TJIIRRS: Number 23

(23 July, 2011, ff)

I have been reading about OFCGs, and I come to some conclusions about the possibility of such a device as a DIY project. Frankly, it is likely to be at or beyond the limits of DIY construction; but that doesn’t make it any less interesting. The OFCG is superficially simple, but it is based on a number of significant advances in laser technology, and is more complex than it appears at first glance.

There are some crucially important building blocks that must be in place before one can hope to make such a device, and I will start with what I am happy to refer to as the “baby steps” at the beginning of the path.

First, it is necessary to construct a mode locked laser. I have been reading and thinking about this, and it seems to me that it should certainly be feasible on a DIY basis, if one is willing to allow certain limits. (We are not going to make our own laser mirrors, for example.) There are lots of ways to mode lock lasers, and lots of types of laser that can be mode locked. [More about this in a bit.]

Second, it is necessary to construct a femtosecond laser. That is seriously tweaky and difficult, and may not be particularly feasible as a DIY project. OTOH, the technology of femtosecond lasers has matured a lot during the past decade, and we may be able to take advantage of that. The fs laser, however, must wait until I can at least produce picosecond pulses.

Right from the outset this diverges into two approaches.

We have a [dead] Coherent CR-599-04 CW dye laser, which I am hoping to resurrect (without its tuning mechanism, which was already missing when we acquired it). I am hoping that the lack of a tuning device will turn out to be a non-issue, as the alignment procedure involves removing it. The CR-599 uses four mirrors, of which only one is present: the pump mirror. The other three (HR rear reflector, “bend” mirror, and OC) are missing. (See section I B, below, when I get it written.)

This laser also requires a pump source. It is, or at least was, common to pump CW dye lasers with large-frame argon ion lasers, and we actually have a 2-watt argon machine here; but argon lasers are large and inefficient, and it appears to be possible to avoid them if one is willing to work within the constraints imposed by other lasers (often, for example, lower power). One hurdle is clearly the threshold pump power that the dye laser requires. If I remember correctly, the lowest threshold pump power that has successfully been used with Rhodamine 6G is about 69 mW from a green DPSSL; and the lowest threshold pump power I have ever heard of was 10 mW from a HeNe, pumping a chilled solution of Oxazine I perchlorate. (This was accomplished in 1986 by a team including Karl Drexhage, in the Chemistry department at Garching. Dr. Drexhage is more famous for his research on dyes; he is a leading authority on them, and IIRC he is particularly expert in the area of dye stability.) I do not have Oxazine I here, but there are other dyes that respond to temperature, and I will be keeping that notion in mind if I end up using a dye with that type of behavior. Unfortunately, the laser that was used in that experiment had a simplified configuration that was optimized for low input power. Even a factor of 50 increase, though, may be achievable, depending on whether I can focus the pump beam tightly enough.

It is relatively easy to acquire fairly powerful green DPSSL modules today, and I have acquired a 200-mW unit because the first optics I have been able to acquire for the CR-599 are designed for the red and orange. If I can get optics that are optimized for yellow and green, later on, I will try pumping with a high-power blue laser diode (Nichia NDB 7352) that I acquired from Cajun Lasers. It is easily capable of putting out 650 mW or more.

The blue laser is not likely to work well with R6G, which has very little absorption around 450 nm, but it should work tolerably well for Fluorol 555 (which I have), and possibly for Fluorescein. There is some slight chance that I could use a mixture of one of these dyes with R6G, and I will be looking into that.

One important question, however, is whether the beam can be focused down to the required spot size (50 microns or less, IIRC). There are two considerations here. First, diode lasers emit beams that are oval, not circular. It is, fortunately, possible to use a cylindrical lens to change the shape to something that is more reasonable. The larger problem, however, is that the available high-power blue diodes are multimode; the beam is not a single spot, the way a HeNe beam is. (See photos, below.)

I am wondering about possible approaches. Is it viable, for example, to move the dye jet slightly out of the beam waist in order to take advantage of a larger pumped region? If so, how much higher does the pump power need to be? (If I can keep the increment to significantly less than one order of magnitude, I have a chance.) Frankly, this may not be possible with the CR-599; but if I end up being obliged to build my own CW dye laser I may make it a feature.

It also occurred to me that I might be able to mode lock the blue diode itself, and I’ve been trying to accomplish that. [More about this after the first set of photos.]

Here, to give you an idea of what diode laser output tends to look like, are six photos of the beam pattern from the blue diode, defocused onto a wall:

                   

                   

From the left, first row: just about at threshold; slightly above threshold; moderately above threshold. Second row: well above threshold; just below the maximum current I am permitting at this point; and at the max, with the laser putting out ~660 mW. As I have pointed out above, this is very obviously a multimode device. Notice, btw, the fact that the room light does not appear as bright in the last photo as it does in the first. The camera is compensating for the brightness of the laser beam on the wall...

(2011.0731, ff)

When I look on the Web, I find that various laser classes mode lock a diode as an exercise; but it seems that they invariably use a diode with an AR coating instead of a mirror on its output face. The claim is that it is necessary to eliminate all reflections back into the amplifying medium except the one from the OC. Although that may be accurate for a clean pulse train, there may be a chance that I can work around it, ...if I can get any sort of locking at all.

I put the diode laser on a plinth, and set a cylindrical lens in front of it to correct the beam shape. I’m using a surplus piece of steel as an optical rail. The setup initially looked like this:

At the opposite end of the rail I have a mirror mount from an argon laser, turned upside down so that I can easily adjust it with the mirror surface as close to the saturable absorber as is practicable. Here it is with the beam going into it, first a normal exposure and then an underexposure so you can more easily see the slit (fabricated from little strips of styrene plastic) that I’m using to clean up the beam a bit:

           

Here is the beam in the air, coming from the diode:

If I block off the mirror, the laser is just below threshold. You can see the slightly brighter region in the middle of the pattern in the photo on the left. If I turn up the current just a bit, it begins to lase...

           

Now I have to figure out how to tell whether I actually get mode locking when I interpose a saturable absorber. (I can certainly get lasing, but that doesn’t necessarily tell me very much, and the beam is not powerful enough for my usual photodetector to pick up.)

(2011.0806, morning; 2011.0813, morning)

As I continue reading, I discover some information that suggests changes to my approach. First, it is possible to do active mode locking of diode lasers by modulating the current through them. I am reserving that approach for when I have time to think about how to implement it relatively simply. Second, because of the brief gain lifetime of the material, diode lasers apparently seem to work best in short cavities, with repetition rates of 1 GHz or higher. The long cavity I was using seemed unlikely to work well, so I moved the external mirror and the dye cell a lot closer to the diode. Third, it is probable that I could tell whether the laser is mode locked by examining the output with a scanning Fabry-Perot interferometer; the output profile of a mode-locked laser should look a lot different from the output profile of the same laser when it is not mode locked. (See Sam’s Laser FAQ for information and links.) I have most of the parts I need to build a crude SFPI, and I should be able to do more work on that after the middle of the month..

(2011.0813, morning)

I have tried a number of times to achieve mode locking with the blue diode, mostly using Erythrosine B as a saturable absorber. It has fairly short fluorescence lifetime, particularly in water, and should be reasonable for this purpose. No luck so far, though. (Perhaps I should say “no obvious luck” — it may not be easy to tell, particularly at the very low power levels I’m getting with the mirror I currently have in place as the external OC on the diode: it passes almost nothing at 445-450 nm...) I also suspect that the saturable absorber solution may have to be moving, though at these power levels I am not seeing any blooming or other distortion of the beam. (I can easily get blooming by turning the blue diode up past its own threshold, and I should probably make a video.)

(2011.0813, evening)

Meanwhile, here are some relevant photos. First, a side view of the cell I built last night, which I’m using to visualize the beam. (For the moment, it contains PPO in 99+% isopropanol.) The beam enters from the left; the blocky gray object at the right is the mirror.

Next, an overview of the setup:

Here is 10^-3 molar Erythrosine B in distilled water, with no other additions:

Finally, here is something that I think is an artifact. This is the visualization cell in use:

Here is a tight crop of the area where the beam hits the mirror. I have circled a bright dot that should be an indication of mode locking; but you’ll notice that it only appears in front of the mirror. If it were real it would be reflected as well, and would be a slightly more elongated spike.

My apologies for the noise in those last two images, btw; they were very long exposures with laser speckle, and I had the ISO speed of the camera set to 400. My guess is that the camera simply can’t deal with the combination of high ISO speed, darkness, and speckle.

On other fronts:

In order to bring up the CR-599 I need to find or acquire a dye pump (I think we have one), clean everything, install the optics, set up the green DPSSL module as the pump laser, and attempt to align everything. If that fails I will probably set up the blue laser as the pump laser, and attempt to focus the beam to a small enough spot on the dye jet.

More as it happens...

Back to the Index

Home

Contact Information:

My email address is a@b.com, where a is my first name (just jon, only 3 letters, no “h”), and b is joss.

My phone number is +1 240 604 4495.

Last modified: Thu Jun 23 16:08:24 CDT 2016