- Email: [email protected]
Formation of repetitively pulsed high-intensity, low-energy silicon ion beams
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
Contents lists available at ScienceDirect
Nuclear Inst. and Methods in Physics Research, A journal homepage: www.elsevier.com/locate/nima
Formation of repetitively pulsed high-intensity, low-energy silicon ion beams Alexander I. Ryabchikov, Denis O. Sivin ∗, Sergey V. Dektyarev, Alexey E. Shevelev National Research Tomsk Polytechnic University, 634050, pr. Lenina 2, bdg. 4, Tomsk, Russia
ABSTRACT This paper investigates the formation of pulsed and repetitively pulsed high-intensity, low-energy silicon ion beams. The formation of a pulsed silicon plasma was carried out by a vacuum arc discharge. Polycrystalline neutron-doped silicon was used as the evaporator’s cathode. A pulsed vacuum arc discharge formed a directed silicon plasma flow with a duration of approximately 350 μs. On the path of plasma transportation at a distance of 35 cm from the cathode surface, a system for forming a ballistically focused ion beam was installed. When a pulsed or a repetitively pulsed negative-bias potential is applied to the extracting grid electrode in the shape of a second-order surface, it ensured the formation of a sheath, the extraction of silicon ions from the plasma, and their acceleration in the layer. The shape of the extracting electrode provided the possibility of ballistic focusing of the silicon ion beam. Repetitively pulsed generation of the bias potential provided the possibility of preliminary plasma injection into the beam drift space between the bias-potential pulses. The paper has studied the features and regularities of the pulsed (with a duration of bias pulses up to 34 μs) and repetitively pulsed (at frequencies of bias pulses of 20 and 100 kHz) silicon ion beams formation with a current of up to 2 A at a maximum ion current density of 0.8 A/cm2 at bias potentials of small amplitudes (0.6–1.8 kV).
1. Introduction Various methods of ion-plasma deposition of coatings and modification of surface and near surface properties of different materials by ion implantation and exposure to various types of radiation play an increasingly important role in the creation of new generation materials with improved performance properties [1–4]. While technology provides many advantages for improving the macroscopic properties of materials’ surfaces by deposition of coatings, there are significant drawbacks associated with the presence of problems with the adhesion of coatings to the base of the material being processed [5,6]. Ion implantation provides a solution to the problems of adhesion, creating ion-doped layers with a given composition within the material itself. The disadvantage of the ion-implantation method is the small projective range of ions and, as a result, a small thickness of the ion-doped layer. The long-range effects that change the dislocation structure at great depths during ion implantation have little effect on the performance properties of structural materials and coatings [4]. To improve the performance properties of various materials by ionplasma deposition of coatings and ion implantation, the use of silicon as a dopant is of particular interest. For example, the presence of silicon in different steels and other composite materials (metal nitrides and carbides) has a positive effect on the mechanical properties of the raw materials [7–9]. As a rule, increased silicon concentration leads to an increase in the material strength and in resistance to tempering, a decrease in grain growth at high temperatures, and recrystallization of composite compounds [10]. Amorphous or nearly amorphous transition-metal-nitride coatings are expected to exhibit unique physical properties that are not observable in their crystalline counterparts. One of the potentially attractive
qualities is the missing grain boundary, which otherwise builds a ‘weak point’ at which gas atoms from the atmosphere, working material or substrate may diffuse [11]. Amorphous transition-metal-Si-N materials have been used as diffusion barriers in microelectronics [12,13] or in coatings with high temperature oxidation resistance [14–17]. The addition of a metalloid such as Si into transition metal nitrides is an effective way to stabilize a nanocomposite or amorphous structure [18–21]. In sufficient content (> 10 at. %) due to its fourfold coordination, Si was shown to severely distort the TiN cubic lattice [22]. Musil et al. reported that the addition of high Si content > 25 at. % into ZrN results in an amorphous structure that is stable up to > 1000 ◦ C [14]. The structural stability, together with the ability of Si to form solid nonvolatile oxide (SiO2 ), promotes the oxidation resistance of transition metal-Si–N systems [15–17]. Recently, the possibility of using high-intensity ion beams with ion current densities from 0.01 to 1 A/cm2 and an ion energy of several keV has been demonstrated for surface modification of various materials by ion implantation. Work [23] presents the results of the formation of modified nitrogen-containing layers with thicknesses of up to 180 μm in steel when processed with a high-intensity nitrogen beam for 60 min. Studies of the formation of intermetallic compounds by high-intensity beams of zirconium ions [24] and aluminum [25] in targets made of titanium and nickel, respectively, demonstrated the possibility of obtaining modified layers with thicknesses of several tens of micrometers. The implementation of research in the direction of deep ion doping of various materials and coatings with silicon requires creating systems and sources that form high-intensity, low-energy silicon ion beams. This paper is devoted to research on the formation of pulsed highintensity axially symmetric beams of low-energy silicon ions with ion
∗ Corresponding author. E-mail address: [email protected] (D.O. Sivin).
https://doi.org/10.1016/j.nima.2019.163092 Received 21 June 2019; Received in revised form 5 November 2019; Accepted 5 November 2019 Available online xxxx 0168-9002/© 2019 Elsevier B.V. All rights reserved.
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
Fig. 1. Simplified scheme of the experimental setup for the formation of pulsed high-intensity silicon ion beams.
current densities of up to several hundred milliamperes per square centimeter.
ballistic focusing when it is transported in the equipotential drift space previously filled with plasma [29]. Extraction and focusing of the beam were carried out by a grid electrode with a cell size of 1.8 × 1.8 mm2 , transparency of 70%, made in the form of a sphere with a diameter of 15 cm. Construction of the grid electrode in the form of a second-order nondegenerate surface with central symmetry implies ballistic focusing of a particle flow, normally rooted relative to the surface of the grid electrode, to some focus. For an electrode as a part of a sphere, the focus is in the geometric center of the sphere at a distance of 7.5 cm. The grid electrode was installed on one of the bases of a solid metal cylinder with a height of 7 cm and a diameter of 14 cm. This construction allowed to allocate some volume, which limited the equipotential space for transporting the ion beam. On the opposite base of the cylinder, various collectors and sensors were installed, whose position relative to the grid electrode along the axis of the system could be regulated. In experiments with measurements of the ion beam current, a flat collector made of tungsten with a diameter of 4 cm was used. The working plane of the collector was located on the axis of the focusing system in the geometric focus of the grid electrode at a distance of 7.5 cm. The bias potential was fed to the system when the plasma flow already reached the ion collector. Thus, the plasma-immersion system was used for the initial formation of the ion flux. The acceleration of ions and their transport in the equipotential drift space was performed using a pulsed bias potential that was applied simultaneously to the grid electrode, solid cylindrical electrode and collector. The potential was formed by a generator of high-frequency short-pulse bias voltages of negative polarity (PG in Fig. 1). The generator could operate both in high-frequency repetitively pulsed mode and in a mode of forming single pulses. In single-pulse mode, the generator allowed the formation of a negative-bias-potential pulse with an amplitude in the range from 0.2 to 1.8 kV and a duration from 5 to 100 μs. In repetitively pulsed mode, the generator provided the formation of a sequence of biaspotential pulses with a frequency from 1 to 100 kHz and a pulse duty factor from 20 to 80%.
2. Experimental setup and methods Studies of the formation of pulsed high-intensity silicon ion beams were conducted with an experimental setup whose scheme is shown in Fig. 1. The vacuum chamber was pumped out to a residual pressure of 10−3 Pa using a turbomolecular pump with a capacity of 1000 l/s. To generate silicon plasma, a vacuum-arc discharge was used; this makes it possible to obtain a sufficiently dense plasma from conducting materials with a high degree of ionization in vacuum [26]. A planar vacuum-arc evaporator with a magnetic coil was installed on the vacuum chamber to control the plasma flow density. The cathode of the vacuum-arc evaporator was a plate of doped silicon with a diameter of 10 cm with an average specific resistance of 0.013 Ohm cm. The material of the cathodes was obtained by the neutron-transmutation doping method of single-crystal silicon in a research nuclear pooltype reactor IRT-T (Tomsk, TPU) [27,28]. The pulsed power supply of the arc was implemented by a forming line with a matched load (SL in Fig. 1), providing a pulse with a duration of approximately 350 μs. The ignition-electrode pulse control and its synchronization with the arc power pulse were performed by the ignition block (TPS) and master generator (MPG). Electromagnetic coils ensured the creation of a longitudinal magnetic field, limiting the plasma angular spread. The high-intensity ion beams formation system, the circuit diagram and the appearance of which is shown in Fig. 2, was installed inside the vacuum chamber opposite the vacuum-arc evaporator on its symmetry axis at a distance of 35 cm from the cathode. The beam formation system consists of several basic elements: a grid electrode, a solid cylindrical electrode, a collector and a disk screen. The operation of the beam formation system is based on the principles of plasmaimmersion extraction, ion acceleration in the sheath and ion beam 2
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
Fig. 2. External view (a) and cross-sectioned scheme (b) of the ion beam formation system.
The parameters of the pulsed bias potential were measured by an active voltage divider. The beam current at the collector and the current in the circuit of the focusing system were measured using Rogowski coils (L1 and L2). Current pulses and bias voltages were recorded by digital multichannel oscilloscopes. Two delay pulse generators (DP1 and DP2) were added to the measurement part of the experimental setup. These generators helped control delay in the negative-bias pulse feeding to the beam formation system relative to the ignition pulse of the vacuum-arc discharge, as well as the synchronization of the measuring oscillographs relative to these pulses. In addition, the delay pulse generators made it possible to control the bias-pulse duration during ion-beam formation. When plasma is created by a vacuum-arc discharge as a result of explosive emission processes on a silicon cathode, a significant flow of macroparticles is generated (droplets, solid fragments ranging in size from fractions of micrometers to tens of micrometers) that reduces the technological capabilities of the ion beam formation system due to contamination of the treated sample surfaces by cathode erosion products [30]. The solar-eclipse effect was used to clean the ion beam of the vacuum-arc discharge without significantly reducing the ion current [29]. On the grid electrode, a disk screen with a diameter of 4 cm was installed along the symmetry axis of the beam-formation system. The screen used in this position is an obstacle to the nearly straight-line trajectories of macroparticles emitted from the cathode to the collector area installed at the system focus. At the same time, due to the ballistic focusing using a spherical grid electrode along straight-line trajectories, the beam ions freely reach the irradiated target area on the collector. The use of an additional screen changes the total ion current in the beam only slightly, since it covers no more than 4% of the total area of the grid electrode (approximately 300 cm2 ) that accelerates and extracts ions from the plasma. Experimental results are presented in [29], demonstrating the almost complete absence of macroparticles in a DC vacuum-arc discharge in the region of the ion beam impact on the target using the ‘‘solar eclipse’’ effect. To study the shape of the ion current-density distribution profile of a pulsed high-intensity silicon ion beam, a 19-channel sensor was installed at the collector site. Fig. 3 shows the appearance of the sensor, which consists of a set of 19 single collectors fixed along one line. Each collector has a size of 2 × 2 mm2 . After beam formation and focusing, the ions fall on the collectors. At the same time, at the beginning of beam formation, the collectors are under the potential of the grid and cylindrical electrodes. To accumulate the charge of the beam ions deposited on the collectors, capacitors were connected to the electrical circuits of each of them. After each pulsed-ion beam formation, sequential polling of voltages from capacitors was conducted using a special device. The ion current
Fig. 3. Top view of 19-channel sensor for ion beam distribution measurement.
density at each point (on each collector) is proportional to voltage U and capacitor capacitance C and inversely proportional to current-pulse duration 𝜏 and collector area 𝑆𝑘 𝑗=
C⋅𝑈 . 𝜏 ⋅ 𝑆𝚔
Thus, measuring the voltages at the collectors, one can determine the integral charge of the ions that fell on each collector and, as a result, calculate the current density of silicon ions on each collector. 3. Features and regularities of high-intensity pulsed silicon ion beams After the start of the MPG master generator, a pulse is generated at its output, which is fed to the TPS ignition-pulse-formation block and to the delayed pulse generators DPG1 and DPG2 (Fig. 1). When supplying a pulsed high voltage to the ignition electrode, a breakdown occurs on the surface of the ceramic between the cathode and ignition electrode. The voltage of the forming line SL ensures the formation and maintenance of a discharge between the cathode and anode of the vacuum-arc plasma generator. The cathode spot formed on the cathode surface generates a plasma flow with a high velocity directed perpendicular to the surface of the cathode, that is, towards the ionbeam formation system [31]. To create the conditions necessary for the efficient transfer of a high-intensity ion beam, it is necessary to preliminarily fill the drift space of the ion beam with plasma [29]. Based on this condition, the delay time to feed a negative high-voltage bias pulse should be chosen from the condition: 𝑡 > (𝐿 + 𝑅) ∕𝜐𝑝 , where L — distance from the silicon cathode to the grid electrode, R — distance from the grid electrode to the collector, and 𝜐𝑝 — plasma flow velocity. 3
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
of a flat accelerating gap with an average density of ion saturation current of 0.011 A/cm2 and a voltage in the sheath of 1.6 kV indicate that the sheath width is approximately 3 mm ((𝑅 + 𝑑)2 ∕𝑅2 = 1.08). After acceleration in the sheath, silicon ions penetrate into the initial equipotential drift space bounded by grid and solid cylindrical electrodes. One of the main conditions for high-intensity ion beam formation is that ion focusing should be conducted under conditions of partial or complete neutralization of the beam space charge when it is transported in the drift space. In the absence of a mechanism for the beam space charge neutralizing, its defocusing will be observed, and a virtual anode can also be formed that will lead to a breakdown in the ion-beam transport. Fig. 5 shows the most characteristic oscillograms of silicon-ion-beam current to a collector and an ion current to a grid electrode for various bias-pulse durations and pulse amplitudes: 0.6, 1.1 and 1.5 kV. The duration of the bias pulses on the beam-formation system varied from 8 to 34 μs. The presented data demonstrate a number of features and regularities of formation and transportation of pulsed silicon ion beams. The oscillograms of the ion current in Fig. 5a show that when the amplitude of the bias potential is 0.6 kV and the pulse duration of the bias is 8 μs, the ion current to the collector has a delay relative to the current to the grid electrode. This delay is due to the final time of sheath formation and ion-beam transport from the grid electrode to the collector. The steepness of the leading edge of the ion-beamcurrent pulse turns out to be significantly less than that of the current to the grid electrode. This behavior of the ion-beam-current pulse is due to the processes of ion-beam space-charge neutralization. When the ion beam enters the drift space, its space charge is not compensated. The presence of cold electrons in plasma, previously injected into the drift space, does not instantly neutralize the ion beam space charge. It takes time to remove plasma ions from the beam volume. As shown by numerical simulation [36], depending on the parameters of the ion beam and plasma, the time taken to neutralize the beam space charge can be several microseconds. As the charge neutralization of the ion beam increases, the conditions for its transport and focusing improve. The leading edge of the ion beam current indicates that this time in this case exceeds 3 μs. The delay of the back edge of the beam-current pulse relative to the end of the bias pulse is associated with the finite ion time-of-flight from the grid electrode to the collector through the drift space. An increase in the bias pulse duration to 17 μs is accompanied by a corresponding increase in the duration of the ion current to the grid electrode (Fig. 5a). At the same time, duration shortening of the ion current to the collector is observed compared with the bias pulse duration. Ion-beam transport disruption occurs. The probability of a break increases with an increase in the current pulse duration of the ion beam from 10 to 15 μs. It is characteristic that a decrease in the ion current to the collector is accompanied by an increase in the ion current to the ion-beam-formation system. Most likely, this effect is associated with ion-beam space-charge decompensation due to the gradual escape of plasma electrons into the accelerating gap through the grid electrode cells. A decrease in the number of electrons leads to an increase in the uncompensated ion-beam charge and the violation of the conditions for its transport. The disruption of the ion beam leads to virtual anode formation in which a significant portion of the ions is affected by the electric field, changing their movement to the opposite direction. Finally, after a small number of oscillations relative to the grid electrode, ions are deposited on it, leading to a corresponding increase in current. The use of longer bias pulses (27 μs and 34 μs) also demonstrates the development of instability in transport of the ion beam in drift space, at which the amplitude of the beam ion current to the collector decreases sharply to some minimum value. However, after some time, with these durations, in addition to reducing the ion-beam-current amplitude, the initial current level to the collector is restored. This means that the conditions for neutralizing the space charge of a high-intensity
Fig. 4. Oscillograms of arc discharge and ion current to the collector.
The experiments used a pulsed vacuum-arc discharge with a maximum current of 75 A and a pulse duration of 350 μs. Fig. 4 shows oscillograms of a pulsed current of a vacuum-arc discharge and a current of silicon ions to a collector installed at a distance (𝐿 + 𝑅) = 42.5 cm from the cathode plane of the vacuum-arc evaporator. The pulse shape of the ion saturation current from the plasma, measured with a constant negative 100 V bias applied to the collector, correlates with the current pulse of the vacuum-arc discharge. The observed amplitude instability of the current to the collector reflects changes in the density of the plasma flow, which is determined by unsteady processes in the cathode spots of the vacuum-arc discharge [32]. Oscillograms show a delay of approximately 20 μs in the appearance of current on the collector relative to the voltage pulse of the vacuum-arc discharge. Expression 1 allows estimation of the plasma velocity 𝜐𝑝 = 2.1 ⋅ 106 cm/s, in agreement with the data of [33], which experimentally shows that for a plasma of a pulsed vacuumarc discharge, depending on the cathode material, the velocity of its directional propagation varies within (1.5–3) ⋅ 106 cm/s. As measured in the experiment, the average density of the ion saturation current from the plasma along the system axis at a distance corresponding to the grid electrode location was 0.011 A/cm2 150 μs after the onset of the vacuum arc discharge. All experimental studies in the single-pulse mode of focused ion beam formation were conducted with a delay of 100 μs between the start of plasma generation and the supply of negative bias potential to the beam-formation system. This provided preliminary filling of the ion beam drift space with plasma. After a negative bias is applied to the grid electrode, a sheath is formed near it. Initially, due to their high mobility, the plasma electrons leave the electrode space, and the socalled matrix sheath is formed. The width of this sheath depends on the amplitude of the bias potential and the concentration and temperature of electrons [3]. Under the action of an electric field in the formed sheath, ions begin to accelerate gradually. The acceleration of ions, in turn, leads to a change in the ion density in the accelerating gap and to the sheath dynamic expansion. The process of plasma emissionboundary stabilization is completed after the sheath expands to a size determined from the condition of ion-current limiting by the spatial beam charge. The limiting ion current density is determined by the Child–Langmuir law [34,35]. Comparison of the grid electrode curvature 𝐾𝑔𝑒 = 1∕𝑅2 , made as part of a sphere, and the plasma emission boundary curvature 𝐾𝑠 = 1∕(𝑅 + 𝑑)2 , where R — grid electrode radius and d — sheath width, allows us to consider the electrode and the emission boundary as flat electrodes under the criterion (𝑅 + 𝑑)2 ∕𝑅2 ∼ 1. An estimate of the emission boundary position in the approximation 4
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
Fig. 5. Oscillograms of the silicon ion beam current and ion current to the grid electrode at different amplitudes and durations of the bias potential pulses: (a) 0.6 kV; (b) −1.0 kV; (c) −1.5 kV.
5
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
ion beam are restored. Because during the pause, the bias potential is applied to the grid electrode, the plasma from the vacuum arc evaporator cannot penetrate into the drift space of the ion beam. This means that reneutralizing of the ion-beam space charge is not provided by the injected plasma electrons. Most likely, plasma production in the beam drift space occurs due to ionization of the residual atmosphere in the vacuum chamber with ions of the ion beam itself. Initially after a current break, there is a gradual very slow increase in the ion current to the collector. This is especially well observed from the oscillograms presented in Fig. 5b. It is characteristic that upon reaching a certain level, a sufficiently rapid increase in the amplitude of the ion-beam current falling on the collector occurs. Another possible mechanism of plasma formation in the drift space should be noted. The ion-beam formation system is under a negative bias potential and is actually a hollow cathode. The injection of ions into the drift space is accompanied by the formation of a region with a positive potential relative to the cathode. A beam collapsing under the action of a space charge ionizes the residual gas both inside the beam and at its periphery. Under the action of the electric field of the ‘‘hollow cathode – space of positive potential’’ system, the peripheral plasma electrons will accelerate in the direction of the positive potential and decelerate when passing through it without reaching the cathode. Electron oscillation in this space will lead to ionization of the residual gas by electron impact. Oscillation of plasma electrons can take place both in the radial direction and along the ion flux distribution. Secondary ion electron emission from the collector can also produce cold electrons. The slow gradual increase of the current to the collector after the breakdown of the ion beam transport most likely indicates its gradual advancing in the drift space. Reducing the distance between the front of the beam and the collector leads to an increase in the amplitude of the limiting ion current, with an uncompensated space charge transported in a large volume of the ion beam. A sharp increase in the current of the ion beam arriving at the collector must be accompanied by an avalanche-like increase in plasma density. However, further research is needed to clarify the type of breakdown. An increase in the bias potential amplitude on the focusing system to 1.0 kV does not fundamentally change the picture and leads to only a slight improvement in the ion beam transportation conditions. The oscillograms in Fig. 5b show an increase in the ion-beam-current amplitude recorded by the collector. This effect is associated with an increase in the sheath width with increasing bias amplitude in accordance with the Child–Langmuir law. At low bias potentials, the sheath width turns out to be comparable to the size of the grid electrode cell. This leads to the curvature of the plasma ion-emission boundary, and a significant part of the ions is accelerated, falling directly on the grid structure elements. As the amplitude of the accelerating voltage increases, the sheath width increases, and the plasma ion-emission boundary, moving away from the grid electrode, gradually straightens. This leads to ion flux formation at the entrance to the focusing system with a smaller angular divergence. As a result, with an increase in the negative-bias amplitude, the ion beam current in the focal plane of the system increases. From a comparison of the oscillograms in Fig. 5a and 5b, it can be concluded that with an increase in the bias amplitude, the pause duration during which the high-intensity silicon ion beam is not transported decreases by almost 30%. A further increase in the bias-potential amplitude on the focusing system up to 1.5 kV provided not only an increase in the amplitude of the ion-beam current to the collector but also a stable silicon-ionbeam formation without transport disruptions over the entire range of the bias potential pulse durations (Fig. 5c). Oscillograms show that the amplitude of the ballistic-focused ion-beam current in the focal region on the collector exceeds 1.5 A. It is obvious that single pulses of a high-intensity silicon ion beam have a limited area of practical application. For example, they can be used in studies of the possible existence of a shock-wave diffusion mechanism for a dopant during high-intensity implantation of
Fig. 6. Oscillograms of the collector current during the silicon ion beam formations from the plasma of a pulsed vacuum arc at a repetitively pulsed bias potential with amplitudes of 0.6, 1 and 1.5 kV (frequency is 20 kHz, duty factor is 50%).
low-energy ions (1–5 keV) at ion current densities of more than 0.6 A/cm2 [37]. To modify the properties of metals and alloys, large fluxes of ion irradiation are required, reaching 1017 –1018 ions/cm2 . Achieving such irradiation fluxes with a single formation of a high-intensity silicon ion beam of microsecond duration seems problematic. With a long pulse of silicon plasma formation, it is possible to significantly increase the ion irradiation dose due to the repeated formation of short-pulse ion beams during a single pulse of a vacuum arc discharge. To establish the features and regularities of formation of a packet of pulsed, ballistic-focused beams of silicon ions, the studies were conducted using repetitively pulsed generators with different pulse frequencies. Fig. 6 shows typical oscillograms of a silicon ion beam current formed from a plasma of a pulsed vacuum arc, with repetitively pulsed bias potentials of amplitudes 0.6, 1.0 and 1.5 kV, a frequency of 20 kHz and a pulse duty factor of 50%. Oscillograms show that for all biases, the durations of current pulses of the ions deposited on the grid electrode and the ion beam formation system immersed in the plasma, in general, correspond to the duration of the bias pulses. The amplitude of the ion current correlates with the change in plasma density during a vacuum-arc discharge pulse (Fig. 4). With a small bias amplitude of 0.6 kV, current pulses characteristic of a breakdown in the transport of an ion beam are observed. 6
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
in the maximum current density in the beam. The data presented in Fig. 8 show that with a bias potential pulse amplitude of 1.8 kV, the ion current density on the focusing system axis was about 0.8 A/cm2 . In addition, the profiles obtained are characterized by a decrease in the beam size at half-height with an increase in the bias potential amplitude. If the beam size at half-height at the system focus was approximately 35 mm at a bias potential pulse amplitude of 1.0 kV, then increases in the bias potential amplitude, for example, to 1.4 and 1.8 kV, reduced the beam diameter at half-height to approximately 20 and 15 mm, respectively. This regularity of change in the ion current density distribution over the cross section of a high-intensity beam is due to the action of two factors. On the one hand, as noted earlier, as the bias pulse amplitude increases, in accordance with the Child–Langmuir law, the sheath width in which the ion flux is formed increases, the angular spread of the ions in the beam decreases and, accordingly, the conditions for the ion beam transport improve. The second factor is due to the feature of transporting a high-intensity ion beam in a drift space. In the case of a focusing system, the conditions for compensating the ion-beam space charge will differ at different places of the drift space. Because ions are accelerated in the high-voltage sheath near the grid electrode, according to the current-continuity law (𝑗 = 𝑧𝑒𝑣𝑛 = const), the density of ions in beam n decreases in proportion to the increase in their velocity. Therefore, the density of the preinjected plasma near the grid electrode in the beam drift space is several times higher than the density of ions accelerated in the layer, which fundamentally makes it possible to completely neutralize the beam space charge with cold plasma electrons. To study the regularities of the high-intensity pulsed silicon ion beam focusing, a series of measurements of the radial profiles of the ion current density distribution at various distances from the grid electrode to the collectors was carried out. The collector system was successively installed at distances from 6.5 to 10.5 cm from the grid electrode with steps of 0.5 cm. Fig. 9 shows the results of measuring the radial profiles of the silicon-ion-beam current-density distribution obtained at bias potentials with amplitude of 1.6 kV and a duration of 5 μs. These figures reflect the dynamics of the ballistic focusing of ions in the beam. At a distance of 6.5 cm, there is a minimum amplitude of the ion current density and a relatively wide beam with two maxima. This form is due to the presence of a disk screen on the grid electrode, which leads to hollow conical ion beam formation if the detector is located close to the grid electrode; as a result, the ion current density in the central part of the beam is absent or decreases. Removing the detector from the grid electrode increases the maximum value of the ion current density. The results reflect the dynamics of the silicon ion beam focusing. A form transformation of the ion current distribution curve from the cross section with two maxima into a ‘‘domed’’ curve with one maximum along the beam axis is observed (Fig. 9). The data demonstrate that for the case of the pulsed silicon ion beam formation with a bias potential amplitude of 1.6 kV, the maximum ion current density measured at a distance of 8.5 cm is 0.66 A/cm2 . This detector position is 1 cm longer than the distance to the geometric focus of the grid electrode. The resulting deviation is due to incomplete neutralization of the beam space charge near the geometric focus of the ballistic focusing system. With a further increase in the beam transportation base, a monotonic increase is observed in the beam width at half-height of the radial distribution of the ion current density. As the ion beam is transported in the drift space, its density increases many times due to ballistic focusing. In our case, the equality of ion density in the beam, taking into account its charge state, and electron density in plasma will be achieved at a distance of slightly less than 5 cm from the beam entrance into the drift space. Further transportation of the beam and its possible focusing with a corresponding increase in the density of the ion current will be accompanied by the appearance of an electron deficiency. The lack of compensation for the space charge of the beam can be revealed primarily in violation of the conditions of ballistic focusing.
Fig. 7. The voltage distribution on the capacitors of the 19-channel sensor at different amplitudes of the pulse bias potential on the beam formation system.
Fig. 8. The profiles of the silicon ion-beam current-density distribution with different bias-potential pulse amplitudes and pulse durations of 5 μs.
4. Study of the profile of radial distribution of the pulsed silicon ion beam current density To study the current-density distribution over the ion-beam cross section, a special 19-channel sensor, a Faraday partitioned cylinder, was installed instead of a collector in the focal plane of the beam (Fig. 3). The averaged ion current density distribution over the beam section was determined by sequential measurement of the voltages on the capacitors included in each collector circuit. Fig. 7 shows the characteristic voltage distributions on the capacitors of the 19-channel sensor. Fig. 8 presents the results of a more detailed study of the profile of the silicon ion beam current density distribution at different amplitudes of the bias pulse. The experiments were performed with bias-pulse amplitudes in the range from 0.4 to 1.8 kV and pulse duration of 5 μs. All experimental profiles of the radial distribution of the ion current density in the beam have axial symmetry, which is due to the axial symmetry of the high-intensity ion-beam formation system. An increase in the potential pulse amplitude leads both to an increase in the total ion beam current (integrated over the entire beam) and to an increase 7
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
Fig. 9. Profiles of silicon ion-beam current-density distribution obtained at various distances between detector and grid electrode (bias pulse amplitude of 1.6 kV and pulse durations of 5 μs).
to 0.8 A/cm2 is demonstrated. As applied to the problems of modifying materials by ion implantation, the possibility of increasing the flux of ion irradiation during single-pulse plasma formation is considered.
The presence of an uncompensated space charge of the beam leads to the appearance of a radial electric field. Under the influence of this field, the beam ions are repelled, and their trajectory deviates from being straight [38]. The deviation of the ion trajectory from the ballistic trajectory explains the shift of the ion current density maximum over the geometric focus of the system observed experimentally. As the bias increases, the ion energy also increases, and as a result, the condition for ion-beam focusing will be improved. In general, the conducted studies allow the conclusion that it is possible to form pulsed and repetitively pulsed high-intensity silicon ion beams with a maximum ion current density about 0.8 A/cm2 . Such a current density of silicon ions with a pulse duration of 5 μs with a single-pulse implantation provides an ion flux of approximately 1.5 ⋅ 1013 ions/cm2 . For the purpose of modifying materials, including semiconductors, metals and alloys and composite materials by the ion implantation method, irradiation fluxes in the range of 1013 – 1018 ions/cm2 are required [39,40]. In pulsed silicon plasma formation, an increase in the flux of ion irradiation by two to three orders of magnitude can be achieved by lengthening the vacuum arc discharge pulse duration using a negative bias pulse generator providing a pulse frequency in the range from 50 to 100 kHz. Therefore, for example, when the arc discharge pulse duration is 10−2 s, the use of a bias generator with a frequency of 105 pulses/s will ensure the formation of 103 pulses, which will increase the flux by three orders of magnitude, making it possible to achieve high fluxes of silicon ion implantation.
Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. CRediT authorship contribution statement Alexander I. Ryabchikov: Conceptualization, Methodology, Project administration. Denis O. Sivin: Writing — original draft, Methodology, Visualization. Sergey V. Dektyarev: Investigation, Resources. Alexey E. Shevelev: Investigation, Formal analysis. Acknowledgment This work was carried out with the financial support of the Ministry of Science and Higher Education of the Russian Federation within the state assignment ‘‘Science’’ (grants No. 3.2415.2017/4.6 and 3.7245.2017/6.7). References
5. Conclusion
[1] J.M. Poate, G. Foti, D.C. Jacobson, Surface Modification and Alloying by Laser, Ion, and Electron Beams [Proceedings of a NATO Advanced Study Institute], Plenum Press, New York, ISBN: 0306413736, 1983. [2] J.S. Williams, J.M. Poate, Ion Implantation and Beam Processing [Academic], Academic Press, Orlando, ISBN: 9781483220642, 1984. [3] P.M. Martin, Handbook of Deposition Technologies for Films and Coatings: Science, Applications and Technology, William Andrew, ISBN: 9780815520313, 2009. [4] A.N.E.V. Kozlov, Yu. P. Sharkeev, et al., Microstructure of the near-surface layers of ion-implanted polycrystalline Cu, Surf. Coat. Technol. 56 (1) (1992) 11–17, http://dx.doi.org/10.1016/0257-8972(92)90190-L. [5] B.N. Chapman, Thin-film adhesion, J. Vacuum Sci. Technol. 11 (1974) 106–113. [6] A. Kinloch, Adhesion and Adhesives: Science and Technology, Chapman and Hall, New York, 1987, ISBN: 0-412-27440_X. [7] X. Liu, Y. Ren, X. Tan, S. Sun, E. Westkaemper, The structure of Ti–Si–N superhard nanocomposite coatings: ab initio study, Thin Solid Films 520 (2) (2011) 876–880, http://dx.doi.org/10.1016/j.tsf.2011.08.003. [8] J. Wang, J. Pu, G. Zhang, L. Wang, Interface architecture for superthick carbonbased films toward low internal stress and ultrahigh load-bearing capacity, ACS Appl. Mater. Interfaces 5 (11) (2013) 5015–5024, http://dx.doi.org/10.1021/ am400778p.
Studies have shown the possibility of the formation of pulsed and repetitively pulsed high-intensity beams of silicon ions. A pulsed vacuum-arc discharge with a current of 75 A formed a plasma flux of silicon with a duration of up to 350 μs. Silicon used as a cathode was previously subjected to neutron transmutation doping, which increased its conductivity. Plasma-immersion formation of an ion flux near the spherical grid structure followed by ballistic focusing of the ion beam in the drift space under conditions of its preliminary filling with plasma ensured silicon ion beam formation with a duration of more than 20 μs and a current of approximately 1.5 A. The regularities of the transport and focusing of a of silicon ion pulsed beam are investigated. It has been established that with bias pulse durations of more than 10 μs with small amplitudes of the accelerating voltage, repetitive transport failures of the high-intensity ion beam occur. An increase in bias amplitude from 0.6 to 1.5 kV contributes to the stabilization of the efficient transport of a silicon ion beam. The possibility of achieving ion current densities up 8
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx [24] A.I. Ryabchikov, E.B. Kashkarov, N.S. Pushilina, M.S. Syrtanov, A.E. Shevelev, O.S. Korneva, A.N. Sutygina, A.M. Lider, High-intensity low energy titanium ion implantation into zirconium alloy, Appl. Surf. Sci. 439 (2018) 106–112, http://dx.doi.org/10.1016/j.apsusc.2018.01.021. [25] A.I. Ryabchikov, A.E. Shevelev, D.O. Sivin, A.I. Ivanova, V.N. Medvedev, Low energy, high intensity metal ion implantation method for deep dopant containing layer formation, Surf. Coat. Technol. 355 (2018) 123–128, http://dx.doi.org/10. 1016/j.surfcoat.2018.02.111. [26] André Anders (Ed.), Handbook of Plasma Immersion Ion Implantation and Deposition, first ed., John Wiley & Sons, New York, ISBN: 978-0-471-24698-5, 2000. [27] A.I. Ryabchikov, V.S. Skuridin, E.V. Nesterov, E.V. Chibisov, V.M. Golovkov, Obtaining molybdenum-9 in the IRT-T research reactor using resonance neutrons, Nucl. Instrum. Methods Phys. Res. B 213 (2004) 364–368, http://dx.doi.org/10. 1016/S0168-583X(03)01592-1. [28] A. Naymushin, Yu. Chertkov, I. Lebedev, V. Boyko, Feasibility study of creating additional experimental channels for silicon doping in IRT-T reactor, J. Ind. Pollut. Control 32 (2) (2016) 424–427. [29] A.I. Ryabchikov, P.S. Ananin, S.V. Dektyarev, D.O. Sivin, A.E. Shevelev, High intensity metal ion beam generation, Vacuum 143 (2017) 447–453, http://dx. doi.org/10.1016/j.vacuum.2017.03.011. [30] I.G. Kesaev, Katodnye protsessy elektricheskoy dugi. [cathodic processes of electric arc], Nauka Publ., Moscow, ISBN: 978-5-518-13837-7, 1968, (in Russian). [31] R.L. Boxman, D.M. Sanders, P.J. Martin, Handbook of Vacuum Arc Science and Technology, Noyes Publications, Park Ridge, New Jersey, ISBN: 9780815513759, 1995. [32] A. Anders, Ion charge state distributions of vacuum arc plasmas, Orig. Species Phys. Rev. E 55 (1) (1997) 969–981, http://dx.doi.org/10.1103/PhysRevE.55. 969. [33] C.W. Kimblin, Cathode spot erosion and ionization phenomena in the transition from vacuum to atmospheric pressure arcs, J. Appl. Phys. 45 (1974) 5235–5244, http://dx.doi.org/10.1063/1.1663222. [34] C.D. Child, Discharge from hot CaO, Phys. Rev. 32 (5) (1911) 492–511, http: //dx.doi.org/10.1103/PhysRevSeriesI.32.492. [35] I. Langmuir, The effect of space charge and residual gases on thermionic currents in high, Vacuum Phys. Rev. 2 (6) (1913) 450–486, http://dx.doi.org/10.1103/ PhysRev.2.450. [36] T.V. Koval, A.I. Ryabchikov, T.M. Kim An, A.E. Shevelev, D.O. Sivin, A.I. Ivanova, D.M. Paltsev, Numerical simulation of high-intensity metal ion beam generation, J. Phys. Conf. Ser. 1115 (3) (2018) 032007, http://dx.doi.org/10.1088/17426596/1115/3/032007. [37] V.M. Anishchik, V.V. Uglov, Modifikatsiya instrumental’nykh materialov ionnymi i plazmennymi puchkami, BGU, Mn, ISBN: 985-445-906-3, 2003, (in Russian). [38] A.I. Ryabchikov, A.E. Shevelev, D.O. Sivin, T.V. Koval, T.M. Kim An, High intensity, macroparticle-free, aluminum ion beam formation, J. Appl. Phys. 123 (23) (2018) 233301, http://dx.doi.org/10.1063/1.5034082. [39] W.D. Davis, H.C. Miller, Analysis of the electrode products emitted by dc arcs in a vacuum ambient, J. Appl. Phys. 40 (5) (1969) 2212–2221, http://dx.doi. org/10.1063/1.1657960. [40] G.Y. Yushkov, A. Anders, E.M. Oks, I.G. Brown, Ion velocities in vacuum arc plasmas, J. Appl. Phys. 88 (10) (2000) 5618–5622, http://dx.doi.org/10.1063/ 1.1321789.
[9] Y.H. Cheng, T. Browne, B. Heckerman, E.I. Meletis, Mechanical and tribological properties of nanocomposite TiSiN coatings, Surf. Coat. Technol. 204 (2010) 2123–2129, http://dx.doi.org/10.1016/j.surfcoat.2009.11.034. [10] J. Tian, P. Xu, J. Chen, Q. Liu, Microstructure and phase transformation behaviour of a Fe/Mn/Si/Cr/Ni alloy coating by laser cladding, Opt. Lasers Eng. 122 (2019) 97–104. [11] M. Pleva, B. Grančič, M. Mikula, M. Truchlý, T. Roch, L. Satrapinskyy, M. Gregor, P. Ďurina, V. Girman, P. Švec Jr., A. Plecenik, P. Kúš, Thermal stability of amorphous Ti-B-Si-N coatings with variable Si/B concentration ratio, Surf. Coat. Technol. 333 (2018) 52–60, http://dx.doi.org/10.1016/j.surfcoat.2017.10.063. [12] J.S. Reid, E. Kolawa, C.M. Garland, M.A. Nicolet, F. Cardone, D. Gupta, R.P. Ruiz, Amorphous (Mo Ta or W)–Si–N diffusion barriers for Al metallizations, J. Appl. Phys. 79 (2) (1996) 1109–1115, http://dx.doi.org/10.1063/1.360909. [13] X. Sun, J.S. Reid, E. Kolawa, M.-A. Nicolet, R.P. Ruiz, Reactively sputtered Ti-SiN films. II. Diffusion barriers for Al and Cu metallizations on Si, J. Appl. Phys. 81 (1997) 664–671, http://dx.doi.org/10.1063/1.364206. [14] R. Daniel, J. Musil, P. Zeman, C. Mitterer, Thermal stability of magnetron sputtered Zr–Si–N films, Surf. Coat. Technol. 201 (6) (2006) 3368–3376, http: //dx.doi.org/10.1016/j.surfcoat.2006.07.206. [15] J. Musil, P. Zeman, P. Dohnal, Ti-Si-N films with a high content of Si, Plasma Process. Polym. 4 (S1) (2007) S574–S578, http://dx.doi.org/10.1002/ ppap.200731408. [16] J. Musil, Hard nanocomposite coatings: thermal stability, oxidation resistance and toughness, Surf. Coat. Technol. 207 (2012) 50–65, http://dx.doi.org/10. 1016/j.surfcoat.2012.05.073. [17] M. Mikula, B. Grančič, M. Drienovský, L. Satrapinskyy, T. Roch, Z. Hájovská, M. Gregor, T. Plecenik, R. Čička, A. Plecenik, P. Kúš, Thermal stability and hightemperature oxidation behavior of Si–Cr–N coatings with high content of silicon, Surf. Coat. Technol. 232 (2013) 349–356, http://dx.doi.org/10.1016/j. surfcoat.2013.05.034. [18] S. Veprek, M.G.J. Veprek-Heijman, P. Karvankova, J. Prochazka, Different approaches to superhard coatings and nanocomposites, Thin Solid Films 476 (1) (2005) 1–29, http://dx.doi.org/10.1016/j.tsf.2004.10.053. [19] P.H. Mayrhofer, Ch. Mitterer, L. Hultman, H. Clemens, Microstructural design of hard coatings, Prog. Mater. Sci. 51 (8) (2006) 1032–1114, http://dx.doi.org/10. 1016/j.pmatsci.2006.02.002. [20] A. Cavaleiro, B. Trindade, M. Teresa Vieira, The influence of the addition of third element on the structure and mechanical properties of transition-metal-based nanostructured hard films part I-nitrides, in: Nanostructured Coatings, Springer Science Business Media, 2006, p. 671, ISBN-10: 0-387-25642-3, ISBN-13: 978-0387-25642-9. ˜ I. Kovács, L. Tóth, B. [21] K.P. Budna, P.H. Mayrhofer, J. Neidhardt, É. Hegedus, Pécz, C. Mitterer, Effect of nitrogen-incorporation on structure, properties and performance of magnetron sputtered CrB2, Surf. Coat. Technol. 202 (13) (2008) 3088–3093, http://dx.doi.org/10.1016/j.surfcoat.2007.11.009. [22] A. Flink, J.M. Andersson, B. Alling, R. Daniel, J. Sjölén, L. Karlsson, L. Hultman, Structure and thermal stability of arc evaporated (Ti0.33 Al0.67 )1 - xSixN thin films, Thin Solid Films 517 (2) (2008) 714–721, http://dx.doi.org/10.1016/j.tsf.2008. 08.126. [23] A.I. Ryabchikov, D.O. Sivin, P.S. Ananin, A.I. Ivanova, I.V. Lopatin, O.S. Korneva, A.E. Shevelev, High intensity, low ion energy implantation of nitrogen in AISI 5140 alloy steel, Surf. Coat. Technol. 355 (2018) 129–135, http://dx.doi.org/10. 1016/j.surfcoat.2018.02.110.
9
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
Contents lists available at ScienceDirect
Nuclear Inst. and Methods in Physics Research, A journal homepage: www.elsevier.com/locate/nima
Formation of repetitively pulsed high-intensity, low-energy silicon ion beams Alexander I. Ryabchikov, Denis O. Sivin ∗, Sergey V. Dektyarev, Alexey E. Shevelev National Research Tomsk Polytechnic University, 634050, pr. Lenina 2, bdg. 4, Tomsk, Russia
ABSTRACT This paper investigates the formation of pulsed and repetitively pulsed high-intensity, low-energy silicon ion beams. The formation of a pulsed silicon plasma was carried out by a vacuum arc discharge. Polycrystalline neutron-doped silicon was used as the evaporator’s cathode. A pulsed vacuum arc discharge formed a directed silicon plasma flow with a duration of approximately 350 μs. On the path of plasma transportation at a distance of 35 cm from the cathode surface, a system for forming a ballistically focused ion beam was installed. When a pulsed or a repetitively pulsed negative-bias potential is applied to the extracting grid electrode in the shape of a second-order surface, it ensured the formation of a sheath, the extraction of silicon ions from the plasma, and their acceleration in the layer. The shape of the extracting electrode provided the possibility of ballistic focusing of the silicon ion beam. Repetitively pulsed generation of the bias potential provided the possibility of preliminary plasma injection into the beam drift space between the bias-potential pulses. The paper has studied the features and regularities of the pulsed (with a duration of bias pulses up to 34 μs) and repetitively pulsed (at frequencies of bias pulses of 20 and 100 kHz) silicon ion beams formation with a current of up to 2 A at a maximum ion current density of 0.8 A/cm2 at bias potentials of small amplitudes (0.6–1.8 kV).
1. Introduction Various methods of ion-plasma deposition of coatings and modification of surface and near surface properties of different materials by ion implantation and exposure to various types of radiation play an increasingly important role in the creation of new generation materials with improved performance properties [1–4]. While technology provides many advantages for improving the macroscopic properties of materials’ surfaces by deposition of coatings, there are significant drawbacks associated with the presence of problems with the adhesion of coatings to the base of the material being processed [5,6]. Ion implantation provides a solution to the problems of adhesion, creating ion-doped layers with a given composition within the material itself. The disadvantage of the ion-implantation method is the small projective range of ions and, as a result, a small thickness of the ion-doped layer. The long-range effects that change the dislocation structure at great depths during ion implantation have little effect on the performance properties of structural materials and coatings [4]. To improve the performance properties of various materials by ionplasma deposition of coatings and ion implantation, the use of silicon as a dopant is of particular interest. For example, the presence of silicon in different steels and other composite materials (metal nitrides and carbides) has a positive effect on the mechanical properties of the raw materials [7–9]. As a rule, increased silicon concentration leads to an increase in the material strength and in resistance to tempering, a decrease in grain growth at high temperatures, and recrystallization of composite compounds [10]. Amorphous or nearly amorphous transition-metal-nitride coatings are expected to exhibit unique physical properties that are not observable in their crystalline counterparts. One of the potentially attractive
qualities is the missing grain boundary, which otherwise builds a ‘weak point’ at which gas atoms from the atmosphere, working material or substrate may diffuse [11]. Amorphous transition-metal-Si-N materials have been used as diffusion barriers in microelectronics [12,13] or in coatings with high temperature oxidation resistance [14–17]. The addition of a metalloid such as Si into transition metal nitrides is an effective way to stabilize a nanocomposite or amorphous structure [18–21]. In sufficient content (> 10 at. %) due to its fourfold coordination, Si was shown to severely distort the TiN cubic lattice [22]. Musil et al. reported that the addition of high Si content > 25 at. % into ZrN results in an amorphous structure that is stable up to > 1000 ◦ C [14]. The structural stability, together with the ability of Si to form solid nonvolatile oxide (SiO2 ), promotes the oxidation resistance of transition metal-Si–N systems [15–17]. Recently, the possibility of using high-intensity ion beams with ion current densities from 0.01 to 1 A/cm2 and an ion energy of several keV has been demonstrated for surface modification of various materials by ion implantation. Work [23] presents the results of the formation of modified nitrogen-containing layers with thicknesses of up to 180 μm in steel when processed with a high-intensity nitrogen beam for 60 min. Studies of the formation of intermetallic compounds by high-intensity beams of zirconium ions [24] and aluminum [25] in targets made of titanium and nickel, respectively, demonstrated the possibility of obtaining modified layers with thicknesses of several tens of micrometers. The implementation of research in the direction of deep ion doping of various materials and coatings with silicon requires creating systems and sources that form high-intensity, low-energy silicon ion beams. This paper is devoted to research on the formation of pulsed highintensity axially symmetric beams of low-energy silicon ions with ion
∗ Corresponding author. E-mail address: [email protected] (D.O. Sivin).
https://doi.org/10.1016/j.nima.2019.163092 Received 21 June 2019; Received in revised form 5 November 2019; Accepted 5 November 2019 Available online xxxx 0168-9002/© 2019 Elsevier B.V. All rights reserved.
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
Fig. 1. Simplified scheme of the experimental setup for the formation of pulsed high-intensity silicon ion beams.
current densities of up to several hundred milliamperes per square centimeter.
ballistic focusing when it is transported in the equipotential drift space previously filled with plasma [29]. Extraction and focusing of the beam were carried out by a grid electrode with a cell size of 1.8 × 1.8 mm2 , transparency of 70%, made in the form of a sphere with a diameter of 15 cm. Construction of the grid electrode in the form of a second-order nondegenerate surface with central symmetry implies ballistic focusing of a particle flow, normally rooted relative to the surface of the grid electrode, to some focus. For an electrode as a part of a sphere, the focus is in the geometric center of the sphere at a distance of 7.5 cm. The grid electrode was installed on one of the bases of a solid metal cylinder with a height of 7 cm and a diameter of 14 cm. This construction allowed to allocate some volume, which limited the equipotential space for transporting the ion beam. On the opposite base of the cylinder, various collectors and sensors were installed, whose position relative to the grid electrode along the axis of the system could be regulated. In experiments with measurements of the ion beam current, a flat collector made of tungsten with a diameter of 4 cm was used. The working plane of the collector was located on the axis of the focusing system in the geometric focus of the grid electrode at a distance of 7.5 cm. The bias potential was fed to the system when the plasma flow already reached the ion collector. Thus, the plasma-immersion system was used for the initial formation of the ion flux. The acceleration of ions and their transport in the equipotential drift space was performed using a pulsed bias potential that was applied simultaneously to the grid electrode, solid cylindrical electrode and collector. The potential was formed by a generator of high-frequency short-pulse bias voltages of negative polarity (PG in Fig. 1). The generator could operate both in high-frequency repetitively pulsed mode and in a mode of forming single pulses. In single-pulse mode, the generator allowed the formation of a negative-bias-potential pulse with an amplitude in the range from 0.2 to 1.8 kV and a duration from 5 to 100 μs. In repetitively pulsed mode, the generator provided the formation of a sequence of biaspotential pulses with a frequency from 1 to 100 kHz and a pulse duty factor from 20 to 80%.
2. Experimental setup and methods Studies of the formation of pulsed high-intensity silicon ion beams were conducted with an experimental setup whose scheme is shown in Fig. 1. The vacuum chamber was pumped out to a residual pressure of 10−3 Pa using a turbomolecular pump with a capacity of 1000 l/s. To generate silicon plasma, a vacuum-arc discharge was used; this makes it possible to obtain a sufficiently dense plasma from conducting materials with a high degree of ionization in vacuum [26]. A planar vacuum-arc evaporator with a magnetic coil was installed on the vacuum chamber to control the plasma flow density. The cathode of the vacuum-arc evaporator was a plate of doped silicon with a diameter of 10 cm with an average specific resistance of 0.013 Ohm cm. The material of the cathodes was obtained by the neutron-transmutation doping method of single-crystal silicon in a research nuclear pooltype reactor IRT-T (Tomsk, TPU) [27,28]. The pulsed power supply of the arc was implemented by a forming line with a matched load (SL in Fig. 1), providing a pulse with a duration of approximately 350 μs. The ignition-electrode pulse control and its synchronization with the arc power pulse were performed by the ignition block (TPS) and master generator (MPG). Electromagnetic coils ensured the creation of a longitudinal magnetic field, limiting the plasma angular spread. The high-intensity ion beams formation system, the circuit diagram and the appearance of which is shown in Fig. 2, was installed inside the vacuum chamber opposite the vacuum-arc evaporator on its symmetry axis at a distance of 35 cm from the cathode. The beam formation system consists of several basic elements: a grid electrode, a solid cylindrical electrode, a collector and a disk screen. The operation of the beam formation system is based on the principles of plasmaimmersion extraction, ion acceleration in the sheath and ion beam 2
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
Fig. 2. External view (a) and cross-sectioned scheme (b) of the ion beam formation system.
The parameters of the pulsed bias potential were measured by an active voltage divider. The beam current at the collector and the current in the circuit of the focusing system were measured using Rogowski coils (L1 and L2). Current pulses and bias voltages were recorded by digital multichannel oscilloscopes. Two delay pulse generators (DP1 and DP2) were added to the measurement part of the experimental setup. These generators helped control delay in the negative-bias pulse feeding to the beam formation system relative to the ignition pulse of the vacuum-arc discharge, as well as the synchronization of the measuring oscillographs relative to these pulses. In addition, the delay pulse generators made it possible to control the bias-pulse duration during ion-beam formation. When plasma is created by a vacuum-arc discharge as a result of explosive emission processes on a silicon cathode, a significant flow of macroparticles is generated (droplets, solid fragments ranging in size from fractions of micrometers to tens of micrometers) that reduces the technological capabilities of the ion beam formation system due to contamination of the treated sample surfaces by cathode erosion products [30]. The solar-eclipse effect was used to clean the ion beam of the vacuum-arc discharge without significantly reducing the ion current [29]. On the grid electrode, a disk screen with a diameter of 4 cm was installed along the symmetry axis of the beam-formation system. The screen used in this position is an obstacle to the nearly straight-line trajectories of macroparticles emitted from the cathode to the collector area installed at the system focus. At the same time, due to the ballistic focusing using a spherical grid electrode along straight-line trajectories, the beam ions freely reach the irradiated target area on the collector. The use of an additional screen changes the total ion current in the beam only slightly, since it covers no more than 4% of the total area of the grid electrode (approximately 300 cm2 ) that accelerates and extracts ions from the plasma. Experimental results are presented in [29], demonstrating the almost complete absence of macroparticles in a DC vacuum-arc discharge in the region of the ion beam impact on the target using the ‘‘solar eclipse’’ effect. To study the shape of the ion current-density distribution profile of a pulsed high-intensity silicon ion beam, a 19-channel sensor was installed at the collector site. Fig. 3 shows the appearance of the sensor, which consists of a set of 19 single collectors fixed along one line. Each collector has a size of 2 × 2 mm2 . After beam formation and focusing, the ions fall on the collectors. At the same time, at the beginning of beam formation, the collectors are under the potential of the grid and cylindrical electrodes. To accumulate the charge of the beam ions deposited on the collectors, capacitors were connected to the electrical circuits of each of them. After each pulsed-ion beam formation, sequential polling of voltages from capacitors was conducted using a special device. The ion current
Fig. 3. Top view of 19-channel sensor for ion beam distribution measurement.
density at each point (on each collector) is proportional to voltage U and capacitor capacitance C and inversely proportional to current-pulse duration 𝜏 and collector area 𝑆𝑘 𝑗=
C⋅𝑈 . 𝜏 ⋅ 𝑆𝚔
Thus, measuring the voltages at the collectors, one can determine the integral charge of the ions that fell on each collector and, as a result, calculate the current density of silicon ions on each collector. 3. Features and regularities of high-intensity pulsed silicon ion beams After the start of the MPG master generator, a pulse is generated at its output, which is fed to the TPS ignition-pulse-formation block and to the delayed pulse generators DPG1 and DPG2 (Fig. 1). When supplying a pulsed high voltage to the ignition electrode, a breakdown occurs on the surface of the ceramic between the cathode and ignition electrode. The voltage of the forming line SL ensures the formation and maintenance of a discharge between the cathode and anode of the vacuum-arc plasma generator. The cathode spot formed on the cathode surface generates a plasma flow with a high velocity directed perpendicular to the surface of the cathode, that is, towards the ionbeam formation system [31]. To create the conditions necessary for the efficient transfer of a high-intensity ion beam, it is necessary to preliminarily fill the drift space of the ion beam with plasma [29]. Based on this condition, the delay time to feed a negative high-voltage bias pulse should be chosen from the condition: 𝑡 > (𝐿 + 𝑅) ∕𝜐𝑝 , where L — distance from the silicon cathode to the grid electrode, R — distance from the grid electrode to the collector, and 𝜐𝑝 — plasma flow velocity. 3
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
of a flat accelerating gap with an average density of ion saturation current of 0.011 A/cm2 and a voltage in the sheath of 1.6 kV indicate that the sheath width is approximately 3 mm ((𝑅 + 𝑑)2 ∕𝑅2 = 1.08). After acceleration in the sheath, silicon ions penetrate into the initial equipotential drift space bounded by grid and solid cylindrical electrodes. One of the main conditions for high-intensity ion beam formation is that ion focusing should be conducted under conditions of partial or complete neutralization of the beam space charge when it is transported in the drift space. In the absence of a mechanism for the beam space charge neutralizing, its defocusing will be observed, and a virtual anode can also be formed that will lead to a breakdown in the ion-beam transport. Fig. 5 shows the most characteristic oscillograms of silicon-ion-beam current to a collector and an ion current to a grid electrode for various bias-pulse durations and pulse amplitudes: 0.6, 1.1 and 1.5 kV. The duration of the bias pulses on the beam-formation system varied from 8 to 34 μs. The presented data demonstrate a number of features and regularities of formation and transportation of pulsed silicon ion beams. The oscillograms of the ion current in Fig. 5a show that when the amplitude of the bias potential is 0.6 kV and the pulse duration of the bias is 8 μs, the ion current to the collector has a delay relative to the current to the grid electrode. This delay is due to the final time of sheath formation and ion-beam transport from the grid electrode to the collector. The steepness of the leading edge of the ion-beamcurrent pulse turns out to be significantly less than that of the current to the grid electrode. This behavior of the ion-beam-current pulse is due to the processes of ion-beam space-charge neutralization. When the ion beam enters the drift space, its space charge is not compensated. The presence of cold electrons in plasma, previously injected into the drift space, does not instantly neutralize the ion beam space charge. It takes time to remove plasma ions from the beam volume. As shown by numerical simulation [36], depending on the parameters of the ion beam and plasma, the time taken to neutralize the beam space charge can be several microseconds. As the charge neutralization of the ion beam increases, the conditions for its transport and focusing improve. The leading edge of the ion beam current indicates that this time in this case exceeds 3 μs. The delay of the back edge of the beam-current pulse relative to the end of the bias pulse is associated with the finite ion time-of-flight from the grid electrode to the collector through the drift space. An increase in the bias pulse duration to 17 μs is accompanied by a corresponding increase in the duration of the ion current to the grid electrode (Fig. 5a). At the same time, duration shortening of the ion current to the collector is observed compared with the bias pulse duration. Ion-beam transport disruption occurs. The probability of a break increases with an increase in the current pulse duration of the ion beam from 10 to 15 μs. It is characteristic that a decrease in the ion current to the collector is accompanied by an increase in the ion current to the ion-beam-formation system. Most likely, this effect is associated with ion-beam space-charge decompensation due to the gradual escape of plasma electrons into the accelerating gap through the grid electrode cells. A decrease in the number of electrons leads to an increase in the uncompensated ion-beam charge and the violation of the conditions for its transport. The disruption of the ion beam leads to virtual anode formation in which a significant portion of the ions is affected by the electric field, changing their movement to the opposite direction. Finally, after a small number of oscillations relative to the grid electrode, ions are deposited on it, leading to a corresponding increase in current. The use of longer bias pulses (27 μs and 34 μs) also demonstrates the development of instability in transport of the ion beam in drift space, at which the amplitude of the beam ion current to the collector decreases sharply to some minimum value. However, after some time, with these durations, in addition to reducing the ion-beam-current amplitude, the initial current level to the collector is restored. This means that the conditions for neutralizing the space charge of a high-intensity
Fig. 4. Oscillograms of arc discharge and ion current to the collector.
The experiments used a pulsed vacuum-arc discharge with a maximum current of 75 A and a pulse duration of 350 μs. Fig. 4 shows oscillograms of a pulsed current of a vacuum-arc discharge and a current of silicon ions to a collector installed at a distance (𝐿 + 𝑅) = 42.5 cm from the cathode plane of the vacuum-arc evaporator. The pulse shape of the ion saturation current from the plasma, measured with a constant negative 100 V bias applied to the collector, correlates with the current pulse of the vacuum-arc discharge. The observed amplitude instability of the current to the collector reflects changes in the density of the plasma flow, which is determined by unsteady processes in the cathode spots of the vacuum-arc discharge [32]. Oscillograms show a delay of approximately 20 μs in the appearance of current on the collector relative to the voltage pulse of the vacuum-arc discharge. Expression 1 allows estimation of the plasma velocity 𝜐𝑝 = 2.1 ⋅ 106 cm/s, in agreement with the data of [33], which experimentally shows that for a plasma of a pulsed vacuumarc discharge, depending on the cathode material, the velocity of its directional propagation varies within (1.5–3) ⋅ 106 cm/s. As measured in the experiment, the average density of the ion saturation current from the plasma along the system axis at a distance corresponding to the grid electrode location was 0.011 A/cm2 150 μs after the onset of the vacuum arc discharge. All experimental studies in the single-pulse mode of focused ion beam formation were conducted with a delay of 100 μs between the start of plasma generation and the supply of negative bias potential to the beam-formation system. This provided preliminary filling of the ion beam drift space with plasma. After a negative bias is applied to the grid electrode, a sheath is formed near it. Initially, due to their high mobility, the plasma electrons leave the electrode space, and the socalled matrix sheath is formed. The width of this sheath depends on the amplitude of the bias potential and the concentration and temperature of electrons [3]. Under the action of an electric field in the formed sheath, ions begin to accelerate gradually. The acceleration of ions, in turn, leads to a change in the ion density in the accelerating gap and to the sheath dynamic expansion. The process of plasma emissionboundary stabilization is completed after the sheath expands to a size determined from the condition of ion-current limiting by the spatial beam charge. The limiting ion current density is determined by the Child–Langmuir law [34,35]. Comparison of the grid electrode curvature 𝐾𝑔𝑒 = 1∕𝑅2 , made as part of a sphere, and the plasma emission boundary curvature 𝐾𝑠 = 1∕(𝑅 + 𝑑)2 , where R — grid electrode radius and d — sheath width, allows us to consider the electrode and the emission boundary as flat electrodes under the criterion (𝑅 + 𝑑)2 ∕𝑅2 ∼ 1. An estimate of the emission boundary position in the approximation 4
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
Fig. 5. Oscillograms of the silicon ion beam current and ion current to the grid electrode at different amplitudes and durations of the bias potential pulses: (a) 0.6 kV; (b) −1.0 kV; (c) −1.5 kV.
5
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
ion beam are restored. Because during the pause, the bias potential is applied to the grid electrode, the plasma from the vacuum arc evaporator cannot penetrate into the drift space of the ion beam. This means that reneutralizing of the ion-beam space charge is not provided by the injected plasma electrons. Most likely, plasma production in the beam drift space occurs due to ionization of the residual atmosphere in the vacuum chamber with ions of the ion beam itself. Initially after a current break, there is a gradual very slow increase in the ion current to the collector. This is especially well observed from the oscillograms presented in Fig. 5b. It is characteristic that upon reaching a certain level, a sufficiently rapid increase in the amplitude of the ion-beam current falling on the collector occurs. Another possible mechanism of plasma formation in the drift space should be noted. The ion-beam formation system is under a negative bias potential and is actually a hollow cathode. The injection of ions into the drift space is accompanied by the formation of a region with a positive potential relative to the cathode. A beam collapsing under the action of a space charge ionizes the residual gas both inside the beam and at its periphery. Under the action of the electric field of the ‘‘hollow cathode – space of positive potential’’ system, the peripheral plasma electrons will accelerate in the direction of the positive potential and decelerate when passing through it without reaching the cathode. Electron oscillation in this space will lead to ionization of the residual gas by electron impact. Oscillation of plasma electrons can take place both in the radial direction and along the ion flux distribution. Secondary ion electron emission from the collector can also produce cold electrons. The slow gradual increase of the current to the collector after the breakdown of the ion beam transport most likely indicates its gradual advancing in the drift space. Reducing the distance between the front of the beam and the collector leads to an increase in the amplitude of the limiting ion current, with an uncompensated space charge transported in a large volume of the ion beam. A sharp increase in the current of the ion beam arriving at the collector must be accompanied by an avalanche-like increase in plasma density. However, further research is needed to clarify the type of breakdown. An increase in the bias potential amplitude on the focusing system to 1.0 kV does not fundamentally change the picture and leads to only a slight improvement in the ion beam transportation conditions. The oscillograms in Fig. 5b show an increase in the ion-beam-current amplitude recorded by the collector. This effect is associated with an increase in the sheath width with increasing bias amplitude in accordance with the Child–Langmuir law. At low bias potentials, the sheath width turns out to be comparable to the size of the grid electrode cell. This leads to the curvature of the plasma ion-emission boundary, and a significant part of the ions is accelerated, falling directly on the grid structure elements. As the amplitude of the accelerating voltage increases, the sheath width increases, and the plasma ion-emission boundary, moving away from the grid electrode, gradually straightens. This leads to ion flux formation at the entrance to the focusing system with a smaller angular divergence. As a result, with an increase in the negative-bias amplitude, the ion beam current in the focal plane of the system increases. From a comparison of the oscillograms in Fig. 5a and 5b, it can be concluded that with an increase in the bias amplitude, the pause duration during which the high-intensity silicon ion beam is not transported decreases by almost 30%. A further increase in the bias-potential amplitude on the focusing system up to 1.5 kV provided not only an increase in the amplitude of the ion-beam current to the collector but also a stable silicon-ionbeam formation without transport disruptions over the entire range of the bias potential pulse durations (Fig. 5c). Oscillograms show that the amplitude of the ballistic-focused ion-beam current in the focal region on the collector exceeds 1.5 A. It is obvious that single pulses of a high-intensity silicon ion beam have a limited area of practical application. For example, they can be used in studies of the possible existence of a shock-wave diffusion mechanism for a dopant during high-intensity implantation of
Fig. 6. Oscillograms of the collector current during the silicon ion beam formations from the plasma of a pulsed vacuum arc at a repetitively pulsed bias potential with amplitudes of 0.6, 1 and 1.5 kV (frequency is 20 kHz, duty factor is 50%).
low-energy ions (1–5 keV) at ion current densities of more than 0.6 A/cm2 [37]. To modify the properties of metals and alloys, large fluxes of ion irradiation are required, reaching 1017 –1018 ions/cm2 . Achieving such irradiation fluxes with a single formation of a high-intensity silicon ion beam of microsecond duration seems problematic. With a long pulse of silicon plasma formation, it is possible to significantly increase the ion irradiation dose due to the repeated formation of short-pulse ion beams during a single pulse of a vacuum arc discharge. To establish the features and regularities of formation of a packet of pulsed, ballistic-focused beams of silicon ions, the studies were conducted using repetitively pulsed generators with different pulse frequencies. Fig. 6 shows typical oscillograms of a silicon ion beam current formed from a plasma of a pulsed vacuum arc, with repetitively pulsed bias potentials of amplitudes 0.6, 1.0 and 1.5 kV, a frequency of 20 kHz and a pulse duty factor of 50%. Oscillograms show that for all biases, the durations of current pulses of the ions deposited on the grid electrode and the ion beam formation system immersed in the plasma, in general, correspond to the duration of the bias pulses. The amplitude of the ion current correlates with the change in plasma density during a vacuum-arc discharge pulse (Fig. 4). With a small bias amplitude of 0.6 kV, current pulses characteristic of a breakdown in the transport of an ion beam are observed. 6
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
in the maximum current density in the beam. The data presented in Fig. 8 show that with a bias potential pulse amplitude of 1.8 kV, the ion current density on the focusing system axis was about 0.8 A/cm2 . In addition, the profiles obtained are characterized by a decrease in the beam size at half-height with an increase in the bias potential amplitude. If the beam size at half-height at the system focus was approximately 35 mm at a bias potential pulse amplitude of 1.0 kV, then increases in the bias potential amplitude, for example, to 1.4 and 1.8 kV, reduced the beam diameter at half-height to approximately 20 and 15 mm, respectively. This regularity of change in the ion current density distribution over the cross section of a high-intensity beam is due to the action of two factors. On the one hand, as noted earlier, as the bias pulse amplitude increases, in accordance with the Child–Langmuir law, the sheath width in which the ion flux is formed increases, the angular spread of the ions in the beam decreases and, accordingly, the conditions for the ion beam transport improve. The second factor is due to the feature of transporting a high-intensity ion beam in a drift space. In the case of a focusing system, the conditions for compensating the ion-beam space charge will differ at different places of the drift space. Because ions are accelerated in the high-voltage sheath near the grid electrode, according to the current-continuity law (𝑗 = 𝑧𝑒𝑣𝑛 = const), the density of ions in beam n decreases in proportion to the increase in their velocity. Therefore, the density of the preinjected plasma near the grid electrode in the beam drift space is several times higher than the density of ions accelerated in the layer, which fundamentally makes it possible to completely neutralize the beam space charge with cold plasma electrons. To study the regularities of the high-intensity pulsed silicon ion beam focusing, a series of measurements of the radial profiles of the ion current density distribution at various distances from the grid electrode to the collectors was carried out. The collector system was successively installed at distances from 6.5 to 10.5 cm from the grid electrode with steps of 0.5 cm. Fig. 9 shows the results of measuring the radial profiles of the silicon-ion-beam current-density distribution obtained at bias potentials with amplitude of 1.6 kV and a duration of 5 μs. These figures reflect the dynamics of the ballistic focusing of ions in the beam. At a distance of 6.5 cm, there is a minimum amplitude of the ion current density and a relatively wide beam with two maxima. This form is due to the presence of a disk screen on the grid electrode, which leads to hollow conical ion beam formation if the detector is located close to the grid electrode; as a result, the ion current density in the central part of the beam is absent or decreases. Removing the detector from the grid electrode increases the maximum value of the ion current density. The results reflect the dynamics of the silicon ion beam focusing. A form transformation of the ion current distribution curve from the cross section with two maxima into a ‘‘domed’’ curve with one maximum along the beam axis is observed (Fig. 9). The data demonstrate that for the case of the pulsed silicon ion beam formation with a bias potential amplitude of 1.6 kV, the maximum ion current density measured at a distance of 8.5 cm is 0.66 A/cm2 . This detector position is 1 cm longer than the distance to the geometric focus of the grid electrode. The resulting deviation is due to incomplete neutralization of the beam space charge near the geometric focus of the ballistic focusing system. With a further increase in the beam transportation base, a monotonic increase is observed in the beam width at half-height of the radial distribution of the ion current density. As the ion beam is transported in the drift space, its density increases many times due to ballistic focusing. In our case, the equality of ion density in the beam, taking into account its charge state, and electron density in plasma will be achieved at a distance of slightly less than 5 cm from the beam entrance into the drift space. Further transportation of the beam and its possible focusing with a corresponding increase in the density of the ion current will be accompanied by the appearance of an electron deficiency. The lack of compensation for the space charge of the beam can be revealed primarily in violation of the conditions of ballistic focusing.
Fig. 7. The voltage distribution on the capacitors of the 19-channel sensor at different amplitudes of the pulse bias potential on the beam formation system.
Fig. 8. The profiles of the silicon ion-beam current-density distribution with different bias-potential pulse amplitudes and pulse durations of 5 μs.
4. Study of the profile of radial distribution of the pulsed silicon ion beam current density To study the current-density distribution over the ion-beam cross section, a special 19-channel sensor, a Faraday partitioned cylinder, was installed instead of a collector in the focal plane of the beam (Fig. 3). The averaged ion current density distribution over the beam section was determined by sequential measurement of the voltages on the capacitors included in each collector circuit. Fig. 7 shows the characteristic voltage distributions on the capacitors of the 19-channel sensor. Fig. 8 presents the results of a more detailed study of the profile of the silicon ion beam current density distribution at different amplitudes of the bias pulse. The experiments were performed with bias-pulse amplitudes in the range from 0.4 to 1.8 kV and pulse duration of 5 μs. All experimental profiles of the radial distribution of the ion current density in the beam have axial symmetry, which is due to the axial symmetry of the high-intensity ion-beam formation system. An increase in the potential pulse amplitude leads both to an increase in the total ion beam current (integrated over the entire beam) and to an increase 7
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx
Fig. 9. Profiles of silicon ion-beam current-density distribution obtained at various distances between detector and grid electrode (bias pulse amplitude of 1.6 kV and pulse durations of 5 μs).
to 0.8 A/cm2 is demonstrated. As applied to the problems of modifying materials by ion implantation, the possibility of increasing the flux of ion irradiation during single-pulse plasma formation is considered.
The presence of an uncompensated space charge of the beam leads to the appearance of a radial electric field. Under the influence of this field, the beam ions are repelled, and their trajectory deviates from being straight [38]. The deviation of the ion trajectory from the ballistic trajectory explains the shift of the ion current density maximum over the geometric focus of the system observed experimentally. As the bias increases, the ion energy also increases, and as a result, the condition for ion-beam focusing will be improved. In general, the conducted studies allow the conclusion that it is possible to form pulsed and repetitively pulsed high-intensity silicon ion beams with a maximum ion current density about 0.8 A/cm2 . Such a current density of silicon ions with a pulse duration of 5 μs with a single-pulse implantation provides an ion flux of approximately 1.5 ⋅ 1013 ions/cm2 . For the purpose of modifying materials, including semiconductors, metals and alloys and composite materials by the ion implantation method, irradiation fluxes in the range of 1013 – 1018 ions/cm2 are required [39,40]. In pulsed silicon plasma formation, an increase in the flux of ion irradiation by two to three orders of magnitude can be achieved by lengthening the vacuum arc discharge pulse duration using a negative bias pulse generator providing a pulse frequency in the range from 50 to 100 kHz. Therefore, for example, when the arc discharge pulse duration is 10−2 s, the use of a bias generator with a frequency of 105 pulses/s will ensure the formation of 103 pulses, which will increase the flux by three orders of magnitude, making it possible to achieve high fluxes of silicon ion implantation.
Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. CRediT authorship contribution statement Alexander I. Ryabchikov: Conceptualization, Methodology, Project administration. Denis O. Sivin: Writing — original draft, Methodology, Visualization. Sergey V. Dektyarev: Investigation, Resources. Alexey E. Shevelev: Investigation, Formal analysis. Acknowledgment This work was carried out with the financial support of the Ministry of Science and Higher Education of the Russian Federation within the state assignment ‘‘Science’’ (grants No. 3.2415.2017/4.6 and 3.7245.2017/6.7). References
5. Conclusion
[1] J.M. Poate, G. Foti, D.C. Jacobson, Surface Modification and Alloying by Laser, Ion, and Electron Beams [Proceedings of a NATO Advanced Study Institute], Plenum Press, New York, ISBN: 0306413736, 1983. [2] J.S. Williams, J.M. Poate, Ion Implantation and Beam Processing [Academic], Academic Press, Orlando, ISBN: 9781483220642, 1984. [3] P.M. Martin, Handbook of Deposition Technologies for Films and Coatings: Science, Applications and Technology, William Andrew, ISBN: 9780815520313, 2009. [4] A.N.E.V. Kozlov, Yu. P. Sharkeev, et al., Microstructure of the near-surface layers of ion-implanted polycrystalline Cu, Surf. Coat. Technol. 56 (1) (1992) 11–17, http://dx.doi.org/10.1016/0257-8972(92)90190-L. [5] B.N. Chapman, Thin-film adhesion, J. Vacuum Sci. Technol. 11 (1974) 106–113. [6] A. Kinloch, Adhesion and Adhesives: Science and Technology, Chapman and Hall, New York, 1987, ISBN: 0-412-27440_X. [7] X. Liu, Y. Ren, X. Tan, S. Sun, E. Westkaemper, The structure of Ti–Si–N superhard nanocomposite coatings: ab initio study, Thin Solid Films 520 (2) (2011) 876–880, http://dx.doi.org/10.1016/j.tsf.2011.08.003. [8] J. Wang, J. Pu, G. Zhang, L. Wang, Interface architecture for superthick carbonbased films toward low internal stress and ultrahigh load-bearing capacity, ACS Appl. Mater. Interfaces 5 (11) (2013) 5015–5024, http://dx.doi.org/10.1021/ am400778p.
Studies have shown the possibility of the formation of pulsed and repetitively pulsed high-intensity beams of silicon ions. A pulsed vacuum-arc discharge with a current of 75 A formed a plasma flux of silicon with a duration of up to 350 μs. Silicon used as a cathode was previously subjected to neutron transmutation doping, which increased its conductivity. Plasma-immersion formation of an ion flux near the spherical grid structure followed by ballistic focusing of the ion beam in the drift space under conditions of its preliminary filling with plasma ensured silicon ion beam formation with a duration of more than 20 μs and a current of approximately 1.5 A. The regularities of the transport and focusing of a of silicon ion pulsed beam are investigated. It has been established that with bias pulse durations of more than 10 μs with small amplitudes of the accelerating voltage, repetitive transport failures of the high-intensity ion beam occur. An increase in bias amplitude from 0.6 to 1.5 kV contributes to the stabilization of the efficient transport of a silicon ion beam. The possibility of achieving ion current densities up 8
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.
A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al.
Nuclear Inst. and Methods in Physics Research, A xxx (xxxx) xxx [24] A.I. Ryabchikov, E.B. Kashkarov, N.S. Pushilina, M.S. Syrtanov, A.E. Shevelev, O.S. Korneva, A.N. Sutygina, A.M. Lider, High-intensity low energy titanium ion implantation into zirconium alloy, Appl. Surf. Sci. 439 (2018) 106–112, http://dx.doi.org/10.1016/j.apsusc.2018.01.021. [25] A.I. Ryabchikov, A.E. Shevelev, D.O. Sivin, A.I. Ivanova, V.N. Medvedev, Low energy, high intensity metal ion implantation method for deep dopant containing layer formation, Surf. Coat. Technol. 355 (2018) 123–128, http://dx.doi.org/10. 1016/j.surfcoat.2018.02.111. [26] André Anders (Ed.), Handbook of Plasma Immersion Ion Implantation and Deposition, first ed., John Wiley & Sons, New York, ISBN: 978-0-471-24698-5, 2000. [27] A.I. Ryabchikov, V.S. Skuridin, E.V. Nesterov, E.V. Chibisov, V.M. Golovkov, Obtaining molybdenum-9 in the IRT-T research reactor using resonance neutrons, Nucl. Instrum. Methods Phys. Res. B 213 (2004) 364–368, http://dx.doi.org/10. 1016/S0168-583X(03)01592-1. [28] A. Naymushin, Yu. Chertkov, I. Lebedev, V. Boyko, Feasibility study of creating additional experimental channels for silicon doping in IRT-T reactor, J. Ind. Pollut. Control 32 (2) (2016) 424–427. [29] A.I. Ryabchikov, P.S. Ananin, S.V. Dektyarev, D.O. Sivin, A.E. Shevelev, High intensity metal ion beam generation, Vacuum 143 (2017) 447–453, http://dx. doi.org/10.1016/j.vacuum.2017.03.011. [30] I.G. Kesaev, Katodnye protsessy elektricheskoy dugi. [cathodic processes of electric arc], Nauka Publ., Moscow, ISBN: 978-5-518-13837-7, 1968, (in Russian). [31] R.L. Boxman, D.M. Sanders, P.J. Martin, Handbook of Vacuum Arc Science and Technology, Noyes Publications, Park Ridge, New Jersey, ISBN: 9780815513759, 1995. [32] A. Anders, Ion charge state distributions of vacuum arc plasmas, Orig. Species Phys. Rev. E 55 (1) (1997) 969–981, http://dx.doi.org/10.1103/PhysRevE.55. 969. [33] C.W. Kimblin, Cathode spot erosion and ionization phenomena in the transition from vacuum to atmospheric pressure arcs, J. Appl. Phys. 45 (1974) 5235–5244, http://dx.doi.org/10.1063/1.1663222. [34] C.D. Child, Discharge from hot CaO, Phys. Rev. 32 (5) (1911) 492–511, http: //dx.doi.org/10.1103/PhysRevSeriesI.32.492. [35] I. Langmuir, The effect of space charge and residual gases on thermionic currents in high, Vacuum Phys. Rev. 2 (6) (1913) 450–486, http://dx.doi.org/10.1103/ PhysRev.2.450. [36] T.V. Koval, A.I. Ryabchikov, T.M. Kim An, A.E. Shevelev, D.O. Sivin, A.I. Ivanova, D.M. Paltsev, Numerical simulation of high-intensity metal ion beam generation, J. Phys. Conf. Ser. 1115 (3) (2018) 032007, http://dx.doi.org/10.1088/17426596/1115/3/032007. [37] V.M. Anishchik, V.V. Uglov, Modifikatsiya instrumental’nykh materialov ionnymi i plazmennymi puchkami, BGU, Mn, ISBN: 985-445-906-3, 2003, (in Russian). [38] A.I. Ryabchikov, A.E. Shevelev, D.O. Sivin, T.V. Koval, T.M. Kim An, High intensity, macroparticle-free, aluminum ion beam formation, J. Appl. Phys. 123 (23) (2018) 233301, http://dx.doi.org/10.1063/1.5034082. [39] W.D. Davis, H.C. Miller, Analysis of the electrode products emitted by dc arcs in a vacuum ambient, J. Appl. Phys. 40 (5) (1969) 2212–2221, http://dx.doi. org/10.1063/1.1657960. [40] G.Y. Yushkov, A. Anders, E.M. Oks, I.G. Brown, Ion velocities in vacuum arc plasmas, J. Appl. Phys. 88 (10) (2000) 5618–5622, http://dx.doi.org/10.1063/ 1.1321789.
[9] Y.H. Cheng, T. Browne, B. Heckerman, E.I. Meletis, Mechanical and tribological properties of nanocomposite TiSiN coatings, Surf. Coat. Technol. 204 (2010) 2123–2129, http://dx.doi.org/10.1016/j.surfcoat.2009.11.034. [10] J. Tian, P. Xu, J. Chen, Q. Liu, Microstructure and phase transformation behaviour of a Fe/Mn/Si/Cr/Ni alloy coating by laser cladding, Opt. Lasers Eng. 122 (2019) 97–104. [11] M. Pleva, B. Grančič, M. Mikula, M. Truchlý, T. Roch, L. Satrapinskyy, M. Gregor, P. Ďurina, V. Girman, P. Švec Jr., A. Plecenik, P. Kúš, Thermal stability of amorphous Ti-B-Si-N coatings with variable Si/B concentration ratio, Surf. Coat. Technol. 333 (2018) 52–60, http://dx.doi.org/10.1016/j.surfcoat.2017.10.063. [12] J.S. Reid, E. Kolawa, C.M. Garland, M.A. Nicolet, F. Cardone, D. Gupta, R.P. Ruiz, Amorphous (Mo Ta or W)–Si–N diffusion barriers for Al metallizations, J. Appl. Phys. 79 (2) (1996) 1109–1115, http://dx.doi.org/10.1063/1.360909. [13] X. Sun, J.S. Reid, E. Kolawa, M.-A. Nicolet, R.P. Ruiz, Reactively sputtered Ti-SiN films. II. Diffusion barriers for Al and Cu metallizations on Si, J. Appl. Phys. 81 (1997) 664–671, http://dx.doi.org/10.1063/1.364206. [14] R. Daniel, J. Musil, P. Zeman, C. Mitterer, Thermal stability of magnetron sputtered Zr–Si–N films, Surf. Coat. Technol. 201 (6) (2006) 3368–3376, http: //dx.doi.org/10.1016/j.surfcoat.2006.07.206. [15] J. Musil, P. Zeman, P. Dohnal, Ti-Si-N films with a high content of Si, Plasma Process. Polym. 4 (S1) (2007) S574–S578, http://dx.doi.org/10.1002/ ppap.200731408. [16] J. Musil, Hard nanocomposite coatings: thermal stability, oxidation resistance and toughness, Surf. Coat. Technol. 207 (2012) 50–65, http://dx.doi.org/10. 1016/j.surfcoat.2012.05.073. [17] M. Mikula, B. Grančič, M. Drienovský, L. Satrapinskyy, T. Roch, Z. Hájovská, M. Gregor, T. Plecenik, R. Čička, A. Plecenik, P. Kúš, Thermal stability and hightemperature oxidation behavior of Si–Cr–N coatings with high content of silicon, Surf. Coat. Technol. 232 (2013) 349–356, http://dx.doi.org/10.1016/j. surfcoat.2013.05.034. [18] S. Veprek, M.G.J. Veprek-Heijman, P. Karvankova, J. Prochazka, Different approaches to superhard coatings and nanocomposites, Thin Solid Films 476 (1) (2005) 1–29, http://dx.doi.org/10.1016/j.tsf.2004.10.053. [19] P.H. Mayrhofer, Ch. Mitterer, L. Hultman, H. Clemens, Microstructural design of hard coatings, Prog. Mater. Sci. 51 (8) (2006) 1032–1114, http://dx.doi.org/10. 1016/j.pmatsci.2006.02.002. [20] A. Cavaleiro, B. Trindade, M. Teresa Vieira, The influence of the addition of third element on the structure and mechanical properties of transition-metal-based nanostructured hard films part I-nitrides, in: Nanostructured Coatings, Springer Science Business Media, 2006, p. 671, ISBN-10: 0-387-25642-3, ISBN-13: 978-0387-25642-9. ˜ I. Kovács, L. Tóth, B. [21] K.P. Budna, P.H. Mayrhofer, J. Neidhardt, É. Hegedus, Pécz, C. Mitterer, Effect of nitrogen-incorporation on structure, properties and performance of magnetron sputtered CrB2, Surf. Coat. Technol. 202 (13) (2008) 3088–3093, http://dx.doi.org/10.1016/j.surfcoat.2007.11.009. [22] A. Flink, J.M. Andersson, B. Alling, R. Daniel, J. Sjölén, L. Karlsson, L. Hultman, Structure and thermal stability of arc evaporated (Ti0.33 Al0.67 )1 - xSixN thin films, Thin Solid Films 517 (2) (2008) 714–721, http://dx.doi.org/10.1016/j.tsf.2008. 08.126. [23] A.I. Ryabchikov, D.O. Sivin, P.S. Ananin, A.I. Ivanova, I.V. Lopatin, O.S. Korneva, A.E. Shevelev, High intensity, low ion energy implantation of nitrogen in AISI 5140 alloy steel, Surf. Coat. Technol. 355 (2018) 129–135, http://dx.doi.org/10. 1016/j.surfcoat.2018.02.110.
9
Please cite this article as: A.I. Ryabchikov, D.O. Sivin, S.V. Dektyarev et al., Formation of repetitively pulsed high-intensity, low-energy silicon ion beams, Nuclear Inst. and Methods in Physics Research, A (2019) 163092, https://doi.org/10.1016/j.nima.2019.163092.