NicoLase

An open-source diode laser combiner, fiber launch, and sequencing controller for fluorescence microscopy

New users please check out the associated open-access publication from PLoS ONE: http://dx.doi.org/10.1371/journal.pone.0173879

Therein you'll find a complete description of the NicoLase platform and performance specifications.

Here in the repository you can find all of the files and supplementary descriptions to build your own fiber-coupled diode laser launch as described in the publication. As the project matures, new features and softwares will be incorporated here.

Motivation

Diode laser units are convenient for fluorescence microscopy and super-resolution microscopy due to their high power and ease of use. A couple dozen lines are avaliable from most makers to cover UV through mid-IR applications. With the addition of direct modulation avaliable on nearly all modern units there's no need for an AOTF or other tuning or power modulation control in addition to the diodes themselves.

Multiple laser lines are almost are a requirement for a modern fluorescence microscope. Rarely can a system produce publication-appropriate results without at least a second laser. Multiple lasers need to be combined and then controlled together for multicolor illumination.

Software and trigger control of the lasers is strongly desired. Microscopy cameras often have a dead time between frames when laser illumination is (almost always) undesired, and at the very least it's nice to have a laser that shuts off when the acquisition simply isn't running. Then, as a further step, it's often necessary to alternate or otherwise sequence lasers for certain experiments (ie ALEX FRET or pseudo-simultaneous multicolor imaging). With digital control of laser emission this needs to be controlled at the laser head.

There are commercial options to achieve nearly all of these. Basically, this system does not do anything you can't buy. But the final cost here is 2/3 to 1/2 of the cost of commerical instruments and can be easily customized in terms of size, number of lines, and final configuration. The system here is cheap (ish), flexible, and quite compact.

Two generations of this project are included in the repo. The first, half-sarcastically called the NicoLase 1500, is a single-mirror coupling system running 5 (expandable to 6) diode lasers. The later, the NicoLase 2400, is a two-mirror system with 4 diode lasers. These share the same footprint, many of the same components, and the same control software and PC interface.

Note: Throughout I use the catch-all term 'lasers' or 'laser outputs' for the end target for TTL outputs of the controller. Of course this same controller and software can be used with any equipment receiving a TTL input pulse that needs to be synced to camera acquisitions. We commonly use an LED illuminator for transmitted light with the controller output, for example.

Hardware

Lasers

The NicoLase is built primarily around Vortran Stradus laser diode units. These are a standardized size for each line, reasonably priced, and produce a large amount of power for their size and cost. Nearly all produce enough power for dSTORM work. These units can be digitally modulated at a high rate (usually up to 2 MHz) and potentially analog modulated if desired. I don't have any financial reason for recommending these units, just a satisfied customer.

A straightforward modification allows you to use Coherent OBIS heads in place of some/all of the Vortran units. Control is nearly identical and aside from some extra screw holes, these are drop-in replacements for the Vortran units.

I haven't had a chance to test any other laser diodes with this system, though there's no reason others with the same characteristics won't work.

Key features for any lasers chosen are :

• 100 mm x 50 mm footprint or smaller
• 19.05 mm or 3/4" beam height, or can be brought to this height with spacers/lifts
• Digital modulation via TTL pulse
• Software control of laser power via USB or RS-232 (for add-on plug-in control)

Heatsinks and Mounting

Diode laser heads require a heat sink for proper operation, which has to at least be a contact with a block of aluminum. We need the laser outputs to be at 39.0 mm to intersect with the steering and coupling optics. Oh, and the laser have to actually bolt down to something.

To solve these issues, a heat sink + mounting block was machined. Three dimensions must be held to strict tolerances for this part. The top must be as flat as possible (0.1 mm specified) with no scratches or machine marks to interface with the bottom of the laser diode for heat transfer. The top and bottom faces must be parallel (0.1 mm specified) so the beams are at the same height. And the total height must be the specified thickness (0.2 mm specified).

Keeping the top flat can be done by using a piece of ground stock and machining other faces. Keeping the top and bottom flat to that tolerance should be straightforward for a decent mill. And we can use a bit of a trick for the total height. This is because we don't actually care about the bottom of the part and can therefore add shims of sheet stock (with clearance holes cut) to bring the total assembly to the desired height. The block can be machined on the bottom face until the thickness is a bit short, then a piece or pieces of sheet stock selected to bring it up to the final thickness. Much simpler and easier to meet the necessary tolerance than hitting the right thickness with machining only.

The heat sink block must be made of either aluminum or copper. Steel, especially stainless, does not have a sufficient heat transfer rate to keep the diodes from overheating. We had our top blocks made from aluminum and the shims from copper sheet, which works great. DO NOT add heat sink paste or thermal grease to the interface between the laser and the heat sink block. That stuff outgasses like mad and gets on your optics, plus the flat surface does an excellent job of heat transfer already.

Mirrors and other optics mount to standard Thorlabs parts except for the fiber coupler itself. This needs a custom pedistal to be machined, but this is a straightforward piece of turning. No dramas there. The original 1500 design uses a flat block for this mount but the 2400 pedistal is a much better design.

All optics are chosen for a 1.5" center height. Mirrors and dichroic mirrors are mounted in Thorlabs Polaris 1" mounts. The single-mirror system requires 3-adjuster mounts, but either 2- or 3- adjuster mounts can be used for the two-mirror system. Hex adjuster are nice for squeezing the mounts closer together, but are a bit more of a pain to adjust. Clearances are just barely there for thumb adjusters and for your fingers. Those more rotundly-digited of us may prefer the hex adjuster mirrors.

Excitation filters can be mounted in standard lens mounts, or options for 3D printed mounts for 1/2" filters designed for the NicoLase is given in my 3D printing repo here. A mount to tie the Aruino to the optical table is given in the same repo, or you can make a mounting plate to bolt in a spare laser slot on the heat sink block (3D printed version forthcoming).

Drawings and pictures of enclosures are provided (one a bit more 'polished' than the other). These keep the lasers away from the unsuspecting eyes of users and the optics protected from bumps and dust. Components for the NicoLase 1500 enclosure are from 80/20 using 25 mm aluminum profile. The 2400 enclosure is made of corrugated plastic sheeting and held together with magnets and 3D printed magnet clips (available here). In both cases the laser power supplies are on the optical table with the breadboard and lasers above, but there's no reason the whole assembly can't be moved to a shelf off of the optical table if you need the space.

Update April 2019 - Plans for an enclosure in the form factor of a 19" rack (that fits under the optical table!) is included in 3600 model folder. This rack can hold the NicoLase breadboard in the upper compartment with additional accessories and equipment underneath. This includes laser power supplies (on their own shelf) with any additional boxes, power supplies, or even a PC underneath. Laser portion can be enclosed for a facility or dark room application. Whole thing sits on sliding feet to fit under the optical table making a very convenient package!

Technical drawings for all machined components are given. These were sufficient for our Uni's machine shop to complete satisfactory work with no additional consultation.

Optics

Optics are relatively straightforward - a few dichroic mirrors, optional filters, and a fiber coupler. This unit is built around mostly standard Thorlabs components. Aside from a T-shirt I got at a conference once, I have no interest in pushing Thorlabs components though they are reliable, self-compatible, and widely available.

Dichroic mirrors are used to combine lasers into a single multi-wavelength beam. A mix of Chroma and Semrock filters are used. All are standard 1" round long-pass filters costing a few hundred USD each. Which ones are used depends on which wavelengths are being combined, but of course you want to pick a cut-on wavelength that reflects the beam being folded in and transmits all others upstream. Picking a cut-on wavelength that falls between adjacent laser lines usually works well, but double-check the provided spectra to confirm there aren't odd peaks or dips in the filter spectrum.

Diode lasers often require an emission clean-up filter to block odd long-wavelength emission from the laser head. We've had particularly bad luck with 488 nm diodes in this sense. At the same time, dio