of material, μ = linear absorption coefficient of the material. The linear absorption coefficient depends on the substance, its density, and the wavelength of radiation. Since μ depends on the density of the absorbing material, it is usually tabulated as the mass absorption coefficient μm = μ / ρ.
An alternative way to produce an X-ray beam with a narrow wavelength distribution is to diffract the incident beam from a single crystal of known lattice dimensions. X-Ray photons of different wavelengths are diffracted from a given set of planes in a crystal at different scattering angles according to Bragg's Law. Therefore a narrow band of wavelengths can be chosen by selecting a particular scattering angle for the monochromator crystal. Crystal monochromators need to have the following properties.
The crystal must be mechanically strong and stable in the X-ray beam.
The crystal must have a strong diffracted intensity at a reasonably low scattering angle for the wavelength of the radiation being considered.
The mosaicity of the crystal, which determines the divergence of the diffracted beam and the resolution of the crystal, should be small.
A variety of geometries are possible for crystal monochromators. Most monochromators are cut with one face parallel to a major set of crystal planes. These monochromators are then oriented to diffract Kα lines from this major set of planes. Some monochromators are cut at an angle to the major set of planes in order to produce a diffracted beam with a smaller divergence. By properly curving the monochromator crystal, the diffracted beam may be focused onto a very small area. This curving may be achieved either by bending or grinding or both bending and grinding. Curved monochromators are usually reserved for special applications such synchrotrons.
Graphite crystals cut on the (0002) face are the most common crystals used as monochromators in X-ray diffraction laboratories. Other special purpose monochromator materials include germanium and lithium fluoride. In all commercially available single-crystal instruments, the monochromator is placed in the incident beam path. Powder diffraction instruments with a point detector typically place a monochromator in the diffracted beam path to remove any fluorescent radiation from the sample. Crystal monochromators systematically alter the polarization of the incident beam, requiring different geometric corrections be applied to the intensity data.
Collimators are objects inserted in the incident- or diffracted-beam
path to shape the X-ray beam. Metal tubes are typically used in single-crystal
experiments. The inside radius of the collimators is typically chosen to be somewhat
larger than the size of the sample so that the sample may be bathed in the incident
beam at all times. Incident-beam collimators are usually manufactured with two
narrow regions. The region closest to the X-ray source carries out the collimation
functions. The second narrow region has a slightly larger diameter than the first
and is used to remove the
parasitic radiation that takes a bent path
due to interaction with the edge of the first narrow region of the
collimator. Diffracted beam collimators only function to remove any stray
radiation from hitting the detector.
The left end of the collimator shown is mounted on the X-ray tube (or
incident beam monochormator). The small yellow-colored region at the left is the
part of the collimator where the size of the beam is determined. The green region at
the right is chosen to have an opening slightly larger than the region drawn
in yellow. This green region removes the
Recently, manufacturers have been selling metal collimators with a single or multiple glass capillaries. These glass capillaries redirect much of the X-ray beam that would otherwise be blocked by the collimator. Such capillary inserts in a collimator have been shown to increase the intensity of the incident beam by a factor of between two and four.
When a very intense and very small point source is needed, such as in protein crystallography, X-ray mirrors may be used to shape the incident beam. Mirrors are sometimes made from materials that act as beta filters for the radiation in use. Mirrors are primarily used with very bright X-ray sources such as rotating-anode generators or synchrotrons.
X-ray mirrors are sometimes used in the incident beam to shape the beam as is done by a collimator. Even with Cu radiation, the spots in protein diffraction patterns are often very close together. The mirrors act to focus the incident beam into an very small cross section producing very sharp spots in the diffraction pattern. Mirrors are often constructed to absorb more of the Kβ radiation than the Kα radiation making the beam approximately monochromatic. Monochromators significantly reduce the intensity of the incident beam; omitting the monochromator maximizes the incident beam flux.
Powder diffraction experiments usually require a line-shaped incident beam that is produced from a pair of parallel knife edges. A set of Soller slits are used in the beam path after the knife edges to remove parasitic radiation that scatters from the edges of the blades. Soller slits are a set of parallel thin foil sheets that absorb nearly all of the X rays not traveling parallel to the metal sheets.
These optics act somewhat as X-ray mirrors that both focus the X-ray beam and selectively absorb the Kβ wavelengths producing an intense beam of Kα radiation.
- Wilhelm Conrad Röntgen,
Über eine neue Art von Strahlen(On a New Kind of Rays) presented to the Würzburg Physical and Medical Society, 1895. Translated by Arthur Stanton, Nature, 1896, 53, 274-276.
- Michael Liang, 1997,
An Introduction to the Scope, Potential and Applications of X-ray Analysisin International Union of Crystallographers Teaching Pamphlets available at: http://www.iucr.org/education/pamphlets.
- R. Jenkins, R. Manne, J. Robin, & C. Senemaud, Pure and Appl. Chem., 1991, 63, 735-746.
International Tables for Crystallography, Vol. C,1995 Kluwer: Boston, p 167.
- M. Otendal, T. Tuohimaa, U. Vogt, and H. M. Hertz,
A 9 keV electron-impact liquid-gallium-jet x-ray source., Rev. of Sci. Inst., 2008, 79, 016102-3. doi: 10.1063/1.2833838
- See web page and associated references at: Grzegorz Jezierski's web page