C. E. Young

Velocity and electronic state distributions of sputtered Fe atoms by laser‐induced fluorescence spectroscopy

Velocity distributions and relative populations in the fine‐structure levels of the a 5DJ ground state of Fe atoms, produced by sputtering with 3 keV argon ions, have been investigated by Doppler‐shifted laser‐induced fluorescence. The laser system employs a single‐mode, scanning ring dye laser, amplified by a sequence of three excimer‐pumped flowing dye cells. Frequency doubling in a KD*P crystal was used to produce high energy (>0.5 mJ) pulses of narrowband tunable UV output near 300 nm. Laser power influence on effective velocity bandwidth was investigated. Favorable light‐collection geometry minimized distortion of the velocity spectra from apparatus‐averaging effects. In impurity flux diagnostic applications in fusion devices, substantial spatial averaging may occur. In the latter case, the narrow velocity bandwidth (70 m/s, transform limit) of the present laser system is particularly useful.

Laser-based secondary neutral mass spectroscopy: Useful yield and sensitivity☆

Abstract Multiphoton ionization (MPI) by pulsed, tunable lasers provides a sensitive means for detection of neutral atoms due to the high efficiency achievable both in the ionization and subsequent detection. Substantial selectivity can be achieved by excitation between energy levels of the atom of interest. This resonant MPI technique can access all atomic states of any particular atom including its ground and metastable levels. In principle all elements may be ionized through judicious selection of the color of the excitation laser light. In practice resonance ionization has been experimentally demonstrated for nearly every element. A variety of problems exist in order to optimally apply resonance ionization spectroscopy (RIS) to the detection of sputtered neutral atoms, however. Several of these problems and their solutions are examined in this paper. First, the possible useful yields obtainable and the dependence of useful yield on various laser parameters for this type of sputtered neutral mass spectrometer (SNMS) are considered. Second, the choice of a mass spectrometer and its effect on the instrumental useful yield is explored in light of the unique ionization region for laser based SNMS. Finally a brief description of noise sources and their effect on the instrumental sensitivity is discussed. That it is possible to combine in one instrument both high useful yields and high sensitivity for the detection of minority species (either very dilute surface constituents or species sputtered in highly excited states) will be demonstrated with results of Fe implanted Si samples in the surface analysis by resonance ionization of sputtered atoms (SARISA) instrument. SARISA accomplishes the necessary noise reduction without signal loss through the extraction of the photoions into a sector-field time-of-flight (TOF) mass spectrometer. In standard, isochronous operation, energy and angular spreads at the point of ionization are compensated in flight to produce well-resolved TOF mass spectra. Noise sources (photons, metastable and scattered atoms) escaping through transparent grids are strongly suppressed.

Laser fluorescence spectroscopy of sputtered uranium atoms

Abstract Laser induced fluorescence (LIF) spectroscopy was used to study the sputtering of 99.8% 238 U metal foil when bombarded by normally incident 500–3000 eV Ne + , Ar + , Kr + and O 2 + . A three-level atom model of the LIF processes is developed to interpret the observed fluorescent emission from the sputtered species. The model shows that close attention must be paid to the conditions under which the experiment is carried out as well as to details of the collision cascade theory of sputtering. Rigorous analysis shows that when properly applied, LIF can be used to investigate the predictions of sputtering theory as regards energy distributions of sputtered particles and for the determination of sputtering yields. The possibility that thermal emission may occur during sputtering can also be tested using the proposed model. It is shown that the velocity distribution (either the number density or flux density distribution, depending upon the experimental conditions) of the sputtered particles can be determined using the LIF technique and that this information can be used to obtain a description of the basis sputtering mechanisms. These matters are discussed using the U-atom fluorescence measurements as a basis. Both the v -parallel (the exciting laser beam is parallel to target surface) and the v -perpendicular (the laser beam is perpendicular to the target surface) number density velocity distributions were measured using the LIF Doppler shifted absorption frequencies of the sputtered ground state U(1) atoms. The results are compared with the collision cascade sputtering theory using the LIF model presented here. The relative sputtering yields for various incident ions on uranium were also measured for the first time using the LIF technique. A surprisingly high fraction of the sputtered uranium atoms were found to occupy the low lying metastable energy levels of U(I). The population of the sputtered metastable atoms were found approximately to obey a Boltzmann distribution with an effective temperature of (920 ± 100) K.

Trace surface analysis: 30 ppb analysis with removal of less than a monolayer. Fe and Ti impurities in the first atomic layer of Si wafers

Abstract Recent trends, exemplified by the stringent demands of the semiconductor industry, demonstrate a need for surface analysis at progressively lower impurity levels. Because the total number of impurity atoms available for measurement in the first atomic layer is severely limited, a successful trace analysis technique must efficiently utilize the impurity atoms available as well as achieve good signal to noise. We report here the results of a new Surface Analysis by Resonant Ionization of Sputtered Atoms (SARISA) apparatus which has a demonstrated ability to measure Ti and Fe impurities in the surface atomic layer of a Si wafer at the 30 ppb level. The Fe results are particularly instructive since Secondary Ion Mass Spectrometry (SIMS), because of the isobaric overlap of Fe and Si2, has achieved no better than 200 ppb with removal of nearly 100 atomic layers. The measurements were made with a 2 mm2 ion spot. A measured collection efficiency of 1 Fe atom counted per 100 Fe atoms sputtered could also be achieved. Apparatus improvements will be described which should increase the collection efficiency to 1 atom counted per 10 sputtered.

Oxygen and titanium sputtering yields as determined by laser fluorescence and auger electron spectroscopy for monolayer oxygen coverage of polycrystalline Ti

Abstract The presence of an oxide layer can strongly influence the charge-state of species ejected from ion-bombarded metal surfaces, as well as the total sputtering yield. These quantities directly affect the influx of metallic impurities from the wall region into the plasmas of fusion devices. Surface coverage can also modify the distribution of sputtered atoms among electronic states and thus the apparent impurity density detected by the laser fluorescence spectroscopy (LFS) technique. The measurements reported here provide LFS data on number density and electronic state populations for the species Ti and Ti + as a function of surface oxygen coverage in a laboratory apparatus providing for direct monitoring by Auger analysis. An ultra-high vacuum chamber reached a base pressure of ⩽10 −8 Pa after 200°C bakeout, making target contamination negligible during data collection. The desorption data for O-atoms by 3 keV Ar ions exhibits four distinct rates, for coverage between 0 and 3 monolayers. Neutral titanium sputtering yield falls approximately linearly with oxygen coverage, reaching 50% of the bare metal value at about two monolayers. All excited multiplets investigated showed enhanced population with oxygen coverage. Agreement with TRIM computer modelling is achieved if oxygen is assumed to occupy some sub-surface sites.

ELECTRON-STIMULATED DESORPTION OF NEUTRALS FROM METHANOL-DOSED AL(111) : VELOCITY DISTRIBUTIONS AND ADSORBATE DECOMPOSITION DETERMINED BY NONRESONANT LASER IONIZATION

Abstract Electron-stimulated desorption (ESD) of neutrals from methanol-dosed Al(111) is studied using laser ionization at 193 nm coupled with time-of-flight (TOF) mass spectrometry. At room temperature and at very low laser intensity, mass spectrometry of the neutral ESD species indicates the presence of desorbing CH 3 O, the methoxy radical. At higher laser intensity, this species is efficiently photolyzed to C + and HCO + fragments. The velocity distributions of these photofragments, indicative of the velocity distribution of the methoxy parent, are measured for methanol dosed onto both clean and pre-oxidized single crystal surfaces. Both of the surfaces yield similar non-Boltzmann distributions with peak velocities of ∼ 900 m/s, corresponding to a peak kinetic energy of ∼ 0.1 eV for the methoxy parent. The similar results may find explanation in terms of oxidation of the Al(111) surface by the initial methanol exposure. The major ionic desorbate observed from this methanol-dosed Al(111) is H + , and its kinetic energy distribution peaks at ∼ 4 eV, a value which is typical of that observed in other ESD studies of ionic desorbates. The order of magnitude difference in kinetic energies between the desorbed ions and neutrals is discussed in terms of possible desorption mechanisms. Neutral ESD, combined with X-ray photoelectron spectroscopy (XPS) is also used as a probe of changes in surface adsorbate composition as a function of temperature and of electron beam dose for methanol/Al(111). The surface concentration of the methoxy species, as monitored via the HCO + photofragment, is found to decrease linearly with increasing temperature. An increase in C + signal at ∼ 470 K is attributed to the formation of a thermal decomposition product with either a higher desorption cross section or a higher laser ionization/fragmentation cross section than the methoxy species. Electron beam damage studies of the methoxy/aluminum system at an electron beam energy of 3 keV give a cross section of 3 ± 1 × 10 −17 cm 2 for loss of methoxy from the surface at this energy.

High resolution continuous wave laser induced fluorescence spectroscopy of sputtered Zr atoms

High resolution laser induced fluorescence (LIF) spectroscopy utilizing a continuous wave ring dye laser was used to measure the velocity distribution of sputtered Zr atoms. The ring dye laser permitted the determination of the sputtered Zr atom velocity distribution with a resolution of 2.4×103 cm/s over the emission range of 0 to 2×106 cm/s. The sputtered Zr atoms were produced by bombardment of the metal target by normally incident 3 keV Kr+. Determination of the velocity distributions was accomplished by tuning the exciting laser frequency through the Doppler shifted absorption frequencies of the sputtered atoms. It was determined that such processes as laser power density and transit time broadening of the sputtered atom’s energy levels, the laser pulse duration and bandwidth as well as the geometry of the fluorescence excitation and detection system were all critical parameters important to the understanding of the experimental LIF velocity distributions. A comparison is also made between the use of...

Electronic excitation of Ti atoms sputtered by energetic Ar+ and He+ from clean and monolayer oxygen covered surfaces☆

Electronic excitation of Ti atoms ejected during energetic ion bombardment (Ar+, He+) of well characterized clean and oxygen covered polycrystalline Ti metal surfaces has been determined. For states with 0 to 2 eV and 3 to 5.5 eV of electronic energy, static mode laser fluorescence spectroscopy (LFS) and static mode spontaneous fluorescence spectroscopy (SFS) were used respectively. These experiments which were carried out in a UHV (<10−10 Torr) system equipped with an Auger spectrometer provide measurements of the correlation between oxygen coverage (0 to 3 monolayers) and the excited state distribution of sputtered Ti atoms. The experimentally determined electronic partition function of Ti atoms does not show an exponential dependence on energy (E) above the ground state but rather an E−2 or E−3 power law dependence.

Laser spectroscopy of sputtered atoms

The use of laser radiation to study the sputtering process is of relatively recent origin. Much has been learned from this work about the basic physics of the sputtering process itself through measurements of velocity and excited state distributions of sputtered atoms and the effects of adsorbates on substrate sputtering yields. Furthermore, the identification, characterization, and sensitive detection of sputtered atoms by laser spectroscopy has led to the development of in situ diagnostics for impurity fluxes in the plasma edge regions of tokamaks and of ultrasensitive methods (ppb Fe in Si) for surface analysis with ultralow (picocoulomb) ion fluences. The techniques involved in this work, laser fluorescence and multiphoton resonance ionization spectroscopy, will be described and illustrations given of results achieved up to now.

Analysis of ion-bombarded and laser-irradiated surfaces of ZnS and Zn via two-photon high-resolution laser-induced fluorescence spectroscopy

Doppler‐shifted two‐photon laser‐induced fluorescence spectroscopy (LFS) has been used to measure the velocity distributions of Zn atoms ejected by Ar+ ion bombardment from both a Zn surface and a ZnS single crystal. Doppler‐shifted spectroscopy in combination with time of flight(TOF), has also been employed to investigate Zn atoms ejected from a ZnS single crystal under irradiation by 308‐nm photons (XeCl excimer laser). The absolute Zn sputtering yields from ion‐bombarded and laser‐irradiated ZnS samples were determined by comparison with LFS measurements of Zn atoms sputtered from pure Zn targets by Ar+ ions since the absolute sputtering yield is known. For the Zn target the velocity distributions of the ion‐induced sputtered Zn atoms were in good agreement with the Sigmund–Thompson distribution based on the known binding energy of 1.35 eV. On the other hand, the velocity distribution of Zn atoms ion sputtered from ZnS showed some deviation from the Sigmund–Thompson distribution and had a lower binding...