Etching a Photograph Into a Silicon Wafer

As part of my research as a graduate engineering student, I work in Cornell's NanoScale Science & Technology Facility (CNF). In Duffield Hall’s cleanroom, I use silicon wafers to create devices with micron (one millionth of a meter) size features.

My parents' 25th wedding anniversary was coming up and I wanted to get them something special. Being a poor grad student, the traditional 25th anniversary gift of silver was definitely not an option; however, while working in the cleanroom, I hit upon an idea for something one of a kind.

Having gone through the photolithographic process hundreds of times, I wondered if I could etch a photograph of my parents into a silicon wafer. I thought such a commemorative gift would be fitting since they have always encouraged my educational pursuits, whether it was taking me to the library when I was a child, buying me a small microscope from Toys “R” Us, or letting me sprawl some new Lego creation over the coffee table. I searched to see whether anyone had done such a thing before (and more importantly how) but I came up empty. So I just had to try it and see if it would work.

The first step was to choose the photograph I wanted to etch into the wafer. I chose a photograph taken at their first Christmas together, just weeks after they had gotten married. To make it extra adorable, I added a baby photo of myself underneath.

Photolithography is a binary process in which the photomask, or pattern-transferring stencil, is written to be transparent or opaque to ultraviolet (UV) light, UV light is either passed or blocked by the photomask, photosensitive material called photoresist is either exposed to or shielded from UV light, and underlying silicon is either etched away or protected. This meant there was no way to transfer a color (analog) photograph into silicon and required I convert the scanned color photographs into monochrome (1-bit) bitmap images where each pixel is either black or white (no in-between or grayscale). I adjusted the brightness and contrast levels of the color image beforehand to ensure the final monochrome image possessed the optimal level of detail.

I then used a software package called LinkCAD to convert the bitmap to GDSII file format, the industry standard for integrated circuit layout. I then used L-Edit CAD software to resize and adjust the image until it would fit on a standard 4 inch wafer.

I uploaded the final CAD file to the machine in the cleanroom that creates photomasks.

Mask Writer: Heidelberg Instruments DWL 2000 Laser Pattern Generator and Direct Writer

I then applied photoresist to a fresh silicon wafer and exposed the wafer to a short burst of intense UV light through the photomask.

Wafer Preparation:

Manual Resist Spinner:

Primer: P-20: 3k rpm, 1 k/s, 30 s

Photoresist: S1827: 3k rpm, 1 k/s, 30 s

Hotplate:

Bake @ 115 C for 1 min

Exposure:

EV620 Contact Aligner:

Mode: Hard Contact

Time: 5 s

After chemical development (similar to darkroom photograph development), the transferred image appeared.

Wafer Development:

Hamatech-Steag Wafer Processor:

MIF 726 Developer @ 2 min

To etch the transferred image into the wafer required a process that would maintain image features without distorting them. Most etching processes suffer from undercutting in which the wafer is etched isotropically in all directions (straight down and to the sides), resulting in distortion of the transferred pattern. Luckily, the CNF has a deep reactive ion etcher (DRIE) which anisotropically etches straight down into the silicon by repeatedly depositing a protective Teflon-like material on the feature sidewalls and then bombarding the bottom of the features with energetic ions to slowly eat down into the silicon. I had to repeat this cyclical process of passivating and etching 500 times to etch 200 microns (two-fifths of a millimeter) into the silicon wafer. This took several hours, requiring me to use the machine in the middle of the night when no one else was around.

When this process was finished, I stripped the remaining photoresist off the wafer using a hot acid bath. I rinsed the wafer and dried it with a nitrogen gun and carefully placed it in a protective case.

I finished just in time and presented my parents (Ray and Cindy) with the wafer on their anniversary. At first they didn’t know what to make of it but after explaining how I made it and walking them through the creation process and, they loved it!