Developing your ion source

Various ions interacting with surface gives rise to a number of interesting phenomena. There are two aspects of the ion-material reaction, chemical and physical. I will only focus on the physical aspect. For the rest, you can assume the ion to be of Argon (guess why?). There are many possibilities here, the ion can remove the atoms from the surface (sputtering), it can scatter back, think of RBS. It can be implanted inside the surface or get adsorbed. What will happen mainly depend on the energy of the ion, think of a stone thrown on the window, it will shatter the glass but a bullet would punch a whole. Typically sputtering starts at hundreds of eV energy, at hundreds of keV it gets implanted deep inside and for lower energy of sub-hundred eV of energy, a number of phenomena specific to the surface properties start to appear. The question here is how do I make an ion source of given energy?

Okay, I think the first step is obvious, you need to ionize the atom of interest, say Argon. There are a number of ways this can be done, all have its pros and cons. It also depends on the form of the source material, for gas, it is easier for solid materials, there are two techniques. One is to vaporize it, okay that is cheating, it's then the same as the gas, but you get the idea. However, it is not always possible easily vapourize some material, especially without vaporizing nearby components. In that case, sputtering. In essence, take a solid material and bombard it with inert ions (in most cases Argon, unless you have a strong reason not to, we will discuss this later in the article.). You don't need a dedicated ion source for this, make a DC discharge of Argon and use the material of interest as the cathode of the discharge, this puts the ions of the cathode materials in the plasma, there you have it. If you happen to have an insulating material as cathode then use RF discharge instead of DC. This works because plasma forms a sheath of hundreds of volts near cathode and ion of the plasma strikes cathode with similar energy, which causes sputtering and sputtered atoms can get ionized in the plasma. Okay, but for gas source materials we can implement a DC discharge or hot filament discharge or inductively coupled RF discharge. I will briefly state the give and take of each. Prefer hot filament over DC discharge as it can operate at much lower pressure then DC discharge, moreover it can operate in a forced mode, which means even if there is not enough space or atoms to form a plasma. A forced mode can, however, lead to increased energy spread, oops I never told you about energy spread, you can unread it! and we will come to it later. When higher lifetime and low downtime is intended, prefer the inductively coupled plasma as it does not have any working electrode in contact with the plasma which leads to increased lifetime, low wear, low contamination, and low energy spread (well, again!).

So, we have ions of our desired material, now we need to accelerate it required energy. Ions are positively charged (well it can be negative as well, but that story is for some other time) so we need to apply some negative potential to say a grid. Before it gates two complicated let us form an idea of potential hills and valleys. Positive ions run down the potential hill from top to the valley.

We need to put our ions on the top of the hill so that it runs down to our ground potential. The energy they will gain is only the difference between the origin of the ion on the hill and our target at the ground, the shape of even the local curvature (hills and valleys) doesn't matter. So we have a basic working ion source, if you have ion source with a given energy, you can put your target at various height in the potential hill and can get different energy.

Any practical ion source has a distribution of energy around the desired (set) energy. This distribution arises because of ions being generated at a different height on the potential hill. But we kept our ion generating plasma at the top of the hill, right? How can then ion generate at different hight? It is all about relative potential differences. A plasma generating discharge device also harbor a small potential, usually in the range of 30 to 600 V. If plasma is well formed, these potentials are confined very close to the electrodes and all ions are generated at the top of the hill. The discharge potentials in such cases only add or subtract a constant shift and it is not a spread. When plasma is not well-formed, a forced mode plasma, the potential would distribute in the entire region of the plasma producing space, this will give rise to energy spread. Clearly, such energy spread cant be higher than the total discharge voltage used to generate the plasma. In most cases where the ions are accelerated to energies in the range of 100's of keV, this small energy spread of 100's of volts can be ignored. In other cases, where it is crucial, it can be controlled by having well-formed and defined plasma volume. Hot filament discharge requires less voltage to for the discharge, 30 to 100 V. Whereas, DC discharge require 300 to 600 V. Easy to see why hot filament is better for lower energy spread. The inductively coupled discharge is the king here, with no active electrode its voltages in the plasma are limited to few tens of V.

Some applications would require exceptional purity of the beam, normally in any operational ion source which deals with various materials, there always is the entire periodic table is present in the beam. And when you use solid source materials in vapor form or sputtering cathode, there always is a working gas present. So we need to isolate the ions of interest from this bunch of different elements present in the ion beam. You can check for yourself that no combination of the electric field can separate ions of different mass having the same energy. This is because electric field deflects lighter ions more, but they are also fast (for the same energy, answer why?), these two effects cancel each other. What about the magnetic field? Fortunately, magnetic field deflects charged particle and though the previous effect appears, it does not exactly cancel (an odd square root of kinetic energy comes to rescue). There are two standard techniques and dipole magnet and Wien filter. There are a number of references which give technical details of both.

I have left more technical details such as the ion optics (think of ions as light and its optics), ion extraction and more involved space charge effects. You can search for all of this and Child-Langmuir law (space charge limited emission), and density (temperature) limited emission, sheath, etc. These factors are extremely crucial in designing the plasma discharge device, ion extraction system, and transport tube. For example, the gaps and size of all optical elements including the extraction system will depend on the requirement of final ion beam energy and current, remember both quantities are required, energy or current alone don't say anything. For example, it may be easy to generate 100 mA ion beam at100 GeV of energy but the same is nearly impossible for 100 eV energy. The beam spot is assumed to be of 1 square cm, which is another very crucial parameter. As you would guess higher current on larger surface area is easier to control. Always think in terms of current density rather than current.

The work-flow of the design is not straight forward, first you need to decide what energy and current density you need, then go on to decide the mass separation techniques, dipole magnet or Wien filter, RFQ, TOF etc,. This mass separator would govern your design of the source and the energy, after the mass separation you may accelerate or decelerate to your final energy. For very low (10-100 eV) energy and moderate current (1 to 100 uA) space charge effect becomes significant and you need to decelerate to this final energy right before your target.

If you think I should have included certain things in more details, please leave comments below.