Sputtering Rates
To use these charts, locate the material for which known conditions are available. Then multiply the rate by the relative factors to arrive at the estimated rate for the new material. For example, with previous data showing 3.5A/s aluminum at l00W, then titanium at similar conditions will generate approximately (0.53/1.00) • 3.5 Å/s ≅ 2 Å/s.
The rates in this table are calculated based on a 500V cathode potential. As the power is increased greater than two times the original rate, the relative rate will drop slightly (up to 10%). For example, aluminum at 250W.
Al250W = 0.9 • AI100W • (P1/P0)
0.9 • 3.5 Å/s • (250/100) ≅ 7.4 Å/s
The rates in the ceramics table assume the use of an RF power supply and account for the partial duty cycle of the RF generator as compared to a DC supply. A pulsed DC supply will yield slightly higher effective rates.
The magnetic materials table shows the rate for DC operation with a new target. As the magnetic target erodes, the influence of the remaining material on the magnetic confinement field will change, leading to variations in sputter rate, operation voltage, and ignition pressure.
This information is for general planning purposes only. The Kurt J. Lesker Company makes no guarantees of the correctness of these numbers in your process. Contact the Kurt J. Lesker Company for specific assistance in setting up your process.
| NON-MAGNETIC MATERIALS* | ||
| Material | Name | Rate |
| Ag | Silver | 2.88 |
| Al | Aluminum | 1.00 |
| Au | Gold | 1.74 |
| Be | Beryllium | 0.21 |
| C | Carbon | 0.23 |
| Cu | Copper | 1.42 |
| GaAs | Gallium Arsenide {100} | 1.03 |
| GaAs | Gallium Arsenide {110} | 1.03 |
| Ge | Germanium | 1.50 |
| Mo | Molybdenum | 0.66 |
| Nb | Niobium | 0.76 |
| Pd | Palladium | 1.77 |
| Pt | Platinum | 1.00 |
| Re | Rhenium | 0.84 |
| Rh | Rhodium | 1.16 |
| Ru | Ruthenium | 0.98 |
| Si | Silicon | 0.60 |
| Sm | Samarium | 1.74 |
| Ta | Tantalum | 0.67 |
| Th | Thorium | 1.31 |
| Ti | Titanium | 0.53 |
| V | Vanadium | 0.50 |
| W | Tungsten | 0.57 |
| Y | Yttrium | 1.53 |
| Zr | Zirconium | 0.88 |
* All rates in this table are relative to aluminum.
| OXIDES AND CERAMICS | ||
| Material | Name | Rate |
| Al2O3 | Alumina | 0.05 |
| SiC | Silicon Carbide | 0.22 |
| SiO2 | Silicon Dioxide | 0.21 |
| Tac | Tantalum Carbide | 0.09 |
| Ta2O5 | Tantalum Pentoxide | 0.39 |
| MAGNETIC MATERIALS | |||
| Material | Name | Mag Moment | Rate |
| Co | Cobalt | Low | 0.73 |
| Cr | Chromium | Med | 0.87 |
| Fe | Iron | High | 0.57 |
| Mn | Manganese | Med | 0.14 |
| Ni | Nickel | Low | 0.86 |
| Ni80Fe20 | Permalloy | High | 0.80 |
There are a few ways that you can increase/ maximize the sputtering rate of materials;
1. Increase power: While each material will be limited in their max power relative to their material properties, the cooling efficiency will allow you to operate the target at the highest possible power density. The first thing you should do is directly cool the target material by utilizing either a bolt-on style or bonded target configuration. This in addition to the aid of a conductive paste or epoxy will maximize the thermal conductivity and allow you to increase the power density to the maximum level attainable by the target material.
2. Decrease source-substrate distance: The closer the target to the substrate, the higher the sputtering rate will be. Generally, the plasma will be contained within 2" above the target surface. Many sputtering applications utilize a 3"-4" source-substrate distance. Assuming a 4" source-substrate distance, the sputtering rate will fall off by approximately 25% for every inch beyond 4". However, the rate will typically increase by approximately 35% for every inch closer you go from 4" away.
3. Lower operating pressures: In sputtering, the more gas in the chamber, the more atom and ion collisions there will be. These collisions will reduce the rate at which material atoms eject from the target surface and deposit onto the substrate. By reducing the operating gas flow, these collisions will be reduced and will have a positive impact on the ultimate sputtering rates that can be achieved.
4. Increase the number of magnetrons in the chamber: Rates will scale linearly by the number of magnetrons that are added to your application. In production applications with specific yield requirements, once the power and source-substrate parameters have been fully maximized, increasing the number of magnetrons is a parameter that can be utilized to enhance sputtering rates.
Uniformity
Enhancing uniformity in sputtering applications involves many variables. Some of these variables are impacted by the magnetron itself, but many are related to the system/ chamber design and flow dynamics, which in some cases cannot be controlled. However, there are a number of techniques that can be done to enhance uniformity in your application. The following list of examples provides some suggestions on what parameters can be adjusted to have a positive impact on uniformity. It is important to first recognize the fact that there are two significantly different deposition configurations that will yield much different suggestions for uniformity enhancement. As a result, we will cover these independently.
Static Substrate Uniformity
Figure 2
Figure 1
When sputtering a static substrate the following parameters will impact the overall coating uniformity;
1. Magnetron to substrate orientation: The magnetron and substrate should be centered on their axis for optimal uniformity.
2. Magnetron Size: The target should be larger than the substrate for optimal uniformity. The typical coating profile will fall off on the edges and be most uniform in the center (as figure 1).
The more overhang you have on the substrate, the more uniform the coating will be.
3. source-substrate distance: If magnetron to substrate overhang is non-existent or limited due to existing chamber design or equipment, increasing the source-substrate distance will help improve uniformity. The further away you get, the more collisions between the argon ions, electrons, and material atoms, which creates randomization on the sputtered film depositing on the substrate and ultimately better uniformity. However, the drawback to this is that the further away you are from the substrate, the lower the sputtering rates will be.
4. Masking: Masking is a technique that can be used to enhance uniformity by blocking or preventing material on certain areas of the target from depositing onto the substrate. For example, it is typical for material build-up to fall off at the edges of the target due to the magnetic field profile, active erosion zone location, and resultant flux profile. By inserting masking at the center portion of the target, you can ultimately flatten out the erosion profile (see figure 2).
Rotating (Single Axis) Substrate Uniformity
When coating a rotating substrate, the following techniques can be used to enhance the uniformity;
1. Off-axis or Confocal magnetron to substrate orientation: The main advantage of rotating the substrate is that you can use a much smaller magnetron to achieve optimal uniformity by off-setting the center line of the target to the substrate. Utilizing a single axis rotation of the substrate is required for the following techniques;
-
A. Off-set: Adjusting the distance from the center of the target to the center of the substrate to an off-set focused on the radius will allow you to enhance uniformity across the entire substrate surface while focusing on only half or partial substrate area.
- i. source-substrate distance: Adjusting the source-substrate distance will also help dial in the uniformity.
-
B. Confocal: In a confocal orientation, the magnetron is angled toward the substrate radius. The ideal parameters for confocal sputtering are as follows;
- i. 30 degree off-axis inclination to the center of the substrate radius
- ii. 4" source-substrate distance
Note: Adjustments to the angle and source-substrate distance may be required based on chamber design and flow dynamics.
In a confocal orientation, the following uniformities can be achieved with 3" magnetrons;
| Substrate Size | Uniformity |
| 4" OD | +/- 1-2% |
| 6" OD | +/- 3-5% |
| 8" OD | +/- 7% |
In confocal or off-axis sputtering, multiple magnetrons can be used for co-sputtering of multiple layers of variable materials and or increasing sputtering rates of the same target material.
Typically, when sputtering a 6" substrate wafer, it is possible to mount up to (4) four sputtering cathodes at a 4" source-substrate distance. It is critical to have the ability to adjust source-substrate distance and the angle of the sputtering cathode to have maximum ability to dial in uniformity.
In off-axis sputtering, having the ability adjust the off-set to the substrate, angle of the source, and source-substrate distance, are all critical parameters in dialing in the uniformity. Below are some examples of how adjusting these variables make a significant impact on both uniformity and rate;
| Configuration 1 |
| Substrate: 6" |
| Off-Set: 3" |
| (Center of target - Center of Substrate) |
| source-substrate distance: 4" |
| Head Angle: 0 deg. |
| Target Material: Aluminum |
| Uniformity: +/- 4.7% |
| Rate: 4.5 Angstroms/ sec. |
| Configuration 2 |
| Substrate: 6" |
| Off-set: 3" |
| (Center of target - Center of Substrate) |
| source-substrate distance: 4" |
| Head Angle: -5 deg. |
| Target Material: Aluminum |
| Uniformity: +/- 2.2% |
| Rate: 3.5 Angstroms/ sec. |
| Configuration 3 |
| Substrate: 6" |
| Off-set: 3.5" |
| (Center of target - Center of Substrate) |
| source-substrate distance: 3" |
| Head Angle: -5 deg. |
| Target Material: Aluminum |
| Uniformity: +/- 3.5% |
| Rate: 5.5 Angstroms/ sec. |
| Configuration 4 |
| Substrate: 6" |
| Off-Set: 3.5" |
| (Center of target - Center of Substrate) |
| source-substrate distance: 4" |
| Head Angle: 0 deg. |
| Target Material: Aluminum |
| Uniformity: +/- 1.3% |
| Rate: 3.6 Angstroms/ sec. |
| Configuration 5 |
| Substrate: 6" |
| Off-Set: 4.25" |
| (Center of target - Center of Substrate) |
| source-substrate distance: 3" |
| Head Angle: 15 deg. |
| Target Material: Aluminum |
| Uniformity: +/- 2.5% |
| Rate: 2.8 Angstroms/ sec. |
| Configuration 6 |
| Substrate: 6" |
| Off-Set: 4" |
| (Center of target - Center of Substrate) |
| source-substrate distance: 6" |
| Head Angle: -5 deg. |
| Target Material: Aluminum |
| Uniformity: +/- 1.9% |
| Rate: 0.9 Angstroms/ sec. |
Note: All configurations above assume single axis substrate rotation.
Increasing Sputter Rates
When sputtering dielectric targets using RF power, it is quite possible for the maximum deposition rate on the substrate to be less than 0.1 Å/sec. That is, depositing a film 100nm thick may take over 2.5 hours. It is no surprise, therefore, that we are frequently asked, "How can I increase the sputter rate?"
Actually, what the questioner wants is to increase the deposition rate, but we're not about to argue semantics with a frustrated researcher.
(But to segue into semantics for a moment: we will use sputters the adjectival form, as in sputter yield, sputter rate, sputter gun, rather than sputtering yield etc.)
In this issue we review ways to increase deposition rates and look at conditions where maximizing one parameter inadvertently affects something else.
While the substrates can be static or rotating, these suggestions apply only to circular sputter guns with flat disc targets and stationary magnet assemblies. Sputter guns with targets of other shapes and configurations, moving magnet assemblies, and linear sputter guns, have their own performance attributes that are not directly addressed here.
The Ar+ ions have sufficient energy to eject target atoms into gas phase
Sputter sources in co-deposition 'action'
Sputter Yield
First, we must understand that each material has its own characteristic sputter yield - the number of atoms (or molecules) leaving the target for each ion that hits it. The sputter yield value depends on: the material; the mass of the incoming ion; the voltage through which the ion is accelerated; and its angle of incidence on the target.
For Ar+ ions striking a target at 45° through a potential of 500eV, the sputter yields of most elements are between 1 - 10, roughly.
Materials that are chemical compounds such as oxides can have much lower sputter yields! For example, Maissel and Glang's book Handbook of Thin Film Technology quotes the sputter yield for SiO2 as 0.13 and Al2O3 as 0.04.
Extending the concept of sputter yield, we will later refer to a material's sputter rate, which is its sputter yield multiplied by the ion current to the target.
Throw Distance Changes
A representation of the effect of deposition rate when the Throw Distance is halved
Reducing the target-to-substrate distance (often called throw distance) is a simple, direct way to increase deposition rate. To fully understand this effect, the angular distribution of sputtered particles must be known. Regrettably, this is a complex subject since material is ejected from a circular 'trench' around the target and terms like over-cosine and under-cosine are used in the literature to describe a sputtered material's flux distribution.
For these notes, however, it is sufficient to understand that the sputtered particles' arrival rate (per unit area of substrate) varies as the inverse square of the throw distance. That is, halving the throw distance quadruples the material's arrival rate at the substrate and the film's thickness grows at 4x the previous rate!
However, it is important to consider the shorter throw distance's affect on the film's (thickness) uniformity. If, for example, material leaves the target in roughly a cosine distribution pattern, then the larger the throw distance, the higher the number of thermalizing collisions between sputtered atoms and sputter gas atoms. These collisions tend to 'flatten out' the cosine distribution making the deposition more uniform across the substrate. Since a shorter throw distance means fewer collisions, film uniformity at shorter distances may be worse.
In addition, at shorter throw distances substrates may see: higher energy sputter particles; more stray electrons; more plasma ions and 'hot' neutrals; and higher thermal radiation heat transfer from the plasma and target surface. So the adverse effects of a shorter throw distance include:
- Excessive substrate outgassing
- Increase in compressive stress in the growing film
- Films beneath the present one damaged by electron bombardment
- Substrate melting!
However, shorter throw distances (and, therefore, higher substrate temperatures) can have beneficial effects too:
- Films may grow as successive monolayers (called Frank-van der Merwe growth, a frequently desirable nucleation mode)
- The film's tensile stress may be reduced
- Film adhesion may improve due to the higher energy of arriving atoms
- Films may be 'densified' by bombardment with higher energy plasma ions and 'hot' neutral
Increasing Power
Doubling the power applied to the target roughly doubles the sputter rate and this always appears to be the 'easy option' when faced with low deposition rates.
Unfortunately, arbitrarily increasing power has many adverse effects.All power applied to the gun must dissipate somewhere in the system. It is claimed that roughly 75% ends up heating the gun's cooling water. That is, 75% of the total power dumped into the target's front face must transfer through the target to reach the water! Clearly, the target's thermal conductivity, thermal coefficient of expansion, mechanical strength characteristics, and melting point, are critical issues.
- Thermal conductivity helps determine the temperature difference between the target's front and rear faces. The larger that difference the higher the thermal stress in the material
- Thermal coefficient of expansion partly determines the mechanical stresses resulting from the thermal stress
- Mechanical strength determines how the mechanical stresses are dissipated (usually by bowing, warping, chipping, or cracking
- Melting point (obviously) determines if the target will melt at the temperature generated by the applied power level - and a molten target can ruin a sputter gun
Another major concern is the 'thermal conductance' of the interface between the target's rear face and the sputter gun's cooling well. Results tabulated in A Heat Transfer Textbook by Lienhard & Lienhard indicate the thermal conductance between two lightly clamped, flat metal surfaces is (a) not very high, and (b) depends significantly on air between the surfaces.
Evacuate that interface - that is, put the sputter gun under operating conditions - and the thermal conductance of the interface between the target and the cooling well may drop to 1/20th to 1/50th of its 'with air' value.
Some target materials are so fragile they crack no matter what sputter power is used. Bonding such materials to copper backing plates may allow their continued use even though cracked. However, if pieces chip off or the cracks become wide enough to expose bonding agent or copper backing plate, the target must be replaced.
Too high sputter power is the most common cause of target and sputter gun damage. Given the target/interface thermal limitations, such damage can be reduced/eliminated by using an appropriate maximum power (see Maximum Power Levels). However, 'appropriate' often equates to 'low' and low power means low deposition rates.
One final point about applying power to a target. Once the appropriate power has been established for a given target/gun, never switch on and immediate increase power to that value! Always increase power slowly to its maximum value through a series of ramps and soaks.
Kurt J. Lesker Company RF power supply for non-conducting targets
Sputter Gas Pressure
Lowering the sputter gas pressure causes a modest increase in deposition rate by a two-fold mechanism:
- Sputtered atoms leaving the target will undergo fewer thermalizing collisions.They are less likely to scatter 'sideways' and a larger percentage will continue to the substrate, slightly increasing the deposition rates
- In power control mode, using either RF or DC power, the plasma-to-target voltage will increase slightly. Ions bombarding the target will, therefore, have a higher energy which slightly increases the sputter yield and consequently the sputter rate
One potential side-effect of lowering the gas pressure is a change in film uniformity. Whether it improves or worsens is typically not predictable because there are many factors involved. But one obvious aspect is a reduction in the number of thermalizing collisions.
An adverse effect of lower gas pressure/higher plasma-to-target voltage combination is the greater likelihood of arcs occurring near the target.
Increasing Target Size
As a method of increasing deposition rate, this option is not easily implemented and is expensive since it requires a new sputter gun, sufficient room to install it in the chamber, and possibly a larger power supply.
For a given power density (see Power & Power Density), the larger the target diameter the higher the sputter rate. The explanation is simple. A larger target diameter means a larger sputter trench area and, for a given power density, increased trench area means increased sputter rate.
Number of Guns
The majority of R&D deposition systems have more than one sputter gun installed. Typically, the user installs different target materials in each gun. However, putting the same target material on two or more guns and operating them simultaneously can double, triple, etc. the sputter rate and resulting deposition rate.
The drawback is, many multi-gun systems were not built for co-deposition work and have just one power supply. Buying additional supplies for simultaneous operation may make this option expensive.
Conclusion
Yes, there are ways to increase deposition rates. Unfortunately the easy winding-up-the-power option, if misused, at best leaves your targets looking a little sad. At worst, your sputter gun splutters to a stop, water leaks into the chamber, or the power supply fries. No, I jest! At worst, all three happen simultaneously.
Cerium oxide target bonded to copper backing plate but sputtered at a power that melted the indium bond and cracked the target
3" diameter Indium tin oxide target sputtered at 1000W (roughly 7x the maximum recommended power)
Aluminum doped zinc oxide target given the 'tough love' treatment of inadequate cooling (at the 10:00 o'clock position) and excessive power
As always, if you have questions or comments email techinfo@lesker.com and they will be forwarded to the author.