No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
Spatial and temporal variation of the plasma parameters in a high power impulse magnetron sputtering (HiPIMS) discharge
... An electric field E z in the plasma (seeFigure 1b) can turn such ions around, increase β, and decrease the deposited fraction δ of the M + ions. Recent probe data [5,6] have shown that in HiPIMS there can be a potential uphill from the sheath edge to the bulk of the plasma that can vary at least in the range 7-60 V, with stronger E z generally observed closer to the target, as well as for stronger magnetic field, for stronger applied power, and during the early stage of the HiPIMS pulse. ...
... fromFigure 5 results in ~ 10 V. This is large enough to significantly increase the M back-attraction and also consistent (within a factor of two) with probe measurements in the same device [5]. ...
... Recent measurements in two different HiPIMS devices [5,6] have shown electric fields z E in the plasma that are strong enough to cause considerable back-attraction of the ionized sputtered species M . The associated potential uphill extended a distance of 5 – 6 cm from the sheath edge into the plasma, and had amplitudes in the range 6 – 70 V. ...
The lower deposition rate for high power impulse magnetron sputtering (HiPIMS) compared with direct current magnetron sputtering for the same average power is often reported as a drawback. The often invoked reason is back-attraction of ionized sputtered material to the target due to a substantial negative potential profile, sometimes called an extended presheath, from the location of ionization toward the cathode. Recent studies in HiPIMS devices, using floating-emitting and swept-Langmuir probes, show that such extended potential profiles do exist, and that the electric fields Ez directed toward the target can be strong enough to seriously reduce ion transport to the substrate. However, they also show that the potential drops involved can vary by up to an order of magnitude from case to case. There is a clear need to understand the underlying mechanisms and identify the key discharge variables that can be used for minimizing the back-attraction. We here present a combined theoretical and experimental analysis of the problem of electric fields Ez in the ionization region part of HiPIMS discharges, and their effect on the transport of ionized sputtered material. In particular, we have investigated the possibility of a 'sweet spot' in parameter space in which the back-attraction of ionized sputtered material is low. It is concluded that a sweet spot might possibly exist for some carefully optimized discharges, but probably in a rather narrow window of parameters. As a measure of how far a discharge is from such a window, a Townsend product ΠTownsend is proposed. A parametric analysis of ΠTownsend shows that the search for a sweet spot is complicated by the fact that contradictory demands appear for several of the externally controllable parameters such as high/low working gas pressure, short/long pulse length, high/low pulse power and high/low magnetic field strength.
... For HiPIMS discharges, electron temperatures above 4 eV have indeed been observed in modeling [32,33,35], and this is expected to represent the electron temperature next to the target surface. Experimentally, however, usually lower electron temperatures are determined, which is typically explained by the fact that they are measured at some distance from the target surface for various target materials [36][37][38][39][40][41]. Dubois et al [12] applied Thomson scattering measurements and determined the electron temperature to be 1.3 eV at 5 mm from a titanium target surface. ...
The high power impulse magnetron sputtering (HiPIMS) discharge brings about increased ionization
of the sputtered atoms due to an increased electron density and efficient electron energization
during the active period of the pulse. The ionization is effective mainly within the electron trapping
zone, an ionization region (IR), defined by the magnet configuration. Here, the average extension
and the volume of the IR are determined based on measuring the optical emission from an excited
level of the argon working gas atoms. For particular HiPIMS conditions, argon species ionization
and excitation processes are assumed to be proportional. Hence, the light emission from certain
excited atoms is assumed to reflect the IR extension. The light emission was recorded above a 100
mm diameter titanium target through a 763 nm bandpass filter using a gated camera. The recorded
images directly indicate the effect of the magnet configuration on the average IR size. It is observed
that the shape of the IR matches the shape of the magnetic field lines rather well. The IR is found
to expand from 10 and 17 mm from the target surface when the parallel magnetic field strength 11
mm above the racetrack is lowered from 24 to 12 mT at a constant peak discharge current.
... The reason is that atoms ionized in the cathode region are likely to be back-attracted to the target due to strong electric fields in the presheath and extended presheath [10,11]. Spatial measurements of the plasma potential in HiPIMS discharges [11][12][13][14] have shown that there commonly is a potential uphill, from the cathode sheath edge and reaching far outside the ionization region (several cm), that can vary in the range 7-100 V. ...
... When the working gas pressure p is increased (while maintaining constant I D,peak and τ −,eff ) the magnitude of I i,sat w decreases. The decrease of I i,sat w when increasing p can be explained by a lower plasma density at the chamber walls [40] due to a slower plasma diffusion at higher p. For p of 0.3 Pa and 0.6 Pa we observe only one peak in the I i,sat w waveform occurring at around t = 35 μs. ...
The plasma potential at a typical substrate position is studied during the positive pulse of a bipolar high-power impulse magnetron sputtering (bipolar HiPIMS) discharge with a Cu target. The goal of the study is to identify suitable conditions for achieving ion acceleration independent on substrate grounding. We find that the time-evolution of the plasma potential during the positive pulse can be separated into several distinct phases, which are highly dependent on the discharge conditions. This includes exploring the influence of the working gas pressure (0.3 – 2 Pa), HiPIMS peak current (10 – 70 A corresponding to 0.5 – 3.5 A/cm2), HiPIMS pulse length (5 – 60 μs) and the amplitude of the positive voltage U+ applied during the positive pulse (0 – 150 V). At low enough pressure, high enough HiPIMS peak current and long enough HiPIMS pulse length, the plasma potential at a typical substrate position is seen to be close to 0 V for a certain time interval (denoted phase B) during the positive pulse. At the same time, spatial mapping of the plasma potential inside the magnetic trap region revealed an elevated value of the plasma potential during phase B. These two plasma potential characteristics are identified as suitable for achieving ion acceleration in the target region. Moreover, by investigating the target current and ion saturation current at the chamber walls, we describe a simple theory linking the value of the plasma potential profile to the ratio of the available target electron current and ion saturation current at the wall.
... R2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 in the pre-sheath and an extended pre-sheath Rauch et al., 2012). Spatial measurements of the plasma potential in HiPIMS discharges (Liebig and Bradley, 2013;Mishra et al., 2010;Rauch et al., 2012;Sigurjónsson, 2008) have shown that there commonly is a potential uphill, from the cathode sheath edge and reaching far outside the ionization region (several cm), that can vary in the range 7 -100 V. Some other mechanisms have also been suggested to contribute to the low deposition rate such as the non-linear yield effect (Emmerlich et al., 2008), sideways transport of the sputtered material , guiding effect of the Bfield and the effect of different ion species on sputter yield (Anders, 2010). Furthermore, it has been demonstrated that regions of intense ionization with locally enhanced potential, referred to as spokes, (see Section X.D.1) have a strong influence on the transport of species towards the substrate and de los Arcos et al. (2014) reported lower deposition rate/power when operating in the spoke-dominated regime than in the dc-like and homogeneous HiPIMS discharge regimes. ...
Magnetron sputtering deposition has become the most widely used technique for deposition of both metallic and compound thin films and is utilized in numerous industrial applications. There has been a continuous development of the magnetron sputtering technology to improve target utilization, increase ionization of the sputtered species, increase deposition rates, and to minimize electrical instabilities such as arcs, as well as to reduce operating cost. The development from the direct current (dc) diode sputter tool to the magnetron sputtering discharge is discussed as well as the various magnetron sputtering discharge configurations. The magnetron sputtering discharge is either operated as a dc or radio frequency discharge, or it is driven by some other periodic waveforms depending on the application. This includes reactive magnetron sputtering which exhibits hysteresis and is often operated with an asymmetric bipolar mid-frequency pulsed waveform. Due to target poisoning the reactive sputter process is inherently unstable and exhibits a strongly non-linear response to variations in operating parameters. Ionized physical vapor deposition was initially achieved by adding a secondary discharge between the cathode target and the substrate and later by applying high power pulses to the cathode target. An overview is given of the operating parameters, the discharge properties and the plasma parameters including particle densities, discharge current composition, electron and ion energy distributions, deposition rate, and ionized flux fraction. The discharge maintenance is discussed including the electron heating processes, the creation and role of secondary electrons and Ohmic heating, and the sputter processes. Furthermore, the role and appearance of instabilities in the discharge operation is discussed. © 2020 The Author(s). Published by IOP Publishing Ltd Printed in the UK
... Spatial measurements of the plasma potential in HiPIMS discharges (Sigurjónsson, 2008, Mishra et al., 2010, Rauch et al., 2012, Liebig and Bradley, 2013 have shown that there commonly is a potential uphill, from the cathode sheath edge and reaching far outside the ionization region (several cm), which can vary at least in the range 7 -100 V, as discussed in Sections 3.3.2 and 7.3.1.1. ...
... The reason is that atoms ionized in the cathode region are likely to be back-attracted to the target due to strong electric fields in the presheath and extended presheath [10,11]. Spatial measurements of the plasma potential in HiPIMS discharges [11][12][13][14] have shown that there commonly is a potential uphill, from the cathode sheath edge and reaching far outside the ionization region (several cm), that can vary in the range 7-100 V. ...
We explored the effect of magnetic field strength | B | and geometry (degree of balancing) on the deposition rate and ionized flux fraction F flux in dc magnetron sputtering (dcMS) and high power impulse magnetron sputtering (HiPIMS) when depositing titanium. The HiPIMS discharge was run in two different operating modes. The first one we refer to as “fixed voltage mode” where the cathode voltage was kept fixed at 625 V while the pulse repetition frequency was varied to achieve the desired time average power (300 W). The second mode we refer to as “fixed peak current mode” and was carried out by adjusting the cathode voltage to maintain a fixed peak discharge current and by varying the frequency to achieve the same average power. Our results show that the dcMS deposition rate was weakly sensitive to variations in the magnetic field while the deposition rate during HiPIMS operated in fixed voltage mode changed from 30% to 90% of the dcMS deposition rate as | B | decreased. In contrast, when operating the HiPIMS discharge in fixed peak current mode, the deposition rate increased only slightly with decreasing | B | . In fixed voltage mode, for weaker | B | , the higher was the deposition rate, the lower was the F flux . In the fixed peak current mode, both deposition rate and F flux increased with decreasing | B | . Deposition rate uniformity measurements illustrated that the dcMS deposition uniformity was rather insensitive to changes in | B | while both HiPIMS operating modes were highly sensitive. The HiPIMS deposition rate uniformity could be 10% lower or up to 10% higher than the dcMS deposition rate uniformity depending on | B | and in particular the magnetic field topology. We related the measured quantities, the deposition rate and ionized flux fraction, to the ionization probability α t and the back attraction probability of the sputtered species β t . We showed that the fraction of the ions of the sputtered material that escape back attraction increased by 30% when | B | was reduced during operation in fixed peak current mode while the ionization probability of the sputtered species increased with increasing | B | , due to increased discharge current, when operating in fixed voltage mode.
... The values are in all cases obtained from IRM modeling, except for Cu. In this latter case, T e was taken from Sigurjonsson [21], who carried out Langmuir probe measurements at close to the same current density in an Ar/Cu discharge (J D ≈1.2 A cm −2 instead of 1.3 A cm −2 ). ...
The combined processes of self-sputter (SS)-recycling and process gas recycling in high power impulse magnetron sputtering (HiPIMS) discharges are analyzed using the generalized recycling model (GRM). The study uses experimental data from discharges with current densities from the direct current magnetron sputtering range to the HiPIMS range, and using targets with self-sputter yields Y SS from ≈ 0.1 to 2.6. The GRM analysis reveals that, above a critical current density of the order of J crit ≈ 0.2 A cm⁻², a combination of self-sputter recycling and gas-recycling is generally the case. The relative contributions of these recycling mechanisms, in turn, influence both the electron energy distribution and the stability of the discharges. For high self-sputter yields, above Y SS ≈ 1, the discharges become dominated by SS-recycling, contain few hot secondary electrons from sheath energization, and have a relatively low electron temperature T e. Here, stable plateau values of the discharge current develop during long pulses, and these values increase monotonically with the applied voltage. For low self-sputter yields, below Y SS ≈ 0.2, the discharges above J crit are dominated by process gas recycling, have a significant sheath energization of secondary electrons and a higher T e, and the current evolution is generally less stable. For intermediate values of Y SS the discharge character gradually shifts between these two types. All of these discharges can, at sufficiently high discharge voltage, give currents that increase rapidly in time. For such cases we propose that a distinction should be made between 'unlimited' runaway and 'limited' runaway: in unlimited runaway the current can, in principle, increase without a limit for a fixed discharge voltage, while in limited runaway it can only grow towards finite, albeit very high, levels. For unlimited runway Y SS > 1 is found to be a necessary criterion, independent of the amount of gas-recycling in the discharge.
... A. Electron energy, electron density, and electrical potentials There have been several studies of the spatial and temporal variation of the electron density in the HiPIMS discharge using Langmuir probe diagnostics. 94,95,135,[140][141][142][143][144][145][146][147][148] Measurements of the temporal and spatial behavior of the plasma parameters indicate a peak electron density of the order of 10 18 -10 19 m À3 , 94,95,140 which during the ignition and growth phase expands from the target as an ion acoustic wave, with a fixed velocity that depends on the gas pressure. 149 This is roughly 2 orders of magnitude higher density than commonly observed in the substrate vicinity for a conventional dcMS discharge. ...
The high power impulse magnetron sputtering (HiPIMS) discharge is a recent addition to plasma based sputtering technology. In HiPIMS, high power is applied to the magnetron target in unipolar pulses at low duty cycle and low repetition frequency while keeping the average power about 2 orders of magnitude lower than the peak power. This results in a high plasma density, and high ionization fraction of the sputtered vapor, which allows better control of the film growth by controlling the energy and direction of the deposition species. This is a significant advantage over conventional dc magnetron sputtering where the sputtered vapor consists mainly of neutral species. The HiPIMS discharge is now an established ionized physical vapor deposition technique, which is easily scalable and has been successfully introduced into various industrial applications. The authors give an overview of the development of the HiPIMS discharge, and the underlying mechanisms that dictate the discharge properties. First, an introduction to the magnetron sputteringdischarge and its various configurations and modifications is given. Then the development and properties of the high power pulsed power supply are discussed, followed by an overview of the measured plasma parameters in the HiPIMS discharge, the electron energy and density, the ion energy, ion flux and plasma composition, and a discussion on the deposition rate. Finally, some of the models that have been developed to gain understanding of the discharge processes are reviewed, including the phenomenological material pathway model, and the ionization region model.
... Furthermore, even a small low-frequency noise, originating from the discharge itself or from the power supply, reduces the signal-to-noise ratio of the measurements and the dynamic range of the measured electron energy distribution function (EEDF), i.e. the ability to analyze the high energy tail of the EEDF is reduced. Employing averaging techniques over whole data sets helps in overcoming this problem to some extent [11]. In this work an average of 30 samples was used. ...
A dual-magnetron system for deposition inside tubular substrates has been developed. The two magnetrons are facing each other and have opposing magnetic fields forcing electrons and thereby also ionized material to be transported radially towards the substrate. The depositions were made employing direct current magnetron sputtering (DCMS) and high power impulse magnetron sputtering (HiPIMS). To optimize the deposition rate, the system was characterized at different separation distances between the magnetrons under the same sputtering conditions. The deposition rate is found to increase with increasing separation distance independent of discharge technique. The emission spectrum from the HiPIMS plasma shows a highly ionized fraction of the sputtered material. The electron densities of the order of 1016m−3 and 1018m−3 have been determined in the DCMS and the HiPIMS plasma discharges respectively. The results demonstrate a successful implementation of the concept of sideways deposition of thin films providing a solution for coating complex shaped surfaces.
... One can see that the density of Ar m at t = 400 µs is higher with increasing pulse width, showing that there is a stronger chance that they will survive longer in the afterglow phase. Furthermore, the increasing second T does not change significantly [37]. This means that the plasma conditions at pulse-off will be different depending on pulse-width, where the short pulses represent the peak of an ordinary HiPIMS pulse characterized by strong sputtering of metal, high ionization rate, and severe rarefaction of Ar, while the long pulses represent the decay phase with decreasing ionization rate, reduced gas depletion and strong gas refill. ...
Time-resolved tunable diode-laser absorption spectroscopy measurements were performed on the argon metastable (Arm) level 3s23p5 4s excited at 801.478 nm, in the dense plasma region in front of the magnetron target in a high power impulse magnetron sputtering (HiPIMS) discharge. From the Doppler profile the evolution of the temperature and density was derived during the pulse as well as during the plasma decay, i.e. in the afterglow. It is shown that the Arm density sharply increases at the beginning of the discharge pulse, followed by a severe Arm depletion along with increasing gas temperature around the peak of the HiPIMS discharge current. The strong dynamics of these parameters involve many elementary processes such as electron-impact excitation, electron-impact de-excitation and ionization of Arm, gas rarefaction, electron temperature increase at the end of the pulse and gas diffusion. These phenomena are discussed with respect to several parameters: distance from the target, peak discharge current during the pulse, pulse length, and gas pressure.
... In this model only energetic contributions due to charge carriers were taken into consideration. It was here assumed that the mean cathode voltage Sigurjonsson[21] using the same setup and under similar discharge conditions, typical values for the different plasma parameters were found, and we assume an From investigations on the ion composition in a Ti-Ar plasma by Bohlmark et al. by integrating the flux intensity curves of the different ions. The same magnetron and deposition system were also used in this experiment.The general expression for the energy flux from eq. (5) can thereby be written as , ...
The total energy flux in a high power impulse magnetron sputtering (HiPIMS) plasma has been measured using thermal probes. Radial flux (parallel to the magnetron surface) as well as axial flux (perpendicular to the magnetron surface) were measured at different positions, and resulting energy flux profiles for the region between the magnetron and the substrate are presented. It was found that the substrate heating is reduced in the HiPIMS process compared with conventional direct current magnetron sputtering (DCMS) at the same average power. On the other hand, the energy flux per deposited particle is higher for HiPIMS compared with DCMS, when taking into account the lower deposition rate for pulsed sputtering. This is most likely due to the highly energetic species present in the HiPIMS plasma. Furthermore, the heating due to the radial energy flux reached as much as 60% of the axial energy flux, which is likely a result of the anomalous transport of charged species present in the HiPIMS discharge. Finally, the experimental results were compared with theoretical calculations on energy flux of charged species and were found to be in good agreement.
The transport of charged particles in a high power impulse magnetron sputtering (HiPIMS) discharge is of considerable interest when optimizing this promising deposition technique with respect to deposition rate and control of the ion acceleration. In this study the internal current densities J (azimuthal direction) and Jz (axial direction) have therefore been spatially and temporally resolved in the bulk plasma region above a cylindrical magnetron using Rogowski coils. From the measurements a phenomenological model has been constructed describing the evolution of the current density in this pulsed plasma. The core of the model is based on three different types of current systems, which characterize the operating transport mechanisms, such as current transport along and across magnetic field lines. There is a gradual change between these current systems during the initiation, build-up and steady state of a HiPIMS plasma. Furthermore, the data also show that there are spatial and temporal variations of the key transport parameter J/Jz, governing the cross-B resistivity and also the energy of the charged particles. The previously reported faster-than-Bohm cross-B electron transport is verified here, but not for all locations. These results on the plasma dynamics are essential input when modeling the axial electric field, governing the back-attraction of ionized sputtered material, and might furthermore provide a link between the different resistivities reported in HiPIMS, pulsed-DC, and DC magnetron discharges.
The plasma probe diagnostic is the primary technique for obtaining information concerning the plasma parameters in low pressure gas discharges. In its classical application, known as the Langmuir technique, the probe method allows us to find the basic plasma parameters: the plasma density and the electron temperature. Once these parameters are obtained, we have enough information for the calculation of plasma hydrodynamic processes. The problem arises when one tries to calculate the rates of non-elastic interactions of the plasma electrons with molecules and ions using the measured electron temperature of the low energy electrons.
We report on the formation of multiple ion acoustic solitons in an unipolar pulsed magnetron plasma.A high density plasma ( ˜ 10^18 cm-3) is created by applying a high power pulse (3 -- 7 J) with pulse length 100 mus and repetition frequency 50 Hz to a planar magnetron discharge. The temporal behaviour of the electron density measured by a Langmuir probe shows oscillations as the plasma density decays. We relate these oscillations to solitons traveling away from the target. The velocity, width, and amplitude characteristics of the soliton are found to be in agreement with the basic properties of the soliton solutions of the Korteweg-de Vries equation. Furthermore, we demonstrate how the decay of the soliton along the axis indicates spherical symmetry of the soliton propogation. The speed of the soliton along the axis of the discharge decreases with increased gas pressure.
We demonstrate the creation of high-density plasma in a pulsed magnetron discharge. A 2.4 MW pulse, 100 μs wide, with a repetition frequency of 50 Hz is applied to a planar magnetron discharge to study the temporal behavior of the plasma parameters: the electron energy distribution function, the electron density, and the average electron energy. The electron density in the vicinity of the substrate, 20 cm below the cathode target, peaks at 8×1017 m−3, 127 μs after initiating the pulse. Towards the end of the pulse two energy groups of electrons are present with a corresponding peak in average electron energy. With the disapperance of the high-energy electron group, the electron density peaks, and the electron energy distribution appears to be Maxwellian like. Following the electron density peak, the plasma becomes more Druyvesteyn like with a higher average electron energy. © 2001 American Institute of Physics.
We report on the creation and propagation of ion-acoustic solitary waves in a high power pulsed magnetron sputtering discharge. A dense localized plasma is created by applying high energy pulses (4–12 J) of length ≈70 µs, at a repetition frequency of 50 pulses per second, to a planar magnetron sputtering source. The temporal behaviour of the electron density, measured by a Langmuir probe, shows solitary waves travelling away from the magnetron target. The velocity of the waves depends on the gas pressure but is roughly independent of the pulse energy.
In this study, electron energy distribution functions (EEDFs) are measured using a Langmuir probe in conjunction with the ac superposition method in the downstream region of a planar and unbalanced magnetron argon discharge and the effects of an anode sheath boundary on the discharge characteristics are investigated. The potential of the anode sheath can be controlled by applying a dc voltage to the substrate, and the nonlinear behaviour of the plasma potential with respect to the dc substrate voltage causes the distinctive evolution of the potential of the anode sheath. It is found that when the potential of the anode sheath reaches a specific value, which is related to the threshold energies of argon for the inelastic collisions, an outstanding EEDF transition from a bi-Maxwellian distribution to a single Maxwellian distribution occurs. We introduce the concept of the total electron bounce frequency as an indicator of how the electron collisions such as the electron–electron collision and the inelastic collisions affect the EEDF features as the potential of the anode sheath changes. This result provides the decisive clue to explaining the appearance of the bi-Maxwellian distribution in magnetron discharges. The results of the spatially resolved measurements of EEDF and plasma characteristics are also presented. From these results, we will discuss the electron transport in the downstream region in detail.
Over the past decade, various magnetron sputtering techniques have appeared that provide a high degree of ionization of the sputtered vapor. Here, ionized physical vapor deposition (IPVD) magnetron sputtering systems are reviewed. The application of a secondary discharge to a magnetron sputtering discharge, either an inductively coupled plasma source (ICP-MS) or a microwave amplified magnetron sputtering, is currently widely used. High power impulse magnetron sputtering (HiPIMS) is a sputtering technique where a high density plasma is created by applying high power pulses at low frequency and low duty cycle to a magnetron sputtering device. Other methods such as the self-sustained sputtering (SSS), and the hollow cathode magnetron (HCM) are discussed as well. Essential to these methods is a high electron density which leads to a high degree of ionization of the sputtered vapor.
Electrostatic probes or Langmuir probes are the most common diagnostic tools in plasma discharges. The second derivative of the Langmuir probe I-V characteristic is proportional to the electron energy distribution function. Determining the second derivative accurately requires some method of noise suppression. We compare the Savitzky–Golay filter, the Gaussian filter, and polynomial fitting to the Blackman filter for digitally smoothing simulated and measured I-V characteristics. We find that the Blackman filter achieves the most smoothing with minimal distortion for noisy data.
This paper presents a critical review of the present status and trends in the sputtering of thin films. A special attention is devoted to the magnetron sputtering, especially to unbalanced magnetrons (UM) and to new sputtering devices utilizing magnetic and/or electric plasma confinement. The last category of devices opens new possibilities in a production of high quality films. Sputtering systems using UM operated in the double‐site‐sustained discharge and in a closed multipolar magnetic field are described. The attention is also devoted to the effects of ion bombardment on the properties of the deposited films. In case of hard TiN coatings it is shown under what conditions dense compact porousless films can be created and how deposition conditions can influence stresses in films. At the end, several new systems which combine advantages of sputtering with other processes like the arc evaporation, the electron cyclotron resonance (ECR) plasma, and the laser irradiation are also briefly described.
We experimentally find that there are hot and cold electron components present in a sputtering magnetron plasma. The density of the hot component is greatest in the magnetic trap and decreases with the distance from the cathode. The cold electron density is negligible inside the trap and is approximately constant outside. The largest cold electron density is nearly as great as the hot electron density inside the magnetic trap.
In plasma-based deposition processing, the importance of low-energy ion bombardment during thin film growth can hardly be exaggerated. Ion bombardment is an important physical tool available to materials scientists in the design of new materials and new structures. Glow discharges and in particular, the magnetron sputtering discharge have the advantage that the ions of the discharge are abundantly available to the deposition process. However, the ion chemistry is usually dominated by the ions of the inert sputtering gas while ions of the sputtered material are rare. Over the last few years, various ionized sputtering techniques have appeared that can achieve a high degree of ionization of the sputtered atoms, often up to 50% but in some cases as much as approximately 90%. This opens a complete new perspective in the engineering and design of new thin film materials. The development and application of magnetron sputtering systems for ionized physical vapor deposition (IPVD) is reviewed. The application of a secondary discharge, inductively coupled plasma magnetron sputtering (ICP-MS) and microwave amplified magnetron sputtering, is discussed as well as the high power impulse magnetron sputtering (HIPIMS), the self-sustained sputtering (SSS) magnetron, and the hollow cathode magnetron (HCM) sputtering discharges. Furthermore, filtered arc-deposition is discussed due to its importance as an IPVD technique. Examples of the importance of the IPVD-techniques for growth of thin films with improved adhesion, improved microstructures, improved coverage of complex shaped substrates, and increased reactivity with higher deposition rate in reactive processes are reviewed.
The electron temperature, ion density, plasma potential and magnetic field are measured in a plasma produced by a magnetron sputtering device. The magnetic field has been mapped by measuring the axial and radial components with Hall probes. The plasma parameters have been measured by a diagnostic system consisting of several Langmuir probes with different collecting areas that can be positioned by a linear translator. The automatic analysis system allows one to take into account non-saturation of the ion current due to sheath thickness variation. The Langmuir probes have been used in various conditions of magnetic field and source power with a gas filling pressure of 1Pa. It has been found that the diagnostic system gives reliable measurements of the plasma parameters in static and inhomogeneous magnetic fields.
A two-dimensional spatial survey is conducted for a magnetron sputtering plasma at two pressures (40 and 5 mTorr) using a Langmuir probe. The plasma density is found to be highest (up to 6.0×1010 cm-3) above the etch region of the cathode, near the magnetic trap. The density drops between the etch regions, near the edges of the cathode and also at distances farther away from the cathode. The floating potential was found to be most negative (down to -12 V) in regions where the highest electron temperatures were observed (up to almost 4 eV) and became less negative (near 0 V) in regions where the electron temperature was lowest (less than 0.5 eV). This complementary trend was consistent in all spatial locations and at both pressures. The plasma potential was found to have very weak dependence, if any, on spatial location and pressure. The relationship between electron transport processes, collision processes and electron temperatures is discussed. Electron energy distribution functions were found to be either Maxwellian or bi-Maxwellian in nature, depending on pressure and spatial location. Maxwellian distributions were found near the magnetic trap or source of the plasma. Bi-Maxwellian distributions were found further away from the source, and it appears they result from Maxwellian distributions bifurcating as they diffuse away from the source. The suitability of the popular models for this bifurcation is discussed.
We have synthesized Ta thin films on Si substrates placed along a wall of a 2-cm-deep and 1-cm-wide trench, using both a mostly neutral Ta flux by conventional dc magnetron sputtering (dcMS) and a mostly ionized Ta flux by high-power pulsed magnetron sputtering (HPPMS). Structure of the grown films was evaluated by scanning electron microscopy, transmission electron microscopy, and atomic force microscopy. The Ta thin film grown by HPPMS has a smooth surface and a dense crystalline structure with grains oriented perpendicular to the substrate surface, whereas the film grown by dcMS exhibits a rough surface, pores between the grains, and an inclined columnar structure. The improved homogeneity achieved by HPPMS is a direct consequence of the high ion fraction of sputtered species.
The electron energy distribution function (EEDF) is calculated from digitized probe current-voltage characteristics by using a finite impulse response (FIR) filter, in which filter horizon and cutoff frequency are flexibly chosen. Good agreement between the measured EEDF and the theoretical one is obtained in application of the filter to a He positive column.
The DC discharge of a planar magnetron was enhanced by twinned microwave electron cyclotron resonance plasma source. The magnetic cusp geometry formed in the processing chamber was used for plasma confinement. The sputtering discharge characteristics was investigated and a combined mode of voltage and current was observed at a pressure as low as 0.007 Pa. Carbon nitride thin films were synthesized using this method. Characterization of the films showed that the deposition rate was high, and the films were composed of a single amorphous carbon nitride phase with N/C ratio close to that of C3N4, with the bonding mainly of CN type.
CONTENTS I. Electron part of probe characteristic with negative probe potential 768 II. Ion part of probe characteristic with negative probe potential 770 III. The method of two probes 777 IV. Use of the probe method under more complicated conditions 781 V. Improvement of probe measurement techniques and errors of the method 785 List of symbols 791 Cited literature 791
When a cylindrical or spherical electrode (collector) immersed in an ionized gas is brought to a suitable potential, it becomes surrounded by a symmetrical space-charge region or "sheath" of positive or of negative ions (or electrons). Assuming that the gas pressure is so low that the proportion of ions which collide with gas molecules in the sheath is negligibly small, the current taken by the collector can be calculated in terms of the radii of the collector or sheath, the distribution of velocities among the ions arriving at the sheath boundary and the total drop of potential in the sheath. The current is independent of the actual distribution of potential in the sheath provided this distribution satisfies certain conditions.
The chapter overviews the role of ionized physical vapor decomposition in integrated circuit fabrication. Ionized physical vapor deposition (I-PVD) is a deposition process in which the depositing species are initially vaporized by physical mechanisms. In addition, the flux of depositing species must be composed of >50% ions. Two other common deposition techniques using ions are known as ion beam-assisted deposition (IBAD) and ion-assisted deposition (lAD). These methods cobombard a film deposited from neutral atoms with inert gas ions. While the ratio of ion flux to the flux of depositing neutrals may be quite large in IBAD and lAD, the depositing species are primarily low-energy neutrals. The key distinction is that I-PVD seeks to deposit the film from ions. The main application for I-PVD is the formation of metal and nitride thin films into the deep, narrow trenches and vias that are found in modern integrated circuits (ICs). I-PVD methods suitable for IC fabrication are discussed in the chapter. Sputtering is preferred because it is relatively easy to create the large, uniform flux of metal needed to coat 200- to 300-mm-diameter wafers.
The electron energy distribution function (EEDF) and plasma parameters in a planar dc magnetron sputtering discharge in argon and krypton were determined using a Langmuir probe. Two groups of electrons are observed in the discharge. The electron temperature of the cold electrons is roughly independent of the discharge pressure, while the electron temperature of the hot electrons decreases with increased discharge pressure. The electron density increases with increased pressure and is roughly a factor of 2-3 higher for a krypton discharge than for an argon discharge.
Interconnects, once the technological backwater of integrated circuit technology, now dominate integrated circuit cost and performance. As much as 90 percent of the signal delay time in future integrated circuit designs will be due to the interconnection of semiconductor devices while the remaining 10 percent is due to transistor-related delay. This shifts the thrust of critical research toward an improved understanding of interconnect science and technology. Shrinking circuit geometries will require high aspect ratio (AR) vias to interconnect adjacent metal layers. By the year 2007 it is predicted that logic circuits will use 6 to 7 interconnected metal layers with via ARs of 5.2:1. Memory will need fewer layers, but ARs as high as 9:1. In this paper, the demands of interconnect technology will be reviewed and the opportunities for plasma-based deposition of vias will be discussed. One promising new method of fabricating high-aspect ratio vias is ionized physical vapor deposition (I-PVD). The technique economically creates a unidirectional flux of metal which is uniform over 200–300 mm diameter wafers. Since metal ejected by conventional sputtering is primarily neutral and exhibits a cosine angular velocity distribution, sputtered metal atoms do not reach the bottom of high AR vias. By sputtering these atoms into a moderate pressure (4 Pa), high-density Ar plasma, however, the metal atoms are first thermalized and then ionized. The ions are then readily collimated by the plasma sheath and directionally deposited into narrow, deep via structures. Experiments have consistently shown that over 80% of the metal species are ionized using I-PVD. The physical mechanisms responsible for ionization will be discussed from both an experimental and modeling perspective and the spatial variation of metal ionization is experimentally determined. © 1998 American Institute of Physics.
We report on electrostatic probe measurements of a high-power pulsed magnetron discharge. Space- and time-dependent characteristics of the plasma parameters are obtained as functions of the process parameters. By applying high-power pulses (peak power of ~0.5 MW), with a pulse-on time of ~100 µs and a repetition frequency of 20 ms, peak electron densities of the order of ~1019 m− 3, i.e. three orders of magnitude higher than for a conventional dc magnetron discharge, are achieved soon after the pulse is switched on. At high sputtering gas pressures (>5 mTorr), a second peak occurs in the electron density curve, hundreds of microseconds after the pulse is switched off. This second peak is mainly due to an ion acoustic wave in the plasma, reflecting off the chamber walls. This is concluded from the time delay between the two peaks in the electron and ion saturation currents, which is shown to be dependent on the chamber dimensions and the sputtering gas composition. Finally, the electron temperature is determined, initially very high but decreasing rapidly as the pulse is turned off. The reduction seen in the electron temperature, close to the etched area of the cathode, is due to cooling by the sputtered metal atoms.
A global (volume-averaged) model is applied to investigate the effect of the electron energy distribution function on the plasma parameters of a high-density, low-pressure, argon discharge. The effective electron temperature increases and the electron density decreases as the electron energy distribution is varied from being Maxwellian to become Druyvesteyn like. The sheath voltage decreases as the electron energy distribution function is varied from being Maxwellian to become Druyvesteyn like for low pressures ({<}2 mTorr) and increases for higher pressures as the electron energy distribution function is varied. Simple global model calculations demonstrated that increasing the operating pressure does not necessary lead to a higher electron density if the electron energy distribution evolves from Maxwellian to become more Druyvesteyn like as the operating gas pressure is increased.
Electron energy distribution functions (EEDFs) have been measured in low-pressure capacitive RF discharges over a wide range of well defined (geometrically and electrically) discharge conditions. Measurements have been made in argon and helium ranging in gas pressure between 3 mTorr and 3 Torr and in discharge current density between 0.1 mA cm-2 and 10 mA cm-2. The measurements show changes in the EEDF due to the occurrence of physical phenomena such as stochastic electron heating and the effect of discharge transition into the gamma mode. Substantial differences in the EEDF in Ramsauer and nonRamsauer gases are also demonstrated and discussed. To achieve these results a higher level of performance was required from the measurement system than had been attained in previous EEDF measurements in RF discharges. EEDF measurements were made using a probe system specifically designed to remove or reduce the severity of many problems inherent to such measurements in RF discharges. The rationale and considerations in the probe system design, as well as many construction details of the probe system itself, are discussed.
In this paper we present a study of how the magnetic field of a circular planar magnetron is affected when it is exposed to a pulsed high current discharge. Spatially resolved magnetic field measurements are presented and the magnetic disturbance is quantified for different process parameters. The magnetic field is severely deformed by the discharge and we record changes of several millitesla, depending on the spatial location of the measurement. The shape of the deformation reveals the presence of azimuthally drifting electrons close to the target surface. Time resolved measurements show a transition between two types of magnetic perturbations. There is an early stage that is in phase with the axial discharge current and a late stage that is not in phase with the discharge current. The later part of the magnetic field deformation is seen as a travelling magnetic wave. We explain the magnetic perturbations by a combination of E × B drifting electrons and currents driven by plasma pressure gradients and the shape of the magnetic field. A plasma pressure wave is also recorded by a single tip Langmuir probe and the velocity (~103 m s−1) of the expanding plasma agrees well with the observed velocity of the magnetic wave. We note that the axial (discharge) current density is much too high compared to the azimuthal current density to be explained by classical collision terms, and an anomalous charge transport mechanism is required.
Zwei Theorien des Niedervoltbogens werden besprochen, und zwar die Theorie von Holst und Oosterhuis und die von Compton und Eckart. Diese letztere wurde etwas erweitert und als die wahrscheinlichste Erklrung des Niedervoltbogens betrachtet. Die Erweiterung besteht in der Annahme einer Energiebertragung von schnellen auf langsame Elektronen; diese Wechselwirkung zwischen Elektronen wird bei groer Elektronenkonzentration (grer als 1012) hufig auftreten mssen. Sodann wurde das Spektrum des Ar-Niedervoltbogens besprochen. An Ar- und Ne-Niedervoltbogen wurden Sondenmessungen vorgenommen, wobei ein Potentialmaximum von ungefhr 11,4 Volt (in bezug auf die Spannung der Kathode) bei Ar und 18,5 Volt bei Ne gefunden wurde, entsprechend der Anregungsspannung der 1s-Niveaus bei Ar (11,7 Volt) und der 2p-Niveaus bei Ne (18,5 Volt). Die hchste Elektronenkonzentration betrgt ungefhr 2. 1012 Elektronen/cm3. Die Sondentheorie von Langmuir und Mott-Smith wird erweitert. Mit der Formel (3) in 5 kann man aus der Sondencharakteristik durch zweimaliges Differenzieren die Geschwindigkeitsverteilung der Elektronen finden.
A Langmuir probe measurement system suitable for characterization of extended, low‐pressure dc or microwave plasmas has been developed around a VAXStation 3200 graphics workstation computer. Both planar and cylindrical probes have been constructed, but the latter has been deemed more satisfactory, and its design developed more fully. A novel interface circuit based on fast high voltage, transformer‐coupled, isolation amplifiers has been constructed to permit probing of regions where the plasma potential is hundreds of volts away from ground, which are typically found in dc glow discharges. This circuit also incorporates a noise suppression feature, using feedback from a second identical probe (compensation probe), which has proven very useful. A great deal of high‐resolution data can be collected quickly using the fast direct memory access (DMA) hardware. Many analysis capabilities and graphical display possibilities are implemented in the FORTRAN control program, but special emphasis has been placed on extracting plasma potentials, electron densities, and electron energy distribution functions (EEDFs) from the first and second derivatives of the probe I‐V curve. These are obtained by numerical differentiation techniques that fully correct for ohmic voltage drops across the current sampling resistor. The capabilities of the system will be illustrated with several examples of probe data and EEDFs obtained in a large dc glow discharge system.
A technique has been developed for highly efficient postionization of sputtered metal atoms from a magnetron cathode. The process is based on conventional magnetron sputtering with the addition of a high density, inductively coupled rf (RFI) plasma in the region between the sputtering cathode and the sample. Metal atoms sputtered from the cathode due to inert gas ion bombardment transit the rf plasma and can be ionized. The metal ions can then be accelerated to the sample by means of a low voltage dc bias, such that the metal ions arrive at the sample at normal incidence and at a specified energy. The ionization fraction, measured with a gridded mass‐sensitive energy analyzer is low at 5 mTorr and can reach 85% at 30 mTorr. Optical emission measurements show scaling of the relative ionization to higher discharge powers. The addition of large fluxes of metal atoms tends to cool the Ar RFI plasma, although this effect depends on the chamber pressure and probably the pressure response of the electron temperature. The technique has been scaled to 300 mm cathodes and 200 mm wafers and demonstrated with Cu, AlCu, and Ti/TiN. Deposition rates are equal to or in some cases larger than conventional magnetron sputtering. A primary application of this technique is lining and filling semiconductor trenches and vias on a manufacturing scale.
The technology is reviewed with emphasis on implementation. PM sputtering is characterized by cathode potentials of 300–700 V and sputtering gas pressures of 1‐15 mTorr (0.1–2 Pa). Deposition rates are proportional to power density, which in turn is primarily limited by the thermal conductivity and cooling efficiency of the target. rf operating characteristics are similar to dc, but plasma plus target impedance is somewhat higher. For both rf and dc PM sputtering higher power (or current) densities are achievable at lower target potentials than for conventional sputtering. Permanent magnets and electro‐magnets have been used to produce closed electron‐trapping field patterns adjacent to the surface of both circular and rectangular planar targets. Plasma intensity and target erosion is a maximum where the magnetic field lines are parallel to the cathode surface. Deposition uniformity can be achieved by substrate motion combined with optimized magnet geometry. For a given material at equivalent deposition rates PM sputtering results in less substrate heating due to reduced bombardment by dark‐space‐accelerated electrons.
Conventional planar magnetrons have been characterized with small (0.01 and 0.02 cm diam) Langmuir probes in the plasma region and also extending into the sheath. The plasma potentials, electron temperatures, and electron densities have been measured at low and intermediate magnetron discharge currents. The low currents reduce the effect of the probe on the discharge by reducing probe heating. The pressure range examined was 1.5–30 mTorr in both Ar and He. With Ar, the plasma potential is relatively constant in the abnormal (bright) glow region of the magnetron, and only begins to drop appreciably in the dark space (≪1 mm thick) near the cathode. The electron temperatures showed a continual increase as the cathode sheath was approached. Temperatures were measured in the 1–5 eV range at 5–30 mTorr Ar, and as high as 22 eV at pressures of 1.5 mTorr Ar. The measured electron densities were also pressure dependent and were highly peaked in the bright glow region near the cathode surface. The densities fell rapidly through the sheath. Significant departures from a Maxwellian electron energy distribution were found for the He plasmas, with a higher proportion of electrons in the high energy tail. The densities, however, were significantly lower than with Ar.
The behavior of argon plasmas driven by time modulated power in ‘‘high density’’ plasma reactors is investigated using a global model. The time evolution of the electron temperature and the plasma density is calculated by solving the particle and energy balance equations. In the first stage of power application during the ‘‘on’’ time, the electron temperature rapidly increases above the steady state value. In this region, charged particles accumulate in the plasma due to the relatively higher power applied than for the continuous wave (cw) case. In the first stage of the ‘‘off’’ time, the electron temperature drops quickly, yielding a smaller particle loss (Bohm) velocity. These effects give rise to higher time‐average plasma densities than for the cw plasma driven by the same average power. The highest average plasma density obtained was more than twice the density of the cw plasma for a duty ratio of 25%. Even higher plasma densities were obtained for shorter duty ratios. The possibility of controlling chemical reactions in the plasma by changing the modulation period is also shown. © 1995 American Vacuum Society
A new deposition technique has been developed which combines conventional magnetron sputter deposition with a rf inductively coupled plasma (RFI). The RFI plasma is located in the region between the magnetron cathode and the sample position, and is set up by a metal coil immersed in the plasma. A large fraction of the metal atoms sputtered from the magnetron cathode are ionized in the RFI plasma. By placing a negative bias on the sample, metal ions are then accelerated across the sample sheath and deposited at normal incidence. Results from a gridded energy analyzer configured with a microbalance collector and located at the sample position indicate the level of ionization is low at a few mTorr and rises to ≳80% at pressures in the 25–35 mTorr range. Optical measurements of metal ion and neutral emission lines show scaling of the relative ionization to higher discharge powers. Significant cooling of the plasma electron temperature is observed when high concentrations of metal atoms were sputtered into the plasma.
We demonstrate the evolution of the electron, energy distribution and the plasma parameters in a high-density plasma in a pulsed magnetron discharge. The high-density plasma is created by applying a high power pulse (1–2.4 MW) with pulse length 100 μs and repetition frequency of 50 Hz to a planar magnetron discharge. The spatial and temporal behavior of the plasma parameters are investigated using a Langmuir probe; the electron energy distribution function, the electron density and the average electron energy. The electron energy distribution function during and shortly after the pulse can be represented by a bi-Maxwellian distribution indicating two energy groups of electrons. Furthermore, we report on the variation of the plasma parameters and electron energy distribution function with gas pressure in the pressure range 0.5–20 mtorr. We report electron density as high as 4×1018 m−3 at 10 mtorr and 9 cm below the target in a pulsed discharge with average power 300 W. We estimate the traveling speed of the electron density peak along the axis of the discharge. The traveling speed decreases with increased gas pressure from 4×105 cm/s at 0.5 mtorr to 0.87×105 cm s−1 at 10 mtorr. The effective electron temperature peaks at the same time independent of position in the discharge, which indicates a burst of high energy electrons at the end of the pulse.
Using a cylindrical Langmuir probe, the temporal evolution of the electron density ne, electron temperature Te and plasma potential Vp has been measured at positions along the centre line in a 100-kHz pulsed magnetron plasma. The duty cycle was fixed at 50% and the Ar gas pressure was 0.27 Pa. At the beginning of the ‘on’ phase, a population of hot electrons was observed (for approx. 1 μs), with Te and ne approximately 50% higher than their time-averaged values. During the remainder of the ‘on’ phase, the electrons heat and then cool, following the rise and dip in the target potential. At the beginning of the ‘off’ phase, ne is seen to fall rapidly by over 60% at all positions and Te rises, however, both parameters recover after 1 μs and then slowly decay with a time constant of not, vert, similar40 μs. During the transition from ‘on’ to ‘off’ phases, Vp, rises rapidly from its usual value, a few volts above ground potential, to a few volts above the cathode potential (e.g. Vpnot, vert, similar+25 V). Some simple explanations for the observations are given.
The spatial electron density distribution was measured as function of time in a high-power pulsed magnetron discharge. A Langmuir probe was positioned in various positions below the target and the electron density was mapped out. We recorded peak electron densities exceeding 10/sup 19/ m/sup -3/ in a close vicinity of the target. The dynamics of the discharge showed a dense plasma expanding from the "race-track" axially into the vacuum chamber. We also record electrons trapped in a magnetic bottle where the magnetron magnetic field is zero, formed due to the unbalanced magnetron.
Measurements of pulsed-power modulated argon plasmas in an inductively coupled plasma source
- S Ashida
- M R Sim
- M A Lieberman
Ashida, S., M. R.Sim, and M. A. Lieberman (1996). Measurements of pulsed-power
modulated argon plasmas in an inductively coupled plasma source. Journal of
Vacuum Science and Technology A 14 (2), 391 -397.
On smoothing of the I-V Langmuir probe characteristic. Memorandum No. UCB/ERL M97/38
- J T Gudmundsson
Gudmundsson, J. T. (1997). On smoothing of the I-V Langmuir probe characteristic.
Memorandum No. UCB/ERL M97/38, Electron Research Laboratory, University
of California, Berkeley.
High power impulse magnetron sputtering discharges and thin film growth: A brief review
- U Helmersson
- M Lattemann
- J Alami
- J Bohlmark
- A P Ehiasarian
- J T Gudmundsson
Helmersson, U., M. Lattemann, J. Alami, J. Bohlmark, A. P. Ehiasarian, and J. T.
Gudmundsson (2005). High power impulse magnetron sputtering discharges and
thin film growth: A brief review. In 48th Annual Technical Conferenece Proceedings, Denver, CO, USA, pp. 458 -464. Society of Vacuum Coaters.
Time-resolved Langmuir probe diagnostics of high-power pulsed dc magnetron discharges during deposition of copper films
- A D Pajdarová
- J Vlcek
- P Kudlácek
- J Lukás
- J Musil
Pajdarová, A. D., J. Vlcek, P. Kudlácek, J. Lukás, and J. Musil (July 15 -20, 2007).
Time-resolved Langmuir probe diagnostics of high-power pulsed dc magnetron
discharges during deposition of copper films. In J. Schmidt, M.Simek, S. Pekárek,
and V. Prukner (Eds.), Proceedings of the XXVIII International Conference on
Phenomena in Ionized Gases, Prague, Czech Republic, pp. 1960 -1962. Institute
of Plasma Physics AS CR, v.v.i.
A new smoothing method for obtaining the electron energy distribution function in plasma by the numerical differentiation of the I − V probe characteristic. Review of
- J I F Palop
- J Ballesteros
- V Colomer
- M A Hernández
Palop, J. I. F., J. Ballesteros, V. Colomer, and M. A. Hernández (1995). A
new smoothing method for obtaining the electron energy distribution function
in plasma by the numerical differentiation of the I − V probe characteristic. Review of Scientific Instruments 66 (9), 4625 -4636.
Electric Probes for Low Temperature Plasmas
- D N Ruzic
Ruzic, D. N. (1994). Electric Probes for Low Temperature Plasmas. New York: The
American Vacuum Society Education Committee.