A creative way of demonstrating the capabilities of microfluidic biosensors, captured on a Raspberry Pi HQ Camera, caught the eye of Rosie Hattersley. This #MagPiMonday, we meet a thoroughly modern Mona Lisa.
Big, brash ideas and enormous displays gain attention, whereas tiny but significant products can be far harder to convey. After helping develop a crucial microfluidic-based biosensor system for use in microscopic research at IBM Zurich, microsystem engineer, and keen YouTuber, Yuksel Temiz and Polytechnique Montréal researchers hit on the idea of using an iconic artwork to show off its possibilities.
The Mona Lisa Fluid Painting project showcases a new microfluidic architecture that allows local and highly parallel dissolution of biochemical reagents in volumes as low as two nanolitres per spot. The image was created from just five microlitres of water ‘painted’ on a 2 cm square microfluidic chip. The microfluidic architecture has potential applications in healthcare diagnostics and life sciences, “opening up new possibilities in studying kinetics and interactions of chemicals in nanolitre volumes and in high-throughput.”
Yuksel spent nearly ten years working at IBM’s labs in Zurich, and says, “In our research, we always needed compact optical, electrical, and mechanical peripherals to read out and control our biosensors. For me, Raspberry Pi has always been the easiest way to implement a minimum viable product, especially for imaging applications.”
When it came to showing off the incredibly tiny droplets, Raspberry Pi soon became the obvious route to take, enabling them to create an impactful showcase to go alongside a press release about their work without investing too much time and money. “Raspberry Pi allowed me to test and optimise the setup very quickly with minimal investment,” enthuses Yuksel.
“Coding is just a hobby for me. But the thing I like about Raspberry Pi is you don’t need to be a professional coder to build functional systems.”
Camera conundrum
More than 10,000 droplets of fluorescent dye in eight different shades were spotted and dried inside microfluidic channels to generate the Mona Lisa image. Once the channels for the droplets had been engraved, the top of the chip was sealed and liquid applied – a process that took just under two minutes. It was important that the dyes were dissolved locally without dragging them. Having painstakingly planned the image, it then faded out after a few hours as the droplets diffused. However, the fluorescence image was captured for posterity, having been filmed using a custom-built setup based on Raspberry Pi’s HQ Camera. Given the impermanence of the image, it was critical that the camera captured everything first time, as there wouldn’t be a chance to reset and repeat things if they went wrong.
Yuksel explains the difficulty of videoing the water filling the channels and revealing the portrait: the imaging area was too large for the field-of-view of their high-end fluorescence microscopes, while the channels and individual droplets of fluorescent dye were too small for an imaging setup with an SLR camera to be considered. “Image stitching was not an option either because the liquid filling was too fast for the motorised stage of our fluorescence microscope.” He realised an imaging setup using Raspberry Pi’s HQ Camera would be ideal because he could easily attach a C-mount magnifying objective lens and fix the camera to a tripod.
Building blocks
Fortunately, Yuksel already had a number of Raspberry Pis in his lab, and had gained plenty of experience of microscopy while designing and building a very impressive Raspberry Pi-based LEGO microscope a couple of years previously.
For the Mona Lisa water painting project, he custom-designed and 3D-printed an adapter for the emission filter to be placed in front of the objective lens for fluorescence imaging. Illuminating the sample was tricky, says Yuksel, because the area was too large for light sources used in classical microscopes, and it had required a specific wavelength for the fluorescent dye. Fortunately, he had a collimated 470 nm LED light from an earlier project, which was able to uniformly illuminate the sample.
Photography-wise, Yuksel had also cut his teeth on a prior Raspberry Pi project: a stop-motion project showing a 3D model boat he’d made in action.
Nonetheless, he was taking no chances, and put together a standby chip just in case. After all, the chip couldn’t be reused once the liquid was applied, so there was just one chance to capture the image-creation process. Thankfully, everything went without a hitch, thanks to his meticulous planning.
