Ambient Light Rejection

Projects General Guidance
1

Hi all,

I am interested in building a fluorometer. I have currently built a light detector with an op amp, with a gain resistor of 10M ohm. My problem is that I am trying to detect a very weak signal with high ambient light, so a small amount of background can saturate the photodiode.

The concept behind the project goes as follows: I shine a blue LED at water with algae in it. The algae contain a pigment called chlorophyll, which glows red when shined with sufficient blue light. My aim is measure how much red light is produced by the chlorophyll molecule when the blue LED is turned on. In a dark room, I am able to obtain a signal with a photodiode with a red color filter and a blue LED I modulate at a rate of ~100 Hz. In a room with light, the photodiode becomes saturated easily and I cannot measure the amount of red light stimulated by the blue LED.

As suggested in many previous forum posts (not linked here), shading the photodiode or using more specific optical filters will not be enough. Also, modulating the LED and filtering out ambient light with the software is not enough. I basically have to figure out how to filter ambient light in the circuit itself.

Fortunately, I have found some documentation on what I need to do, but I am not exactly sure how to implement or modify it. The circuit I am interested in comes from this thesis (linked here; see Figure 2.1). This circuit is very close to what I want, but I do not want to modulate the signal at ~16 kHz. Rather, I want to modulate the LED through the digital pin on the Arduino at ~100 Hz.

In the diagram attached, I have presented the circuit from the thesis (panel A) and the functional circuit I have built (panel B).

fluoro_pic
fluoro_pic1040×629 113 KB

What I think I need to do:
I think I need to basically copy the circuit from the thesis, but replace the 10k Ohm resistor (or R3) with a resistor of ~6M Ohm and switch the op amp with a TLC27M2ACP. I got this 6M ohm value from Equation 15.

I have tested this and the circuit doesnt appear to work and so I am looking for feedback.

2

I strongly doubt you can and still expect to make quantitative measurements.

The best approach, taken by essentially all fluorescence spectrometers, is to have the sample chamber enclosed to completely exclude ambient light. Pretty easy for a DIY project, as well.

To further reduce detection noise, fluorescence emission is almost always collected at right angles to the excitation beam, and often with an excitation filter between the sample and the photodiode.

the circuit doesn't appear to work

Which circuit? Please post a complete schematic diagram of the circuit you are currently testing, with component identification. The one on the right is lacking decoupling capacitors, which are absolutely required. The 10 nF feedback capacitor is enormous, given the high impedance circuit.

Finally, it cannot be built on an open breadboard, with loose wiring, and be expected to work.

3

You reference a rather old paper. I think you will find the math utilized in this paper of interest as well as the approach to eliminate ambient light noise.

PDF

image
image997×652 153 KB

4

Try putting your sensor in the end of a tube, that will kill most of the ambient and is quick and easy to test.

5

Without thinking too deeply into it, this sounds like a great candidate for a Lock-In amplifier, which you can probably do entirely in software.

Research left up to the student :slight_smile: