Abstract
According to the invention there is a method of depositing SiO.sub.2 onto a substrate by pulsed DC reactive sputtering which uses a sputtering gas mixture consisting essentially of oxygen and krypton.
Claims
1. A method of depositing SiO.sub.2 onto a substrate, comprising: supporting the substrate with an electrically conductive platen within a chamber of sputtering apparatus; providing a sputtering gas mixture consisting essentially of oxygen and krypton in the chamber of the sputtering apparatus by introducing the oxygen and krypton into the chamber at respective flow rates, wherein the ratio of the flow rate of the krypton in units of sccm to the flow rate of the oxygen in units of sccm is in the range of 0.1 to 0.9; providing a pulsed DC power to a target of the sputtering apparatus; and applying RF power to the platen to produce a DC bias; wherein the target comprises silicon, silicon is sputtered from the target by the sputtering gas mixture consisting essentially of krypton and oxygen, sputtered ions are accelerated to the substrate on the platen by the DC bias, and SiO.sub.2 is deposited on the substrate as a result.
2. A method according to claim 1 in which said ratio of the flow rates of krypton to oxygen is in the range 0.2 to 0.8.
3. A method according to claim 1 in which the RF power applied to the substrate is in the range 20 to 150 W.
4. A method according to claim 1 in which the sputtering gas mixture is present in the process chamber, in which the deposition of SiO.sub.2 takes place, at a pressure in the range 1 mTorr to 20 mTorr (0.13 Pa to 2.67 Pa).
5. A method according to claim 1 which deposits the SiO.sub.2 so as to have a density of 2.35 gcm.sup.−3 or greater on the substrate.
6. A method according to claim 1 in which the SiO.sub.2 is deposited as a film.
7. A method according to claim 1 in which the substrate is positioned on a substrate holder, and during the deposition of SiO.sub.2 the substrate holder is at a temperature of less than 100° C.
8. A method according to claim 1 in which the pulsed DC power is provided by a magnetron.
9. A method according to claim 1 in which the RF power applied to the substrate is in the range 20 to 125 W.
10. A method according to claim 1 which deposits the SiO.sub.2 so as to have a density on the substrate of 2.40 gcm.sup.−3 or greater.
11. A method according to claim 1 in which the substrate is positioned on a substrate holder, and during the deposition of SiO.sub.2 the substrate holder is at a temperature of less than 70° C.
12. A method of depositing SiO.sub.2 onto a substrate, comprising: supporting the substrate with an electrically conductive platen within a chamber of sputtering apparatus; introducing a sputtering gas mixture consisting essentially of oxygen and krypton into the chamber of the sputtering apparatus, wherein the percentage of the krypton in the sputtering gas mixture is at least 30%; providing a pulsed DC power to a target of the apparatus; and applying RF power to the platen to produce a DC bias, wherein the target comprises silicon, and silicon is sputtered from the target by the sputtering gas mixture, sputtered ions are accelerated to the substrate on the platen by the DC bias, and SiO.sub.2 is thereby deposited on the substrate so as to have a density on the substrate of at least 2.35 gcm.sup.−3 as a result.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
(1) Embodiments of methods and apparatus in accordance with the invention will now be described with reference to the accompanying drawings, in which:
(2) FIG. 1 is a semi-schematic diagram of apparatus of the invention;
(3) FIG. 2 shows SiO.sub.2 film density versus platen RF for SiO.sub.2 films deposited in Ar/O.sub.2 and Kr/O.sub.2 environments;
(4) FIG. 3 shows SiO.sub.2 refractive index versus platen RF for SiO.sub.2 films deposited in Ar/O.sub.2 and Kr/O.sub.2 environments;
(5) FIG. 4 shows SiO.sub.2 deposition rate versus platen RF power for Ar/O.sub.2 and Kr/O.sub.2 environments;
(6) FIG. 5 shows hysteresis curves of (a) target voltage and (b) target current for a silicon target with Ar or Kr mixed with oxygen as a function of oxygen flow;
(7) FIG. 6 shows secondary electron emission as a function of bombardment energy and gas type for ions incident on Mo and W targets (taken from Glow Discharge Processes, Chapman, Wiley and Sons, 1980);
(8) FIG. 7 shows SiO.sub.2 film density versus percentage inert gas for SiO.sub.2 films deposited in Ar/O.sub.2 and Kr/O.sub.2 environments;
(9) FIG. 8 shows SiO.sub.2 film stress versus platen RF power for SiO.sub.2 films deposited in Ar/O.sub.2 and Kr/O.sub.2 environments; FIG. 9 shows electrical breakdown data (current density versus field) for silicon dioxide films sputtered in Ar/O.sub.2 and Kr/O.sub.2 environments; and
(10) FIG. 10 shows FTIR spectra for SiO.sub.2 films sputtered in Ar/O.sub.2 and Kr/O.sub.2 environments.
DETAILED DESCRIPTION OF EMBODIMENTS
(11) FIG. 1 shows an embodiment of an apparatus of the invention, depicted generally at 10. The apparatus includes a vacuum chamber 8 within which is located a wafer platen 2. An upper portion of the chamber 8 includes the target 9, which can be formed from silicon or silicon dioxide. A pulsed DC power supply 1 is provided to apply pulsed DC power to the target 9. Typically, the apparatus will be in the form of a pulsed DC magnetron reactive sputtering arrangement, in which case the apparatus will further include a magnetron. This is not shown in FIG. 1, but in practice the magnetron would be located behind the target 9, as is well-known by the skilled reader. In use, the platen 2 supports substrate, typically a wafer, in opposition to the target 9. The platen 2 is formed from a conductive material which is biased with an RF signal provided by an RF power supply 3 through a capacitive coupling circuit so that the platen 2 can act as an electrode. The RF bias in the presence of a plasma produces a negative DC bias to develop on the platen 2 so that sputtered ions are accelerated towards the substrate. Sources of oxygen 4, argon 5 and krypton 6 are provided. Oxygen, argon or krypton are selectively admitted into the chamber 8 using mass flow controllers 7 as part of an appropriate gas manifold. Oxygen gas reacts with silicon sputtered from the target 9 to form a layer of silicon dioxide on the surface of a wafer positioned on the platen 2. It is understood that the argon source 5 is provided solely for the purpose of performing comparative experiments between reactive sputtering achieved using Ar/O.sub.2 environments and reactive sputtering achieved using Kr/O.sub.2 environments. In other embodiments an argon source 5 would not be provided and a simpler gas manifold might be used.
(12) Experiments were performed using target powers of 1.25 and 2 kW, a platen temperature of 50° C., with various RF powers being applied to the platen 2. Experiments were performed in Ar/O.sub.2 environments using flow rates of 30 sccm Ar and 80 sccm O.sub.2. Experiments were also performed in Kr/O.sub.2 environments using a Kr flow rate of 30 sccm, and an oxygen flow rate of 80 sccm. With an Ar/O.sub.2 environment and a target power of 2 kW, the SiO.sub.2 film density is 2.26 gcm.sup.−3 at a film thickness of 1000 nm and no RF power applied to the platen. When RF power is applied to the platen, it is found that the film density increases, and at a platen RF power of 120 W, it is found that the film density increases by ˜3% to 2.33 gcm.sup.−3. It is believed that RF biasing of the platen helps to densify the silicon dioxide film by energetic ion bombardment. However, overly high DC bias (producing highly energetic ions) can also produce ion-irradiated defects and voids, actually resulting in a lower film density. This may explain why the SiO.sub.2 film density appears to be saturating at the medium-high platen RF powers shown in FIG. 2 for deposition using Ar/O.sub.2 environments. We have shown that with Ar/O.sub.2 environments, the film density actually decreases at very high applied platen RF powers (data not shown here). FIG. 2 also shows SiO.sub.2 film densities obtained using a Kr/O.sub.2 reactive sputtering process. It can be seen that the use of a Kr/O.sub.2 environment in the reactive sputtering process can result in the SiO.sub.2 film density being increased to over 2.40 gcm.sup.−3. We find that the film density is essentially independent of the target RF power, but increases with RF power applied to the platen.
(13) FIG. 3 shows that SiO.sub.2 films reactively sputtered with Kr/O.sub.2 have a higher refractive index than those reactively sputtered with Ar/O.sub.2. This is consistent with the increased SiO.sub.2 film density associated with Kr/O.sub.2 reactive sputtering and may be indicative of a higher atomic packing density. In FIG. 3 for 1.5 kW pulsed DC power we can see a substantial increase in refractive index for higher bias powers.
(14) FIG. 4 shows SiO.sub.2 film deposition rate as a function of platen bias power for both Ar/O.sub.2 and Kr/O.sub.2 environments. Data were obtained for target RF powers of 2 kW and 1.25 kW. It can be seen that Kr/O.sub.2 reactive sputtering provides a significant increase in deposition rate compared to Ar/O.sub.2 reactive sputtering. More particularly, the deposition rate associated with the Kr/O.sub.2 reactive sputtering process is ˜16% higher than with the Ar/O.sub.2 process, and this is essentially independent of the platen RF power. From this we conclude that the process of the present invention may have significant advantages in terms of fabrication throughput.
(15) Without wishing to be constrained by any particular theory or conjecture, it is believed that the increase in deposition rate of silicon dioxide in the Kr/O.sub.2 reactive sputter gas environment compared to Ar/O.sub.2 can be partly attributed to the increased sputter yield of Kr in Si compared to Ar. However, the fact that film density is independent of deposition power tends to argue against simple momentum transfer as being the cause of the density increase for SiO.sub.2 films reactively sputtered with Kr/O.sub.2. In addition, the increases in deposition rate and film density are not accompanied by an increasing target voltage as shown in Table 1.
(16) TABLE-US-00001 TABLE 1 Comparison of IV characteristics for different gas mixes, target RF power 2 kW Target Gas Mix V (Volts) Target I (Amps) Pressure (mT) 100% Ar 386 5.1 8.3 100% Kr 488 4.1 8.4 30 sccm Ar:80 sccm O.sub.2 206 9.66 7.7 30 sccm Kr:80 sccm O.sub.2 205 9.77 7.8
(17) Although a krypton only plasma exhibits a much higher voltage and a lower current than an Ar only plasma at the same pressure, when sufficient oxygen is introduced to poison the target, the voltage and current for Ar/O.sub.2 and Kr/O.sub.2 plasmas are very similar. FIG. 5 shows IV characteristics for the reactive process with Ar and Kr. The target RF power is 2 kW. It can be seen that both the V and I characteristics of Ar/O.sub.2 and Kr/O.sub.2 converge as the O.sub.2 fraction is increased and the target reaches the poisoned regime.
(18) A low current for the Kr only plasma is consistent with a low secondary electron emission for krypton compared with argon working gas. From the known data shown in FIG. 6, it can be seen that Ar has approximately twice the level of electron emission compared with Kr over a wide energy range. Therefore, sputtering with Kr instead of Ar will result in fewer secondary electrons being emitted from the target and a lower ionisation fraction. In addition, low energy secondary electrons emitted from the target can be captured by both inert ions (Ar and Kr) and the electronegative gas (oxygen) to form neutral recoil atoms and negative ions which can be accelerated away from the target. Both species can arrive at the substrate with appreciable energy and be incorporated into the growing SiO.sub.2 film. Hence, it is believed that secondary electron emission has an important influence on SiO.sub.2 film properties. Sputtering with krypton produces fewer reflected neutral recoil atoms than sputtering with argon. This can be expected to lead to an increase in film density for SiO.sub.2 film sputtered using argon and not oxygen, due to a reduction in the number of inert gas atoms incorporated during film growth. Secondary electron emission is not strongly dependent on momentum transfer of ions to the target, which is consistent with the observation that silicon dioxide film density is not strongly dependent on target power.
(19) FIG. 7 shows the relationship between silicon dioxide film density and the ratio of inert gas flow rate to oxygen flow rate (in sccm) for krypton and argon with 2 kW pulsed DC to the target and with a 80 W RF bias to the wafer. It can be seen that reactive sputtering with Ar/O.sub.2 exhibits a reduction in film density as the argon fraction is increased. This is consistent with increased argon being trapped in the growing SiO.sub.2 film. However, SiO.sub.2 film sputtered with Kr/O.sub.2 show an increase in film density as the Kr fraction is increased.
(20) FIG. 8 shows SiO.sub.2 film stress as a function of platen RF power for SiO.sub.2 films were actually sputtered using Kr/oxygen and Ar/oxygen. A target RF power of 2 kW was used. The SiO.sub.2 film stress does become increasingly compressive in nature with increased platen RF bias, but there is not a direct correlation between SiO.sub.2 film stress and film density. SiO.sub.2 films reactively sputtered using krypton are more compressive with no platen RF bias compared to SiO.sub.2 films sputtered using argon at zero platen RF bias. However, there is little difference in stress between Ar/O.sub.2 reactively sputtered films and the higher density films associated with Kr/O.sub.2 reactive sputtering at the same platen RF powers. Most authors conclude that compressive stress is the result of film bombardment by energetic particles. Inert gas entrapment in sputtered films is often observed, even with ion bombardment, but it is not necessarily causative of stress generation. Indeed, it had been demonstrated in the literature that stress and neutral gas incorporation are independent quantities. Without wishing to be bound by any particular theory or conjecture, we note that the different dependence of silicon dioxide film stress and density on applied RF power to the platen tends to point to neutral gas entrapment playing a significant role in SiO.sub.2 film density.
(21) The properties and quality of the deposited silicon dioxide film have been investigated. FIG. 9 shows electrical breakdown behaviour for SiO.sub.2 films of similar density (ca. 2.35 gcm.sup.−3) sputtered with argon and krypton. Comparable electrical breakdown behaviour is observed for films of similar density. FIG. 10 shows FTIR absorption spectra of SiO.sub.2 films deposited with Kr/O.sub.2 and Ar/O.sub.2. The absorption spectra of the deposited films show peak intensities due to Si—O—Si stretching and bending modes at 1090 cm.sup.−1 and 812 cm.sup.−1 respectively. The position and the shape of the main Si—O vibrational band at 1085 cm.sup.−1 is consistent with the behaviour of a pure stoichiometric silicon dioxide film. There is no significant difference in the spectra obtained for the Ar/O.sub.2 and Kr/O.sub.2 reactively sputtered films, indicating that they are structurally similar. The data shown in FIGS. 9 and 10 indicate that SiO.sub.2 films deposited using Kr/O.sub.2 reactive sputtering of the invention are of high quality and have good performance characteristics.
(22) The Kr/O.sub.2 based reactive sputtering of SiO.sub.2 provided by the present invention has been demonstrated to produce films showing good electrical breakdown performance and comparable wet etch rates to films obtained using prior art Ar/O.sub.2 reactive sputtering techniques. The present invention can produce SiO.sub.2 films having higher densities than the films produced using prior art techniques, but otherwise appear to be structurally indistinguishable from the prior art films. Additionally, significant improvements in film deposition rate can be achieved, with improvements of up to 20% being observed. Improvements are observed over a wide range of process conditions. The skilled reader will be able to optimise the reactive sputtering technique of the invention to suit any given end application or to satisfy desired fabrication criteria by routine experimentation. For example, an RF bias may be applied to the substrate holder in order to achieve an acceptable trade off between film density and deposition rate.
(23) SiO.sub.2 films produced in accordance with the invention have a wide range of possible end applications. The improved properties associated with the invention make the SiO.sub.2 films particularly attractive for optical and acoustic end applications such as with SAW, BAW and TCSAW filters.
