Patterned layer epitaxy

From apm
Jump to: navigation, search
This article is a stub. It needs to be expanded.

Hydrogen Depassivation Lithography (HDL)

Removing material

In "Hydrogen Depassivation Lithography" (HDL) hydrogen atoms that seal (aka passivate) a surface are selectively removed / broken off from a (flat) workpiece surface by "injecting" electrons into the bonds between the hydrogen atoms and the underlying surface. (Full electron shell H- ions repulsed away ?). The electrons are injected via a (briefly cranked up) tunneling current that originates from the "needle like" tip of a scanning tunneling microscope.

In high performance HDL the stripping off of hydrogen atoms (aka depassivation) can be done one at a time with atomic precision. This allows the creation of precise digital patterns.

The depassivated surface atoms have open broken bonds (dangling bonds) sticking out. Such dangling bonds (aka chemical radicals) are extremely reactive. They "itch" to react and bond with any colliding molecule or atom that is not extremely uncreative (which is all of air except argon and other noble gasses). Thus HDL needs to be performed under a very good vacuum (today in big expensive UHV systems only found in dedicated labs).

Adding material

By injecting (in a controlled manner) pure gases of a single species of molecule these molecules collide, stick and bind to the previously depassivated pattern. This is a statistical thermodynamic gas phase process. Molecules are added at random spots inside the depassivated area. Once bound, the molecules repassivate the spots they landed on. Thus after a while exactly one mono-layer is added (exponential saturation). This way atomic precision gets restored again despite having a statistical process involved.

HDL with atomic precision has been demonstrated experimentally on silicon surfaces. Here the molecule species added are silane molecules (the silicon analogons to hydrocarbons) Other molecule species (e.g. phospine) can be used to (sparsely) include these other elements; (atomically precise doping). (Side-note: these compounds are usually highly toxic - a lab safety concern)


High performance HDL with atomic precision (one specifically selected hydrogen atom at a time) is necessarily rather slow. It is possible to crank up the STM tunneling current even further and have a more crude strewing depassivation tool. This can be used to quickly carve out large areas that shall be fully depassivated. The sharp corners need to be done more slowly and carefully.

With HDL there is no way to hydrogen-REpassivate an accidentally depassivated spot. If the hydrogen atoms gone thats it, theres no way back. (One can repassivate all depassivated areas at once with hydrogen).

Limitations of HDL (state 2017):

  • no overhangs
  • no structures with strained bonds
  • only demonstrated with silicon surfaces (?)
  • still a quite high error rate
  • almost 2D -- not very high structures (yet) -- a general STM limitation (likely tip sample crash due to slow feedback loop)

(TODO: Is there work on trying HDL on diamond instead of silicon (or other surfaces entirely)?)

Related: Site activation assembly in foldamer based systems


External links