Photonic Fence: Dosing Experiments

One of the challenges of the Photonic Fence project has been figuring out how much laser we need to kill mosquitoes. We know in principle that we can combine low-cost sensors, laser technology, and software to identify, track and kill mosquitoes in flight.  However, the choice of which laser we use may have big systems implications, so we wanted to spend extra time zooming in on this question.

The Photonic Fence is capable of identifying the discriminating characteristics of a mosquito (e.g., wing beat frequency, shape, size, and airspeed), training a laser on it, and delivering enough photonic energy to kill or incapacitate it. Not only can the system distinguish between mosquitoes, butterflies, and bumblebees, but it can even determine whether a mosquito is male or female (only female mosquitoes bite). Once the software establishes that the insect is a valid target, it tracks the mosquito in flight, runs a safety check to ensure no innocent bystanders are in view, and then activates a laser to zap the mosquito. 

A recent series of experiments focused on exactly what happens when a mosquito is hit with a laser. How energy is imparted to an organism can be a very interesting question… For example, a 9mm slug carries about 500J of energy when leaving the gun muzzle; this bullet would convey the same energy of impact a 50kg person experiences when falling from 1m. Which would you rather have?  

To figure out which method of energy delivery mosquitoes would definitely prefer not to have, we designed a set of experiments exploring across a five-parameter space:

  1. Laser power: This is probably better stated as laser fluence – how many photons hit the mosquito per unit area. Presumably, more photons should be bad for the mosquitoes, but bigger lasers may cost more money.
  2. Pulse duration: The total amount of energy imparted to the mosquito is a function of how fast energy is delivered, and for how long.
  3. Spot size: When the same amount of energy is imparted over a smaller area, its ability to affect multiple areas of a mosquito may go down, but the localized effect of exposure may be more pronounced.
  4. Laser wavelength: Different materials absorb and reflect different wavelengths of light at different rates.  At the same time, some wavelengths (colors) of laser are considerably less expensive than others, generally because of how they are used in commercial products like home movie players.
  5. Targeted body part: An interesting question is whether some body parts resist laser energy, and some are more susceptible. Do they have a soft underbelly, so to speak?

To run an experiment, the team started with a cage of 84 live female Anopheles stephensi mosquitoes. These subjects were then anesthetized with carbon dioxide, and manually manipulated into an organized grid on the bottom of the cage. Next, the team shot each individual mosquito with a laser pulse. Twenty-four hours after laser exposure, the team looked for signs of life, documenting the results. Each of the five parameters was varied systematically so we could see its effect on mortality.

Data from ten weeks of these experiments map nicely into seven mortality curves:

Each curve was generated from a minimum of five data points, and as many as fifteen. The data set used to generated the mortality curve is shown below. 

For each experiment, we were able to clearly see how varying one parameter, like wavelength, affected when 90% of the mosquitoes would die in the following 24 hours (we refer to this as the “lethal dose”, or LD90).  Over a 10 week period, the team anesthetized, manually manipulated, shot with a laser, and then assessed for signs of life 24 hours after exposure, a total of 6,092 mosquitoes.  

When you sprinkle in the handling and robustness controls (mosquitoes used to verify that the subjects were comparable from week to week), the GRAND TOTAL goes to more than 10,000 Anopheles stephensi

While this data is very … illuminating… there is still more work to do, and more mosquitoes to zap.