Infrared eye tracking - safe levels of infrared exposure on the eyes

Question

I'm making an infrared eye tracker where an infrared light will be shone on one of the user's eyes, and a camera will capture the user's gaze, and I have a 940mm IR LED for this purpose. I have the LED connected to an embedded camera board that provides a 3.3V up to 2A, connected through a current limiting resistor in series. I've picked 5 values of current limiting resistors: 56Ω, 100Ω, 150Ω, 180Ω and 220Ω, which have the corresponding measured current draw (I_f) of 30mA, 15mA, 10mA, 7.9mA and 7.8mA respectively (rounded off). I am faced with the problem to select from any of these as well as vary the distance and see which one provides the best illumination within safe limits, and does not have to be placed at an absurd distance away from the eyes, like 20cm away for example.

Here's what I can control: forward current (I_f) of the infrared LED (through changing the resistor), and distance the LED is placed away from the eyes (assume LED is normal to the eye).

Here's what is fixed to me: the LED itself (and characteristics) and the supply voltage of 3.3V.

Here's what I need to know: What's a safe amount of time of exposure to the infrared LED per session at a given forward current and distance? What is the period of time the user must rest between these sessions? How close can I place the IR LED to my eyes?

I now need to draft up a risk assessment to justify my choice of distance, power used, time per session and time of rest between sessions. I've read up on some literature regarding safe levels of infrared on the eye and I've been trying to find thresholds or guidances that provides me answers to these, and here are the most relevants one I could find: https://assets.publishing.service.gov.uk/media/5a7efbaf40f0b62305b8466a/HPA-CRCE-016_for_website.pdf https://www.mouser.com/pdfDocs/AB191-4-luxeon-ir-family-eye-safety-application-brief.pdf

Some of these equations require spectral irradiance which is not provided from the data sheet. How do I obtain the spectral irradiance in W.m^-2.nm^-1 in order to fit these calculations?

Answer 1

The information is actually in the datasheet, but it takes some application of solid angle geometry and calculus to get it to the form you need.

You have the radiant flux in mW/sr on page 2. From this you can calculate the total power reaching the eye given the distance. Either making a simplifying assumption that the apparent size of the eye seen from the LED is small enough and the LED is pointed straight at it, or using the spatial distribution graph on page three. This gives you a power in W, you can get the irradiance in W/m2 if you divide by the area of the subtending circle.

Then take the graph which plots relative radiant intensity against wavelength (page 3). This is the spectral irradiance, only it's been normalised so the peak is 1. Integrate the area under the curve and rescale the Y axis so that the integral is equal to the value in mW/m2 that you obtained at the previous step, and you've got your result.

Now, this is for the irradiance on the surface of the eyeball. If you need it on the surface of the retina, then you've got to do the calculation differently to account for the lens, but the same general principles apply.