Investigations of Atmospheric Aerosol Systems (let us abbreviate AAS) have now spread to many laboratories around the world, mainly because of the important role they assume in the air pollution problem.

In most cases, the studies on AAS attempt to evaluate the so-called “size-distribution function”, through which the general behavior of such a system is commonly described.

During recent years energy dispersive x-ray fluorescence (EDXRF) has been used to measure the elemental content of atmospheric aerosols. The code described here is used to reduce EDXRF data and determine the elemental composition of samples collected on membrane filters. The program has been specifically written for EDXRF analysis of size-fractionated aerosols collected by a dichotomous sampler.

The x-ray fluorescence spectrometer used in our laboratory employs a pulsed-mode x-ray tube and a lithium-drifted silicon detector. Pulse-height spectra are produced for elements ranging in atomic number from Z = 13 to Z = 82 (corresponding to an energy range from 1.4 to 32.1 keV). Approximately uniform x-ray production is attained by producing independent spectra from three secondary targets (Ti, Mo, Sm).

For the past several years the Boeing Aerospace Company has been implementing advanced nondestructive chemical analysis methods to improve product reliability and reduce material inspection costs. Previous testing of incoming material for conformance to vendor test reports or of production materials for verification of alloy composition, had consisted of either time-consuming destructive testing or nondestructive chemical spot testing, which often was insensitive to differences between alloys of similar chemical properties. Beginning in 1974, development of EDXRF techniques was initiated to provide a rapid nondestructive analysis capability for both laboratory and factory use. For materials containing elements easily excited by EDXRF methods, costly destructive sampling and testing can be avoided. Generally, chips, wire, barstock, sheet or plate can be analyzed using an annular radioactive source. The uniformity of the X-ray flux diminishes sample geometry and surface roughness effects.

In the event that nuclear fuel from light water reactors (LWR) is reprocessed to reclaim the uranium or plutonium, several analytical techniques will be used for product accountability. Generally, the isotopic content of both the plutonium and uranium in the reprocessed product will have to be accurately determined. One plan for the reprocessing of LWR spent fuel incorporates the following scheme. After separation from both the fission products and transplutonium actinides (including neptunium and americium), part of the uranium and all of the plutonium in a nitrate solution will merge together to form a coprocessed stream. This solution will be concentrated by evaporation and sent to a hold tank for accountability. Input concentrations into the hold tank could be up to 350 g U/ℓ and nearly 50 g Pu/ℓ. The variation to be expected in these concentrations is not known. The remaining uranium fraction will be further purified and sent to a separate storage tank. Its expected stream concentration will be about 60 g U/ℓ. These two relatively high actinide stream concentrations can be monitored rapidly, quantitatively, and nondestructively using the technique of energy-dispersive x-ray fluorescence analysis(XRFA).

A procedure for determining niobium in uranium-niobium binary alloys was developed. The procedure requires only that a compact be made by pressing turnings, and wet chemical treatment is avoided entirely. By eliminating the wet chemistry, time is reduced by a factor of at least three compared with previous practice.

Standardization was accomplished by separating each batch of machined turnings from specially cast uranium-niobium alloys into individual strands. A composite portion of each strand was analyzed for niobium by wet chemistry. The balance of the turnings from each batchy was pressed into a compact. Seven standard compacts resulted. The average Intensity ratio of niobium-to-uranium vs. the percent niobium for fourteen determinations on each standard compact was used to calculate a correlation coefficient of 0.999.

The standard deviation of the Intensity ratios between duplicate compacts (average of 52 samples) was less than the standard deviation found between aqueous duplicates (average 125 samples) by a factor of 4.

Occupational exposure to lead is common in many industrial applications and hence it is of considerable medical interest to control the body-burden of lead in living man. More than 90 % of the lead in the body is concentrated in bone and hence in vivo measurements of the lead in the skeleton should give the most satisfactory way for estimating the body-burden. The routine method used today for checking on lead contamination is that of measurements on blood samples. However, since the concentration of lead in the blood is a sensitive function of the actual exposure conditions, this method provides only a poor indication of the total body-burden and the integrated lead exposure.

The P-L2,3 emission spectra of the phosphorus compounds were obtained using a secondary excitation. We found that the spectra depended upon the chemical state, the main peak of L2,3 emission spectra shifts to a higher energy, and the peak of intensity of higher energy increases with an increase of oxidation number. It should be noted that the phosphorous spectra of the phosphate anions scarcely reveal influences of the phosphate compounds, except that of H3PO4. We concluded that the spectral feature for the phosphorus compounds is influenced by the condition of surrounding atoms but is not influenced by the bond condition.

In this article we present the experimental data on L3-absorption spectra of metallic tin. and tin in SnO2 . We have used (i) a bent crystal spectrometer and (ii) a high resolution flat crystal spectrometer with a microfocus target as the source for the continuum radiation. We have observed a “white line” on the low energy side of the primary L3-absorption edge of Sn in SnO2. Similar “white lines” in absorption spectra were reported earlier by M. Siegbahn in rare earth elements with f-vancies, who designated such lines appearing immedlately on the low energy side of the primary absorption edge as “white lines” in absorption spectra. It is interesting to note that metallic tin does not reveal the “white line”, while Sn in SnO2 clearly revealed a sharp white line on the low energy side of the primary absorption edge.

In order to measure changes in composition as a function of distance (macrosegregation) in directionally solidified two phase samples, a well collimated incident x-ray beam is required for XRF analysis. This is accomplished using Bragg diffraction of AgKα radiation from a highly perfect Si crystal. Because the incident beam is also monochromatic, additional advantages are realized: a) the backgrounds caused by Compton and thermal diffuse scattering (TDS) of the incident beam are well localized in the energy spectrum and do not interfere with the fluorescent peaks, b) the TDS can be used as a monitor of the incident photon flux and hence eliminates often substantial errors caused by incident beam intensity fluctuations.

Using several prepared standards, the ratio of PbL counts to TDS counts was found to be a function of the total Pb content of the two phase microstructure, with a reproducibility determined only by counting statistics. Furthermore, the function was found to be nearly linear over a wide range of compositions. Standard methods of absorption or enhancement correction can be employed using this ratio. The spatial resolution, determined by profiling a sharp discontinuity between two metals, was 0.5 mm.

Macrosegregation data is presented for Pb-Sn two phase alloys whose compositions range from 35 wt % Pb to 70wt % Pb. Comparison of compositions with those determined by a titration method agrees to within 2 wt % for most of the metallurgical structures present in the work. Somewhat larger deviations were found for samples with high Pb contents with extremely coarse two phase microstructures.

Interest in the analysis of inks on paper has come primarily, although not exclusively, from forensic scientists. Two classes of problems are encountered: comparison of inks in different regions of a single document, to detect additions (such as adding a “ty” to your handwritten check for “nine” dollars) made with a visually similar but hopefully compositionally distinct ink; and verification of authenticity in documents (including stamps, stock certificates, currency, etc.) by the presence of specific tag elements in certain of the inks present.

It is feasible to investigate the thickness of oxide layers on silicon wafers by X-radiation in the 0.1-10 nm thickness range. For example, X-ray photoelectron spectroscopy (XPS) is a well applicable technique, with information depth of a few nm. Fig. 1 presents the principle of this method. An impinging characteristic X-radiation hν (e.g. Al Kα) count rate ejects Si 2p photoelectrons from the Si-substrate (d), with count rate n2, which, on their way to the electron spectrometer, have to pass through the SiOx-interface (c), the SiO2-layer (b) and the contamination overlayer (a), whereas Si 2p photoelectrons ejected from the SiO2-layer, with count rate n2 have just to penetrate the contamination overlayer. The Si 2p electrons originating from the SiOx-interface, for the situation shown in Fig. 1, can be added to the substrate count rate.

Primary beam transmission filters in energy dispersive X-ray fluorescence (EDXRF) analysis are used to shape the spectral output of the X-ray tube. The effective use of these filters allows the optimization of excitation conditions for each different analysis. Filters are used in two basic ways in EDXRF; either as edge filters or as white filters. The proper choice of filter and excitation conditions optimizes the analysis of a particular element or spectral region by shaping the primary radiation to reduce background and to maximize excitation.

Many x-ray tubes, used by crystallographers and others, operate with the aid of a tungsten filament in the region of 2500°K, The high operating temperature results in evaporation of the filament material with two serious consequences. The first is a finite but relatively short lifetime. The second is contamination of the target and windows with tungsten. In addition, if the tube is of the demountable type, connected to an oil-diffusion pump and a mechanical fore-pump, carbonaceous deposits can be a problem. In a typical tube, the filament is mounted within a centimeter or two of the target. The resulting radiant heating of the target presents additional cooling problems especially with low melting-point targets. Many if not all of the above objectional features are circumvented by a plasma controlled x-ray tube using a low pressure atmosphere of helium and a cage-like cathode fabricated from nickel wire-mesh. An experimental model has been operated for several hours at 15 kv and 10 ma on an aluminum target. Scaling up of the apparatus will permit power dissipations in the kilowatt range limited mainly by the available power source or vaporization of the target material.

We have previously reported on the uniqueness and potential of room-temperature spectrometry of low-energy x-rays with a mercuric iodide (HgI2) detector (1,2,3). In this paper we emphasize the use of HgI2 detectors for x-ray fluorescence (XRF) analysis.

Because no vacuum plumbing or cryogenic cooling is required, the design of a mercuric iodide room-temperature x-ray spectrometer is extremely simple. Our present design consists of coupling a detector directly to the first-stage FET in a modified Tennelec 161 D preamplifier and making the configuration “light-tight”. Aside from providing a suitable entrance window, there are no other requirements for routine spectroscopy.

Geological samples prepared for analysis by X-ray fluorescence (XRF) are particularly prone to intrasample and intersample phase variations creating significant inhomogeneity and variations in matrix effects. These features may “best be eliminated or diminished by fusion of the sample with, a suitable flux (1,2). Presentation to the X-ray spectrometer may be as a briquet of the comminuted fusion head (1,3,4,5) or as a molded glass disc or wafer (l,2,5,6). Because of the stability of glass discs and the necessity to eliminate possible contamination during grinding and making briquets, molded glass discs appear to be the best and easiest means of presentation.

Element mapping has been so far an analytical field covered by those investigation methods using a scanned particle beam on the excitation side of the specimen. These methods include localized interaction processes, as in the electron microprobe, secondary ion mass spectrometer and Auger electron spectrometer. On the other hand, methods which use a stationary radiation beam on the input side (non-localized interaction) usually give summarized information about the sample area (X-ray fluorescence spectrometry, photoelectron spectrometry, …). This paper deals with an approach to extending the application field of X-ray fluorescence analysis (XRFA) to element mapping. Fig. 1 shows the main components of the X-ray imaging system. Since modern SRF systems often work under computer control, the only additional hardware components of the system are a scanned sample holder and an X-ray source producing a line-shaped X-ray beam. The computer is usually not used for ease, speed or convenience of operation, but is used to generate the element image and process all the collected fluorescence data.

Identification or recognition of an unknown is a problem encountered in areas as diverse as X-ray diffraction, signature and voice verification, fingerprint identification, and spectroscopy(mass), infared(IR), ultraviolat(UV), Auger). Features uniquely characteristic of the unknown, which are used in the identification process, are called the “pattern” of that unknown. In a typical identification, the one or more sub-patterns (components) making up the unknown pattern are identified by a search through a set of known (standard) patterns to yield potential matches. This searching and matching can be done sucessfully by hand, especially if the number of patterns in the file of standards is not too great. When the number of phases which consititute the unknown pattern becomes extremely large the use of computers becomes a practical necessity if the search is to be conducted in a reasonable time. A comparison of X-ray search-match approaches with each other and with search-match capabilities existing in other fields may lead to better X-ray phase identification techniques.

The use of the computer controlled powder diffractometer for the qualitative and quantitative analysis of multi phase powder mixtures is becoming relatively common place, and we ourselves have described in some detail the hardware and software components of typical systems 1,2) of which the AFD 3600 is the latest version 3). Although great success has been achieved in the use of the minicomputer for machine control, data acquisition and display, plus some degree of data workup, less success has been gained in search/matching of unknown powder patterns, although computer programs have been available for a number of years which utilize relatively large computers for automatic search/matching e.g. 4,5) the success with which these programs have been applied has been somewhat disappointing. Much of the difficulty arises from the marginal quality of data in the JCPDS index and from uncertainties in the accuracy of the experimental patterns.

Modern (>1950) data analysis techniques are commonly applied in many scientific fields but their careful application in other fields, including X-Ray diffraction, has been notably lacking or generally confined to the environment of university research. Perhaps this lack of widespread application in XRD can be blamed on the absence of sophisticated computer controlled instrumentation, and, if this is so, the situation should change rapidly, since today's automated diffractometers are driven by very powerful minicomputers with large firm disks and sophisticated operating systems. In such an environment it is entirely practical to implement, on-line, state-of- the-art data analysis techniques, and in this paper we present one such example.

Indexing of powder-diffraction patterns by computer techniques has advanced to the state that it is now often possible to determine unit-cell dimensions and crystal system for an unknown material solely from its powder-diffraction data. This indexing is fully automated, proceeding directly from positions of observed diffraction lines as input, with decision-making steps being made by a computer. Ease of indexing depends on quality of data (accuracy and completeness), volume of the unit cell, and symmetry of the crystal system. In general, a powder pattern of a triclinic compound with a large unit cell requires a more accurate and complete data-set for successful indexing than does a cubic material having a small unit cell. Fortunately, data from a well-aligned diffractometer or Guinier camera ordinarily suffices for computer indexing. Because of systematic errors in the low-lying diffraction lines, data from the Debye-Scherrer technique usually are not adequate for computer indexing (except for the simpler cases). A brief review of the strategies/algorithms of some of the computer indexing codes now available is given. Criteria for assessing the reliability of a particular computer- assisted indexing are discussed. Finally, attention is directed toward future developments such as by automating the collection of powder- diffraction data, analyzing data by computer data processing, and increasing the speed and reliability of computer indexing.