Joss Research:
The Molectron Nitrogen Laser
Tuning a Nitrogen-Pumped Dye Laser

[Note: This page is largely about a commercial laser. If you are interested in building your own, please see the “DKDIY” page instead.]

In the Spring of 2002, we bought a used Molectron N2 laser on eBay. Here are a few photos of the case & controls; if you want more detail you can click any of the small images for a 640x480 px version, or use the text links below the small images.

1600 x 1200 px             1600 x 1200 px             1600 x 1200 px             1600 x 1200 px

It proved trivial to bring this machine online, when I finally got to it, in the Spring of 2005 (argh): I removed a broken cable-tie from inside the case, powered it up, and adjusted the thyratron reservoir voltage and the mirror alignment.

Here are two fairly representative traces, taken on 3 June, 2005 using a Tektronix 7104 scope, a 7A26 (if I remember correctly) plugin, and a nice fast photodiode sent to me by Howard Davidson a while back:


[Note, added on 2006 January 22 and 24: I have taken more photos of ’scope traces, this time using a 7A29 plugin, which is faster than the 7A26 or the 7A19. There is more electrical noise, or perhaps the faster plugin is simply doing better at showing us the noise that has been there all along. Here, anyway, is a representative single trace, and an exposure showing perhaps a dozen; these demonstrate a little problem, which I’ll say more about in a moment.


My apologies for the fuzziness. I’m not sure I had the focus quite right, and the camera shook a bit when I pressed the shutter button. When I get a chance, I may try to improve upon these.

Note the pulsewidth, btw, which appears to be approximately 8 nsec FWHM. I am not really sure about this; the rather long fall time worries me. Note, also, the fact that the trace at the right edge is almost half a box higher than it is at the left edge. I am beginning to think that there is an impedance matching issue here.

(24 January, 2006)

I found a piece of RG-58A/U cable about 8 inches long, and replaced the previous cable with it. RG-58A/U is a 50-ohm cable, and should match both the detector head and the scope input impedances fairly well. Here are the results of the change:


As you can see, the long tail has disappeared, and the trace at the right edge of the screen is at the same height as it is on the left edge. The FWHM pulsewidth is now revealed to be no more than about 6.5 nsec. [Let that be a proper caution to you.]

In any case, once I had the machine up and flying I decided to run the usual dyes with it, in the process of which I learned a few things....

Tuning I: Setup and Adjustments

(mid-May, 2005. NOTE: You can click any of the small images to get an 800-x-600-px enlargement, except for the photo of Milan Karakas’s dye cell, which is 1024 x 768. For any of my photos, you can get a 2272-x-1704-px image by changing ".8c." in the filename of the 8x6 large image to ".22c."; Milan’s 1024 x 768 is the largest size currently available.)

Here is a photo of the dye laser setup, with the grating moved off to the left a bit, and the nitrogen beam focused to a line on the front of the dye cuvette. (Ordinarily, I position the grating considerably closer to the cuvette.) Yes, the cuvette holder is rude, crude, boorish, and socially unacceptable. It’s a lot better than it was, though, and I’m still working on it.

The mirror and grating are tilted to align them with the region of dye that is being pumped by the N2 laser, rather than with the walls of the cuvette. If I remove either or both of them the dye will continue to lase, using the reflections from the cuvette walls as its feedback source.

Here’s a photo of a dye cell that Milan Karakas built, being pumped by one of his nitrogen lasers (the blue glow in the background is the nitrogen laser; we’re looking at the back of the dye cell). You’ll notice that the windows are several degrees away from perpendicular to the dye laser beam, and that some reflections are visible. I believe that the dye here is R6G.

When I have time, I will provide a set of photos illustrating mirror adjustment, which is most easily performed with the grating removed. In the meanwhile, here’s a slightly different setup (upper), showing the cuvette lasing with no external feedback (left lower) and with a mirror (right lower):


(“Mr. Hip” cuvette holder notion provided by the fiendish and unregenerate Miss LisaJulie. Mr. Hip’s footstool and couch by Meubles Bazilians de Paris, depuis 1614. Next up, micrometer adjustment of Mr. Hip’s butt for precise control of the cuvette tilt angle...)

I think the defocused vertical band of green in the lower two photos may be superluminescent lasing, in which the medium operates with such high gain that it has substantial single-pass output — it lases without any feedback. (A nitrogen laser without any end mirror and with misaligned end windows operates in this mode. The reason why the nitrogen laser produces a beam instead of a cloud like the one you see here is that the channel of the nitrogen laser is many times longer than it is wide.) I stacked the deck for these photos by adjusting the position of the focusing mirror (not visible in any of these photos) so that the excited region of dye was only a few millimeters long and at most a millimeter across and was, at least vaguely, in the shape of a very short horizontal stripe. Because of that, the superluminescent lasing [assuming I’ve identified it correctly; see Dr. Rüdiger Paschotta’s excellent encyclopedia] was brightest across from the cuvette and at about the same height off the bench, or perhaps a bit higher — the excited region was only approximately horizontal, and its shape was not a very straight line. Also, as I say, it was quite short.

The bright lower dot is the dye using the cuvette walls for feedback, and the not-so-bright dot above it in the third photo is essentially how I tell when I have the mirror correctly aligned, before I put the grating into place. The reason why the spot isn’t particularly bright is that the light has only passed through the dye twice. As you can see, however, even just a second pass is enough to produce a more-or-less-directed beam, though that’s partly because the mirror is a few millimeters away from the excited region. (In the abstract, putting a [perfect] mirror in the center of a fully symmetric excited region wouldn’t change the output at all.) The further away the mirror is, the more it narrows the beam; but of course with N2 pumping you don’t have much time, so you want to put the mirror as close to the dye as you can, within reason.

Adjusting things to achieve the desired condition (mirror aligned with cuvette and excited region; grating or second mirror aligned with cuvette, excited region, and first mirror) can be tricky, and I hope to put a more detailed set of photos in place when I have time.

Note, btw, that none of these kinds of lasing actually prevents any of the others from occurring: in the third photo, the dye is lasing all three ways at the same time. This has implications for tuning — you want to provide as much feedback as you can, in the hope that tuned lasing will use as much of the available excitation as possible. This also helps explain why any attempt to tune a pulsed dye laser by injecting a HeNe beam or a laser pointer beam into it is guaranteed to fail: the dye is putting out thousands of watts of “junk” light, and it will casually ignore your feeble attempt to distract it from amplifying its own spontaneous emission. It also explains why you want to put the mirror and grating as close to the cuvette as you can, if its walls are parallel to each other: the round-trip time for light inside the cuvette is less than 100 psec. The more bounces it can make the higher the gain will be, and the less energy will be available for tuned output. This is an argument for making your own cuvette, so that you can deliberately misalign its walls slightly. (I have built cuvettes of that sort, and when time permits I will try to add a set of photos to demonstrate the differences.)

The cuvette, btw, contains Fluorescein and a small amount of 7-Diethylamino-4-Methyl-Coumarin, in 91% isopropanol to which I have added a drop or two of very strong ammonia. Fluorescein on its own doesn’t absorb particularly well at 337 nm, so even though it has very good quantum efficiency, it’s difficult to lase with nitrogen laser pumping. I use the Coumarin, which absorbs the pump light extremely well and emits at a wavelength that is more readily absorbed by the Fluorescein, to help it along. This technique, while of limited utility, clearly works with some dye pairs.

Here are some photos to illustrate that. First, Fluorescein in 91% isopropanol. The solution doesn’t lase. Second, I add a drop of concentrated aqueous ammonia, and the fluorescence gets a bit brighter, but the absorption depth is still too large. Third, I add more Fluorescein. The absorption depth decreases, but it is still too big, so the solution doesn’t lase, or perhaps just barely begins to reach threshold.


Fourth, I add a small amount of 7-Diethylamino-4-Methyl-Coumarin, possibly not even enough to lase by itself, and the resulting solution definitely lases — note the laser speckle at the bottom of the image. (This last photo is actually a different batch of solution, which shows the effect more clearly than the original batch did.)

Notice that the middle of the spot is white and thoroughly overexposed in several of these, particularly the last one. Even though I was looking at the cuvette from an angle, the sensor in my camera was damaged by the beam. “Do Not Stare [Or Even Look] Into Laser With Remaining Camera.”

(24 January, 2006)

Here is a pair of oscilloscope photos, showing the pulse from fluorescein, but without any external mirrors. This is a mixture of superluminescent lasing and reflection from the cuvette walls.


This is a very old dye solution, and probably isn’t fully representative. Even so, it shows risetime of about a nanosecond and three quarters, and pulsewidth of perhaps 3 nsec FWHM, which is close to what you would expect if the dye is lasing only when the nitrogen laser is near the peak of its output. (Remember that as long as the dye is lasing we are not going to see its lifetime. Lasing depletes the population more quickly than fluorescence does, at least with the common dyes.)

For the sake of comparison, here is the nitrogen laser’s output (purple) superimposed on the dye laser’s output (green). Because the triggering was the same for both of the original photos (it was derived from the output of the photodiode), the dye laser pulse was originally too far to the left. I have moved it, but of course I had to guess at where to place it, so you should take this image with a grain or two of salt. I had to trace over both photos by hand, btw, because they were not bright enough to select with the Gimp; this accounts for the slightly jagged look.

(Here, if anyone cares, is an earlier version, for which I also tweaked the timing:

I suspect that the dye trace [again, green] is still a bit too far to the left in this photo.)

The next set of photos illustrates vertical alignment and misalignment. You can see the tuned spot most sharply in the middle image, where the vertical axes of the grating and the mirror are aligned well with each other and with the pumped stripe on the cuvette. (The dye here is Rhodamine 6G.)


You can see that when the mirror and the grating are aligned best, the untuned spots are dimmest. This will be visible again in the tuning curve below.

Tuning II: Results

Here’s a visual tuning curve for Rhodamine 6G:


As mentioned above, you can see that at the peak of the tuning range, shown in the 4th through 6th photos, the untuned beams above and below the tuned beam (which are from reflections off the walls of the cuvette) are much dimmer than they are at the ends of the tuning range. If I built a setup that provided even more feedback the effect would be stronger, and the tuning range would be at least slightly broader.

Note, also, that the relative position of the tuned beam changes with respect to the positions of the untuned beams. This and other effects make it necessary to adjust the alignment of the grating as I tune, and sometimes makes it necessary to tweak the horizontal alignment of the mirror slightly in order to get maximum output.

A final note, added in proof (as it were) on 10 June, 2005: I went to take this laser off the bench so I could put a different one into place (see this page if you want more info about that project), and noticed that the nitrogen hose was lying on the floor. When I put it into the port on the laser, it fell right back out again. This strongly suggested that I’d been running the laser on air. (I was wondering why I got best operation at the lowest pressure the control electronics would allow!)

I put the hose in and tightened it up. Sure enough, I now get best operation around 45 Torr rather than 32, and best operation with Fluorescein appears significantly brighter by eye than it did when I was using air, so I’m probably getting two or three times as much power out as I was. Such is life. (It pays to notice these things early, rather than late, but it’s certainly better to notice them than not.)

To the top of the LASERs section

To the top of this [updated] mirror

To my current research homepage


This work was supported by
The Joss Research Institute

Contact Information:

Email:, where you can replace a with my first name (just jon, only 3 letters, no “h”) and b with joss.

Phone: +1 240 604 4495.

Last modified: Tue May 9 12:50:33 EDT 2017