It seems silly to me but some of the most basic things in sputter deposition technology keep coming up in discussions with our extended customer base around the world. I cannot possibly address all of these issues here in the Blog section of Plasmaterials.com but maybe it makes sense to touch on at least a few of these topics. Maybe I can make a mini “Blog Series” over the next several months to touch on some of these issues.
For example, “What is the proper partial pressure for the working gas during deposition”? Unfortunately this isn’t quite as straight forward as I would like it to be. So, like always, the answer is, “It depends”.
What it “Depends on”, for the most part, is what power supply configuration the customer is using, i.e. conventional direct current (DC) or radio frequency (rf) and what is the deposition process, i.e. reactive deposition or pure argon working gas? For a conventional DC or AC power supply used in conjunction with “Conductive” elemental or alloyed sputtering targets, the general rule of thumb is to operate the partial pressure of the working gas (typically inert argon) between around 5-7 Milli Torr (Microns). Of course, this is just a starting point. Depending on various iterations associated with the desired properties of the resultant films, it may be necessary to adjust this level upward or downward to achieve the desired results. It is also important to balance the incoming flow rate of the monitored gas with the pumping speed of the system being employed. This is typically done with some form of conductance restriction of the orifice connecting the pumping stack to the chamber. It may be done with a variable orifice plate or an intermediate location (stop) for the linear gate valve (isolation valve). The basic idea is to restrict the pumping speed of the vacuum system during deposition once the ultimate vacuum level has initially been attained so that the amount of working gas being bled into the chamber need not be excessive. Obtaining this equilibrium flow is a bit of a trial and error procedure since every vacuum system has its own unique pumping characteristics. The vacuum level during deposition should be monitored with some form of intermediate vacuum level control gauge such as a capacitance monometer, ion gauge tube, etc. that is accurate in the mid-vacuum level range, say 10-2 to 10-4 Torr. Neither the vacuum gauge nor the gas inlet should be physically located near the throat of the pumping stack as that would produce a false reading due to turbulence in conductance between gas inlet and gas outlet. Both should be located somewhere in the chamber wall close to where the plasma is to be generated but not actually in the plasma itself.
For nonconductive target materials such as oxides, nitrides, carbides, etc. it is not possible to generate a plasma with a conventional AC or DC power supply. For these materials it will be necessary to produce the plasma with either an rf generator with an associated matching network or a midrange frequency pulse generated AC/DC supply such as the Advanced Energy Pinnacle Plus unit. In this case it requires a bit less gas to generate an effective plasma, usually between 2-5 x 10-3 Torr. When using an rf generator, where the frequency is usually fixed by the FCC at 13.56 mega Hertz, it is necessary to match the impedance of the cathode/target assembly being employed with that of the power supply. This is done with a matching network built into the system. This matching network may, or may not, have an auto tune drive network to automatically optimize the forward power but in either case, it must always be tuned in such a way as to minimize the reflected power going into the cathode assembly. Any reflected power turns into secondary electrons which generates heat at the target surface. It is always necessary to minimize the surface heat on any target surface. There is enough heat being built up on the target surface by the momentum transfer between the charged particles (ions) generated in the plasma striking the atoms/molecules on the target surface and breaking the atomic bonds exothermically that you don’t want any additional heat. All surface heat must be dissipated through the target, through the bond layer, through the backing plate and finally through the water cooling channel of the cathode assembly. Heat is a bad actor.
For the conductive, as well as the nonconductive, sputtering target deposition scenario if it is not possible to ignite the plasma by simply applying power slowly to the cathode assembly, and if the system is not equipped with a plasma igniter, it may be necessary to increase the partial pressure of working gas initially to start the plasma. Simply raise the gas flow controller slowly and monitor the vacuum pressure gauge. The plasma should ignite within an order of magnitude or so of pressure increase. Once it ignites, slowly reduce the gas flow back to the initial settings and wait for the pressure to equilibrate back to the same vacuum gauge reading as suggested above. It may be necessary to slightly alter the flow rate to return to the same equilibrium conditions and desired partial pressure. If the plasma fails to ignite by say, 2-7 x 10-2 Torr, return the argon gas flow rate back to the initial setting and explore other possibilities for problematic behavior in the system other than gas pressure.
