The PRA LN-1000 Nitrogen Laser

(24 May, 2004)

The PRA LN-1000 nitrogen laser operates at room pressure, and puts out pulses that are more or less 800 psec long. (At least, that was what the mfr stated. Most room-pressure nitrogen lasers seem to have pulsewidth between about 600 psec and 1 nsec, and I do not yet have an easy way to check, so we’ll leave it as given for now.) The laser is rated to deliver about 1.5 mJ, which corresponds to more than 2 MW peak power. Here are two views of it:

We won this laser on eBay, some time ago. When it arrived I found the key broken off in the lock (fortunately all the way in, so I could turn the switch with a screwdriver); some of the screws were missing from the case, and it was clear that not all was well within.

I dusted out the HV section and tried running the laser. It mostly self-triggered, emitting various snorts and barks, and it only occasionally lased; but it was clearly a real machine and not just a pile of scrap. I disassembled the primary spark gap, cleaned it, and adjusted one of the electrodes. While doing this, I noticed that the back end of the housing was pitted and eroded from sparking, as was the aluminum plate it mated to. I wire-brushed these to clean them, but it was clear that I needed some highly conductive material between them to prevent a recurrence. There are some conductive pastes and greases, typically heavily loaded with silver, but I don’t think I have any. Fortunately, Gordon Garb was visiting, and he had some PMC (“Precious Metal Clay”) with him, in several forms. I used a small amount of the “glop” version, which comes in a tiny syringe, as conductive paste. PMC is well over 90% metal, in this case silver, and it worked quite nicely.

The trigger system quit while I was testing, and I eventually discovered that one of the chips (it uses two MC1455 timers and one other old-style CMOS device) had shorted out. I’ve got a temporary replacement in place, and the laser is now operational. I have been running it at about 15 kV, with about 175 psi of pressure at the inlets. (It uses nitrogen to pressurize its spark gaps.)

First, using. Here’s what happens when I focus it into a cuvette of 7-Diethylamino-4-Methyl-Coumarin (left) or Rhodamine 6G (middle, focused on the cuvette, and right, focused on the target in the background):

In the first diagram, we have something that I will politely describe as “an artist’s representation” of an ordinary low-pressure nitrogen laser pulse (heavy dark line), and the output of a dye laser driven by that pulse (blue line). The timescale is in nanoseconds — a low-pressure nitrogen laser pulse is typically 5 to 10 nsec long — and the output power units are totally arbitrary.

[Note, added on 24/25 January, 2006:

Here is an actual pair of oscilloscope traces (drawn over by hand because I was unable to select them in the Gimp), showing the output of a low-pressure nitrogen laser (magenta), and the output of an untuned 4-Methyl-Umbelliferone dye laser being driven by it (green). The oscilloscope was triggered by the output of the photodiode in both cases, which means that the dye laser pulse is shown a bit early here, and should actually be displaced a little to the right of its current locaation.

The dye here is using one of the cuvette walls as its output coupler; I was able to position a high-reflectance mirror on one side of the cuvette, but did not have time to make a full-blown tuning setup. The dye pulse has a much sharper risetime than I guessed when I drew the diagram on the left, above, and it ends sooner. So much for guesswork.]

The physical dimensions of the dye laser are such that the light inside it gets to make as many as 15 to 20 transits while the laser is above threshold. (Remember, light travels roughly 1 foot per nanosecond in air, so if the dye laser is 4 inches from end to end, light can make one full transit in 2/3 of a nanosecond or so.) This provides quite a bit of feedback, which is what allows you to tune the output wavelength — without feedback, the dye lases at whatever wavelengths it lases at, and you can’t do much about that except to tweak the concentration and maybe the solvent. With nitrogen laser pumping, you don’t even get to do a whole lot with the concentration, because if you don’t get the dye solution to absorb most of the pump pulse in a shallow region, you typically fail to reach threshold.

I should probably point out that a dye laser emits thousands of watts of spontaneous output. This utterly swamps anything less than thousands of watts of feedback. (If you’ve ever tried to tune a pulsed dye laser by injecting the output of a HeNe, you will have observed that there was no effect at all. Even if you put the dye cell inside the HeNe’s optical cavity, where there are watts of circulating power, it still won’t work.)

In the second diagram, we have a room-pressure nitrogen laser pulse in black, and a corresponding dye laser pulse in green. Mind you, I haven’t actually collected a dye laser pulse from one of these with a ’scope, so I’m partly making this up (“artist’s representation”), but it makes sense to me — even though the dye lifetime is 6 to 10 nsec (for things like Fluorescein and Rhodamine 6G), we’re continually extracting energy from the dye laser, so it won’t run for very long after the nitrogen laser turns off. The pulse from a room-pressure nitrogen laser is appreciably less than 1 nsec long, however, which means that the entire dye laser pulse is about as long as 2 or 3 transits of the cavity if the dimensions are the same as they are in the example at left. This fails to provide adequate feedback, and the dye laser’s output is not tuned.

There are two obvious ways out of this.

• 1. Use a low-pressure nitrogen laser. We have two of these, so it’s a viable option, but it would mean putting the LN-1000 on the shelf for now, and I’m not quite ready to do that yet.
  
• 2. Build a very small dye laser. I am seriously tempted. (In fact, as of May 26th, I’m working on this. See below for results.)
  
• 3. Nobody expects the Spanish Inquisition! (...Uhh, ahem. Excuse me. Couldn’t resist.) Here’s the real #3: Live with the untuned output. Under certain circumstances, this is just fine; but I think I want to be able to tune, so I’m likely to go with #1 or #2, or perhaps both. (Note, added later: for now, I’m going with #2.)

I should also note that it was wretchedly difficult to align the parts of the dye laser; but that’s not a show-stopper, just a pain in the neck. I’ve succeeded in the past, so I know it can be done. (Note, added later: in fact, the narrow laser proved to be fairly easy to align. Part of this is the fact that one of the mirrors gets aligned during construction, and ceases to be an issue once it’s right.)

Milan Karakas asked whether I could take a photo of the guts of the laser head. Unfortunately, it is sealed inside a metal box. (See the “Innards” section, below, for a look.) What I can do is take a photo of the reflection from the front of the cuvette, which provides an image of the plasma inside the laser and the profiles of the electrodes. Unfortunately, these things are a lot easier to see in person than in a photograph. This is about as good a picture as I’ve been able to obtain, so far:

This is a multiple exposure — the camera held its shutter open for a full second, and I had the laser doing perhaps 5 pps. A single pulse doesn’t provide enough brightness. What you’re seeing here, btw, is the fluorescence of the paper target, not the actual UV reflected from the cuvette. While I can’t really show it here, the brightest region in the plasma seems to move around a bit from shot to shot.

Here’s an untuned dye setup, and two pictures of the output — one taken in roomlight, and one in the dark:

As you might guess, the mirrors are not fully aligned. Even so, the dye is lasing fairly brightly.

I’m thinking about the design of a small cuvette with integral "max ref" mirror (actually a piece of overcoated aluminum front-surface mirror, probably with about 95% reflectance) — I figure that if I can get the transit time inside the dye laser down to 150 psec or so, I have a fair chance of tuning it. Whether I can actually bring that off remains to be seen.

(29 May, 2004)

Well, that worked. Here are some photos. First, the dye cell, with a grating standing to its left. It’s rude and crude, but it works. (I’ve revised it since this photo was taken; the page with the photographic tuning curve for 4-MU shows the new setup.) Second, the output with no grating — we’re using the first-surface mirror and the dye (2 or 3 cc of 95% ethanol, in which I have dissolved a small amount of 4-Methyl-Umbelliferone and a drop or two of strong ammonia-water); the window opposite the mirror is deliberately misaligned, so it can’t provide much feedback.

Now a few tuning photos. First, here’s what happened when I put Rhodamine 6G into the cuvette:

What I think happened here is that the gain in the Rhodamine is so high that it was able to lase in superfluorescent mode, in a range of wavelengths at once. The grating splits the output. (These photos were taken yesterday; I now have things arranged so that the grating is probably in its first order, and when I get a chance I’ll try Rhodamine again.)

Next, I’ve put a sort of photographic tuning curve for 4-Methyl-Umbelliferone on another page.

(02 June, 2004)

Last night, I aligned the grating. I hope I’ll be able to add some diagrams of what I think is going on with grating alignment, because it is a cute puzzle. I only just figured out how to think about this today, though, and haven’t had time to generate decent pictures yet. In any case, at this point I only have to do minor tweaking on the other axis when I change the tuning. This is a distinct improvement.

I suspect that the round-trip transit time in the dye laser is on the order of 150-200 psec (it depends somewhat on the position of the grating, but the thing is about 3 cm across), and I’m guessing that the dye continues to lase for some hundreds of psec after the end of the pump pulse. I don’t yet have a way to check that, but eventually I hope to. It will take a very fast photodiode and a very fast oscilloscope, though...

(02 June, 2004)

We have some Scientech power meter heads that we acquired on eBay, and today I put one in front of the LN-1000. It registered a small voltage, too small to look at easily, so I put my little instrumentation amp on it. (This is a 100X amplifier that I built some time ago; it lets me read a type S thermocouple with an ordinary digital multimeter, and I use it when I’m firing my gas test kiln.) With the amp in place I found that I could get a decent reading on the oscilloscope. After a few dry runs, and after I tweaked the laser for best output at 4 pps, I brought the head to the same output level with a power supply, simultaneously measuring the voltage and current with a pair of meters. It took 0.419 V and 10.23 mA, which says I’m getting about 1.07 mJ per pulse. (The laser was probably rated to produce 1.45 mJ when it was new.) Considering the fact that this laser was nonfunctional when we got it, and had real problems, I think we’re doing reasonably well.

At some point I may clean the second spark gap and the laser channel, but for now I think I’ll let things stand as they are. The output is well over 1 MW peak power, and it is just fine for driving the dye laser I’ve constructed.

I had to massage one of these (completely overexposed), and as I mention, I did serious tweaking on the shot into the end of the laser, showing the electrodes and what I think is the preionizer. I must also apologize for the greenish color of two of these — I had the camera’s white-balance set wrong.

The spark gap that I’m showing in several of these, btw, is the primary one; the secondary one is hiding behind the laser, at center rear. I haven’t opened it up, and I’ve discovered that the laser will operate through a fairly wide range of settings on it, so I’m basically leaving it alone for now. If you look at the end of the primary gap where it meets the aluminum plate, you can see the PMC I smeared on the joint to prevent further sparking and erosion.

The laser head in this machine appears to be nearly identical to the one in the PTI GL-3300 laser; here’s a link to their online manual for it. I think the spark gaps may be a bit different, and the rest of the design seems to have significant differences, but the head is clearly the same design.

Background and References:

My page about the Sci Am laser includes several relevant references.

Contact Information:

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

My phone number is +1 240 604 4495.

Last modified: Sun Oct 14 02:02:08 EDT 2007