Some experiments with the blue LD's

This is a forum to share experiences and ideas about holography.
Thieu

Some experiments with the blue LD's

Post by Thieu »

Hello everyone,

Last week I did some experiments with external feedback on the blue 445nm diodes. I found that the diodes are so sensitive to external feedback that it's possible to lower the threshold current and change the longitudinal mode structure even by just putting a microscope slide (no coverslip) in the *uncollimated* beam, right in front of the diode. In one of my tests, the diode operated almost exclusively SLM from threshold to around 300 mA this way.

Even though operating the diode like this is by no means comparable to a real ECDL with a TEC and high stability current source, it might be interesting for those of us who just want to give it a try without expecting excellent results. Given the expected short exposure times at this wavelength, the stability of a simple good current source fed diode that has reached thermal equillibrium may well be enough to produce holograms better than those obtained with green pointers for about the same cost.

For those interested, more details about the measurements and videos of the mode structure development during current ramps are on my blog ( http://piepklein.blogspot.com/ ).
Colin Kaminski

Some experiments with the blue LD's

Post by Colin Kaminski »

What is the resolution of the grating you use to spread the lines? I am wondering how sensitive this simple feedback system would be to thermal length changes.
Tom B.

Some experiments with the blue LD's

Post by Tom B. »

Excellent work!

For those that might be interested, Nichia's site has data sheets on this and other interesting diodes such as lower power single transverse mode blues (NOT SLM), some UV laser diodes and an an announcement that they will be shipping samples of their new 50 mW green diode in August.

http://www.nichia.co.jp/en/about_nichia/index.html

Here's a company selling 1W blue diodes pulled from new Casio projectors for USD $63.50:

http://www.stagelightingeffects.com/ind ... par=163849 (6th item down)
(I have not dealt with them, buyer beware etc. but they look legit.)

Interesting times!
Thieu

Some experiments with the blue LD's

Post by Thieu »

Colin Kaminski wrote:What is the resolution of the grating you use to spread the lines? I am wondering how sensitive this simple feedback system would be to thermal length changes.
Hi Colin,

This is the monochromator I use to view the mode spectra:
monochromator
monochromator
monochromator small.jpg (210.05 KiB) Viewed 16282 times
The light enters from a slit at the lower right-hand corner and after being collimated by a concave mirror it reflects at two gratings that are coupled mechanically in the original system. After being focused again it falls onto the webcam. I measured the angles of the gratings at different settings which gave about 600 lpm (15000 lpi) for the gratings. I'm not 100% sure, but the deflections at each grating probably add up, giving effectively 1200 lpi for the combination. I calculated the wavelength resolution for the complete system to be about 0.013 nm/pixel in the final image on the webcam. Since the gratings are pretty large, the ultimate resolution in the diffraction limit is probably quite a bit higher, but it's hard to reach that mainly because of astigmatism introduced by the off axis use of the mirrors.

But the spectrograph is not involved in the feedback setup. The feedback setup is just this:
microscope slide feedback
microscope slide feedback
microscope slide feedback small.jpg (147.97 KiB) Viewed 16282 times
Nothing more than the slide resting on the part of the heatsink that sticks out in front of the diode. The distance between the slide and the front window of the diode is a couple of mm's.

The sensitivity to thermal length changes will be in the order of a fractions of a wavelength. In practice, it may work to just glue the slide directly to the diode. This will keep the system compact and it will settle down into thermal equillibrium together with the diode itself. Thermal expansion coefficients are in the order of 10 microns per meter per kelvin, so a small mm sized system will be stable to the nm if the temperature is kept stable to a fraction of a degree. For a system that has settled down in a quiet room this will be no problem at all, especially because it will only have to be stable for the duration of the exposure.

The above is theory though. Because I don't have a TEC controller, I can't measure the spectra at different temperatures. Maybe wler could make a scan?
Thieu

Some experiments with the blue LD's

Post by Thieu »

Maybe it's good to elaborate a bit more on the thermal sensitivity, to avoid some misunderstanding.

Expansion coefficients of common materials are in the order of 10 microns per meter per kelvin. So if you have a path length difference of a meter in your setup, and the setup heats up by 0.025 degrees you will have a 250 nm phase shift along this path. If this happens halfway during the exposure, the labda/2 phase shift will cause the fringes of the first part to be overwritten by exactly the opposite pattern during the second half, so you'll get no hologram.

If you have a diode laser with a cavity of 1 mm, this will expand to (1 mm + 0.25 nm) with the same 0.025 degrees temperature change. This 0.25 nm will seem very small and of no concern compared to the wavelength, but beware that it results in the same situation as above. The wavelength of this diode laser will be such that the relative phase will be 0 after each pass of a cavity length. So after a meter of travel, the original diode will be at exactly phase 0, but the expanded one will have it's phase 0 point at 1 meter + 250 nm. It will lag behind half a wavelength at the 1 meter point. So even if the cavity length difference is only 0.25 nm and the corresponding wavelength shift is only a fraction of a pm (!) , you will still get a complete loss of fringes after 1 meter path difference.

So the same rule applies to the setup in general and the laser in particular.

Formulated as a rule of thumb:

Temperature stability required during exposure (in degrees C) = 1 / maximum path length difference in setup (in cm)

So for simple holograms of say 10 cm depth, and low demands on quality, 0.1 of a degree should be enough. This is consistent with the fact that there is no TEC in for instance the diode laser that's part of the Integraf and many other diode laser holography kits.
dave battin

Some experiments with the blue LD's

Post by dave battin »

Thieu wrote:So for simple holograms of say 10 cm depth, and low demands on quality, 0.1 of a degree should be enough. This is consistent with the fact that there is no TEC in for instance the diode laser that's part of the Integraf and many other diode laser holography kits.
This is correct Thieu, with low demands on the diode itself, (mw instead of watts), this diode easily accepts positive feedback thru it's output window, then its just a matter of allowing it to come to a thermal equilibrium at the same time. By adjusting the current while the laser is in an interferometer set-up,i can achieve the longest coherent length possible, i then allow it to run a while to thermally adjust then check the coherent length and make an even smaller adjustment to the current again, and i repeat this cycle until im happy with the lasers fringe structure......

i made two one beam holograms last night (w/8inches of coherent length),and only passive cooling.
a_k

Some experiments with the blue LD's

Post by a_k »

Hello Thieu

Thanks for sharing your results and congratulations for the success. Depending mainly on the amount of feedback there are different regimes possible each with distinct impacts on the bahaviour of the LD. This paper contains a description of the feedback levels required and the effects to be expected for each of them: http://www.photodigm.com/pdf/Photodigm_ ... edback.pdf

The glass plate provides a weak feedback (Regime 1) and depending on the distance of the glass plate / phase of the feedback the linewidth might get better or worse.

Did you check if the transverse mode structure did show changes when the cover plate was at an optimum position? It would be interesting to learn if there was an improvement as well.

Ahmet
Thieu

Some experiments with the blue LD's

Post by Thieu »

Thank you Ahmet, that's a very interesting article. From the descriptions given in the article about the behaviour of the laser I'd say that I'm in regime 2 at the moment though. A quick simplified (non gaussian optics) calculation of the amount of expected feedback also supports this. The beam grows to a size of about 50 mm * 1000 mm after travelling 1m in air. So if it emanated from a 100 um * 1 um facet, it will be 150*1000 um wide after travelling 1 mm. Together with the 4% reflection on a glass interface, this gives about a 45 dB reduction. Tomorrow I'll see and try if I can explore the other regions as well by varying the distance to the slide.

With the slide there is little effect on the higher order transverse modes other than that the threshold current at which they come into play decreases. I can get them to flash on and off with the current set just below the threshold.

With the other setup though, with the glass plate in the focal point of the focussed beam, the effects on the transverse modes were huge! All kinds of modes flashed by, even some that aren't visible at all when the laser is operated without external feedback. I'll see if I can make a video of that as well, it was very interesting to watch.

I also saw by the way that indeed the transverse and longitudinal modes are coupled, their relative intensities in the spectrum change when I put the entrance slit at different locations in the far field pattern. Some modes completely disappear exactly in the center of the beam.
wler

Some experiments with the blue LD's

Post by wler »

Thieu wrote:[ Maybe wler could make a scan?
Oh sure, that's very interesting news indeed ! So far I was playing around with grating feedback and got single mode operation beyond 200-250mW, but often somewhat unstable and as of yet I coulnd't get a system that would work realiably without concurrent mode monitoring (BTW for the blue diodes just checking noise in the laser output seems sufficient to identify SLM or non-SLM operation, in contrast to other diodes).

I did notice that the 15% or so feedback of the grating I used seemed a bit on the high side, and what you say confirms this, namely just a few percent may be better. Indeed also for the red diodes I found feedback between 5 and 10% generally sufficient. So I will try a glass plate in a couple of days, actually a small wedge may be better in order not to have the competition between two surfaces (laserwolfer sells on ebay such wedge plates and now I am glad I got one, though it wasn't suitable for my original purpose).

For the last few weeks I was tied up with redesigning my LD-TEC combo driver to be optimized for the blue diodes, now it can handle 500mA and the prototype is very stable, to a few mK over an extended period of time. I still need to have the PCBs manufactured, assemble them and do the final testing to be sure. And this time it won't need special ICs.

Best,
Wolfgang
a_k

Some experiments with the blue LD's

Post by a_k »

Thank you for the detailed answer.

I'm not sure about being in Regime 2. For the Nichia diodes the emitting point is 1.9mm from the base. From the photos i would say your mounting plate is about 4-5mm thick and the distance from the emitting point to the glass plate would be 2-3mm. Let's say it's 2, this would be a total travel of 4mm for the feedback. The divergence of the beam, assumed to be 12 and 40 degrees, the size of the reflected spot at the plane of the emitting point would be about 5mm^2. I think the plate would need to be much closer to the front facet for exploration of the other regimes and maybe the can would need to be opened to get close enough.

I asked about the transverse modes because there is some hope that a common feedback would synchronise them and force SLM operation to higher levels than just above threshold and the fact that despite lowered threshold you were observing single mode behaviour up to 300mA would support this assumption. The video you mentioned would be highly appreciated.

With the plate at the focal point of the beam the possible feedback is much higher and could almost be enough for strong feedback and depending on how much the plate is tilted i think all the regimes could be scanned through.

Ahmet
Locked