Energy-resolved Mass Spectrometer Research Articles

Experimental characterization of the deposition of silicon suboxide films in a radiofrequency magnetron reactive sputtering system

Sputter deposition of silicon suboxide films in a radiofrequency (rf) magnetron discharge using oxygen gas as a reactive species has been studied by a combination of plasma and film characterization techniques. The deposited films are silicon suboxides (SiO x , 0≤ x≤2). Using an energy-resolved mass spectrometer we found an impingement rate of SiO + ions on the deposition surface, which is comparable in magnitude to that of Si +, so a significant contribution of SiO to the Si and O incorporation in the suboxides cannot be ruled out. Increase of the oxygen flow beyond the point of SiO 2 deposition is characterized by a strong decrease of the Si + ion arrival rate on the growth surface. At the same time, the SiO 2 + arrival rate is increased, indicative for the increased oxidized state of the cathode. Number of atoms of the various kinds in the deposited layers have been determined by using Rutherford backscattering (RBS) and Elastic recoil detection (ERD) and correlated with the results of the plasma characterization.

Characterization of the plasma in a radio-frequency magnetron sputtering system

In order to understand the fundamental mechanisms in a radio-frequency magnetron sputtering system, the main properties of the argon plasma used in the process have been measured. A complete three-dimensional map of the ion density, electron temperature, and plasma potential has been obtained using a Langmuir probe. The electron temperature as well as the ion density have been found to increase in the region of the so called race track at the cathode. Furthermore, from the plasma potential map, the time-averaged local electric field has been obtained, pointing out the race track as the region where the most intense ion bombardment takes place. Besides, only the ions produced near the race track are accelerated towards the cathode, whereas those produced in the remaining volume move towards the anode. Finally, the dependence of the plasma quantities on the incident radio-frequency power and deposition pressure has been studied. The plasma potential measured using the Langmuir probe has been found to agree with that determined using an energy resolved mass spectrometer in all studied conditions.

Investigation of plasma parameters in the DC planar magnetron in balanced and unbalanced mode

DC planar magnetrons are quite often used for deposition of thin layers. Hence, the diagnostics of magnetron plasma and application of magnetron in surface technology are of great interest at present time. The investigated magnetron (type Vtech75 by Gencoa) is operated in balanced as well as in unbalanced mode. A Ti target was used. Argon or a mixture of argon and nitrogen are used as process gases. Discharge power and pressure are varied. The plasma is investigated by a radially movable Langmuir probe positioned in the plane of the substrate. Electron mean energy, electron density and the plasma potential at different radial probe positions are estimated. Moreover, measurements with an energy resolved mass spectrometer are performed. The flow densities of ions to the substrate are obtained. The results from both investigation methods are discussed.

高周波‐直流結合形マグネトロンスパッタ法におけるＡｒ｀＋´及びＣｕ｀＋´イオンのエネルギー分布

Energy distribution of 40Ar+ ions and 63Cu+ ions incident to the substrate surface in rf-dc coupled magnetron sputtering with which an rf source and a dc power were simultaneously applied to a copper target in order to control the incident ion energy on the target were measured by an energy-resolved mass spectrometer. The target power (rf+ dc) was varied from 2.0 to 5.0 kW at an Ar gas pressure of 1.0 Pa. The mean energy of 40Ar+ ions and 63Cu+ ions shifted from about 11.0 to 3.0 eV with the increase of the ratio of dc power to rf power. The number of 63Cu+ ions significantly increased with the increase of the target power, whose value was about 500 times the number of 40Ar+ ions at the target power of 5.0 kW.

The distribution of ion energies at the substrate in an asymmetric bi-polar pulsed DC magnetron discharge

Using an energy-resolved mass spectrometer and a time-resolved Langmuir probe, the distribution of bombarding ion energies, their fluxes and energy fluxes at a substrate in an asymmetric bi-polar pulsed DC magnetron have been determined. The discharge was operated in Ar at a pressure of 0.53 Pa with a Ti target and pulsed DC frequencies of 100 and 350 kHz with a range of duty cycles (from 50 to 96%). At 100 kHz, the Ar+and Ti+ time-averaged ion energy distribution functions (IEDFs) reveal three peaks, which are at low energy (<10 eV), in a mid-range (20-50 eV) and at high energy (60-100 eV). We correlate these peaks with distinct phases of the discharge voltage. At 350 kHz the IEDFs show four peaks reflecting a more complex voltage waveform. The low-energy ions are generated in the `on' phase when the plasma potential is typically a few volts above ground. The Ti+ energy spectra show a remnant of the original sputter-neutral energy distribution function. The mid-range ions are produced in the quiescent region of the voltage reverse phase, when the plasma potential is raised globally a few volts above the cathode potential, typically 10-30 V. The high-energy ions are generated in a period of ~0.3 µs, during the discharge voltage overshoot, when the target potential rises to typically over +140 V. However, given the time resolution of the Langmuir probe (0.5 µs), it is not possible to determine if plasma potential is lifted globally to this high potential or only close to the cathode. At 350 kHz, these `fast' ions make up to about a quarter of the total ion flux at the substrate and an upper bound transient power flux of about 2.5 times the maximum delivered in the `on' phase. The total power flux to a substrate in the sustained phase of the discharge is found to increase with frequency and reverse time.

Modifying the IEDFs at a plasma boundary in a low-pressure RF discharge using electron beam injection

Using an energy-resolved mass spectrometer situated behind a specially designed electrically isolated target boundary, the bombarding ion energy distribution function (IEDF) in an electron beam-plasma discharge at low pressure (<0.26 Pa) has been determined. The collimated electron beam (mean energies 50-180 eV and beam currents up to 3 mA) incident upon the collisionless sheath of the target boundary was found to sink its floating potential down to approximately the voltage equivalent of the beam energy. The measured IEDFs were found to agree well with predictions of broad distribution functions with long low-energy tails obtained from a modified Tonks-Langmuir model of the collisionless beam-plasma boundary.

The measurement and control of the ion energy distribution functionat a surface in an RF plasma

Using an RF-driven collecting surface, mounted on the front endof a high-resolution energy-resolved mass spectrometer, the ion energydistribution functions (IEDFs) within a 13.56 MHz argon discharge have beenmeasured and controlled. A technique of RF signal feedback has beendeveloped, in which the RF amplitude and phase (fundamental and first twoharmonics) in the sheath are varied, so manipulating the mean bombardingion energies and widths of the IEDFs. For high RF sheath potentials, theIEDFs are broad and bi-modal. However as the sheath potential drop isreduced (with imposed matched RF signal) the IEDFs narrow, eventuallybecoming single peaked when no RF potential drop is present in the sheath.The observed widths in the IEDFs broadly agree with simple modellingpredictions for the `high frequency' RF sheath regime; however wecalculate, from the ion spectral data, marginally thinner sheaths thangiven by the Child-Langmuir law (derived using the mean DC sheathpotential).

Ar addition effect on mechanism of fluorocarbon ion formation in CF4/Ar inductively coupled plasma

Using a Langmuir probe and an energy-resolved ion mass spectrometer, gas-phase kinetics of fluorocarbon ions has been investigated as a function of the Ar percentage in a mixed CF4/Ar plasma. Spatially resolved electron energy distribution function, plasma potential, and ion density are measured in an inductively coupled plasma. As the Ar percentage increases, the average electron energy decreases while the electron density remains flat. The ion density also stays constant at low Ar percentages but increases over the Ar percentage larger than 63% mainly due to the increase of the Ar+ density. The plasma potential decreases as a result of the increase of Ar partial pressure, and this is confirmed by measuring the ion energy distributions of argon and fluorocarbon ions using the ion mass spectrometry. With the mass spectrometry, it is found that the most prominent ions, CF3+ and CF+, are formed predominantly by a process of dissociative ionization or radical ionization while CF2+ ions are formed dominantly by a process of charge transfer. As a practical application of this study for SiO2 etching, the densities of the fluorocarbon ions and radicals are correlated with the SiO2 etch rate and its selectivity to photoresist. Microtrench profile is also investigated as a function of Ar percentage and it is observed that the microtrench tends to be suppressed with the increase of the Ar percentage. This tendency is correlated with changes in the plasma chemistry as the Ar percentage increases.

Effects of Ar pressure on ion flux energy distribution and ion fraction in r.f.-plasma-assisted magnetron sputtering

In r.f.-plasma-assisted sputtering, the ion fraction of the sputtered metal is thought to be enhanced by increasing Ar pressure. In this study, effects of Ar pressure on ion energy distribution and ion fraction of the metal flux have been investigated. The experiments have been performed by using a 55 mm diameter Ti magnetron cathode and a 60 mm diameter copper r.f. coil. The ion energy distribution of the depositing flux was measured by an energy-resolved mass spectrometer . The experimental results show that the number of Ti + ions increases with Ar pressure, whereas that of Ar + ions decreases. The ion fraction of the Ti flux reached 80% at an Ar pressure of 4.0 Pa. The electron temperature decreased with increasing Ar pressure as a result of plasma quenching induced by the ionization of the Ti metal atoms that have a lower ionization potential . The results emphasize the strong effects of Ar pressure on the ion fraction and plasma conditions. The Penning ionization is thought to be the dominant ionization process for Ti atoms in the r.f.-plasma-assisted sputtering.

Diagnostics and modeling of a hollow-cathode arc deposition plasma

Methane or acetylene added to an argon or helium arc plasma is dissociated and/or ionized effectively in order to deposit amorphous carbon (a-C:H) layers. The plasma can be confined or spread out by a homogeneous or cusp magnetic field. The issue of the present study is whether the magnetic confinement has an influence on the type of ions governing the deposition and their energy. To this end, an energy-resolved mass spectrometer (plasma monitor) was used, screened off by a μ-metal cage. It appears that, for strong magnetic fields of up to 10 mT, besides CH3+ and C2H2+ ions (from the methane and acetylene admixture, respectively), also relatively more dissociated species like CHx+ (x=2, 1, 0) and CyHz (y=2, 1, 0; z=1, 0) arise. Especially with the C2H2 admixture, atomic C+ ions appear to become more abundant with increasing magnetic field. The energy of the ions appears to decrease with increasing magnetic field. With Langmuir cylindrical electric-probe measurements it was confirmed that this is related to a decreasing plasma potential. By means of a numerical model, spatially resolved density profiles of the most important species were calculated. Depending on the reactor parameters, relevant ion densities are of the order of 109–1010 cm−3, resulting in deposition rates of up to some μm h−1.

Effects of coil dc potential on ion energy distribution measured by an energy-resolved mass spectrometer in ionized physical vapor deposition

Energy distribution of Ar+ and Ti+ ions and plasma conditions have been investigated for various dc potentials of the rf inductive coil in ionized physical vapor deposition. The sputtering cathode used in the experiment is a conventional magnetron sputtering source with a Ti target (55 mmφ) which is coupled with a rf coil (60 mm φ, made of Cu). The mass spectrometer used to measure ion energy distributions is an energy-resolved type plasma monitor. The coil dc potential is controlled by changing the resistance of the resistor in the termination inductance-capacitance-resistance (LCR) circuit connecting the coil to the ground. By increasing the resistance of the LCR circuit, the peak of the Ar+ energy spectra shifts to a lower energy. In addition, it is found by the probe measurements that the plasma potential decreases with increasing the termination resistance. The changes in the peak energy in the ion energy spectra show a good agreement with the change in the plasma potential; as a result of the change in the plasma potential, the energy distribution of ions accelerated toward the grounded substrate by the sheath shifts to a lower energy in accord with a plasma potential change. In this experiment, plasma potential is strongly affected by the dc self bias of the rf inductive coil induced by a current flow from the ground to the coil through the termination resistance because in the apparatus used the coil was placed inside the chamber.

Ion energy distribution in ionized dc sputtering measured by an energy-resolved mass spectrometer

Ion energy distribution of sputtered Ti particles in ionized dc sputtering has been measured by an energy-resolved mass spectrometer. The energies of the Ti + and Ar + ions were measured by a Balzers PPM421 plasma monitor. The experimental results showed that the energy of sputtered Ti particles was enhanced from 2 or 3 to about 30 eV as a coil rf power increased from 0 to 200 W, for a constant magnetron cathode current of 0.3 A. On the other hand, when a cathode current increased from 0.1A to 0.5 A for a constant coil rf power of 200 W, the mean energy of Ti + decreased, as a result of the plasma quenching caused by the increase in the number of the sputtered Ti particles. In addition, it was found that the number of Ti + ions was saturated for cathode dc currents higher than 0.3 A. The results obtained in this study demonstrate that the ratio of the coil rf power to the magnetron cathode power or current is crucial to obtain a proper ionization ratio and energy of sputtered particles.

Energy and mass spectroscopy studies during triode ion plating of TiN films in two different layouts

Abstract Plasma properties are compared in two triode ion-plating systems using a low-voltage arc discharge. The systems differ in the configuration of the chamber and magnetic field. Mass and energy distributions of positive ions during deposition of TiN films were measured by an energy-resolved mass spectrometer. Plasma parameters were determined also by Langmuir probe diagnostics. In both systems, single and multiple charged ions of Ar and N2 gas and the evaporated Ti were detected. The system with a strong magnetic field confinement exhibits a higher plasma potential (Up=55 V) under standard deposition conditions, the peak ion energy is higher and the energy distribution is narrower than in the system without magnetic confinement (Up=12 to 35 V). The probability for the multiple charged ions in this system is much higher than in the system without the magnetic confinement, so the population of Ti++ in this system is even higher than that of Ti+. The presence of high-energy neutral Ti is also observed.

高周波支援マグネトロンスパッタリング法におけるＡｒ｀＋´及びＴｉ｀＋´のエネルギー分布のＡｒ圧力依存性

Effects of Ar pressure on the number of the 48Ti+ ions incident to the substrate have been investigated in inductively coupled rf plasma enhanced magnetron sputtering. The energy distributions of the 40Ar+ and 48Ti+ ions incident to the substrate were measured by an energy-resolved mass spectrometer. In addition to the energy distribution measurements, plasma conditions have been measured by a Langmuir probe technique to investigate the effects of Ar pressure on plasma. The experimental results of the ion energy distribution measurements show that the number of 48Ti+ ions arriving at the substrate increased as Ar partial pressure increased and that the number of 40Ar+ ions arriving at the substrate increased to an Ar pressure of 1.0 Pa and decreased for further increase in Ar pressure. From electron temperature measurement results, it was found that electron temperature decreased with increasing Ar pressure. The main process of Ti ionization is ascribed to the Penning effect, the ionization of Ti by the excited neutral Ar, resulting in a decrease of plasma temperature. It is concluded that an increase in Ar pressure enhances the ionization probability of Ti sputtered atoms.

高周波支援マグネトロンスパッタリング法におけるＡｒ｀＋´及びＴｉ｀＋´のエネルギー分布のカソード電力依存性

Energy distribution of Ar+ and Ti+ ions incident to the substrate surface in dc magnetron sputtering with an inductively coupled rf plasma excitation have been measured by an energy-resolved mass spectrometer. In addition, dc self bias of the inductive coil, cathode potential, and plasma parameters have been examined in order to discuss the correlation between these parameters and the energy distribution. In this experiment, the cathode current was variation from 0.1 A to 0.5 A for a constant coil rf power of 200 W at an Ar partial pressure of 0.3 Pa. With increasing the cathode current, the width of the energy distributions of Ar+ and Ti+ ions incident to the substrate surface became narrower and the peak energies of the energy distributions shifted to low energy. It was found that the change in peak energies of 43 to 18 eV with increasing cathode current was in good agreement with the change in the plasma potential of about 46 to 18 V.

Diagnostics of SF6 plasmas by energy-resolved mass spectrometry: influence of the electrode material on internal plasma parameters

Abstract A detailed understanding of physical and chemical processes in plasmas and plasma sheaths can be achieved by measuring the energy distributions of all ions (IED) at the place where the plasma species interact with a substrate. In order to measure IEDs in the reactive ion etching processes an energy-resolved quadrupole mass spectrometer (E-QMS) was assembled underneath the powered electrode (cathode) of a diode reactive ion etcher. To examine the influence of the electrode material, we used brass, steel and aluminium as cathode materials. The plasma ions reach the QMS through an orifice in the cathode with a diameter of a 100 μm. The ion energy distributions of SF + x ( x =0…5), F + and F + 2 ions from sulfur hexafluoride (SF 6 ) plasmas in the pressure range 0.1–5 Pa at different DC bias potentials between −50 and −250 V and a frequency of 13.56 MHz were investigated. Marked differences in the IEDs of all ionic species appeared for the three different electrode (cathode) materials. A model is developed which explains these experimental results with different secondary electron coefficients of the electrode materials.

Effects of ionization power on ion energy distribution in ionized r.f. sputtering measured by an energy-resolved mass spectrometer

Abstract In an ionized sputtering technique, it is crucial to give a proper energy to the particles in order to improve film properties or to enhance directionality of the sputtered particles without causing disorder or other undesired damages to the growing film. In this study, ion energy distribution has been investigated by an energy-resolved mass spectrometer for ionized Ti sputtering in order to discuss effects of coil r.f. power and magnetron cathode r.f. power on energy distribution of Ar+ and Ti+ ions arriving at the substrate. The cathode used in the experiment was a magnetron type with a 55-mm diameter Ti target. Ion energy distribution was measured by PPM-421 Plasma Monitor (Balzers) whose orifice to the ion analysis optics was set in front of the sputtering cathode with a distance of 200 mm. The coil r.f. power and the cathode r.f. power were varied up to 200 W. The experimental results show that energy distribution of Ti+ ions was enhanced from a few tens of eV to more than 100 eV as the coil r.f. power increased. The energy increase of Ti+ ions by an r.f. coil plasma was more drastic for a lower cathode r.f. power. As the cathode r.f. power increases, the energy of Ti+ ions decreased, as a result of quenching of the r.f. coil plasma. The quenching is thought to be induced by the increase in the number of Ti atoms passing through the r.f. plasma region. Energy distribution of Ar+ ions showed similar tendency to that of Ti+ ions.

Energy-resolved mass spectrometry and Monte Carlo simulation of atomic transport in magnetron sputtering

Here we report the energy distributions of ions measured by an energy-resolved mass spectrometer during magnetron sputtering of Cu in Ar. The effects of the discharge power on the distributions are described. Both Cu+ and Ar+ ions exhibit non-equilibrium energy distributions with a main maximum (corresponding to thermalized species) and a high-energy tail, with energies ranging up to about 50 eV. The experimental data are compared with the results of a direct simulation Monte Carlo computer simulation. The elastic collisions of neutral species are calculated in a quasi three-dimensional geometry with two-dimensional cells and a featureless third dimension. The collision cross-section and scattering angles are calculated using the Biersack–Ziegler interatomic potential. The Sigmund–Thompson distribution is used for the sputtered atoms. The calculated energy distributions of Cu and Ar atoms as a function of discharge power are correlated with the experimental data. Momentum transfer in elastic collisions with the sputtered particles may be the main effect contributing to the formation of the non-equilibrium energy distribution of the gas molecules. The calculated average energy of the Ar atoms increases proportionally to the discharge current, and reaches values of above 1000 K.

Energy-resolved mass spectra of ions and neutrals in triode ion plating of Ti(C,N) coatings

Ti(C,N) coatings have been deposited by triode ion plating in a mixture of Ar, N 2 and C 2H 2. The energy spectra of selected ions and neutrals during deposition were measured by an energy-resolved mass spectrometer. The observed energy spectra of ions and neutrals are discussed. The energy spectra of neutral species were measured as the energy spectra of ions produced by crossed-beam electron ionization in the mass spectrometer. Both the energy spectra of Ti + ions and Ti atoms show wide energy spectra, ranging to about 40 eV, Properties of the Ti(C,N) coatings are reported, especially the stress-free lattice parameter and the residual macroscopic stress measured by X-ray diffraction. It is observed that the compressive stress varies from −6 to −21 GPa with the different deposition conditions used. Effects of the total pressure, the substrate voltage and the current in the low voltage arc on the energy spectra of ions and neutrals are discussed.

Energy distribution of ions in an unbalanced magnetron plasma measured with energy-resolved mass spectrometry

The energy distribution of ions in the plasma during unbalanced magnetron sputtering was measured with an energy-resolved mass spectrometer. Typical energy distributions of the main ion species were measured for sputtering of a Cu target in an Ar atmosphere in the pressure range from about 0.1 to 10 Pa. The effects of the discharge pressure and of the distance between target and mass spectrometer on the ion energy distribution were studied. A low energy (thermalised) peak in the ion energy distribution function was observed at an energy corresponding to the plasma potential. In addition, a high energy tail was observed at energies of about 3–30 eV above the low energy peak, not only for sputtered ionised atoms but also for the Ar ions. The ion energy distribution of this high energy tail can be fitted with an exponential function; the mean energy was about 1–5 eV. The integrated intensity of the high energy tail was a function of the pressure-distance product p × d, where d is the distance of the mass spectrometer from the magnetron. The function is close to exponential decrease with increasing p × d product for the high energy tail of the Cu + energy distribution. A similar exponential decrease was observed also for the high energy tail of the Ar + energy distribution with the exception of lowest pressures (<0.2 Pa) where the intensity of the Ar + high energy tail increases with increasing pressure. An explanation of these observations is proposed based on energy transfer from sputtered atoms to Ar gas atoms during collisions.