Oops.  When we’re making nano-devices, chaos is usually bad.  I named this spot “The Barrens”.

It’s supposed to be a single straight waveguide (basically, a pipe for light) stretching off into infinity.  Instead, this spot got scratched partway through the fabrication process, leaving behind a chaotic landscape that resembles a desert of mesas, monuments, and mountain ranges. They look similar because the process that created them is similar - something eats away at the landscape (in our case, it’s a high-energy etching plasma) and leaves behind areas that were protected by tougher materials (in this case, the material is nanogunk, one of the most resilient materials known to humankind).

*nanogunk is not actually a scientific term. But gunk is maddeningly persistent on our samples sometimes.

A strange landscape with an even stranger sky.

This is a microscope view of the edge of a smooth chunk of silicon, coated with a thin clear plasticy layer of photoresist.  Just like the colors in a soap bubble, this colorless thin layer produces rainbow colors due to the wave nature of light.  Waves of light reflecting off the top and bottom of the film interfere with each other and add to each other’s brightness, or cancel each other out, depending on the film’s thickness. Since each color of light corresponds to a wave of different size, at different film thicknesses you get different colors of light disappearing from the spectrum as they cancel each other out.

The film above has areas where the surface is relatively smooth (areas with very widely spaced bands of color).  Where the film mounds up near its edges or near the black dust specks, the colors change rapidly along with the film’s thickness. Nonuniform films like these are usually used just to protect a material’s surface between fabrication steps.  To use the film for material processing, the nonuniformities - and with them the rainbow colors - would need to be removed.

The crazy-huge mountains of the nanoworld!  The strange waves and scallops are what is left of the protective mask I used to shield the semiconductor material below from a high-energy etching plasma.  The mask held up to the plasma, although it was probably damaged a bit - and then I damaged it further by a couple of rounds in another plasma, that we use to get rid of polymer buildup.  This is what the sample looked like under the microscope after the last plasma cleaning step - fluffy mountains made of thick residual mask. I didn’t care about the fate of the mask - all I cared was that it had done its job, and that I could tell that the etch worked on the unprotected bits of semiconductor.  It did work, as evidenced by the sharp cliff.  Acetone later washed the mask away, leaving the upper surface of the cliff smooth and clean.

A mini-monument, made of semiconductor laser material.  It looks to me a bit like Devil’s Tower.  It’s much, much smaller, though.  Scaling this little nano-tower (600nm high) to the height of Devil’s Tower (386m high) would be like scaling up an average-sized human (~1.7m) to about 1.1 billion meters tall.  This is, as Wolfram Alpha helpfully informs us, about 3x the distance between the earth and the moon.  Every time I do a calculation like this, it still boggles me how tiny these nanostructures are.

So, what is the nanotower doing here?  This is an image of a plasma-etched wafer of semiconductor, the sort we like to make lasers out of.  I was testing the plasma-etching process, so I was just etching simple structures into this wafer.  I did get etching - the top of the tower is where the top surface of the wafer used to be; the tower itself was formed by a tiny dust particle that happened to land on a surface, shielding a bit of the wafer from the plasma.  Unfortunately, I also got this strange fluffy surface on the wafer, signs of problems with the etching gases, or possibly contamination in the chamber.

This entire view would fit easily inside a single cell.

The “lake” is a crater with glass at the bottom - being an insulator, glass tends to build up charge that deflects the electron beam my microscope fires at the sample, so very little bounces back… it appears dark.  The “snow” is silicon, brighter since it’s more conductive than glass.  The “trees” are some sort of nanofluff that’s not supposed to be there, probably the residue of laser material that I tried to etch away.

To give you a sense of scale, if your average virus were exploring this landscape, it would lumber like a dinosaur, taller than the trees.