The Defense Department wants to turn plants into sensors

When writing about connected tech that’s applied to animals, insects, and plants, it’s a time-honored tradition to label the resulting connected system the internet of — for example — cows, bees, or trees, respectively. But in most of these cases, what’s actually happening is a researcher or farmer is placing a sensor on something and monitoring how it reacts or behaves.

The water treatment plant in Minneapolis, Minn., for example, has attached sensors to mussels in order to track when treated water needs more attention. The plant runs the treated water through a tank of mussels, each with a sensor on their shell, and if the mussels close, it means the water isn’t pure enough. In Warsaw, they track their water quality by monitoring a tank of clams.

But while the internet of shellfish is a fun concept, and I personally love the idea of using the natural world to monitor water quality, the Defense Advanced Research Projects Agency (DARPA) wants to take such biomonitoring to a new level by turning the plants themselves into sensors — and doing away with the electronics entirely.

Dr. Blake Bextine is a program manager and acting deputy director of the Biological Technologies office in DARPA, itself the R&D arm of the Defense Department. One of the projects he oversees is attempting to engineer plants that can detect something — be it pollutants in the soil, toxic gases, or landmines — and then indicate the presence of that something from a distance. The plants also need to be able to grow from seed so as to make their deployment in a particular area easy.

The DARPA project wants to enable someone looking to detect arsenic in the soil to scatter the seeds of a specifically engineered plant, wait for those plants to grow, and then use a drone or maybe a walk through the field to see if the soil contains high levels of arsenic. The plant might indicate the presence of a compound by turning a different color, changing the humidity in the air, or by generating oddly shaped leaves or buds.

For anyone who has been thinking about IoT for the past few years, using plants as biomonitors is almost like the idea of smart dust that can cheaply and easily share information about the environment. Plants don’t require power, so there would be no batteries to change. They don’t throw off toxins in the environment or lead to e-waste. (One can engineer sterile plants to prevent cross-contamination.) And a single plant could detect multiple chemical compounds, which is much more efficient than traditional sensors.

Bextine says the research could be applied to boost crop yields, which would be great because food instability is a leading source of conflict. It could also be used for alerting soldiers to landmines or toxic gases.

But building such a sensor is hard. Researchers working on the project have been at it for the last two years, but their work remains in the second phase and has so far only focused on plants’ sensing capability. Building a plant that can sense different pathogens, toxins, or whatever else is difficult. It’s tough to get some type of sensing into plant cells without harming the plant so it wouldn’t grow.

The second phase is focused on reporting. That means engineering the plant to change colors, sprout more leaves, or otherwise indicate that it has detected the presence of something. The key with reporting is that the plant’s change must be viewable from what Bextine called a “standoff distance.”

This is where traditional IoT sensors might need to be introduced, as detection could involve sensors that measure the humidity of a particular area, or temperature changes from a plant opening or closing its stomata. (Stomata are the small pores in plants’ leaves that open and close in order to let the plants exchange oxygen and carbon dioxide.) It might also involve changes in a plant’s color or leaf structure that a camera on a drone could detect. Or it might even be something obvious enough that people could see it with their naked eyes. Bextine gave an example of plants that would grow inside hotels or office buildings and change color in the presence of a virus.

The second stage of DARPA’s research was supposed to be finished by now, but the pandemic slowed it down by a few months. Bextine expects the agency to share its results in the next few months. The third phase of research will be focused on figuring out how to replicate the plants and easily deploy them.

I know that I spend most of this newsletter focusing on electronics, but the idea of bringing living organisms into the IoT is compelling. We already use plants and animals as sensors, but the DARPA research takes the idea of biomonitoring to an entirely new level — one that could provide good insights, and at a low cost to the environment.