Confocal Microscope

Description

This is an attempt to build a simple confocal microscope based on a laser scanner made for laser light shows. It also integrates widefield imaging and a CMOS camera for easier sample alignment and focusing.

A confocal microscope can achieve better detail and narrower depth of field than a conventional microscope by scanning a laser spot across a sample, and having the scattered light from the sample pass through a small pinhole after it is collected by the objective lens. The pinhole allows only light within a very shallow sample depth to pass through to the detector, because light below and above the field will not be in focus. Only light in the field plane will be focused to a small enough spot to pass through.

Below is the optical layout, showing only rays for the confocal imaging path.

The HeNe laser first heads through a 5X beam expander which widens the beam to approximately 4mm across. It then heads through a beamsplitter and into the X/Y galvo assembly.

The galvo has two mirrors that rotate in perpendicular directions and are arranged so that rotation of one causes the beam output to tilt horizontally, while the other tilts it vertically.

In order for a confocal microscope to perform efficiently, all of the light from the laser should be coupled into the objective lens entrance pupil regardless of the angle of the galvanometer mirrors. In this microscope, this is achieved with two achromatic doublet lenses with equal focal lengths (f=40mm). These were chosen to minimize spherical aberration and allow the use of different laser wavelengths, and allow for this system to be used for fluorescence microscopy.

The lenses are positioned 80mm (2f) apart, with the galvo 40mm behind the first and objective 40mm after the second. The first lens will cause the laser to be focused to a point in the image plane midway between the two lenses. The second lens collects the light passing through the image plane and directs it into the exit pupil of the objective. You can see from the ray diagram that the light will reach the exit pupil regardless of the galvo angle, because both the galvanometer mirrors and the exit pupil are in planes that are conjugate to the image plane.

The red lines show the beam path when the galvo is set to (0,0) (no deviation), and the orange lines show the path when the galvo points off-center.

Between the second lens and the objective is a beamsplitter which is used for the widefield optical path. This allows the sample to be viewed like it would be in a conventional microscope, and makes it easier to see what the confocal path is looking at before running a time-consuming sample acquisition using the confocal system.

This path includes a f=180mm (approx.) tube lens, CMOS camera with ND filter, and an LED with condenser for epi-field illumination. Since the system uses infinity-corrected objective lenses, a tube lens is required to image the sample onto the camera sensor. The ND filter is to protect the sensor from overexposure or damage from the focused confocal laser beam when the galvo is not moving.

Below is the view of the Thorlabs RDS12P reticle target through the widefield path using a 5X objective.

The number 1 in this picture is about 250um tall.

The red square is the result of having the X and Y galvos scanning the sample at high speed (60Hz for Y, 1.2kHz for X)

The first confocal picture was taken with the pinhole opened about 4mm, using a BPW34 photodiode connected to a DMM in current mode for the acquisition.

The galvo was controlled by a DG1022 arbitrary waveform generator with SDM3065X DMM. The LED was turned off to avoid polluting the image.

I made an amplifier board previously to use the galvo as a laser projector with the 0-3V output of an Arduino Metro M4’s DAC. This board was repurposed to amplify the signal from the AWG and convert it from single ended to differential.

The signal generator was set to a very low frequency (0.004 Hz for Y and 0.4Hz for X and the DMM was set to acquire 80 samples per trigger, with the trigger source set to the X sync output of the AWG.

This is the first picture:

This was with the AWG voltage set high enough that the scan area exceeded the visible area of the widefield path, but this was before I added the CMOS camera which is what limits the visible area.

It’s clear that there is some jitter in the galvo output, particularly around the edges.

I also saw this strange behavior later on when running at low scan speeds:

This picture shows the full optical path with supporting electronics as of 4/19/23:

And here is a picture of the actual setup with the confocal (red) and widefield (blue/green) paths overlaid:

I’d like to upgrade this setup by first getting the galvo fixed, then swapping out the BPW34 for a photomultiplier tube. Even with a transimpedance amplifier (the one from this project) the signal is very small and noisy, and a PMT would provide much better gain and frequency response.

I’m also planning to add a laser shutter, objective turret, and eventually some other laser sources, but I’ve already burned one objective by sending a laser that was too powerful through it and don’t want to ruin another one…