The results of recent work of the Millikan Laboratory of Physics on the physics and application of ultra s oft X-radiation s are summarized. These investigations include (1) the compilation of ultrasoft X-ray interaction coefficients. (2) the development of high-intensity, low-voltage X-ray tubes (100 to 1000 v), (3) the application of pulse-height analysis in flow proportional counter measurement of 10 to 100 A radiations, (4) a detailed study of the photographic action of the ultrasoft X-radiations, and (5) the application of these long wave length X-radiations to micromass and microchemical analysis by mïcroradiographïand total-reflection techniques.

Three beta-excited X-ray source types (the compound, apposition, and mixture source) have been designed, and the photon spectra compared. Each source in a series was assembled using a single target element with the concentration of pure beta emitter held constant for the series. The apposition source was studied with varying target thicknesses. The compound sources were crystalline precipitates of molybdate and tungstate groups compounded with the nonearrier-free beta emitter. The mixture source consisted of a fine-milled target powder with the beta emitter adsorbed on it. Assay of the X-ray source emission has been done on the Radiation Instrument Development Laboratories (RIDL) lOQ-channel analyzer with a 3-in. Nal detector. Variation in the spectra exists between the source types of a series. Anomalous peaking occurs with the Sr90 MoO4 source emission at 29 key. The compound source, potentially productive, demonstrates poor characteristic spectra.

A recently developed Dutch commercial projection X-ray microscope will be described. The principal features are: simplified focusing by using reflected electrons, four different target materials exchangeable during operation, and hlnocular viewing of either the specimen itself or the fluorescent image. The specimen and film chamber is evacuated and the voltage is variable between 5 and 20 kv. Some results will be shown.

An X-ray absorption method was utilized to determine rapidly and precisely the thickness and weight of inorganic coating layers applied to paper. The mass absorption coefficient μ/ρ was calculated for kaolins and substituted in the following formula

where I is the X-ray intensity after passing through a medium whose thickness is x centimeters, I0 Is the X-ray intensity before passing through a medium, μ. is the linear absorption coefficient, and ρ is the density of the absorbing medium.

The film thickness was calculated and related to coat weight. Coated sheets of variable basis weight, utilizing different adhesives and coat weights, were used. The method is rapid and coating thicknesses can be determined with an accuracy of plus or minus 0.5μ and coat weight of plus or minus 0.2 lb. It is believed that the method can be adapted for production control on the paper coating machines.

X-ray absorption, having been earlier demonstrated to be a unique tool for the accurate measurement of density profiles within gas-fluidized solid systems such as those employed in the chemical and related industries, is being applied to the measurement of steam voids in water flowing through electrically heated rectangular stainless steel channels under conditions of pressure, temperature, flow rate, and heat flux corresponding to “local boiling” (subcooled water) as well as “bulk boiling” (saturated water) in the coolant channels of water-cooled nuclear reactors.

A balanced dual-beam photometer system is employed. A dual-beam X-ray tube and two scintillation-crystal-photomultiplier-tube detector probes are mounted in a manner permitting remote-operated traversing of the test channel in the horizontal as well as vertical directions. The photomultiplier tubes are used in a difference circuit with one of the tubes monitoring the test channel and the other a remoteoperated reference wedge. The difference-measuring connection offers common made rejection so that any differential output Is due to a change in attenuation of the testchannel beam. This output is fed through a differential amplifier to a strip-chart recorder.

With a test channel consisting of 0.020-in-thick stainless steel walls, backed by 0.062-in. beryllium windows for strength, void concentrations within a 0.250-in, thickness of water currently can be measured with an absolute accuracy of 2% or better (which is far superior to that obtainable with previous techniques) and further improvements in accuracy appear likely. The test channel is enclosed in a heavy steel housing with windows, permitting void measurements to be made at pressures up to 2000 psia.

Calibration procedures and data are presented for plating thickness measurements of gold, silver, tin, and nickel on copper, brass, and steel basis metals. Both emission and absorption techniques are described, and examples are given of measurements on thicknesses from 10 to 600 millionths of an inch. The use is shown of masking and integrating techniques for curved or unusual shapes.

For the special problem of duplex platings, which have two metals each varying in thickness, the Kα/Kβ absorption ratio and the Lα/Lβ emission ratio are compared with conventional absorption and emission techniques.

Optical and chemical determinations are shown to be less precise as well as considerably slower than X-ray measurements on the same piece. Examples are given of the use of X-ray measurements to study fundamentals of plating processes, such as barrel-plating thickness distributions and variation of thickness across a surface, as well as to study processes such as wear or porosity.

Emission-concentration proportionality Is one of the main features of electronprobe mlcroanalysis. It can be interpreted by taking simultaneously Into account the Z/A factor of Webster's law of deceleration and the variation with atomic number of the backscattering ratio. The baclcscattering ratio itself can be deduced approximately from experimental distribution curves of thecharacteristicemisslonjejqpertments on copper-gold alloys show good agreement with the theoretical predictions. An important practical consequence is that regulation of the electron beam falling onto the specimen leads to better results than regulation of the current passing through it. The different corrections are discussed, with special reference to the fluorescence induced by the general continuous spectrum. The case of insulating materials is considered separately; preliminary experiments seem to show that their behavior is not essentially different from that of conducting materials.

Pulse-height analysis is a valuable tool in X-ray fluorescence analysis, both for qualitative and quantitative purposes. The elimination of higher order interferences permits determinations that would otherwise be very difficult or impossible, The systematic application of pulse-height analysis in qualitative work greatly simplifies the interpretation of complex spectra. In certain cases one can apply nondispersive analysis, relying on the pulse-height analyzer alone for separating energy levels of X-ray photons. Technique and limitations of pulse-height analysis will be discussed.

Selected substances have been examined by X-ray fluorescence to determine the effect of particles or surface irregularities on the fluorescent intensity. The powdered substances examined were segregated into particle size classifications by the use of standard and micro sieves and ocular microscopic measurements. The substances thus classified were Iron, germanium, and titanium metals, copper furnace slag, and a cobalt molybdate catalyst. Graphical representations are used to characterize the relation between particle size and intensity. In certain cases the influence of particle size is expressed in terms of a calculated absorption length of the substance. The extent to which particle size in a powder affects the fluorescent intensity is a function of the absorption coefficients of the substance for the exciting and fluorescent X-rays, The effect on intensity of successive powder dilution of a high atomic number witha low atomic number substance has been measured. The results are of assistance in providing a concept of the structure necessary for maximum intensity. Little difference in the fluorescent intensity of an element is exhibited as between its dilution in a dry powder and in an aqueous solution, if the absorption properties of the two diluents are very similar. The effect of small surface irregularities in a metal exhibits no systematic relation to the fluorescent intensity.

Fluorescent X-ray spectrographic analyses have been made for the oxides Fe2O3 caO, MgO, K2O, TiO2, and MnO on samples of amphibolite rock. In addition, the above oxides and SiO2 and Al2O3 have been determined in a suite of constituent hornblendes from these rocks.

Classical wet chemical analyses of materials of the same minera logical composition as the samples are used to establish the working curves. This eliminates the necessity for Internal standardization or other treatment of the samples. The only sample preparation involves extremely fine grinding before analysis.

Precision of the analyses is equivalent to, or better than, that achieved by wet chemical methods.

With modern X-ray instruments, precision is generally satisfactory for the analysis of MgO, Al2O3, SiO2, CaO, and Fe2O3 in finished and raw-mix cements and other similar materials. Accuracy, that is, agreement between cheraicaland X-ray values, quite frequently is much worse than precision. Some typical examples illustrating this point are reviewed.

In the soft region, generally, it has not been possible to assign the cause of iuaccuracy to interelement effects. The hypothesis is formulated that in the soft region this problem could be due to mineralogical differences and inhomogeneity on a micro scale. Theoretical calculations support the plausibility of this argument.

To improve accuracy, present techniques have been based on type standardization, acid digestion fusion using a high flux-to-sample ratio, etc. In general, these methods have certain basic limitations. In order to retain useful intensities and lineto- background ratios for the soft region, 8 new fusion technique has been developed. The experimental data based on this method of reducing micro-inhomogeneity and eliminating minaralogical differences indicate improvements in accuracy. In the case of finished cements the improvements in accuracy have been quite modest. With rawmix cements, startling improvements have been obtained for the analysis of SiO2 and CaO.

A highly simplified multielement calibration systemfor semiquantitativeanalyses of mineral samples by fluorescent X-ray spectrography was discussed at the Eighth Annual Conference on Applications of X-ray Analysis. This system relates scattered radiation intensity (background) with peak intensity as measuredon a chart recording to determine concentrations of several elements in a sample.

A continuation of the study reveals the effects of operating variables such as: (1) sample preparation and choice of sample type, (2) operating power and target choice of the X-ray tube, (3) collimation ratio, (4) goniometer scanning rate, (5) choice of method for measurement of background intensity, and (6) control and adjustment of detector, sealer-ratemeter, recorder, and other electronic circuits.

The over-all effects of some operating variables are negligible owing to the compensatory nature of the calibration system. The net effects of others can be directly attributed to particular operating conditions, and these conditions can be controlled to achieve an optimum balance of effects to yield the desired results for accuracy, time required to complete the analyses, and other important requirements in the analyses of minerals.

A method is presented for the X-ray spectrometric determination of copper, tin, and uranium in a bronze heat-treating material. The method is based on the irradiation of an acid solution of the sample, to which indium and zinc have been added as internal standards. Quantitative determinations of copper and tin (each in the 40 to 60% range) are calculated from the ratios of the CuKα/ZnKα and SnKα/In Kα intensities. Uranium concentration (in the 1 to 10% range) is calculated from the ratio of the ULα/ In Kα intensities, A correction for the interference of the second-order In Kβ1,3 doublet with the first-order ULα peak is described. The over-all limits of error for single determinations at the 95% confidence level are ±0,58% copper, ±0,56% tin, and ±0,31% uranium.

An X-ray fluorescence method is presented for the analysis of tungsten-molybdenum solutions. Tungsten and molybdenum metal powders, produced by a fused-salt electrolysis procedure, are dissolved and analyzed for tungsten and molybdenum by using the W La1, Pt La1 intensity ratios. The platinum spectral line from the X-ray tube target serves as a self-internal standard. The sensitivity limit is about 0,05% and in the concentration range of 0.5 to 100%, the accuracy of analysis is within about 2% standard deviation. The technique is also applied to the analysis of tungsten and molybdenum in fused salts and to the determination of several other Impurity elements in these metals.

A double detector modification for the General Electric X-ray unit is described. The detector consists of a double-window flow proportional counter tube backed up by a sealed krypton-filled counter tube. This allows counting of either soft or hard radiation at maximum efficiencies without the necessity of changing counter tubes. Advantages in specific applications are discussed.

The proper choice of an analyzing crystal sometimes makes It possible to suppress second-order reflections which interfere with X-ray fluorescence analysis. Some of the problems associated with the analysis of small amounts of hafnium in zirconium, using a silicon crystal and a pulse-height analyzer, are discussed.

Techniques were developed and statistically evaluated for the determination of both the homogeneity and the elemental compositions of columbium alloy ingots. X-ray fluorescent spectrography and automatic data handling systems are utilized. Sample preparation is discussed in detail from the consumable-electrode arc-melt process to the final physical form suitable for X-ray fluorescent spectrography. Arrangements of apparatus for obtaining, recording, processing, and interpreting data, with specific attention to limitations for various matrices, are treated at length.

The method has been applied to specific cases with, for example, a correlation of the degree of variation of homogeneity of columbium alloy ingots with remelting practice. An attempt is made to provide a preliminary basis for determining the value of the technique as a possible means of quality control.

An adapter for a standard Norelco spectrographte attachment will be described which permits the examination of a selected area of the specimen by X-ray fluorescence analysis. The size of the area can be varied depending upon the choice of pinholes with sizes available dovwi to 0.010 in. or 250μ. The sample is placed into a specially modified sample holder so that it can be moved at orthogonal directions under the beam from outside the spectrograph while the sample is under X-ray irradiation. Dam on applications will also be given.

New Instruments for X-ray or neutron diffraction or for analyses by X-ray spectrographic methods are described. They incorporate a variety of new features which open a broad new scope of experimental techniques to the investigator. The diffractometer can be used with the diffraction plane either horizontal or vertical. The X-ray tube is conveniently supported by the diffractometer, allowing one to adjust the take-off angle from 0 to 10° without disturbing the precise alignment of the instrument. The 2θ angle bisecting mechanism functions throughout a 360° range and allows independent rotation (ω) of the specimen about the common axis. It is designed specifically for fully automatic remotely programed operation in addition to the usual manual modes.

The constant-potential generator is rated up to 50 ma at 50 kvcp and up to 40 ma at 60 kvcp, A monitor-type switch selects combinations of kilovolts and milliamperes in a manner such that exact reproducibility of settings is assured.

The dual-function strip-chart recorder incorporates both linear and three-cycle logarithmic slide wires so that each uses the output from the same linear rate meter.

The Integrated Radiation Analyzer consists of a pulse-height analyzer, electronic timer, high-speed sealer, precision linear ratemeter, power supply, printer, and diffractometer control. Integration of these components has made possible unique programing for X-ray analysis.

A flat-crystal spectrometer is described that is fully evacuated to permit analysis of light elements. A uniselector in conjunction with a programing unit allows the crystal to be moved to any series of angles with stopping intervals as close as 0.01°. Detectors are a scintillation counter and a flow proportional counter driven at twice the angular velocity of the crystal to maintain them in correct position, A further detector is used to monitor the primary X-ray beam. Programing is achieved by means of a patchboard, and each board allows up to 20 elements to be measured at one time and in a sequence as desired, Pinal results appear in tabular forrn on an electric typewriter. Provisions are also provided for continuous scanning, the X-ray spectrum being plotted on a strip chart recorder. The sample is brought into position with a slide arranged so that the spectrometer may be kept continually evacuated and only a small pocket of air need be pumped. Alternately, helium may be used with resultant low consumption per sample.

Uniformity of particle size is most critical iti both X-ray spectrographic and diffraction samples. This is especially true when there are both hard and soft materials in the sample. The soft material reduces to a smaller particle size and therefore presents a great density and area to the X-ray beam. Accurate quantitative determinations are most difficult under these clrcum s tances. This paper decribes a means of grinding to a more uniform particle size and presents graphic data on the effects of particle-size variation.