Gold Absorber Research Articles

Transparent X-Ray Lithography Masks

With a view to producing a rigid and optically-transparent mask for use in X-ray lithography, the membrane strength was theoretically analyzed as a function of the residual stress in films. In silicon nitride films, reduction of the residual stress resulted in an increase in membrane strength accompanied by a decrease in optical Transparency. Silicon nitride film was then combined with transparent silicon dioxide film and an Si3N4/SiO2/Si3N4 sandwich-structure membrane was developed. In this membrane, the compressive film force approached the tensile film force closely, keeping the membrane under tension. This produced an optically-transparent, rigid, smoothly stretched-out membrane. An X-ray lithography mask with 0.8 µm thick gold absorber patterns on this membrane was successfully used for four-level multiple-pattern replications using an optical auto-alignment system.

Masked ion-beam lithography: A feasibility demonstration for submicrometer device fabrication

Masked ion-beam lithography (MIBL) is a high-resolution pattern-replication process that offers high throughput capability with submicrometer resolution. In MIBL, a collimated beam of protons is directed through a mask to expose a resist-covered wafer in proximity to the mask. We report here on four key MIBL technology issues: (1) mask technology, (2) radiation-damage effects, (3) resist materials, and (4) fabrication of NMOS devices using MIBL. We have established a process for fabricating thin (0.7 μm) silicon-membrane masks with submicrometer e-beam-generated gold absorber patterns. We have evaluated mask-induced beam scattering (<0.4° for channeled 175-keV-protons) and mask in-plane thermal distortions (<0.1 μm over 1 cm2 for 0.1 W/cm2 beam power) which are the principle factors affecting resolution and throughput. Radiation-damage effects were studied by simulating MIBL fabrication of MOS test chips. No statistically significant radiation effects were found after annealing at 450°C. We have demonstrated submicrometer resist patterning in both positive (PMMA) and negative (PS) resists, and have used these resists in the fabrication of submicrometer-gate NMOS test devices using MIBL exposures on all levels.

Determination of mercury emissions from a municipal incinerator

A cold-vapor atomic absorption system is described for the determination of volatile mercury in the stack gases of a municipal solid-waste incinerator. The gases are washed in a K/sub 2/CR/sub 2/O/sub 7//HNO/sub 3/ solution. By oxidation of the mercury in the washing bottle, a 99% collection efficiency is obtained. Total collection is guaranteed by adding a gold-coated sand absorber to the sampling train. After sampling, aliquots of the diluted absorbing solution are analyzed according to the reduction-aeration method with SnC/sub 12/. The released mercury is concentrated on a gold absorber. After thermal desorption from the gold, the mercury is measured in a modified atomic absorption spectrometer. The detection limit of the combined method is 1 ng. Calibration is achieved by injection of elemental mercury onto the gold. Generally, the reproducibility is better than 3%. The mercury concentrations found vary from 121 to 163 mu g of Hg per cubic metre of flue gas.

Preparation of x-ray lithography masks with 0.1 μm structures

A process for fabricating x-ray lithography masks having structures as small as 0.1 μm is described in detail. The patterns are made by electron-beam lithography in PMMA resist laid on 1-μm-thick polyimide membrane. The fact we use a very thin substrate strongly reduces the influence of electron backscattering on the resolution. That way it is possible to transfer 0.1 μm lines and spaces in 0.25-μm-thick PMMA resist. Gold absorber patterns are formed using electroplating, this technique being able to reproduce the smallest details of the lithography. On the other hand it allows high-aspect ratios because electroplated gold may be as high as the resist layer. Examples of 0.25 μm lines with 0.05 μm spaces in 0.25-μm-thick gold and 0.2 μm spatial period gratings in 0.15-μm-thick gold are given.

Polyimide membrane x-ray lithography masks—Fabrication and distortion measurements

A process for fabricating polyimide membrane x-ray lithography masks is described in detail. Thin membranes of polyimide are formed by spinning polyamic acid on Corning 0211, or other glass substrates, and polymerizing in situ. The glass substrate acts as a holder and an efficient heat sink during formation of gold absorber patterns on top of the polyimide, a procedure that usually includes ion beam etching. A support ring is then bonded to the polyimide, and the glass etched in dilute HF. Optically smooth polyimide membranes with thicknesses from 0.5 to 5 μm are readily achieved. Measurements of the distortion of polyimide membrane x-ray lithography masks have been made using a holographic moiré method and a method based on measuring the spread in the angle of diffraction from gold-grating patterns. These indicated a distortion less than two parts in 105 over the central 7.5-mm-diam area of a 19-mm-diam membrane.

Fabrication of micro- and submicron-bubble memory devices by a mask transfer technique with subsequent getter-ion etching

In order to fabricate magnetic bubble memory circuits of high bit density using x-ray lithography, it is necessary to form accurate gold absorber patterns for x-ray masks and accurate permalloy patterns for the bubble memory circuits. The combined processes of mask transfer technique and getter-ion etching instead of conventional ion etching in pure argon gas were used to make patterns in these metal films. Pattern width accuracies within ±0.1 μm and the angles of edge slope up to 70° were obtained by using these processes. Our new method was applied to the microfabrication of gold x-ray mask patterns and permalloy patterns of the circuits for magnetic bubbles of diameters 0.95 and 1.5 μm.

A magnetic moment measurement of the first-excited 2+ state of22Ne

The magnetic moment of the first-excited 2 + state of 22Ne has been measured with the time-differential recoil-into-vacuum technique with the use of a plunger. The state was populated with the reaction 4He(19F, p)22Ne at E(19F) =41.4 MeV. Two helium targets prepared with a thickness of approximately 4/ag. cm -2 by implantation into nickel at 30 keV were used. Protons were detected at forward angles between 0 ~ and 10 ~ in a silicon surface-barrier detector shielded by a 90/~m gold absorber. Gamma rays were detected in coincidence with protons in four 12.7 cm X 12.7 cm NaI(T1) detectors at angles +45 ~ +90 ~ -90 ~ and -135 ~ and in an 125 cm 3 Ge(Li) detector at an angle of 0 ~ with respect to the beam direction. The outgoing 22Ne ions (o/c = 0.048) were stopped in a 30/~m thick stretched gold foil. The unperturbed "y-ray angular correlation on recoil into silver showed an 85% m = 0 substate population. An increasing linear background under the Pl peak in the particle spectrum due to the reaction 12C(19F, pn)29Si was caused by carbon deposited on the stopper and the target foil during the experiment. The target foil was continuously monitored for flatness and stability by a laser interferometer, which was also used to determine the distance of closest approach (6/~m) as described in an earlier paper [1]. In a preliminary analysis including contributions from the various atomic states of the free ion, the g-factor of the first-excited 2 + state was determined to be [g[ = 0.32 -+ 0.03. The lifetime of this state deduced from the Ge(Li) detector data was in good agreement with the compiled value [2]. Recent shell-model calculations [3] yield values of the g-factor between 0.30 and 0.35. The same reaction on a He-implanted Fe foil was used to measure the transient field precession at a high recoil velocity. This measurement yielded an integral precession angle 4)/g = 17 -+ 4 mrad, which is 15 -+ 4 mrad higher than the LindhardWinther (LW) prediction [4]. An earlier experiment [5] at low recoil velocity (u/c = 0.017) yielded dp/g = 2.9 -+ 0.7 mrad in fair agreement with the LW theory. The deviation from the LW value at high recoil velocity can be understood as capture of polarized electrons from Fe into ls vacancies of the moving 22Ne ion, see ref. [6].

Abstract: New x-ray mask of Al–Al2O3 structure

This paper describes the fabrication and performance of a newly developed x-ray lithography mask consisting of an aluminum substrate and an Al2O3 film which was grown on the aluminum substrate by anodization. Transparent membranes of Al2O3 film were made by chemically etching parts of the aluminum substrate beneath the film on which gold absorber patterns were fabricated. Gold micropatterns were replicated successfully by Al Kα (8.34 Å) x rays. Properties of this mask include ease of fabrication, transparency, feasibility of realignment by optical means, and high mechanical strength allowing the capability of larger window size.

Energy-Straggling Measurements of Heavy Charged Particles in Thick Absorbers

Energy-straggling distributions of 20- and 49-MeV protons and 80-MeV helium ions have been measured for energy losses of 10 to 92% of the initial kinetic energy. Aluminum and gold absorbers were used, and the residual energy spectra were measured with lithium-drifted silicon semiconductor detectors. Experimental results show very good agreement with the theories of Payne and Tschal\"ar, provided that broadening due to atomic binding is taken into account, and disagreement with the Bohr theory for energy losses greater than 20%. In particular, as the particle beam slows down in the absorber, the variance of the energy distribution increases faster than the absorber thickness traversed, and a degree of skewness (i.e., a low-energy "tail") is introduced at very large energy losses.

Mean multiple scattering correction to the energy loss of heavy particles

The mean increase in energy loss δT m due to multiple scattering of heavy particles is treated for thick absorbers. Using Molière's theory with the improvements by Nigam et al., the mean square and mean fourth power of the reduced scattering angle are calculated for all absorber materials, for incident energies larger than a few MeV, for energy losses up to 80% of the incident energy, and for circular detectors with angles of acceptance up to ±35°. A method is given for computing the mean energy correction δT m from the incident and outgoing energies, from the stopping power of the absorber, and from the detector angle of acceptance. A typical value for δT m of 650 keV is obtained for 40 MeV protons loosing 32 MeV in a gold absorber and seen under a detector angle of ±29°.

Some measurements on depolarization and scattering of electrons in thick absorbers

Measurements are reported on depolarization of 280 to 1750 keV beta particles passing through thick absorbers; for electrons emerging parallel to the incident beam the degree of polarization is measured with a calibrated Mott analyser. The depolarization proves to be strongly Z dependent. In Lucite ( Z = 3.6) , little depolarization is observed, even for an energy reduction from 1.75 to 0.26 MeV in a 5.2 mm thick absorber. In gold, the observed depolarizations with moderate energy reduction in relatively thin absorbers are already of 40 to 50 %. These large values are not explained by the present theories. The angular distributions of 270 keV electrons scattered and slowed down by a thick Lucite and a thick gold absorber are also different, yielding a most probable scattering angle of 32° in thick Lucite and of 38° in gold.

Rate of Energy Loss and Ranges of Carbon and Oxygen Ions in Solids

The rate of energy loss of carbon and oxygen ions passing through carbon, aluminium, nickel, silver, and gold absorbers has been measured in the energy interval of 0.36 to 3.2 MeV. Ranges of the carbon and oxygen ions in these stopping materials were obtained by numerical integration of the rate of energy loss against energy curve. The relative accuracy of the results is to about 2% and the absolute accuracy is to better than 5%.

Measurement of Neutron Absorption Cross Sections with a Pile Oscillator

The following paper deals with the measurement of thermal neutron cross sections by a technique referred to as the pile oscillator. In this method, a neutron absorber is moved back and forth in a field of thermal neutrons such as that existing in a chain reactor. In the vicinity of the absorber there is a depression in the neutron flux and the motion of this depression past a nearby ion chamber produces an oscillating signal whose amplitude is proportional to the total neutron absorption cross section of the sample. In the instrument considered here, the oscillating signal generated in the ion chamber is amplified, rectified, and fed to an integrating circuit. The absorption cross section of an unknown sample is determined by comparing the charge accumulated in the integrating circuit with that produced by a standard gold absorber. The instrument will detect a total absorption cross sestion of about ${10}^{\ensuremath{-}3}$ ${\mathrm{cm}}^{2}$ and hence it is suitable for measuring cross sections of separated isotopes which are available only in small quantities. One of the major difficulties in making this instrument useful for absorption measurements has been the elimination of the effects of scattering by the sample. It was found that the signals caused by absorption and scattering are slightly out of phase with each other, and by properly phasing the rectifier, absorption cross sections which are no greater than 1 percent of the corresponding scattering cross sections can be measured. A number of thermal cross section measurements have been made with this instrument, most of which will be published separately; as illustrations, the following thermal cross sections are given: In, $191.2b$; ${\mathrm{Ag}}^{107}$, $29.9b$; ${\mathrm{Ag}}^{109}$, $83.7b$.