PRECIOUS METALS
Q: I want to deposit a thin layer of Pt onto a moving substrate and would like an FFE magnetron. Is this your suggestion?

A: In your circumstances and as the Pt as the layer is very thin I would suggest an narrow magnetron to reduce the amount of precious metal used. The FFE magnetron will use >50% of the target, but the target needs to be at least 125mm or 5" wide. So the target material mass is much higher. Also, to obtain a certain uniformity on a moving substrate, the wider the target the longer the target overhang and the further away the target should be from the substrate.
Also, since the Pt is a precious metal, the reclaim value is very high, so target use is not as important as with say expensive ceramic targets.
The best choice would be a 40mm wide target. For a 200mm long substrate and 2.5% uniformity, the target length can be 325mm and the target to substrate separation should be 25-30mm.
This reduces material costs and the amount of material on shields etc.

CARBON DEPOSITION
Q: What is the recommended arrangement to deposit thick carbon layers?

A: As the Carbon is thick, it will take time to deposit that due to a low sputter yield of the material. A weaker than standard magnetic field will increase the target voltage (600-800 volts) and increase rates to some degree for carbon. The wider the target, the more power that can be applied and hence the more material produced. Graphite targets can be clamped or bonded, a clamped target will take less power than a bonded target, but bonding graphite is not easy.

HIGHER TARGET EROSION AT ENDS - TURN AROUND

Q: For all magnetrons I commonly observe areas near the end of the target corners that erode faster. What is the reason for this and is there a solution?

A: There are several reasons why this can occur. If you look at the EXB drift pdf for download on the right column, it explains some of the factors. Once the pumping and gas problems are eliminated, the physics of the electron movment in the plasma will always create a slightly varied erosion rate at opposite corners. It cannot be elimiated completely but it can be minimized by careful magnetic design. 3D modelling of the magnetic system and variation of the strength of the magnetics at the turn-around will minimize the problem. Magnetrons in close proximity to each other also enhance the cross-corner effect as can the use of AC dual magnetron arrangements.

MAXIMUM POWER LEVELS ON THE TARGET

Q: What principles determine the maximum power I can apply to my target materials?

A: The main aspect to what power the target can withstand is down to the material itself and also the means of heat removal from the target or backing plate. For good thermal conductors such as Cu and Al which also have good mechanical properties, max power densities can be high (20-50 W/cm²).
If the target is bonded to a backing plate (as is the case for more fragile or soft materials), then the bond material maybe the point of weakness due to a low melting point (in the case of indium). There are higher melting-point bonds, but for the family of bonded targets the maximum power density is generally 10-20 W/cm²).
Thermal shock is also a problem. Brittle target materials will crack if the power is too high or the thermal heating and cooling cycle is too agressive. For these materials generally the maximum is 5-15 W/cm².
For low melting point target materials, the heat removal will be the most important, and special high water flow magnetrons are recommended. Power levels for such materials are 5-20 w/cm² depending upon the cooling mode, efficiency and the material.

Once the power density that the material can withstand has been assessed its important to always bear in mind the coating quality required. Even though high powers maybe possible, the demands of the product may limit the effective power that can applied.

PREVENTING COATING DELAMINATION FROM THE ANODE

Q: How can I prevent the deposited material on the front of the anode from flaking away and causing an arc or electrical short?

A: Once the material gets on the anode its important to prevent it from coming away until a cleaning process is used. We have designed our anodes to be below the target level so that the amount of re-deposit on it is much reduced but the problem still needs to be managed. Usually, by alumina shot blasting of the surface the roughness can be increased to a level that should prevent flaking. Another thing you could do is to connect up the water cooling on the earthed magnetron rear plate, that will remove some of the heat from the anode and reduce the thermal cycling.

Once you have blasted your own anodes give them a good clean. This could be an ultrasonic, but a high pressure steam/water type blaster is good (can use a car jet washer at a garage if you don't have such a thing). You can also 'scrub' them well in hot soapy water as well. Blow the excess water off with compressed air and let them dry - don't use paper towels as it will transfer to the anode.

That will remove the 'dust' and help adhesion of the coating to the parts. Any Gencoa replacement or new anodes are already blasted, cleaned and vacuum packed.

For very highly stressed material on the anode such as very thick silicon dioxide it maybe required to roughen the surface even more by means of a flame-sprayed aluminium layer. This is standard practise in the semiconductor industry where its critical to prevent defects from the chamber fixtures.

MAXIMUM RF POWER

Q: How much RF power can I apply safely to a 6" magnetron?

A: The magnetron itself can take 3-5kW of RF power in terms of its coolant capacity. The limiting factor however will be the target material (cracking at high powers) and the power feed mode. For RF the cable (RG 393) will take 400-500w with strong dielectrics such as Al2O3, SiO2. HIgher than this power can impose too much heat on the cable, so it will be a safety hazard if touched. We can fit two RG 393 cables for 1 kW RF. For more than 1 kW we need to make a direct RF connection from the matching unit - and ideally mount the matching unit on the rear of the cathode or mounting tube.
For metal targets with RF power 1-2 kW per cable is possible as the capacitance of the targets are much lower than the dielectrics. Also, as a general principle for RF power transmission via cables, keep the cables as short as possible.

SiO2 DEPOSITION FROM DOUBLE MAGNETRONS

Q: Why is it recommended to use a double cathode for reactive deposition of SiO2?

A: The reason a double cathode is best is a result of the anode becoming
covered with SiO2. This is an insulating layer, so after a period of
time the process becomes unstable. Double magnetrons use AC power that
make one target positive as the other is negative, hence there is always
an anode. The plasma is like an electrical circuit the negative charge
on the cathode needs a positive return path to complete the circuit -
i.e. the earth.

If you look at this web page and download the files in the right column
it has more information.
http://www.gencoa.com/1,31-Process-tips.htm

A single cathode can be used as long as they have either a 'hidden'
anode or the metallize the anode after a period of time - i.e. run
without oxygen. The TiO2 process is less of a problem than the SiO2.

 SETTING THE MAGETRON TO SUBSTRATE SEPARATION

Q. What is the operating standard range of distances from target to substrate?

A. As a general rule, the target to substrate separation should be more than the target width / 2. So if its a 150mm wide target 80mm t-s is good. Any closer and the substrate will be in the magnetron plasma and could become charged electrically leading to substrate heating, arcs or thin film device damage. Further away and the uniformity drop-off starts at the ends starts to increase. Sometimes the distance is greater to reduce heating effects on temperature sensitive materials.