Note that the nm/kev relationship is inverse because wavelength and energy are inversely related.

4 X-ray Spectroscopy 4.1 What are X rays? X rays are electromagnetic radiation: another region in the spectrum, along with ultraviolet, infrared, visible etc. They are characterised by very short wavelengths, meaning high frequencies and high energies. The wavelength range of the X ray region is 0.01 to 10 nm, compared to 400 800 nm for the visible region. This makes X ray photons up to 80000 times more energetic than visible radiation. There are two other units commonly used in reference to X rays: another wavelength unit, the Angstrom (Å) and the electron volt, a measure of energy (expressed as kev). Their relationship is shown in Eqn 4.1. 10Å = 1 nm = 1.24 kev Eqn 4.1 Note that the nm/kev relationship is inverse because wavelength and energy are inversely related. X rays are produced in three ways: 1. when high energy electrons collide with a surface, and are slowed down by the collision; the loss of energy produces a continuous band of X ray wavelengths 2. emissions from decaying radioactive nuclei 3. emission of X rays from matter which has been irradiated with a X ray beam; the emitted X rays are single wavelengths, and of different energies to the excitation beam Processes 1 & 2 are employed in X ray sources, while process 3 produces fluorescent radiation, exploited in the important analytical technique, X ray fluorescence, which this chapter concentrates on. 4.2 Dangers associated with X ray use The high energies associated X rays mean that they are potentially harmful to living organisms, because of the chance that the energy will be absorbed by molecules in the organism. This can cause destruction or alteration to the molecules absorbing the energy. The damage may be direct the photon breaks bonds in an important molecule, or indirect the photon splits water, forming the very reactive hydroxyl radical, which then reacts with an important biochemical substance. If the species affected is genetic material, i.e. DNA, then the chance exists for mutated cells to be produced, leading to the formation of cancers. The effect is cumulative: this is why there is a limited number of body X rays that a person can have with in a set period of time. Too many increases the number of damaged molecules and therefore, the chance of serious health consequences. There are a number of ways that limit the possible dangers associated with X ray use: instruments using them are thoroughly sealed (usually with lead) so that no radiation should escape the seals are checked on a regular basis where the instrument has removable panels which can, if opened, allow the X rays to escape, then microswitches are employed which will trip out the X ray source within milliseconds, eliminating substantial risk of accidental exposure operators of X ray instruments wear radiation badges which are developed each month to check the levels of exposure

4.3 X ray spectroscopic techniques X rays are used analytically in three different techniques: X ray fluorescence (XRF) where the emission of characteristic wavelength X rays from matter irradiated by an X ray beam is used to identify and quantify particular elements X ray diffraction (XRD) where X rays are deflected through characteristic angles (diffracted) by the crystal structure of the matter; this can be used for spectral fingerprinting, especially of minerals X ray absorption (XRA) which is the familiar use of X rays to laypersons in medicine and materials testing, where the varying density of matter causes variation in the absorption of the X ray beam. Bones are denser than tissue and absorb more radiation, hence giving high absorption, while defects (holes) in a metal component are picked by a X ray absorption scan because the X rays pass unabsorbed through them Only XRF most useful for chemical analysis will be examined in this course. 4.4 X ray fluorescence As you might expect from the name, X ray fluorescence is a emission technique, where the radiation measured from the sample is of different (and higher) wavelengths than the incident radiation. A little bit of history (non examinable) The 1895 discovery of X rays by Wilhelm Roentgen was also the discovery of X ray fluorescence, though Roentgen did not do anything with this aspect of his work. In 1913, Henry Mosley showed that the generated X rays were characteristic of particular elements, and that the wavelengths were related to atomic number. It was this discovery by Mosley that caused the re ordering of the elements in the periodic table away from atomic mass, and removed some of the anomalies in Mendeleev s table. In the following decade, a number of physicists, including Mosley and the Bragg brothers, struggled to design equipment that could efficiently generate fluorescent X rays, using electron beams as excitation. In 1928, Glocker and Schreiber showed that using a beam of X rays to excite fluorescence was a better way and actually carried out quantitative analysis of real samples. However, this discovery did not lead to immediate developments because of problems in separating different wavelengths (monochromation) and detecting the radiation. In 1948, Friedman and Birks took some of the components of an X ray diffraction instrument and built the first XRF spectrometer as would be recognised today. The first commercial instruments were released in the 1950s, and the next two decades were dominated by improvements in detection systems, particularly the silicon drift semiconductor detector discovered in 1970. Generation of fluorescent X rays As with other forms of spectroscopy that you are familiar with eg UV/VIS, AAS, flame emission XRF deals with the electrons. In those other spectroscopic techniques, the electrons were simply moved up and down in their orbitals, whereas in XRF, electrons are actually ejected from the atom. A second difference is that in XRF it is the inner shell electrons that are affected, whereas in the others, it is those in the outer shell. The emitted radiation is produced after an inner shell electron (most commonly from one of the first two shells the old symbols K&L tend to be used rather than 1st and 2nd orbitals in an atom) is ejected entirely from the atom by incident X rays (Figure 4.1(a)). AIT 4.2

This creates a "hole" in the electron shells, which is filled by electrons in higher shells. A cascade effect can occur since an electron dropping down a level to fill a hole creates a hole which in turn, is filed from a higher shell again (Figure 4.1(b)). Since the electrons are moving to lower energy levels, they lose the excess energy in the form of radiation (Figure 4.1(c)). M shell L shell M & L electrons drop down X ray K electron ejected X rays emitted K shell (a) (b) (c) FIGURE 4.1 Electron transitions involved in X ray fluorescence The transitions in XRF are given labels which identify them when a spectrum is recorded. The labels indicate: the shell from which the electron was ejected K or L, very rarely M the shell from which the filling electron came the Greek letters,, and indicate that the filling electron came from 1, 2 or 3 shells respectively, from the hole Transitions from KL, MK, ML and NL shells produce radiation of X ray energy for most elements (except the very light ones). CLASS EXERCISE 4.1 (a) What are the two transitions shown in Figure 4.1? (b) Draw the transitions corresponding to the labels K and L. N M L K AIT 4.3

Why are these labels important? There are important trends in wavelength and intensity that can be followed for these transitions (shown in Table 4.1), which can help identify the presence of species in a sample, and also are a shorthand way of referring to certain characteristic emissions. TABLE 4.1 Energy and intensity trends for X ray fluorescent photons Trend Behaviour Energy for a given element Intensity for a given element K > K >> L > L K > K >> L > L Energy for a given transition Increases with increasing atomic number (see Figure 4.2 & Table 4.2) CLASS EXERCISE 4.2 Explain the order of energies for the four transitions. TABLE 4.2 Energies (kev) of selected element XRF lines Element K K L L Ca 3.69 4.01 0.34 0.35 Cr 5.41 5.95 0.57 0.58 Fe 6.40 5.06 0.71 0.72 Sn 25.19 25.48 3.44 3.66 Pb 74.22 84.92 10.55 12.61 100 80 K 60 kev 40 20 L 0 10 20 30 40 50 60 70 80 90 Atomic Number FIGURE 4.2 Energies of K and L transitions vs atomic number AIT 4.4

Given that the K line is the most intense for each element, there might seem to be little use for the other lines. Qualitatively they help in making a definite identification, but they also have use quantitatively because for reasons of excitation and detection, there are limits on which elements can be analysed by their K line. This is covered in more detail later. 4.5 XRF instrumentation There are two basic designs of XRF instruments, mainly due to the development of the wavelength selective semiconductor detectors described above. These two instrument types are known as: wavelength dispersive the older type, using a conventional monochromator based design to separate the fluorescent wavelengths (Figure 4.3) energy dispersive uses a special detector which doesn't require a monochromator (Figure 4.4) sample holder emitted radiation single excitation X-rays many s detector path X-ray source collimator dispersing crystal detector FIGURE 4.3 Schematic diagram of a wavelength dispersive XRF instrument (with moving detector) The wavelength dispersive instrument uses a monochromator where the emitted wavelengths are diffracted by a crystal at different angles. Conventional spectroscopic instruments, such as UV VIS spectrophotometers, are designed so that the diffraction medium rotates and one wavelength at a time leaves the exit slit. In XRF instruments which are designed for high resolution work, both the diffracting crystal and the detector unit move around a circular track (known as a goniometer), picking up one wavelength after another. ICP emission instruments also use this design. FIGURE 4.4 Schematic diagram of a energy dispersive XRF instrument AIT 4.5

The energy dispersive instrument has a key component that makes it very different to the older wavelength dispersive design: its semiconductor detector has the unique ability to be able to distinguish between photons of different energy withoutt the need for physical separation. It is thus a single detector multichannel instrument, with the advantages that such a configuration brings: simplicity speed cost size Energy dispersive instruments can be small benchtop designs, or even hand held battery powered 4.5). They are not XRF devices, capable of being used in the field or in the factory (as shown in Figure as accurate as a laboratory device, but have obvious convenience advantages (think about an ion items selectivee electrode vs a HPLC for measuring nitrate). They can also be used to analyse large which cannot be fitted into a conventional instrument. Examples of this include valuable works of art and ancient artefacts unearthed in archaeological excavations (see Figure 4.6). However, wavelength dispersive instruments are still important because they are more sensitive and have better resolution than energy dispersives. The best XRF instruments in terms of absolute performance remain wavelength dispersive, and can cost in excess of $300,000! Radiation source The X ray tube consists of a heated wire cathode, which emits electrons. They are accelerated towards to the anode a block of metal knownn as the target and the collision releases energy in the form of a continuous spectrum of X rays. A schematic diagram of the tube is shown in Figure 4. 7 below. The tube is evacuated to avoid complications caused by ionisation of gas molecules. FIGURE 4.5 Portable energy dispersive XRF instrument (courtesy of InnovXsys) FIGURE 4.6 Portable energy dispersive XRF instrument used to analyse a religious painting in position (Kriznar et al, 9th International Conference on NDT of Art) AIT 4.6

window X-rays target electrons cathode FIGURE 4.7 Schematic diagram of an X ray tube There are three variables that affect the output of an X ray tube (as shown in Figure 4.8): applied voltage between the electrodes current the material used on the surface of the target exposed to the electron beam. Target anodes are made from a block of copper, with a coating of another metal on the surface which takes the electron impact. The block is water cooled to remove the heat which accumulates as part of the collision process. Metals commonly employed as the contact surface are rhodium, tungsten and molybdenum. Heavier elements generate more radiation, as can be seen in Figure 4.8(c). FIGURE 4.8 Effect on X ray tube output of (a) tube current (b) applied potential and (c) target element (from Bertin, Introduction To X ray Spectrometric Analysis, Plenum). AIT 4.7

CLASS EXERCISE 4.3 Complete the table below, to summarise Figure 4.8. Tube Variable Wavelength Range Intensity Voltage Current Target element The continuous spectrum shown in Figure 4.8 is not quite accurate, as the real output has two or more peaks superimposed. These are due to X ray fluorescence processes occurring in the atoms of the target. These may cause an interference in XRF that must be allowed for or removed, or can be used as a monochromatic source of X rays, if appropriate filtering is available, as occurs in XRD. Rhodium, for example, the target metal in the XRF instrument in this department, emits lines at 2.7, 2.8, 3.0, 20.2 & 22.7 kev. The only option for many years to the X ray tube was a radioactive source, which generated a range of X ray wavelengths, without the need for electrical power. However, the problems of radioactive sources and the lower output intensities reduced the usability of these sources. Traditionally, X ray tubes require large quantities of electrical power: typical operating conditions may be 35kV and 30 ma (1.05 kw), necessitating connection to the mains power supply. Low current X ray tubes (10 100 ua) have been recently developed which have allowed batterypowered portable devices for field analysis and small desktop instruments, such as the Minipal we have. CLASS EXERCISE 4.4 What would be the main disadvantage of these low current X tubes? Voltage is the most important variable in X ray fluorescence because it is determines the energy of the exciting photons, and therefore which elements will be excited. The kev unit for X ray energy is related to the voltage needed to excite a particular element line. An electron accelerated towards a positive electrode with a voltage equal to that kev value (in kv) has the energy contained in that X ray. To excite the element that generates that fluorescent X ray typically requires a tube voltage twice the kev value. EXAMPLE 4.1 The K line of calcium has an energy of 3.69 kev. What tube voltage needed to excite this line? 2 x 3.69 = 5.38 kv. AIT 4.8

CLASS EXERCISE 4.5 The X ray tube in our instrument has a maximum voltage of 30kV. What is the maximum atomic number for an element that can have its K line excited (see Figure 4.2)? Detectors There are three common X ray detectors: semiconductor gas filled/flow/proportional scintillation Semiconductor detectors are used in energy dispersive devices, while the other two are used together in wavelength dispersive instruments. The reason that two detectors are required in WD instruments is that neither covers the full range of X rays. Their mechanism of operation is described below. SEMICONDUCTORS The most recent development in X ray detection is the silicon drift detector (using semiconductor technology). The precise means by which X rays are detected by such a device are beyond the scope of this course. The most important aspect of this detector is its ability to produce an output (a current pulse) that is proportional to the energy of the incoming photon. It can do this for polychromatic radiation equally well, when linked with pulse height counting equipment. This sums the number of pulses of each amount of current (e.g. 1, 1.1, 1.2 ua) and so the final spectrum is really a plot of number of pulses vs current. A correlation between energy and current must be made by some internal setting in the detector, eg a current pulse of 1.2 ua is created by a photon of 2.5 kev energy. This unique ability means that energy dispersive instruments do not need a monochromator with the all the multi channel type advantages that brings. The two major limitations are: resolution the ability to distinguish between photons of similar but not identical energy is not perfect sensitivity there is a maximum count rate (in our instrument, it is 60,000 per second) above which the detector is overloaded It is not clear why similar detectors have not been developed in other regions of the electromagnetic spectrum, especially the UV VIS where so many instruments operate. PROPORTIONAL (OR FLOW) COUNTERS X rays have sufficient energy to ionise atoms. This ability is exploited in a number of detectors, which use a chamber filled with an inert gas, such as argon, and electrical plates which collect the ions that are produced. The chain reaction may generate hundreds of electrons and ions from the one photon. The degree to which the chain reaction proceeds depends on the applied voltage between the collector plates and the energy of the photons. The most commonly used voltage range is 800 1000V, where the response of the detector is approximately linear which voltage. This region is known as the proportional region, and a detector using this behaviour as a proportional counter. AIT 4.9

A typical proportional counter is shown below in Figure 4.10. A small flow of gas is normally passed through the detector to keep it at a constant pressure. The gas is 90% argon/10% methane. These detectors are also known as flow counters. Proportional counters are most suited to X ray photons from 1.5 25Å (because these detectors are used in WD instruments, wavelength is the usual unit), which in terms of X ray fluorescence, come from the lighter elements. X-rays gas in anode FIGURE 4.9 Schematic diagram of a proportional flow counter cathode SCINTILLATION COUNTERS The term 'scintillation' means the generation of multiple photons of visible light from matter, which has absorbed a higher energy photon, e.g. an X ray. Certain crystalline substances possess this property. The most widely used scintillation crystal is sodium iodide, which has been doped (contaminated) with a about 0.2% thallium iodide. Some organic substances, often in solution, are also used. Each X ray photon that enters the scintillation chamber can produce thousands of visible wavelength photons. These are measured by a conventional photomultiplier tube. Scintillation counters are best used for shorter wavelength X rays, from heavier elements, and so complement proportional counters in their range of operation. Figure 4.10 shows the ranges available from each type of detector. Where they overlap, neither works perfectly, and unfortunately, this is the region in XRF where the commercially important metals such as iron, chromium and manganese produce lines. Å() 0.5 1.0 1.5 2.0 2.5 5 10 20 Scintillation Flow FIGURE 4.10 Detector ranges AIT 4.10

4.6 Qualitative analysis XRF is an ideal qualitative technique for elements above sodium, because of its limited sample preparation. Solids are in fact more easily analysed than solutions. Granular solids need only be ground into a powder, while fixed shape materials only require a polish. The technique is non destructive which means that valuable samples can be analysed without risk of damage. This allied with the portability of an energy dispersive instrument means that works of art and archaeological items can be scanned in situ. Emission lines (weak and strong) for elements are well known, and it is easy to identify the components of a mixture. The rules of intensity trends for the various lines must be followed (Table 4.1). If a line which corresponds to the K emission of a particular element is found, then the K line must also be present, and at about 20 40% of the intensity of the K line. XRF is a surface technique: only the top 100 um (or so) of the sample is analysed. Therefore, for a true overall picture of the sample, it must be made homogenous. However, for a sample with intentional layers, e.g. chrome plated steel, only the surface would be analysed, whereas a technique which relied on solution or powder samples, would require that the coating layer be separated. It is not the most sensitive technique, but the newer instruments are able to detect much lower concentrations than before. Typically it will be able to detect concentrations of 10 mg/kg for elements with K lines present, and about ten times this for those with only L lines sufficiently well for qualitative purposes. The technique does not discriminate at all between different forms of the same elements. Iron gives exactly the same set of wavelengths, regardless of whether it is in steel, iron ore, aqueous solution or an ionic salt. Figure 4.11 shows an XRF spectrum, recorded using an energy dispersive instrument, of a dust sample. It illustrates the ability of the technique to identify the presence of a wide range of elements. FIGURE 4.11 XRF spectrum of a dust sample (from Christian & O'Reilly, Instrumental Analysis, Allyn & Bacon) AIT 4.11

CASE STUDIES OF SOME UNUSUAL APPLICATIONS FOR XRF 1 The authentication of Victoria Crosses The medals for the highest military honour in Commonwealth countries are made from two melted cannons from the Crimean War (1860s). Because of the value (monetary and emotional) of the medals, they can only be analysed by a totally non destructive technique. The cannons have a very specific elemental content, and this allows potential forgeries to be identified. Medals in a number of New Zealand collections were analysed, compared with those of known origin in the Australian War Museum and authenticated. The examination of ancient artefacts from an archaeological excavation XRF was used to determine the composition of gold figurines, weapons, paint from wall murals and clay pots excavated on the Greek island of Thera in the remains of the 1500 BC city Akrotiri which was destroyed by a volcanic eruption in similar circumstances to the Roman city of Pompeii. Analysis of the pigments used in a medieval religious painting To assist with future restoration, a large painting from the 15th century in a church in Seville, Spain, was analysed on location by XRF to identify the pigments. 4.7 Quantitative analysis In many respects, quantitative analysis by XRF is the same as other techniques. Standards are prepared, an optimum line (preferably K) chosen to maximise sensitivity and calibration graphs drawn to find the answer. However, XRF is a technique which suffers very severely from matrix interference. As stated above, the iron K peak may occur at exactly the same position, regardless of the matrix, but its intensity will be very different from sample to sample. This is particularly the case when analysing solids, since the matrix is very concentrated. The causes of the interference include: matrix components absorbing the excitation radiation so that the amount of radiation available to excite the analyte is reduced matrix components absorbing the analyte fluorescence particular prominent where elements a few atomic number lower than analyte are present, eg when analysing nickel (atomic number 28) in steel, the high concentration of iron (AN 26) will reduce the Ni K line intensities significantly, compared to a sample of the same concentration of Ni with no iron fluorescence of matrix components, particularly from heavier elements, can cause additional excitation of the analyte, particle size variations, which cause differences in the contact area between sample and incoming X rays and also the scattering of the fluorescent photons To counter the matrix interference, a number of familiar (and not so familiar) techniques have been employed, but by far and away, the most important is matrix matched standards. The normal drawback with this method is the availability of certified reference materials (standards of exactly known composition), but for some of important XRF using sectors (metals, minerals, ceramics) this is not really a problem as such materials are readily available. 1 The complete articles are downloadable from the subject webpage. The details are not examinable but you may be required to give an outline of the study and why XRF was useful. AIT 4.12

For the other areas (eg soil) there are the normal methods of dealing with interferences: standard addition the most obvious method, but not as easy to do with solid samples; generally a borax melt (see Sample preparation) will have to be used to ensure homogeneity internal standard not usually employed to deal with matrix interference; in this case, the properties choice is an element ± 2 in atomic number from the analyte scattered X ray standardisation where a wavelength of excitation radiation (eg the Rh line) scattered to the detector is used in the same way as an internal standard There is a new development from one instrument manufacturer Panalytical known as standardless analysis, where a large database of many different types of materials (of known composition) is built into the software. A highly sophisticated mathematical modelling process compensates for interferences by the elements detected in the scan. It is not perfect in terms of accuracy, but the great majority of samples and analytical purposes, it is sufficiently good. 4.8 Practical aspects SAMPLE PREPARATION As mentioned above, variations in particle size can affect intensity. Therefore, all standards and samples need to be in a similar physical form and treated equally. Metals should be cleaned, granular material ground into a fine powder. Powdered material is often pressed into a disc, most requiring some type of binding agent to hold them together. A common method for powdered material is the borax fusion disc, where an approximately 10% mixture of sample in borax (Na2B4O7) is heated to melting (above 1000C) and poured into a metal tray to cool. The effect is to break down the sample matrix to some extent, and provide a more homogeneous mix. It does, however, dilute the sample and reduce line intensities. USE OF HELIUM Lighter elements than calcium require a helium atmosphere for best sensitivity because the low energy photons generated by these elements are absorbed by air. FILTERS These are placed between tube and sample, and absorb a particular range of X rays which would otherwise interfere with the sample spectrum. The tube lines can be removed by this method, but at the expense of a good deal of sensitivity. Trial and error is the best way to determine whether a filter is necessary, and if so, which one. TUBE VOLTAGE & CURRENT CONTROL The importance of tube voltage has been discussed above. With current which only affects intensity, it might be expected that the highest current possible is the best choice. However, this is not the case, as a current that is too high will generate a large broad background bump in the spectrum. This is particularly the case if the sample has high concentrations of light elements. AIT 4.13

Revision Questions 1. What procedures are used to ensure the safety of operators of X ray spectrometers? 2. Describe the effect on the output of an X ray tube of (a) current and (b) voltage? 3. How does the spectrum from a X ray tube with a tungsten target differ from that of one with a copper target, assuming the peak intensity is equalised? 4. Why do X ray spectrometers commonly use both flow and scintillation detectors? 5. Draw a diagram showing the transitions occurring that resulting in the production of K and L photons. 6. How do wavelength and energy dispersive XRF instruments differ? 7. Why is XRF such a good qualitative technique? 8. What quantitative technique would you recommend for the analysis of lead containing rocks? Explain your answer. Answers on following page What You Need To Be Able To Do define important terminology explain how X rays are generated outline the health and safety aspects of X ray use distinguish between the three types of X ray spectroscopy describe common X ray sources outline the principles of operation of common X ray detectors draw schematic diagrams and explain the function of each component for typical energy and wavelength dispersive XRF instruments explain the electronic transitions which produce X ray photons by fluorescence describe practical aspects of qualitative analysis by XRF describe practical aspects of quantitative analysis by XRF list advantages and disadvantages of XRF AIT 4.14

Answer guide for revision questions Where the answer can be found directly in your notes, a reference to them will be provided. 1. p4.2 2. Figure 4.2 3. Different positions of the Ka & Kb lines superimposed on the broad continuous spectrum 4. p4.9 5. Lb Ka 6. Wavelength dispersive: two detectors, moving along circular path; energy dispersive: single detector, no moving parts 7. p4.11 8. Matrix matching impossible, so use of source emission line as internal std possible if intense enough; otherwise std addition or added internal std AIT 4.15