fringe locking & laser diodes

Joe Farina wrote:

Hello Everyone, As I can see from comments on this forum, people are having some troubles when using laser diodes to make holograms. On the other hand, I have heard some success stories also. But my own experiences with laser diodes for holography have been unsatisfactory. I guess the problem is due to wavelength shifts over the length of the exposure. Maybe these problems would be lessened (or perhaps eliminated in some cases) if exposure times are kept short, in the neighborhood of a few seconds or less. But these short exposure times don't fit in with the kind of holography I want to do. It seems there are many complicated schemes for stabilizing the wavelength of laser diodes. At first, I thought active temperature control would be the answer (as is so often stated in the literature). Perhaps this is true, but after I built an active temperature control device for the Mitsubishi diode, the results weren't what I was hoping for. I used good components and did a fairly careful job, but the fringes on my interferometer were drifting somewhat when I turned on the temperature control circuit. If I turned it off, and simply relied on the passive heatsink, the fringes had better stability. Maybe I was doing something wrong, but I think it's safe to say that active temperature control is rather complicated (not to mention expensive) and I am quite skeptical of its effectiveness for stabilizing laser diodes for holography. I am wondering if there's a better way. It seems to me that if we want to stabilize laser diodes for holography, we need some kind of "reference" system in the device which relates specifically to the kind of stability we need. The first thing which comes to my mind is fringes on an interferometer. I am wondering if the wavelength of a laser diode can be stabilized (in a fairly simple manner) by attaching it to a "reference" interferometer. Some time ago, I built a finge-locking device which was very effective for stabilizing fringes. I was surprised how simple and inexpensive it was: all it requires is a cheap op-amp (costing about $2), a small speaker, a pair of photodiodes, a couple of potentiometers, and some resistors. My problem is that I'm having a hard time "connecting" the fringe locking idea to a laser diode driver. My idea is to use the signal coming in from the fringe locker (which would normally be connected to a speaker or transducer which moves a mirror back and forth to change path lengths, so the fringes would be stabilized). But in this case I would want the signal to go through the laser driver to change the current going to the diode. If there is a way of doing that, then I think the wavelength could be stabilized cheaply and effectively. But since I don't know much about electronics, maybe it's not possible.

Colin Kaminski - Sat, Dec 15, 01 10:03:23 PM

Colin - Sun, Dec 16, 01 12:03:10 AM
This is a web site about linking diodes together for more power.

http://optics.caltech.edu/billgr/phaselock/phaselock/index.h tm 64.170.194.24

Tom B. - Sun, Dec 16, 01 04:55:04 AM

Joe Farina - Sun, Dec 16, 01 01:31:38 PM

Colin - Sun, Dec 16, 01 04:50:58 PM

Jonathan - Sun, Dec 16, 01 04:58:25 PM
With laser diodes it looks like the biggest obstacle to bright holograms with CW lasers (i.e. movement) has a new rival - frequency instability.

Over a year ago I replaced my He-Ne with a Mitsubishi 658/35 diode and a D&S driver. I'm using an overhead frame system and mount both a 13" parabolic mirror and a spatial filter overhead, therefore physical stability has always been a concern to me. I mounted the diode in a small metal clamping device, with no particular attention paid to temperature dissipation other than making sure air cooling wasn't impeded more than necessary.

However, the problems (arc-shaped banding in the real image, moire lines in the virtual image) began to occur with increasing regularity. The problem we now identify as frequency shifting or mode-hopping was intermittent, but it's regularity seemed to increase over the course of the last year. Possibly this means that diodes become more prone to mode-hopping with age, if no temperature stabilization is present? Has anyone found the same thing? My exposures weren't getting any longer - still in the 5-10 second range.

Anyway, at first I thought the moire lines were due to a physical instability in the system. So I reinforced, rebuilt, blocked every draft, improved the draft exclusion canopy, located my shutter release outside the lab, and damped every component and structural frame component I could. All with feedback from the interferometer of course, so I could see what was effective. Thanks to Colin, who suggested damping and explaining how it worked. Turned out damping provided as much fringe stability improvement as all the other efforts combined.

The result of all this was better (brighter) holograms WHEN they were free of banding, BUT the banding still occurred intermittently regardless. So obviously this particular problem wasn't movement, but rather frequency shifting of the diode, as the discussions on this forum have revealed.

Here's where the discussion on solutions becomes blurred. If we are talking about fringe-locking solutions, using feedback from an interferometer, then things can be done to compensate for fringe movement caused by path length movement. This is described well in the article Tom posted above. But for frequency shifting of a diode, what is the mechanism, besides doing one's best to stabilize the diode temperature?

Optima uses a thermocouple device in their BTC-2000 temp. controller. No mention of any other "frequency locking" mechanism. The Samourai laser (Yves Gentet) specifically locks the frequency of the diode (and they use four different types ranging from 639 to 687 nm). If you look at the site Joe posted above you can see the following description:

"TEMoo & Electronicaly locked single frequency mode: the coherence length is extremely long (several meters) without frequency changes over several hours. We have controlled the contrast of fringes and movement at 2 meters depth in an holographic set-up (diff between ref and object beam): over 30 minutes exposure, no fringe movement could be seen. A green LED indicates you when the laser is frequency locked after 20 minutes warm-up time."

My feeling is that temperature stabilization should be the first thing to do, and if the heat sink is large passive heat control may be enough. I don't know. The experiences posted above seem to suggest that. Joe did you find that your results were still unsatisfactory even with a good heat sink, but no TEC?

Obviously, the laser diode (Samourai) from Yves Gentet must have the electronics in the power supply to somehow control frequency shifting beyond just temperature control. His laser is "current regulated" but that is all the information provided.

Can anyone shed some light on how frequency locking can be accomplished with a red laser diode, above and beyond providing stable temperature control? Perhaps we may have to wait until the technology evolves with better semiconductor materials, fabrication techniques, or designs. 209.90.161.49

Jonathan - Sun, Dec 16, 01 06:49:31 PM
For anyone interested there's a technical paper on this subject at:
http://opticb.uoregon.edu/~mosswww/FrequStabi/Cover.html

Points worth noting are:
- it's robust and can be built relatively inexpensively
- the control method is applicable to any laser system but in this paper is described using an external-cavity diode laser
- the results given are based on using "thermo-electric temperature stabilization to better than 10 mK" 209.90.161.142

Joe Farina - Sun, Dec 16, 01 08:17:48 PM

Jonathan - Mon, Dec 17, 01 12:58:58 AM
Joe, I think you are right about there being two distinct problems here, when it comes to frequency stability, namely "broad linewidth" and "mode-hopping".

Unfortunately, I've spent quite a lot of time with my table set-up as an interferometer, instead of producing holograms. Perhaps more than most because my laser is being used without a heat sink. I intend to correct that. But what is interesting is that there seem to be three distinct fringe behaviours that one can observe.
1) complete wash-out of the fringes (for a few seconds, more common in first 1/2 hour)
2) partial wash-out of the fringes (for a few seconds, and roughly to the same degree of contrast loss each time)
3) a discrete "jerking" of the fringes, where they change position suddenly without losing contrast (not talking about vibration induced jerking - this is quite different)

I offer a hypothesis. My feeling is that the "wash-out" of the fringes is a function of a broadening of the linewidth, where there is no longer a dominant wavelength sufficient to provide the necessary coherence length. And the discrete movement corresponds to what is being called "mode- hopping", where there is, in the timeframe of one exposure, actually two or more periods, each with a dominant wavelength.

The symptoms in the holograms are noticeably different too, where in the case of 1) above, there wouldn't be an image. And in the case of 2) the image would be dim. For 3) you'd see a "double-exposure" effect, where there is a secondary interference pattern visible in the film, and the virtual image (transmission) shows the contour lines, or moire pattern. I'm not sure, but this may be the sliced-bread effect that some people have described.

The contour lines problem is by far the most common condition when the hologram is bad, which isn't all the time. About half are actually symptom free and bright. What I find somewhat inconsistent with my observations is that I've seen relatively FEW of the discrete fringe "jumps" with the interferometer, compared to complete or partial "wash-outs", but then you have to be watching more carefully for them since they are so quick. 209.90.161.50

Tung H. Jeong - Mon, Dec 17, 01 08:35:00 AM

Colin - Mon, Dec 17, 01 10:01:16 PM

M S Valera - Tue, Dec 18, 01 12:50:24 PM

Tom B. - Wed, Dec 19, 01 04:53:12 AM

Colin - Fri, Jan 04, 02 02:09:22 AM

Tom B. - Fri, Jan 04, 02 11:40:49 PM

paul - Fri, Oct 25, 02 12:12:44 PM