Developments during the past few years in solid-state radiation detectors and low-noise electronics employing field-effect transistors operated at cryogenic temperatures have resulted in the availability of high-resolution energy-dispersive spectrometers for a variety of applications in x-ray spectrometry. Using pulseamplitude analysis techniques, these spectrometers make it possible to obtain multi-elemental analyses on a routine laboratory basis employing x-ray fluorescence techniques. The combination of these spectrometers with small, inexpensive on-line computer data systems makes It possible to obtain rapid on-line qualitative and quantitative analysis of samples in the laboratory and in special field applications. A general review of the present state of development in detectors, electronics and on-line data systems will be presented together with descriptions of applications of such equipment in the laboratory.

The need to detect low intensity beams of resolved ions with keV energy is encountered frequently in mass spectrometry. It is most readily accomplished by using some form of pulse-counting detector in which ions are allowed to strike a target surface (conversion dynode) where each ion releases a pulse of low energy secondary electrons. The secondary electrons undergo further amplification in an electron multiplier or scintillator type of ion detector which is specially designed to detect single ions. Ion-to-electron conversion, amplification of the secondary electrons, efficiency of ion detection by pulse counting and dead-time corrections to observed random counting rates are all statistical processes which must be understood so that observed data can be corrected properly to obtain precise and accurate measurements. These processes are reviewed, along with some of the basic properties of ion-induced secondary electron emission from an amorphous Al2O3 thin film target. Ion detection in the ion microprobe is interpreted in light of these basic considerations, and several mass spectra are given for electron multiplier dynode surfaces.

This paper is a review of automation of electron microprobe and x-ray fluorescence instrumentation. Such a review seems timely because of the great increase in the application of computer systems in this field over the past decade. Some of these applications have been conceived to meet true technological needs while in other cases they have “been undertaken to “keep up with the Joneses.” I would like to show not only what automated systems are now feasible but also when and how they should he employed. The “when” and “how” of automation are largely dependent upon the application being considered; in this study, x-ray applications have been divided into the following classes;

• (1) on-stream process-control,

• (2) off-line quality assurance,

• (3) routine service laboratory,

• (4) general purpose analytical laboratory.

The use of on-line computers for control and acquisition of data from x-ray and neutron diffractometers has continuously improved and expanded. Systems vary from a small 4K core computer to a time-sharing system with a medium or large computer. The choice of a single time-shared computer or an individual standalone system must be based on one's own particular environment. As large high-speed electronic computers became available, increasingly complex chemical and magnetic structures have been analysed and solved; this has created a demand for rapid, reliable, and versatile means of obtaining diffraction data. Since small computers have been developed at reduced cost and with increased storage capacity, they must be considered for use in diffraction experimentation. Therefore, in x-ray and neutron scattering, small computers are needed for data acquisition and large computers are needed for data analysis.

A unique approach has been used in the construction of a vacuum-two-crystal spectrometer. The instrument uses digital shaft encoders directly coupled to the analysing crystal axes to give accurate absolute angular measurement over a 360° revolution of each crystal. Each axis is driven directly by a torque motor tachometer combination which allows precise speed and position control. The complete two-crystal carriage assembly rotates about the first crystal axis, thus permitting the x-ray tube to be fixed to the vacuum chamber. The carriage and detector rotations are driven through antibacklash gearing by smaller shaft encoder, torque motor, tachometer assemblies. Because the complete unit is housed within the vacuum chamber, there are no alignment changes on system pump-down. The spectrometer is part of a completely automated computer controlled system. The computer is used for data acquisition and analysis as well as for controlling the angular position of the analyzing crystals, detector, and carriage.

The instrumentation and software for performing X-ray intensity measurements with a paper tape controlled diffractometer are described. The hardware includes two addressable axis positioners which control Slo-Syn stepping motors on the 2θ and ω) axes of the diffractometer, an addressable scaler-timer, a multiaxis programmer and a Teletype, in addition to the normal counting electronics. This system may be manually controlled with front panel switches or with instructions entered on the Teletype. In the automatic off-line mode instructions for motor speed, motor direction, starting angle, final angle, angular increment and scaler preset (time or counts) punched on paper tape are read and executed in sequence. A Teletype output of 2θ and ω angles, time and counts is obtained at each step. This off-line system was used for the measurement of austenite in H12 hot work die steel austenitized at various conditions and which contained a maximum of 13% austenite. A helium chamber was used to extend the limit of detection to 0.4% austenite. The X-ray analysis involved measuring the areas of the (200) austenite diffraction line and the (200) martensite line. For each of these lines, the system was programmed to integrate the counts over angular intervals corresponding to a low-angle background, the peak and a high-angle background using the ability of the axis positioner to stop the scaler-timer at the end of each angular interval. The additional capability of slewing rapidly between the various diffraction lines reduced the time required for automatic data collection. The present off-line system can be used to simplify other types of X-ray diffraction analysis such as residual stress and microstrain/particle size determinations, since the manual data handling can be eliminated with the computer compatible punched paper tape output. Future development of this instrumentation includes direct computer control of the diffractometer and computerized data reduction, with the advantage of a paper tape back-up system.

An IBM 1800 time-sharing system is used in our X-ray laboratory to control a four-circle diffractometer for structure research, several powder diffractometers, a pole-figure goniometer and a microdensitometer along with other instruments outside the diffraction area. A survey of the computer system is given and the hardware necessary to automate the diffractometers is discussed. The computer supervision ranges from simple data-logging with a minimum of control to complete control of all actions depending on the diffractometer and the requirements of the experiment. Also described is the use of the computer to process the data and to perform background jobs.

The Differential Omega Scanning Technique (DOST) was developed to improve topographs obtained from silicon wafers which generally possess process induced flexure during the manufacture of semiconductors products. A commercially available Lang camera has been electronically instrumented to provide an automatic adjustment of omega motion of the specimen on the goniometer axis. The position of the specimen relative to the primary beam (theta) is continually altered by DOST to assure maximum, or near maximum, diffraction intensity with changes of flexure of the wafer as the specimen is scanned. The resulting topographs have near complete images and appear to have improved quality. Using techniques developed within the x -ray laboratory, routine acceptable topographs are obtained in only the time required to scan the diameter of the specimen one time.

A modular x-ray and electron beam system has been constructed to accomplish most standard types of x-ray analysis automatically. The system includes a vertical powder diffractometer, a vacuum spectrograph, a four-axis single-crystal diffractometer and an electron microprobe with four spectrometers. The system is controlled by a single computer with auxilliary drum storage. The powder diffractometer has automatic goniometer drive, automatic sample changing with computer cataloging of samples and automatic variable divergence slit. The x-ray spectrograph is a standard unit with added 100 KV capability, automated goniometer drive, eight or 32 position sample changer, six-analyzing-crystal changer, collimator changer, and detector changer. All functions including sample number cataloguing are computer controlled. The four-axis single crystal diffractometer has all functions computer controlled and includes peak search option. The electron microprobe/scanning electron microscope has all spectrometers, x, y, z stage motions and sample changer controlled by computer. X-Ray channels, beam current and sample current are automatically read-out. Control of probe functions and correction of mechanical errors in stage operation, are aided by a digital beam scanner with a light pen. Counting channels and computer interfacing are modular, and channel components are interchangable by the operator. Computer hardware may be expanded and the software is modular. Any combination of instruments up to eight may be operated.

The computer operates all instruments simultaneously in a time-sharing mode. Programs consist of operating system programs, individual instrument programs and data processing programs. The operating system drives all x-ray hardware, fetches programs and data from the drum, allocates data storage, queues final results for printing and performs other necessary executive functions. For qualitative analysis, spectra for each instrument are collected automatically in several modes and are stored on the drum. A variety of methods are used to remove background, strip unwanted peaks and find peak position and intensity. A.S.T.M. diffraction, spectrographic and special files are maintained on the drum. Quantitative analysis is executed automatically using a version of "MAGIC" on the electron microprobe and a multiple linear regression method for the spectrograph and diffractometer.

An automated- electron microprobe system using an on-line 8,000 word minicomputer has been designed. The general principles and need for such a system is discussed. The microprobe-computer interface hardware consists of computer addressable axis positioners, scalers, timer and a digital to analog converter. The axis positioners are used to position the spectrometers and specimen stage, whereas the digital to analog converter is used to position the electron beam in a 1000 × 1000 point matrix. The digital to analog converter is also used to drive a X-Y recorder for computer plotting of calibration curves, pulse amplitude distributions, X-ray line profiles and wavelength scans. System software is written in the interactive CLASS language which was designed specifically for laboratory instrument control. The language permits large computer programs to be chained through a minicomputer while passing variables from one function to the next. The operating software consists of automatic peak location, sequencing of data collection, statistical analysis of data, and correction of data for fluorescence, absorption and atomic number effects. An automated quantitative analysis of several copper-gold alloys and a computerized test for specimen homogeneity is described to demonstrate system operation. Future automation considerations are also discussed.

An x-ray fluorescence analysis method applicable to the case of fluorescent spectra excited with monoenergetic x-rays has been developed. The technique employs a minimum number of calibration steps using single element thin film standards and depends upon theoretical cross sections and fluorescent yield data to interpolate from element to element. The samples are treated as thin films and corrections for absorption effects are easily determined- Enhancement effects, if not negligible, are minimized by sample dilution techniques or by selective excitation.

Two years ago a paper on a new method of X-ray fluorescence analysis was presented. This method uses a variable take-off geometry and permits a quantitative analysis without calibration curves or the knowledge of the primary X-ray spectrum. As references either chemically analysed samples or pure elements are used. In case that only primary excitation exists, under the aspect, mentioned above, the method is to be regarded as “absolute”. For secondary excitation a simple parabolic approximation has been introduced, which requires a calibration. This paper gives an outline on the progress obtained during the last two years.

Although isotope excited non-dispersive X-ray analysis has shown considerable promise for use in portable instruments and in on-line control applications, its use as a general analytical tool in the chemical laboratory has not developed as well as expected. Much of this is thought to be due to lack of flexibility in the instruments available and to difficulties with their use by unskilled operators. Wide element coverage is essential and many analytical problems arise at the light element end of the periodic table. While developing on-line techniques for light elements it was found that a measuring head using a very compact source-sample-detector geometry was well suited to a number of routine analytical laboratory problems, and enabled high precision to be obtained with minimum sample preparation for rapid single element analysis. Sensitivities in a light matrix from 100. p.p.m. for silicon to 10 p.p.m. for heavy elements were attainable. An instrument specifically for chemical laboratory use was therefore designed to accept solid, liquid and powder samples for single element analysis of silicon and all heavier elements. The instrument can give direct numerical readout of concentration and achieves high stability and precision. In its basic form it is equipped with a range of isotope excitation sources but will also accept a small X-ray tube which has been specially developed for it to provide higher sensitivity and precision. The measuring head may be removed for remote use, and multiple sample changing facilities with printout of results can be added, permitting unattended automatic routine analysis.

Energy dispersive x-ray analysis has become an extremely useful analytical tool. The technique provides for the direct observation of x-ray emission spectra, eliminating the need for a dispersive crystal. The purpose of this reported investigation was to study the use of the technique with a simple pulse height analyzing system and to develop a routine method for correcting Interferences due to adjacent element spectral overlap and matrix effects.

The analyzing system consists of a radioisotope source, a lithium drifted silicon detector, a preamplifier, an amplifier, two single channel analyzers and two digital ratemeters. In order to obtain results suitable for quantative measurement, a two-step empirical method was employed for the correction of peak overlapping and matrix effects. If two peaks in a spectrum overlap at their tails, one can set up a channel width of the analyzer to a region where there are no overlapping pulses. It is then possible to calibrate the ratio of the intensity obtained from this channel to that obtained from the whole peak in its pure state, i.e. without the appearance of a neighbor peak. The actual intensity of the peak in the overlapping spectrum is, therefore, the observed counts multiplied by the ratio. The next step is the correction of matrix effect by means of conventional empirical methods using standard samples. Two types of the samples, Zn-Cu powder mixtures and Ee-Cu in aqueous solutions, were studied to illustrate this method. The usefulness of applying the analyzing system and technique to industrial measurements, either on-line or batch, will also be discussed.

Previous investigations on the quantitative determination of sulfur, chlorine, potassium, calcium, scandium and titanium in aqueous solutions by a radioisotopic excited fluorescent spectrometer has been extended to include other elements which are very difficult to separate and determine quantitatively by chemical methods. Six elements taken for the investigation and some of the results to be presented in this paper are:

• (1) zirconium,

• (2) hafnium,

• (3) niobium,

• (4) tantalum,

• (5) molybdenum and

• (6) tungsten.

In the earlier investigations which have been reported in this conference, lighter elements (atomic numbers ranging from 16 to 22) were used for the investigation. In the present studies, however, comparatively heavier elements have been used. Therefore a radioisotope such as iron 55 used earlier is not suitable because it cannot excite the K x-ray of these elements. To excite the K and L of these elements, we use the radioisotope iodine 125. The advantage of using this radioisotope is that it is inexpensive and commercially available although its half-life is comparatively short.

The spectrometer used with further improvements has been described and presented earlier. We used a multi-channel analyzer of 1000 in the present investigation, A liquid cell was specially designed for thisr study. Chemicals used for preparation of solutions were of reagent grades. Some of them had to be specially prepared. For example, hafnium, often contaminated with zirconium, was specially prepared and checked spectroscopically. Some difficulties have been encountered in preparing concentrated solutions such as niobium and tantalum due to the inherit characteristics of these elements to form insoluble compounds. Procedures will be described for the preparation of these solutions. Instruments used and results will be presented in this paper.

X-ray fluorescence spectroscopy has been in use since the early days of the twentieth century, when Moseley confirmed the order of the chemical periodic table. However, fluorescence spectroscopy until recently has depended on diffraction methods to obtain sufficient resolution. Intrinsic resolution of ionization chambers, scintillation detectors, and proportional counters is inadequate for discrimination o f lines due to adjacent elements of low atomic number. The advent o f solid-state detectors, especially those using lithium-compensated silicon and low-noise electronics, has recently brought intrinsic energy resolution to the point where lines from adjacent elements as light as carbon and nitrogen can be resolved in theory; and detection of K radiation from elements as light as sodium is practical. Thus the solution to the long-standing problem of an adequate detector is at hand, and energy-dispersive spectrometers are now feasible.

X-ray diffraction patterns using continuous radiation from copper and tungsten target x-ray tubes and detected with a Si(Li) energy analysis system are presented. Errors caused by a misaligned diffractometer and x-ray penetration into the sample are shown to be more difficult to correct and larger in magnitude than errors arising from energy calibration. All these errors can be minimized by mixing a standard with the unknown sample.

The energy resolution of the detector influences the breadth of the diffraction peaks more strongly than the standard slit systems available with commercial diffractometers. Thus, to reduce the recording time and maintain the same standard deviation for the data, one should increase the sizes of the front and receiving slits including the Soller slits. X-ray energy diffraction patterns can be recorded with standard deviations less than +0.001 Å in the d spacing with only 200 sec measurement time using the standard diffractometer slit system. Copper targets are probably as useful as tungsten even though the continuous intensity is about three times.less. Copper has fewer interfering characteristic lines, and its use permits convenient conversion to normal θ scanning diffractometer operation.

A system, has been designed and tested for rapid energy dispersive diffractometry and simultaneous fluorescence analysis. A turntable composed of the sample chamber with attached air-cooled x-ray tube allows the 2θ angle to be varied with respect to the stationary Si(Li) detector. Data for most analyses can be obtained in one minute per sample. Results are stored in the memory of a multichannel analyzer and are read out on a CRT, strip chart recorder or tabulated in digital format by a computer.

Fast X-ray fluorescence analysis with radioisotope excitation requires intense sources to produce reasonable counting rates. The inconvenience of handling such sources and the small number of suitable radioisotopes places severe limitations on their use.

We have explored the possibility of using low-power X-ray tubes as exciting sources for energy-dispersive fluorescence analysis. The principal advantage to X-ray tubes is the ability to produce X-ray fluxes to three orders of magnitude higher than those obtained with convenient radioisotope sources while dissipating only a few watts in the tube. Furthermore, the variety of possible anode materials and range of currents in the tube make possible optimum choice of exciting energy and intensity for particular applications.

We have designed and tested such tubes in a variety of anode configurations suitable for fluorescence excitation. Using either X-ray filtering techniques or multiple fluorescence geometries It is possible to significantly reduce the Bremsstrahlung background relative to characteristic radiation.

As compared with normal radioisotope-target assemblies, excitation of a sample by the X-ray tube results in comparable sensitivity in only a tenth to one hundredth of the time.