Helium-3 activation analysis of oxygen in silicon nitride films on silicon wafers

Oxygen in silicon nitride films on silicon wafers was analyzed by activation with the¹⁶O(³He, p)¹⁸F reaction. By³He bombardment of samples propertly arranged under consideration of the¹⁸F recoil effect, total oxygen was reliably determined and its predominant part was estimated to be located whether on film surface, in film interior, or on film-substrate interface. Sample films with 0.1 to 2 μm thicknesses were found to contain 0.2 to 2 μg/cm² of oxygen in locations varying with preparation conditions. This method has been compared with ESCA and other methods for surface analysis.     

Determination of oxygen in fluoride glass by proton activation analysis

Oxygen was determined in three kinds of ZrF₄-based fluoride glass [ZrF₄–BaF₂–GdF₃–AlF₃ (ZBGA), ZrF₄–BaF₂–LaF₃–YF₃–AlF₃–LiF–NaF (ZBLYALN) and ZrF₄–BaF₂–LaF₃–YF₃–AlF₃–LiF (ZBLYAL)] used for fabricating optical fiberby¹⁸O(p, n)¹⁸F reaction without significant nuclear interference. The main long life⁹⁶Nb nuclide was produced by the⁹⁶Zr(p,n) reaction in a non-destructive analysis of ZBGA-fluoride glass and reduced by using a coincidence system with Ge(Li) and NaI(T1) detectors. Substoichiometric separation of¹⁸F was also used to determine oxygen in fluoride glass, especially in glass containing yttrium as a component element because the⁸⁹Zr produced by the⁸⁹Y(p,n) reaction is a positron emitter, the same as¹⁸F. It was confirmed that the oxygen concentration in fluoride glass was 13–2460 ppm related to the loss by scattering.     

Determination of oxygen impurity in high purity materials by charged particle activation analysis using alpha projectiles

40 MeV -particles have been used to determine oxygen impurity at ppm levels in silicon, copper, and stainless steel, through the radiochemical separation of¹⁸F from the matrix. The separation of¹⁸F has been carried out by two techniques, viz.(1) distillation of H₂SiF₆ and (2) precipitation of KBF₄ and some modification has been applied in the separation, depending on the nature of interferences from the matrix. Instrumental approach was also carried out to determine the oxygen impurity at 100 ppm in Si matrix because this approach is not possible in Cu and stainless steel samples due to matrix activity.     

Oxygen depth profiling by activation with the16O(3He,p,)18F reaction

A technique has been developed for oxygen depth profiling in a thickness range of 5 to about 100 mg cm⁻² by activation with the¹⁶O(³He, p)¹⁸F reaction. An apparatus was set up for the activation of oxygen with a uniform probability along the depth, and a method has been devised for step-wise etching of the activated sample. This technique has been applied to the study of oxygen behaviour in heat treatment of silicon under various conditions.     

Substoichiometric precipitation of fluoride with lanthanum

The substoichiometric precipitation of fluoride with lanthanum was studied by using¹⁸F and¹⁴⁰La tracers and it was found that fluoride could be precipitated substoichiometrically with lanthanum and the reaction ratio between fluoride and lanthanum was 3∶1. The pH range at which fluoride can be separated substoichiometrically with lanthanum is between 2 and 8. Barium and indium interfere in the precipitation of fluoride, sodium, copper and manganese, however, not. Fluorosilicate can also be precipitated substoichiometrically by using lanthanum as a precipitant and the reaction ratio between fluorosilicate and lanthanum was 1∶2. This separation was applied for the determination of oxygen in silicon crystals. The concentration of oxygen measured in some silicon crystals was between 5 and 27 ppm and in good agreement with those by non-destructive method and infrared spectrophotometry.     

Radiochemical determination of oxygen traces in selenium

A sensitive method for determination of oxygen in selenium is described. Oxygen is converted into the gaseous compound sulfur dioxide labelled with³⁵S. The β-radiation of³⁵S is measured by liquid scintillation counting. The detection limit is 0.02 ppm. The efficiency of vacuum distillation in the purification of selenium, and the influence of air, dry oxygen and water on the absorption of oxygen by selenium have been investigated.     

Determination of isotopic composition of lithium by neutron activation analysis; comparison with mass spectrometry

An activation analysis method is described for routine determination of⁶Li-abundances in various lithium compounds on the basis of the reaction sequence⁶Li(n,α)T and¹⁶O(t, n)¹⁸F and the measurement of¹⁸F. Irradiations of diluted equeous solutions of samples containing CrO₃ as internal flux monitor were carried out at a thermal neutron flux density of ϕ≤10¹¹ n·cm⁻²·sec⁻¹. Interferences from variations in neutron self-shielding oxygen concentration and triton range do not exceed 0.5% when using the dilution technique. The results for⁶Li abundances from 3.52 to 7.60% with standard deviations of 1 to 2.5% were confirmed by mass spectrometric measurements.     

The problem of diffusion during charged particle activation analysis; the case of18F in pure germanium single crystals

During ion bombardment, thermal diffusion and radiation enhanced diffusion of atoms occur. These phenomena may be a source of error in ion beam analysis, particularly in radioactivation analysis if contaminant surface atoms are present. It is shown here that penetration of¹⁸F (derived from surface oxygen by nuclear reaction) in germanium single crystals, does not extend appreciably farther than the maximum range of the recoiling¹⁸F nuclei. Since the analyzed depth is over an order of magnitude larger than the recoil, the validity of charged particle activation analysis at the ppb level in the present case (oxygen→¹⁸F in germanium), is clearly demonstrated.     

Determination de L’oxygene et du fluor au moyen de l’activation par les photons γ et le comptage des neutrons retardes emis par17N

The activation of¹⁸O and¹⁹F by the reactions¹⁸O(γ, p)¹⁷N and¹⁹F(γ, 2p)¹⁷N was determined as a function of the maximum energy of the bremsstrahlung beam. By the use of this method it is possible to measure oxygen or fluorine by counting the delayed neutrons emitted by¹⁷N with a half-life of 4.2 sec.      

Alpha-activation analysis of nitrogen in boron-doped silicon single crystal

The rapid dissolution of silicon and the substoichiometric separation of¹⁸F were studied in order to determine nitrogen in a boron-doped silicon single crystal by activation analysis via¹⁵N(,n)¹⁸F. The silicon dissolution method was developed by using a nitric acid-based solution containing a known amount of ammonium fluoride, instead of hydrofluoric acid, as the¹⁸F carrier. The silicon could be dissolved to a depth of 100 m with a 10 mm Ø in 9 minutes. The conditions for¹⁸F steam distillation and substoichiometric precipitation as lanthanum fluoride were improved. 50% of the fluorine could be separated substoichiometrically even though the carrier amount was increased from 6–12 mmol to more than 16 mmol in order to add it to the dissolving solution. This dissolution method and improved substoichiometric separation were used to determine nitrogen in a boron-doped silicon single crystal. The nitrogen concentration was found to be less than the detection limit (50 ppb).     

Study of diffusion of impurities in semiconductor silicon by activation analysis and nuclear reaction methods

Diffusion of impurities is a basic operation in the technology of modern semiconductor devices especially integrated circuits. A knowledge of diffusion parameters such as concentration profile, diffusion depth and diffusion coefficient makes the optimization of a technological process possible. A destructive method of the step-by-step removal of the layers from the silicon single crystal slices irradiated in a nuclear reactor with subsequent determination of impurity contents was used in the diffusion studies of phosphorus, arsenic and gold. Phosphorus was determined by the beta counting of³²P after a separation, arsenic and gold were determined by means of a Ge(Li) gamma-ray spectroscopy without any chemical separation. The diffusion of boron was studied by the nondestructive method of the deconvolution of the energy spectrum of prompt alpha radiation from¹⁰B(n, α)⁷Li after thermal neutron irradiation.     

Reactions parasites rencontrees lors du dosage de l’oxygene dans le silicium au moyen des particules alpha

Activation analysis by charged particles may encounter great difficulties when ‘parasitic’ reactions induce in the matrix itself the same radionuclide as the one requested by the reaction for analytical aim. In this investigation we study reactions inducing¹⁸F activity in silicon and we show the importance of these reactions for determining low levels of oxygen in silicon by the reactions \(O\mathop \to \limits^\alpha {}^{18}F\) . For irradiations carried out with 54 MeV α-particles, the profile of¹⁸F production in silicon presents an activity decreasing very rapidly from the surface to depths of 500 μm without any connection with the \(O\mathop \to \limits^\alpha {}^{18}F\) activation curve. The diffusion of¹⁸F nuclides created at the surface from the oxide bed does not at all explain that result. The differences between these curves are attributed to parasitic reactions inducing¹⁸F activity in silicon. In order to confirm the reaction \(Si\mathop \to \limits^\alpha {}^{18}F\) we also investigated the profile of²⁴Na and²⁹Al production by another parasitic reaction in silicon. It is pointed out that the choice of incident energy is quite primordial, for analyses of oxygen in silicon with 54 MeV α-particles are not quite suitable. The best energy range appears to be around 35 MeV.     

Determination of silicon dioxide in iron and steel-making slags and fluorspars by 14 MeV neutron activation and high-resolution gamma-ray spectrometry

Silicon as silicon dioxide in iron and steel-making slags and fluorspars was analyzed by 14 MeV neutron activation high-resolution γ-ray spectrometry. Silicon was detected by measuring the 1.78 MeV γ-ray of²⁸Al, the product of the²⁸Si(n, p) reaction, using a 30 cm³ coaxial Ge(Li) detector. A modified TPA method was used for the calculation.²⁸Al is also produced from phosphorus by the³¹P(n,α) reaction, and from aluminium by the²⁷Al(n,γ) reaction. The contribution from the former reaction could be corrected experimentally when the P₂O₅ content of the sample was known, while the latter reaction could be neglected in this neutron energy region. The experimental correction coefficient for phosphorus agreed well with the theoretical value calculated from the nuclear properties of silicon and phosphorus. Yields of²⁸Al from SiO₂, P₂O₅ and Al₂O₃ of the same weight percentages were calculated as 1, 0.426 (experimentally 0.44) and 0.0022, respectively. The results of this method agreed well with the results of the usual chemical methods. The limit of detection of SiO₂ in iron and steel-making slags and fluorspars was calculated as 0.07%. The coefficient of variation of repeated experiments was compared with the statistical one.     

Determination of fluorine in organic compounds by epithermal neutron activation analysis

Classical activation analysis of fluorine by thermal neutrons has a limited application because of frequent interference from chlorine, the short half life²⁰F (11.4 s) and too high dead time of detectors. A procedure is described for fluorine determination using¹⁹F (n,p)¹⁹O reaction. Use of a boron carbide shield has no effect on the activity of¹⁹O (boron ratio −1) but considerably reduces background and interference due to¹⁸O (n, γ)¹⁹O reaction. The technique has been successfully applied to the determination of fluorine in organic compounds even in the presence of large amounts of chlorine and oxygen.     

Photon activation analysis of carbon in glasses for fiber amplifiers by using the flow method for the rapid separation of 11C

Summary We have studied the photon activation analysis of carbon in InF₃-based fluoride, chalcogenide and tellurite glasses for fiber amplifiers, and especially the nuclear interference from a matrix produced by (γ,n), (γ,2n), (γ,p) and (n,γ) reactions and a flow method for the rapid and simple separation of ¹¹C. A chemical separation technique is indispensable for determining carbon, because seventeen radionuclides are observed in these glasses. The flow method can sequentially be undertaken, the fusion of an irradiated sample with an oxidizer, the conversion of ¹¹C into ¹¹CO₂, and the absorption of ¹¹C in ethanolamine solution. We used a mixture of Pb₃O₄ and B₂O₃ as the oxidizer. There is interference with the ¹¹C measurement because ¹⁸F and ⁶⁸Ga are produced in fluoride and chalcogenide glasses by the ¹⁹F(γ,n), ²³Na(γ,αn) and ⁶⁹Ga(γ,n) reactions, respectively, and also absorbed in ethanolamine solution. Therefore, this flow method can only be applied to tellurite glasses. The chemical yield provided by the flow method was close to 100% when determining carbon in Japanese Iron and Steel Certified Reference Materials (JSS) by using lithium carbonate as a standard sample. We determined that the carbon concentrations in four kinds of tellurite glass were 7 to 14 ppm.     

Determination of silicon in aluminium by 14 MeV neutron activation analysis

A method has been developed for determining silicon in aluminium by fast neutron activation. It is based on the separation of two gamma lines by a Ge(Li) detector: the 1.73 MeV line from the product of²⁷Al(n, α)²⁴Na and the 1.78 MeV line from the²⁸Si(n, p)²⁸Al reaction. In the case of aluminium-silicon alloys 100 μg silicon can be determined, with an error of 10% in an aluminium sample of 1 g. This work was supported in part by the International Atomic Energy Agency.     

Charged-particle activation analysis studies on carbon,nitrogen and oxygen mainly in semiconductor silicon

High utility value of charged-particle activation analysis is exemplified by the study on carbon, nitrogen and oxygen in semiconductor silicon. A summary of the present authors' works is shown, and the equilibrium of carbon and oxygen in a silicon melt with ambient carbon monoxide is discussed. Also, a note is given about the chemical separation of¹⁸F for the³He activation analysis of various matrices.     

Proton microprobe determination of fluorine depth distributions and surface multielement characterization in dental enamel

Particle induced X-ray emission analysis (PIXE) and nuclear reaction analysis have been applied to multielemental analysis of human dental enamel in a proton microprobe. The PIXE technique was excercised for the characterization of multielemental lateral distributions on and just below the sample surface. For the assessment of depth distributions of fluorine the resonant nuclear reaction¹⁹F(p,aλ)¹⁶O was used. The combination of two analytical techniques comprising both atomic and nuclear interactions in a microprobe was applied to enamel samples of different origins. One sample constituted a healthy enamel and the other one enamel from a restored tooth. The microprobe in the combined mode was demonstrated to permit the establishment of lateral, concentration gradients of elements heavier than phosphorus with a resolution of 15 μm in enamel and simultaneously of depth distributions of fluorine better than 0.7 μm. The detection limits approached, 10–25 ppm for most of the elements considered. Supported by the Swedish Natural Science Research Council     

Application of bremsstrahlung activation analysis for the determination of carbon and nitrogen in sodium

Carbon and nitrogen are determined by means of the photonuclear reactions:¹²C(γ n)¹¹C and¹⁴N(γ, n)¹³N. This article presents a simultaneous chemical separation method for the isolation of¹¹C and¹³N. It is based mainly on combustion of the sample, after irradiation and etching, in a mixture of oxidising acid fused salts (B₂O₃ and Pb₃ O₄) containing a dispersing agent (NaCl and KCl) and the respective carriers. Each analysis takes about 40 min altogether. Minimum carbon and nitrogen contents of a few 10⁻² μg·g⁻¹ can be measured. The influence of the main competitive nuclear reactions on neighbouring elements, including the sodium itself, is examined.     

The use of charged particle bombardment for studying oxygen self-diffusion in oxides

The various nuclear techniques which have been used to study oxygen self-diffusion in oxides are discussed. Results are given for measurements using resonance capture in the¹⁸O(p, α)¹⁵N and¹⁸O(p, γ)¹⁹F reactions and the different techniques are compared.