DOC

Novel high plasma density source

By Juanita Hernandez,2014-02-11 08:22
11 views 0
The invention of sputtering using the first DC diode-type arrangement led to huge innovation. However, the technique was very limited in most respects, due to the very low plasma density available. Evolution of the technique resulted in the invention of magnetron sputtering, which enhanced the sputtering rate to reasonable levels. Today, the ever-increasing demands of modern products for processing and manufacturing accuracy and efficiency have pushed magnetron sputtering technology to its limits, thus limiting the products themselves and their production lines. We have now developed a brand new technology for sputtering that does not suffer from the natural limitations of magnetron sputtering, and offers new opportunity by further broadening the scope of sputtering technology. Our technology solves the key barrier stoppers associated to magnetron sputtering, namely (1) low target utilisation. (2) Process instabilities and rate drifts associated to the racetrack formation around the target. (3) Instabilities, low deposition rates, and thin target limitations when sputtering from magnetic materials due to the short-circuit of magnetic flux from the magnetron by the target. (4) Target poisoning instabilities when reactive sputtering, arising from the simultaneous presence of different target erosion rates, near and away from the racetrack. (5) Low deposition rate when reactive sputtering, due to the small surface area that the racetrack offers for the chemical reaction, and (6). Localisation of the plasma to a small region next to the racetrack, imposing limitations to the effect of plasma assisted film densification, and increasing the chances for target poisoning when reactive-sputtering.

    Plasma Quest Limited

    Plasma Quest Limited

    High Target Utilisation Sputtering (HiTUS)

    a summary of PQL’s novel technology

     1

    Plasma Quest Limited

    Novel high plasma density source used for a high target utilisation (HiTUS) sputtering system

M.Thwaites*, B. Holton, J.V. Anguita, P. Hockley and S. Rand, J. Ratteit

Plasma Quest Limited,

    Unit 1B Rose Estate,

    Osborn Way,

    Hook,

    Hampshire RG27 9UT,

    United Kingdom,

* Address for correspondence: mike.thwaites@plasmaquest.co.uk

ABSTRACT

    The invention of sputtering using the first DC diode-type arrangement led to huge innovation. However, the technique was very limited in most respects, due to the very low plasma density available. Evolution of the technique resulted in the invention of magnetron sputtering, which enhanced the sputtering rate to reasonable levels. Today, the ever-increasing demands of modern products for processing and manufacturing accuracy and efficiency have pushed magnetron sputtering technology to its limits, thus limiting the products themselves and their production lines. We have now developed a brand new technology for sputtering that does not suffer from the natural limitations of magnetron sputtering, and offers new opportunity by further broadening the scope of sputtering technology. Our technology solves the key barrier stoppers associated to magnetron sputtering, namely (1) low target utilisation. (2) Process instabilities and rate drifts associated to the racetrack formation around the target. (3) Instabilities, low deposition rates, and thin target limitations when sputtering from magnetic materials due to the short-circuit of magnetic flux from the magnetron by the target. (4) Target poisoning instabilities when reactive sputtering, arising from the simultaneous presence of different target erosion rates, near and away from the racetrack. (5) Low deposition rate when reactive sputtering, due to the small surface area that the racetrack offers for the chemical reaction, and (6). Localisation of the plasma to a small region next to the racetrack, imposing limitations to the effect of plasma assisted film densification, and increasing the chances for target poisoning when reactive-sputtering.

I. Introduction

    Sputtering is a technique used to deposit thin coatings of materials on valuable substrates [1,2]. It is one of the most fundamental processes exploited in any semiconductor research or production plant. Sputtering is also used to deposit thin coatings on other delicate substrates such as lenses or aerospace structures, in order to promote their optical and/or wear-resistant properties. The technology is now increasingly being used also in the food packaging industry.

    Today, sputtering is a preferred high vacuum method for depositing thin films. The high quality and uniformity of the depositions achieved this way is due to the high level of process control and purity. The process is also environmentally friendly, whereas other chemical methods or electroplating are not, and industry is constantly being persuaded to stop their operation. Sputtering is a suitable alternative that offers good process control, target and substrate cleaning, and also film densification through ion bombardment.

    The basic technique of sputtering is very mature and well understood [1-5]. The early arrangement for sputtering consisted of a DC diode-type arrangement [1]. This technique was very limited in most respects, due to the very low plasma density available. The requirement for faster processing in mass production led to the evolution of the technique, resulting in the invention of magnetron sputtering. This device increased the plasma density at the target, thus enhancing the sputtering rate to reasonable levels, allowing for industrial production [6-8].

Since its invention in the 1960‟s [9], the basic technique of magnetron sputtering is still today the industry

    standard method for sputtering. Modern magnetron design and technology refinements using pulsed methods have optimised this technology. However, the ever-increasing demands of modern products for manufacturing accuracy and efficiency have stretched magnetron technology to its limits, thus setting limits to the products

     2

    Plasma Quest Limited

    themselves and their production lines. It is hence required to investigate new breaking technologies that do not suffer from the natural limitations of magnetron sputtering, and allow for new opportunities both in the specification of the products, and in the cost and efficiency of their production lines. Magnetron sputtering is not the ideal process for sputtering, due to several constraints:

    (1.) The magnetron confines the plasma to a characteristic ring-shaped region next to the surface of the target. This leads to a strong target erosion rate localised to this ring-shaped region on the target, and thus, to the eventual formation of the characteristic annular “race-track” around the target, which is where most of the

    sputtering process occurs. The concentration of the target erosion to this small race-track region, causes a short operating lifetime and low utilisation of the target material. Typically, only about 20% of the material from the target gets sputtered before it needs replacement. This increases the costs significantly, especially for precious metal targets, and increases the shut-down and conditioning time of the sputtering system.

    (2). Magnetron sputtering suffers form process instabilities. These are associated to the change of the shape of the target through its lifetime, as the race-track region of the target evolves during its operation. This has a strong effect on the process, since it is from this race-track region that most of the sputtering processes occur. On top of this, as the race-track grows deeper into the target, this brings the sputtering surface closer to the magnetron, hence the magnetic field strength at this surface rapidly gets stronger. This brings about an ever-increasing plasma density with the utilisation of the target, and also a change in the working distance. The net effect of this is an observed drift in the deposition rate with the operation of the target through its lifetime.

    (3). Targets of magnetic materials, such as iron, nickel and cobalt, act as a short-circuit for the magnetic flux from the magnetron, causing the magnetic field strength at the surface of the target, hence the deposition rate, to be much reduced. To overcome this problem, very thin targets of these materials must be used. However, this greatly aggravates the two problems mentioned above.

    (4). Further process instabilities arise from the fact that the erosion rate is non-uniform over the surface of the target. This problem is particularly aggravated during reactive sputtering using a reactive gas, since target-material rich and target-material deficient materials will be formed at the surface of the target at the same time, depending on the local erosion rate. This gives rise to local target poisoning, which can often be seen in the central region of the target as a discoloured area, even when sputtering using argon only. Target poisoning usually leads to plasma instabilities, because the plasma is run from the target, therefore changes in the status or electrical conductivity of the target will affect the intensity of the plasma, and hence the process. The proximity of the plasma to the target enhances the probability of diffusion of reactants from the plasma to the target surface, thus maximising the rate of target poisoning. This proximity is also not ideal for chemical reactions in the plasma phase, thus loosing huge potential in the sputtering capabilities.

    (5). Plasma-cleaning of the substrate prior to the deposition is made difficult, since the substrate table must be made into an electrode itself. Also, a plasma must be struck using the substrate table and without using a magnetron (for uniformity reasons). This makes necessary to use a low plasma density and a high substrate voltage, in order to run the plasma. A high substrate voltage is linked to high-energy ion bombardment on the substrate, and hence to crystalline lattice damage and detrimental performance or even destruction, in semiconductor devices. It is desired to have a uniform and high-density plasma present on the substrate to allow for a thorough cleaning, whilst maintaining the substrate voltage as low as possible, to maintain any plasma-related damage as superficial as possible. Substrate cleaning is vital to ensure the adhesion and quality of the coating, and to minimise spread in the performance of the coated product, in a manufacturing environment.

    The problems mentioned above have two origins: (a) the confinement of the plasma to the race-track region, and (b) the fact that the plasma is driven from the target. In this paper, we describe an alternative system that offers similar sputtering rates and uniformity when using the same working distance, but does not involve the use of a magnetron. Instead, our system uses the entire surface of the target homogeneously for sputtering, and does not involve a race-track, hence eliminating points 1 to 4 above. Also, the plasma is generated away from the target, therefore eliminating the plasma instabilities associated to the status of the target (points 4 to 5). The uniformity in the erosion rate at the target make this system particularly well-suited for reactive sputtering, reaching sputtering rates significantly greater than those of commercial magnetron systems (up to 20 times faster), and with higher break-down field characteristics. One of the key advantages of our system is its high degree of target utilisation (above 90%), which renders its name as high target utilisation (HiTUS) system ?.

     3

    Plasma Quest Limited

II. Description of equipment

    The general description of the Hi-TUS system has been reported previously [10]. Briefly, a standard diode-type sputtering arrangement is adopted. The deposition rate is enhanced by focussing a remotely-generated high-density plasma onto the target, by means of two electromagnets. A schematic of the system is shown in Fig. 1. The intensity of the plasma is magnetically enhanced using the “launch magnet” (see Fig. 1). The plasma is then bent and directed towards the target using the “steering magnet”.

    Even with the plasma running, sputtering at the target does not occur unless a negative DC voltage is applied to it, either using a DC power supply, or an additional rf generator. The sputtered material then deposits on the electrically earthed substrate, where growth occurs. A shutter is used in the conventional way to sputter-clean and condition the target prior to the deposition. Also, the substrate can be sputter-cleaned using the high density plasma. This is done simply by reversing the polarity of the “steering” electromagnet by pressing a switch (see

    figure1). This causes the plasma to be “repelled” from the steering electromagnet, and causes it to bend in the opposite way, towards the substrate. We have found that this form of substrate clean dramatically improves the adhesion performance of the coating. The vacuum chamber is evacuated using a turbomolecular/rotary pump assembly, and the gasses (argon, oxygen or nitrogen) are fed using mass flow controllers (MFC) into the chamber.

    Several (six) different targets are mounted on a large single copper backing plate. This plate is water-cooled, and it can be rotated to change the particular target used, without breaking the vacuum. Only the target being used is exposed to the plasma, and the rest of the targets and backing plate are hidden inside an earthed housing. This method is particularly cost-effective when compared to a multi-target magnetron system, since the savings in multiple magnetrons are considerable.

    The design and functioning of our high-density plasma source is based around the energetics of the argon atoms. The key objective of the design was to excite a high density of electrons to an energy level between 40 and 100eV [11, 12]. These energies correspond to a maximum in the ionisation probability of argon. A 13.56MHz radio-frequency (rf) source and matching unit were connected to a suitably designed antenna, which was adjusted to couple the rf into the argon gas inside a quartz tube. The mechanism by which the electrons gained energy this way is through electron “trapping” into the plasma waves generated [13-16], giving rise to

    Landau damping.

    The flow of current in the two electromagnets was adjusted to give the highest target current, while running an argon plasma. This current is of the order of 10 Amps, yielding about 200-300 Gauss. It is observed that the region of the plasma close to the rf antenna appears pink in colour. This colour accounts for the high concentration of non-ionised excited argon atoms in the plasma. However, the region of the plasma close to the launch electromagnet appears intense blue. This accounts for a very high concentration of argon ions, which greatly exceeds that of non-ionised excited argon atoms. The curved region of the plasma up to the target also shows an intense blue colour, indicating a highly ionised plasma. The colour of the plasma can be seen photographed in our website at www.plasmaquest.co.uk.

     4

    Plasma Quest Limited

III. Results

    The optical emission spectra (OES) from the various regions of the plasma are shown in Fig 2. Fig 2(a) shows

    the OES from the pink region of the plasma, and Fig. 2(b) shows that from the blue region. The bands labelled

    in the figure occur in the UV, and therefore do not contribute to the visual appearance of the plasma. Fig. 2(a)

    exhibits the strongest emission band at 420nm, which corresponds mainly to emission from non-ionised excited

    argon atoms. The emission at 434.8nm correspond to that from ionised argon. Using the correct oscillator 10-3strengths [17, 18], the ion density (n) of the pink region of the plasma was estimated at about 10 cm. Fig 2(b) i

    shows a very strong increase in the ion emission band at 434.8nm for the blue region of the plasma. For this 14-3region, n was estimated at about 10 cm, some 4 orders of magnitude denser than in the pink region. i

    Further evidence for the high density of our plasma is obtained from the target current density values. We have 2achieved a current density (j) of 70mA/cm at the target. This current density is uniform over the surface of the target, since the target erosion is uniform over its surface. From this, the ion density (ni) of the plasma can be

    estimated using the following equation [12].

    encii j?i4

where e is the electron charge, cis the average speed of the argon ions, and j is the ion current density. Using ii

    413-35 ×10c cm/s for , [12], n yields values close to 5×10 cm. It should be noted that j was measured at the iii

    target, away from the plasma source (some 30cm). It is expected that the ion density at the plasma source

    would be higher than this calculated value.

The relationship between target current and target voltage is plotted in Fig 3, for four different values of launch -4tube rf power. The argon gas pressure was maintained at 26×10 mbar. The target was driven using a DC

    power supply. The figure shows that the target current increases almost linearly with target voltage up to about

    50 volts. It is observed that increasing the target voltage to about 100 Volts causes the target current to saturate

    to a maximum level (I) predetermined partly by the rf power. S

The fact that the target current remains unchanged with increasing the target voltage from 100 V to 1000 V,

    means that the contribution from the secondary electron emission to the target current is negligible. Therefore,

    our measurement of j is a true figure for the argon ion current density arriving at the target. This validates our i

    estimate of n. i

Figure 4 shows the saturation target current (I) as a function of argon gas pressure, for four values of launch rf Spower. These are the main parameters that control the value of I when sputtering using argon. At the low S-4pressures (4 - 12×10 mbar), I increases quickly with the pressure, indicating a tendency towards a pressure-S

    limited process for the production of argon ions. At the higher pressures, I increases less strongly with pressure, Sbut more strongly with rf power, indicating a tendency towards an energy-limited process for the production of

    argon ions.

    Figure 4 shows the deposition rate for nickel and iron as a function of target voltage. The rf power and pressure -4of the plasma were fixed at 500 W and 20×10 mbar respectively. The working distance was fixed at 15cms. Figure 4 shows that the deposition rate varies almost linearly with the target voltage from 0 to a maximum value.

    There is a minimum target voltage before any deposition occurs, which roughly corresponds to the binding

    energy of the target material, and also, the voltage needed to achieve I. The deposition rate tails off slightly at Sthe higher target voltages.

Reactive sputtering has been performed to deposit aluminium oxide from a pure aluminium target, using an

    argon/oxygen gas mixture. Pure argon gas was introduced to the chamber by placing a diffusion ring close to

    the aluminium target. This was done to maintain the target in a “metallic mode” even when high doses of oxygen

    were introduced into the chamber. Keeping the target in a “metallic mode” is key, because otherwise, excess

    oxygen will “poison” the target, and the deposition rate will collapse. Also, adding the highest possible amount of

    oxygen into the chamber is necessary to avoid producing aluminium-rich aluminium oxide, in order to maximise

    the break down (BD) electric field of the deposit. For this reason, the oxygen was fed into the chamber though

    another diffusion ring placed as close as possible to the substrate.

All samples grown using the HiTUS process show refractive index (RI) values ranging from 1.66 to 1.68, with a

    uniformity of better than +/- 1% over a 4 inch wafer, using substrate rotation. These RI values are higher than

    for ion beam deposited material (1.64) supplied by Nordiko. As alumina has an „ideal‟ RI of over 1.74 [17], this

     5

    Plasma Quest Limited

    may indicate improved quality from the HiTUS process, however metal rich samples also give high RI. To determine which of these is the case, BD electric field measurements were made. These measured about -1-1, and in some cases measured up to 22MVcm. It must be emphasised that we do not posses a 15MVcmclean-room environment, and therefore the surface of the wafers were exposed to dust before and after the alumina deposition. The presence of these dust particles would reduce the BD field strength by acting as -1preferred pathways for the electrons. This means that the highest measured values (above 20MVcm) are more

    likely to be closer to the true BD field of the alumina.

IV. Discussion

    The development of our high density plasma source, and the magnetic confinement required to direct the plasma to the target has been the key breakthrough for the HiTUS technology. The plasma source was designed to create maximum ionisation of argon atoms by exciting the electrons to the energies that correspond to the maximum ionisation cross-section of argon. Figure 4 shows that for the higher pressures, I increases Svery quickly with rf power, but does not reach a maximum saturation limit. In fact, our measurement of n was i

    done without reaching the limit of the plasma source itself, but instead, it was limited purely by the power limitations of the 2kW rf power supply. Further experimentation is to proceed using a more powerful 5kW rf power supply.

    Many of the disadvantages of using a magnetron to enhance the sputtering rate have been successfully overcome using the HiTUS system instead. The HiTUS system can yield similar deposition rates as commercial magnetron sputtering systems when using pure metal targets, and has a much enhanced rate when reactive sputtering. It has been explained how HiTUS technology does not suffer from the key constrictions mentioned in the introduction, pertaining magnetron sputtering. Most of these constrictions are attributed to the following three reasons that do not apply to the HiTUS system: (1) The formation of a racetrack region, (2) the driving of the plasma from the target, and (3) The localisation of the high-density plasma to a region next to the racetrack region of the target.

    Attempts to solve point (3) above have been partially solved using type II unbalanced magnetrons, where the plasma is stretched from the racetrack region all the way to the substrate, in order to enhance the quality of the coating. However, the plasma density still peaks at regions closest to the target, which is where most of the plasma-phase chemical reactions would occur. The HiTUS system can be arranged to concentrate the plasma even further away from the target, or even close to the substrate, thus minimising target-poisoning, and further optimising film quality.

    One of the main advantages of the HiTUS system is the possibility of cleaning the substrate uniformly using a very high-density plasma. The fact that the substrate table is earthed means that there is minimal plasma damage to the substrate. However, the high density of the plasma ensure a thorough cleaning of the entire surface of the substrate. We have ample evidence that demonstrates a much enhanced adhesion performance of coatings when performing such cleaning on the substrate, and these results will be published in a separate paper.

    It is important to realise that the reason why the deposition rate varies almost linearly with the target voltage is that the plasma is sustained by a remote source, therefore the intensity of the plasma is independent from the target voltage. This is the only way to deconvolute the effects from purely changing the target voltage or purely changing the intensity of the plasma, on the properties of the deposit. This deconvolution effect allows us to explore other regions of the parameter-space that would be otherwise impossible to explore. This is the case for depositions carried out either at very high target voltage and low plasma intensity, or at high plasma intensity and very low target voltage, both independently of the gas pressure.

    The latter case is particularly useful for two reasons. Firstly, very low target voltages can be used, whilst maintaining a very stable high-powered plasma. Figure 6 shows the deposition rate as a function of target voltage for the very low target voltage values, using a 4 inch cobalt target and a 2kW rf plasma. The figure shows that the sputtering rate can be controlled to extremely low values, about 10Å/minute. The ability to control such a low growth rate process is very important for new technologies requiring ultra-thin coatings, such as very small transistor gate oxides, magnetic RAM (MRAM) devices [19], or quantum tunnelling devices. Another advantage of using very low target voltages is that the damage to the target is minimised. This is important when using delicate targets such as those made of organic materials, intended for use as the source material for organic photoluminescent devices such as flat panel displays, or organic semiconductors. Such low target voltage is not possible when using magnetron sputtering, because it is necessary at least several hundred volts to initiate and maintain a stable plasma. Reducing the target voltage to 50 volts or less would lead

     6

    Plasma Quest Limited

    into the extinguishing of the plasma, thus these two applications are made very difficult using standard magnetron sputtering.

    Figure 6 shows that there is a minimum target voltage needed to initiate the sputtering process, even when running a high powered plasma. This corresponds to about 50 Volts for cobalt and nickel, and about 100 Volts for iron. There also exists a region of voltage where the sputtering rate increases only very slowly with increasing target voltage, until a point at about 100 to 120 Volts, where the sputtering rate then increases at a faster constant rate with target voltage, up to several hundred volts. We are unsure about the origin of this initial trend.

    The reactive sputtering process for alumina was studied in detail. It was observed that this process was stable for many hours, without the need to adjust any of the deposition parameters. We do not use any form of feed-back control system from the optical emission (OE) of the plasma, nor pulse the operation of the target. Also, we were able to deposit good quality alumina in the “high growth rate” region of the hysterisis curve, far from the runaway point. We believe that this is partly because we are using the full surface of the target for erosion, and for the reactive process to take place. However, we also believe that a stronger influence for this is the fact that the sputtered material from the target has to travel through a very high density plasma before it can be deposited on the substrate.

    This very high density plasma contains excited or activated oxygen available to react with aluminium, and we believe this is the region where most of the reactive process occurs. This means that the alumina forms mainly in the high density plasma region, away from the target. The effect of this is to minimise the problems associated with “target poisoning” which lead to the “runaway” point in the hysterisis curve. This is where the surface of the target becomes oxidised, and thus the electrical properties of the target change, leading to process instabilities, or a very low deposition rate when sputtering using r.f.. The oxidation of metallic aluminium in the plasma phase is the only explanation to the fact that we observe a strong increase in the deposition rate with increasing oxygen flow when running the aluminium target using a DC power source, before meeting the runaway point, and obtaining very high quality aluminium oxide films.

    The advantage of being able to deposit good quality alumina films maintaining the target stable in the “metallic mode” (that is, away from the “target poisoning” or “runaway” point in the hysterisis curve) is threefold. First, the deposition can be performed at high deposition rates. Secondly, the target does not need a preconditioning stage prior to sputtering, and hence the deposition can be stopped and re-started as often as wished during a process without any drift in the deposition rate due to instabilities from target oxidisation. Thirdly, There are no issues from target oxidisation, and therefore the process is stable. There is no need to adjust the process parameters using feed-back systems that monitor the OE from the plasma, or pulsed systems. Further process stability is gained from the fact that the plasma is not run form the target (unlike magnetron systems), therefore the plasma does not depend on the oxidation state of the target.

     -1The relatively high BD fields measured for the alumina depositions (>20MVcm) and the optical transparency of

    the films suggest that the alumina films are not aluminium rich, and are close to their stoichiometric value. This means that the higher RI values observed are an indication of true film densification, rather than an artefact from the ellipsometric measurement.

    The higher density of our alumina films is probably due to a densification effect of the plasma. The very high density of the plasma would force most of the particulates to deposit on the substrate in highly energetic or activated states. The extra energy of these particulates would allow them to settle in lower energy configurations.

     7

    Plasma Quest Limited

Conclusions

    Magnetron sputtering offers great advantages over diode sputtering, but also introduces some disadvantages, such as the ones listed in the introduction section. The origin of the main dissadvantages has been identified to be either 1) The formation of a racetrack around the target, and 2) The driving of the plasma through the target.

    We have developed a sputtering system that either completely eliminates or minimises these disadvantages. The system uses a remotely-generated plasma source, and magnetic confinement to guide the plasma to the target. This way, the entire surface of the target is uniformly used for erosion, thus eliminating points 1) and 2) above. The system offers over 90% of the target material utilisation, so it is called high-target utilisation system (HiTUS).

    The development of a very high density plasma source allows high sputtering rates, similar or sometimes higher to the ones offered by commercial magnetron systems used in production lines, but without the instabilities and low target use associated with the formation of a racetrack.

    The advantage of using a high density plasma source for reactive-sputtering has been explained in detail for the case of aluminium oxide, the advantages being faster growth rates, and higher film densification, and higher break-down electric fields.

    The advantages of using the HiTUS system for substrate plasma-cleaning prior to coating have been explained in terms of a more thorough clean, and more superficial or no crystalline lattice damage. This offers enhanced adhesion performance.

    The lower cost of the HiTUS system compared to a magnetron sputtering system has been explained for a multiple target system.

     8

    Plasma Quest Limited

    References

[1] S. C. Brown, Introduction to Electrical Discharges in Gases, Wiley, New York and London (1966).

    [2] J. W. Coburn and E. Kay, Appl. Phys. Letters 18, 10, 435 (1971)

    [3] P. T. Smith, Phys. Rev. 36, 1293 (1930).

    [4] E. Eser, R. E. Ogilvie, and K. A. Taylor, J. Vac. Sci. Tech. 15, 2, 199 (1978). [5] D. B. Medved, P. Mahadevan, and J. K. Layton, Phys. Rev. 129, 2086 (1963). [6] R. W. Berry, P. M. Hall and M. T. Harris, Thin Film Technology, Van Nostrand (1968). [7] B. N. Chapman and D. S. Campbell, J. Phys. C. (Solid State Phys.) 2, 200 (1969). [8] B. N. Chapman, J. Vac. Sci. Tech. 11, 106 (1974).

    [9] B. D. Fraser, Thin Film Processes, Academic Press, New York and London (1978).

    [10] Liz‟s paper

    [11] O. A. Popov, High Density Plasma Sources, Noyes Publications, Massachusetts (1995).

    [12] B. Chapman, Glow Discharge Processes, Wiley, New York (1980).

    [13] P. Zhu and R. W. Boswell, Phys. Rev. Lett. 63, 2805 (1989).

    [14] P. Zhu and R. W. Boswell, Phys. Fluids B, 3, 869 (1991).

    [15] M. Light, I. D. Sudit, F. F. Chen and D. Arnuch, Phys. Plasmas, 2, 4094 (1995). [16] E. W. McDaniel, Collision Phenomena in Ionised Gases, Wiley, New York and London (1969). th ed. London and Tokyo, (1993-1994). [17] D. R. Lide, Handbook of Chemistry and Physics, 74[18] I. P. Zapesochnyi and P. V. Feltsan, Optics & Spectroscopy 20, 291 (1966) [19] J. C. Mallinson, Magneto-Resistive Heads, Academic Press, San Diego, (1995)

    Figure Captions

    1. Schematic diagram of a multiple-target HiTUS sputtering system

    2. OES obtained from the plasma, from (a) the pink region of the plasma, close to the antenna, and (b) the

    blue region of the plasma, close to the magnetic field.

     -43. Ion current versus target voltage for four rf power levels, at an argon pressure of 26?10 mbar, using a

    10cm diameter Chromium target.

    4. Ion current versus process pressure for four rf power levels, using a 10cm diameter Chromium target.

    5. Deposition rate for nickel and iron targets (38mm ? 38mm) as a function of target DC voltage, using an -4 argon pressure of 20?10mbar, and 0.5 kW rf power.

    6. Deposition rate as a function of DC target voltage, for very low target voltage values, using an argon -4pressure of 26?10 mbar, and 2.0 kW rf power, for a 4 inch cobalt target.

     9

    Plasma Quest Limited

    Figure 1

     Rotating substrate table and

    shutter assembly Launch electromagnet

     Quartz tube

     PLASMA

     r.f. antenna Pumps

     Multiple target backing plate: 8 target materials or more Steering electromagnet

    Figs 2(a) and (b)

     2500025000 a)

     2000020000

     1500015000

     434.8 nm434.8 nm420.0 nm420.0 nm1000010000 CCD countsCCD counts++438.8438.8ArArArAr 437.3437.3 50005000 435.8435.8 434.4434.4 00432.9432.9

    b) 431.4431.4

     429.9429.9 2000020000428.4428.4

     426.9426.9

    1500015000 425.5425.5 424.0424.0 CCD countsCCD counts1000010000 422.5422.5

     421.0421.0 50005000419.5419.5Figures 2 (a) and (b)

     418.0418.0

    00416.5416.5 10 415.0415.0

    Wavelength (nm)Wavelength (nm)

Report this document

For any questions or suggestions please email
cust-service@docsford.com