Professor Mark Csele's Homebuilt Lasers Page

The TEA Nitrogen Gas Laser

Nitrogen TEA Laser Firing
A homebuilt TEA nitrogen laser in use in the Niagara College Laser Lab. The entire laser is 'open air' and is built into an aluminum chassis. The triggering spark gap is visible in the upper left corner as is the transverse discharge along the electrodes.


This class of lasers produces pulses of light using a Transverse Elecrical discharge in gas at Atmospheric pressures. Since atmospheric pressures are used, no vacuum system is required making this an inexpensive laser to construct. TEA configurations are used for some nitrogen as well as carbon-dioxide lasers. For a nitrogen TEA laser, it is possible to use 'open air' as the lasing gas since air is 78% nitrogen!


It is well known that nitrogen will lase in a TEA configuration (Transverse Electrical Excitation at Atmospheric pressure). A TEA nitrogen laser resembles a 'traditional', low-pressure N2 laser in having two long parallel electrodes with capacitors on either side and a spark gap however this laser operates without a vacuum pump (in some cases, such as a TEA CO2 laser, the laser discharge occurs at several atmospheres of pressure!). The gain of the laser is so high that in many cases the laser can use air as the lasing gas since air is 78% nitrogen anyway - the price is lower efficiency. As background reading be sure to view the Nitrogen laser page on this site first since this laser shares many of the same design principles.

The first issue with the TEA nitrogen is the pulse width. The molecular nitrogen laser is self-terminating ... upon discharge the nitrogen enters the upper lasing energy level then decays to a lower state emitting light at 337.1 nm in the UV. In a low-pressure N2 laser (e.g. the design outlined in Scientific American by James Small), about 20nS after the discharge begins the population of nitrogen molecules at the lower state exceeds that of the upper state and lasing terminates since population inversion, required for lasing, is no longer maintained. The lower level has a much longer lifetime than the upper level so lasing cannot be maintained continually like HeNe or other gas lasers.

In the case of nitrogen the problem is that the lifetime of the upper state, which really sets the width of the laser pulse in this case, is inversely dependent on pressure according to the formula [1]:

t = 36/(1+p/58)

where t is the lifetime of the upper lasing level in nanoseconds and p is the pressure in the discharge tube in torr. (ref: Patel, Rev. Sci. Inst., below)

Hence in a 'normal' N2 laser operating at 30-60 torr the lifetime of the upper lasing level is about 20 nS. In a TEA laser the higher pressure reduces the lifetime of the upper energy state to about 2.5 nS. The pulse width of the TEA laser is hence reduced to about 1 nS. Now, construction of a 'normal' N2 laser is difficult enough given that particular attention must be paid to keep inductances very low. Consider the capacitors used - although tempting one cannot use regular 0.045" thick PCB for an N2 laser since the self-inductance of the board is quite high which slows the discharge and hence it fails to work. In the TEA nitrogen laser the situation is ten times worse so that extreme measures must be taken to keep inductances low! Consider the characteristic impedance of a Blumlein transmission line given by [2]:

Z = (377/e1/2) * (s/L)

where Z is the impedance of the line in ohms, s the thickness of the dielectric and e the dielectric constant of the capacitor material, and L the width of the conductors in the transmission line.

The ultra-fast discharge required for a TEA laser dictates that we keep the dielectric as thin as possible which, of course, leads to problems with dielectric breakdown due to the high voltages required.

Using the first formula we find that the lifetime of the ULL is 2.5ns so that logically our transmission line must discharge within that time. This dictates a maximum length for the transmission line based on the speed at which it discharges [5]:
L = (c * tp) / e1/2

where c is the speed of light (3*108 m/s) and tp the length of the pulse. Since the later parameter must be kept to 2.5ns or less this dictates that the transmission line must not exceed 43cm where a dielectric with an e of 3 is used (Of course if polyethylene, mylar, or other materials are used for the caps this number will change accordingly). High dielectric materials will dictate a shorter width of the transmission line. Anything over this figure will result in energy being wasted since population inversion will end with the ULL lifetime.

Aside from keeping inductances low the second concern is distribution of the discharge. At low pressures, the discharge tends to smooth-out over the entire length of the electrode as required but at atmospheric pressures it tends to form a single spark or 'hot spot' at one point along the electrode. A fast-rising discharge waveform does help to a large extent however careful design if required to further reduce arcing. Several schemes exist to reduce this effect ranging from careful alignment of smooth electrodes via a micrometer [3] to a multi-segmented electrode using up to 27 small electrodes with 27 small capacitors instead of one electrode [4]. This is actually a reincarnation of a very old approach - one developed for the original N2 laser in the 1960's. The original N2 laser (developed by Heard) employed an electrode composed of many segments. Preionization may also be employed (see the bottom of this page). Commercially, many lasers use a preionization technique in which a discharge through a small spark gap prior to the main discharge ionizes gas in the laser channel.

Finally, electrode spacing on a TEA laser is necessarily small in order to keep the E/p ratio (the ratio of electric field to pressure and discharge length, measured in V/cm-torr) to a target value. Research [9] shows that the E/p ratio of nitrogen is between 150 V/cm-torr to 300 V/cm-torr. The actual E/P depends upon the pulse length (actually, the product of pulse-length and pressure called the Ptp parameter) - the E/P rises as pulses become shorter so the expected E/P ratio for a TEA laser would be larger than that expected for a low-pressure design with a longer pulse length.

Assuming a value of 200 V/cm-torr (an "accepted" value), an applied voltage of 7500 volts, and a partial-pressure of nitrogen of 608 torr (80% of 760 torr), the electrode spacing would be only 0.6 mm! The same voltage, employed with an electrode spacing of 1cm, implies a pressure of around 40 torr (as employed in a low-pressure design, outlined on another page on this site). It might be added that the E/p ratio of air is higher than pure nitrogen - possibly as high as 385 V/cm-torr. A TEA nitrogen laser running on air would be expected to have a smaller electrode spacing than the same laser employing pure nitrogen while a 'low pressure' design employing air (and having fixed electrode spacing) would likely perform better at lower pressures than the same laser using pure nitrogen gas.

For wider electrode spacings, the E/P ratio could be modified by the addition of helium (an inert buffer gas). Commercial TEA lasers frequently use gas mixtures of only 2% to 5% nitrogen (with the balance helium) keeping the partial pressure of nitrogen low (so the E/P of the nitrogen component is matched) allowing the use of high pressure gas mixtures with wide electrode spacings (up to 5 cm) ... see the description of such a laser at the bottom of this page.

A number of amateur laser constructors have sent me accounts of their TEA laser experiments. I have constructed two such lasers: the first a ridgid 'textbook' approach and the second a 'quick and dirty' approach made with aluminum foil and plasitc film which took only 30 minutes to construct. Bizarre as it may seem, the 'quick and dirty' laser outperformed the 'better built' one. If you are looking for a 'scratch built' laser project this laser is certainly the cheapest however don't be surprised if it fails to lase on your first attempt as it is somewhat 'touchy' to align the components. Construction and materials also must be done carefully to eliminate stray inductances which will slow the discharge and cause problems. The 'traditional' nitrogen laser is considerably more forgiving. As well, if you are planning on pumping a dye laser consider that the beam width is quite narrow. It is possible, but trickier than with a low-pressure laser as a pump laser.

Prototype Design

The TEA laser resembles the low-pressure Nitrogen laser in design with the exception that the entire transmission line must be optimized for a much faster discharge. The figure below outlines the cross section of the laser which may be compared that of the low-pressure Nitrogen laser.

Cross section of the TEA laser
Figure 1: Cross section of the prototype laser

This 'quick and dirty' prototype laser, which worked well as seen in the photo below, was assembled in about 30 minutes from commonly available components. It is known that the electrodes need to be very smooth and parallel so I chose to cut these from a long piece of aluminum floor edging. This product is sold in Canada at Canadian Tire stores under the brand name 'Shur-Trim' (# 344P) made by Drummond Metal Products. It is a 25mm wide strip of aluminum with bevelled edges - perfect for keeping the discharge elevated away from the capacitor underneath.

TEA electrodes

Two electrodes, each 24cm long, were cut from the edging. The sharp corners which result were filed smooth and the entire edge was sanded with 600 grit emery cloth to ensure it was absolutely smooth.

The electrodes were held to the capacitor below it using two 5 lb. scuba weights. To adjust the gap two plastic alignment rods were used (you do NOT want to get too close to the charged capacitors as they will throw a painful arc).

TEA Laser - Top View

The capacitors were made from aluminum foil and polyethylene film. The poly film was 6 mil thick: it was left-over 'Super-Six' vapour-barrier (required by Ontario building codes) available from Home Depot. A piece of film about 40cm square was cut and the foil attached by masking tape. The bottom plate was 30cm wide by 45cm long and protruded from one side for connection to the spark gap. The top plates were each 30cm long by 9cm wide. Again these were simply taped to the plastic film. The foil was regular Reynolds aluminum foil. In terms of materials, I suspect regular foil is a better bet as it is more flexible and tends to bend easily when electrostatic force pulls it towards the plastic film (when high voltage is applied). Thicker foil, or even aluminum plates, might not be a benefit here as they will undoubtedly leave air gaps which lower capacity and increase inductance of the transmission line.

TEA Spark Gap The gap was made from workshop leftovers: a small steel "L" bracket and an 8-32 screw with round-ball nut (The kind used to hold lamp parts together). It was taped to the top of one capacitor. The other end of the gap was the aluminum foil which protruded from the bottom of the capacitor. It is important that the inductance of the gap itself be minimized in order to obtain the fastest possible discharge [6] - as such sharp edges must be avoided like they are here.

The spark gap was set to about 3mm. It is important that it be set reasonably small since this gap sets the voltage at which the laser fires. If the gap is set too wide the voltage will exceed the breakdown of the rather thin dielectric and the laser will quickly fail.

TEA Spark Gap The power supply for this laser is a very simple one consisting of a 7.3kV/5mA transformer from an old Xerox copier, a variac to regulate the AC input, a high-voltage diode from an old TV, and a 100K, 13W wirewound charging resistor.

Results of the Prototype

The laser is seen operating here. It was observed to operate, with the spark gap set small enough, at a minimum voltage of about 5kV but operated much better when the gap was widened slightly and the firing voltage set to about 6kV. The gap between the parallel electrodes was set to 1.2mm on the front, 1.1mm on the rear. It is interesting to note that the laser would NOT work if the gap was too narrow ... at least 1mm of gap was required to allow it to lase.

Asymmertic output was found between emission from the rear of the laser and the front. The front output was estimated visually to be about 4 times more powerful. No mirror was used but a single mirror at the rear of the laser will certainly improve the output.

The operating laser as seen from the top. The gap between the two electrodes must be carefully adjusted so that the discharge occurs along the entire length of the gap, not just at one point. This is quite critical and requires a bit of 'fiddling' to get it to work. At first, a single spark appears atone end. Open thatend very gently until the spark begins to move towards the other end of the laser, distributing itself evenly along the distance. A bright blue spot should then occur on white paper placed in front of the laser (fluorescence from the paper itself: the actual output of this laser is UV and as such, is invisible).

The spark gap, btw, should be shielded as it is quite noisy and produces intense light. In the photo above a rag was used to shield it temporarily to avoid having the video camera lose focus.

Download a short movie showing the laser operating PROTOTYPE TEA LASER FIRING (467K file size). As the movie plays you'll hear it firing many times but will only see the output (on a white notepad) flash periodically. This is because the fast pulse may be missed on many frames as the camera records. A second video shows te PROTOTYPE TEA LASER FIRING - TOP VIEW (346K file size). Visible in the clip are the lead weights holding the laser together. The spark gap, on the right-side of frame, is shielded by a rag.

Improved Design

The first problem encountered was longevity of the capacitor dielectric. A single sheet of 6 mil polyethylene lasted only about 100 shots before breakdown occured. Two sheets of this material were tried as well. While this improved lifetime (it never did fail) it also decreased output quite drastically. This is due to the decrease in capacitance (and hence energy storage) as well as increased self-inductance of the capacitors. So far, the best performance has been obtained by using drafting mylar (polyester) film. A single 2.5 mil thick sheet has been used quite successfully at a low firing voltage of 5kV. The tradeoff to use a low voltage was made in order to allow the thinnest dielectric possible. By trial and error it has been found that the best performance has been obtained using this thinner dielectric (and hence lower voltage). It has also been determined that the optimal capacitor width is about 9cm on the side nearest the spark gap, and wider at 15cm on the opposite side. Increasing capacitance of the side opposite the gap produced a notable increase in output power but increasing the 'gap' side did little to affect the output. The gap was located near the rear of the cap.

A second laser has been constructed using the design principles above. It was built into a 17" by 10" by 2" aluminum chassis box. The box accomodates capacitors of 9cm and 15cm width and will features decent mechanical compression via multiple plastic screws to hold the electrodes firmly to the capacitor plates. Electrodes of length 24cm and 36cm can be accomodated. Longer electrodes improve output. As well there is provision for a rear mirror via a three-screw mount as part of the case. The small 5kV supply will be built right into this box to keep it compact and safe.

If attempting to build a TEA laser like this, use thin dielectric but one thick enough to prevent breakdown (this must be determined experimentally). As well, electrodes must be kept exceptionally flat to ensure it operates ... the chassis in the design seen here was bolted to a 3/4" piece of particle board to keep everything flat - the prototype was built directly onto a piece of flat board. Use a voltage of around 5kV and remember that alignment of the electrode gap is critical!

Top View of the second laser

The 5.0kV 200uA power supply was scavenged from an old laser printer and is tucked into the upper-left corner of the chassis. Taken from an old HP Laserjet, the module uses a 24 VDC supply (at 0.15A) and produces 5kV originally used for charging the drum. The spark gap is built from two pieces of plexiglas with a brass terminal which attaches to one capacitor. The gap was set to about 3mm (until it fired about twice per second) and the electrode gap adjusted for maximum output. The optimal gap was found to be 1.5 mm at the front and 1.4mm at the rear. Again, gap adjustment was crucial for proper operation. When properly adjusted any 'hot spots' were spread down the entire length of the electrodes as much as possible (although the photograph at the top of this page still reveals shows several 'hot spots'). For the most part, a nice dull purple glow was evident down the entire length of the electrodes. The electrodes are held down to the aluminum foil by many plastic screws.

Download a short movie showing the laser operating TEA LASER FIRING (633K file size). As the movie plays you'll hear it firing many times but will only see the spark gap and output (on a white business card) flash periodically. This is because the fast pulse may be missed on many frames as the camera records.

Improvements and Updates

Like the low-pressure laser on this site, one may argue that the laser should have been designed in a symmetrical manner with both capacitors of equal size. Assuming that the impedance of the spark gap is much higher than the transmission line (which it most certainly is for a laser such as this), each side of the laser should be equal in terms of size! This concept seems to be reinforced by some researchers [7] (e.g., Compact High-Power N2 Laser: Circuit Theory and Design, Schwab et al, IEEE Journal of Quantum Electronics QE-12, No. 3, March 1966, p.183) where the author essentially states that "Considering rise times of about 25ns, which are inherent to many reported N2 lasers, the travelling-wave concept becomes obsolete.". Of course this research was referring to a low-pressure design in which timing is far less critical but it does outline a potential flaw in 'overoptimizing' the design to include a triangular, smaller, capacitor on the sparkgap side of the laser.

Milan Karakas [5] (his page is on my LINKS page) is currently working to optimize his TEA design by modelling the transmission line. I will present results here as they are shared with me and develop

As well as the 'Blumlein' design presented here a number of amateurs have successfully built lasers using a Marx bank generator as the pump. See the LINKS section of this site for details

Alternative Electrode Options

The single biggest problem with the approach of using two single, long electrodes at atmospheric pressures are 'hot spots' which resemble an arc in the discharge. Critical alignment may reduce these effects but not eliminate them. Three approaches to fix this problem are (i) preionization of the gas between the laser channel, (ii) addition of helium in the discharge, and (iii) the use of multi-segmented electrodes. The first approach is used in commercial excimer lasers (see below) in which multiple spark gaps fire before the main discharge occurs. UV light emitted from the sparks preionizes gas in the laser channel. Milan Karakas (see the LINKS section) has used this technique on his homebuilt TEA laser, apparently with excellent results. The second approach, of adding helium, is also employed in excimer lasers when using nitrogen gas. Lumonics excimer lasers, for example, utilize a mixture of only 2% nitrogen and 98% helium. This is mentioned in the article by Bastings, et al, [2]. The third approach, using multi-segment electrodes, is described by Patel [1] in which the author uses an electrode with 5 segments. The impedance of his capacitors is stated as 0.0527 ohms/m. The dielectric is two sheets of 0.125mm thick material. Reported output power is 1.0 MW with 0.6 nS pulse widths. An article by Herden [4] presents a similar design utilizing 25 electrode segments i a 27cm long laser. Capacitors and electrodes were fabricated from self-adhesive copper strips and mylar foil 0.2 mm thick. Power outputs of 400kW with a FWHM of 0.3ns are reported using open air as the lasing gas.

Dye Laser Pumping

This laser is approaching the level where it can be used as a pump source for a dye laser. The main problem here is one with optics: the thin pump beam is difficult to focus onto a dye cell. Given the wide dispersion of the beam though it may not be as tough as once assumed. At a distance of 46cm from the front of the laser the beam is observed to have a width of 4mm. A lens from an HP Laserjet was used to focus this beam to a thin line which appears to be suitable for pumping a dye. The next step is to build a suitable dye laser with a small cell. This design will be based on my original design detailed on another page on this site.

Commercial TEA Lasers

Lumonics Excimer-500 Laser
While TEA nitrogen lasers are available commercially one of the easiest lasers to convert to a nitrogen TEA is an excimer laser. In the Niagara College laser lab I have done this with a Lumonics Excimer-500 which now runs from a mix of nitrogen and helium gases. Real excimers require a halogen gas (fluorine or chlorine) which is toxic and expensive - in our lab we allow students to mix their own gases using nitrogen and helium which is much safer and cheaper while giving the same experience as mixing a real excimer (i.e. nitrogen, like the halogen in an excimer laser, is used in small quantities and introduced using a 'boost bottle' arrangement). The SOP (Standard Operating Procedure) for this laser can be found on the SOP Repository. The SOP contains photos of the laser, gas mixing valves, as well as procedures for safe operation. This particular laser is a class-IV laser (given that the output of 400mW is in the UV region) and we use this laser to pump dye lasers (see the dye laser page on this site).

Lumonics-500 excimer circuit, Copyright John Wiley & Sons, 2004

The Lumonics Excimer-500 laser, like most commercial excimers, uses a single large capacitor (C1) which is discharged through a transverse electrode via a hydrogen thyratron. The laser features multiple preionizing sparks along the transverse electrodes shown as the gaps with capacitors C2. UV light from the sparks ionize gas in the laser channel which conducts current from capacitor C1.

The normal mixture for this laser is 2% nitrogen and 98% helium with an operating pressure of 30psig. Lasing the N2+ ionized molecular species at 428nm can also be accomplished by lowering the concentration of the nitrogen to 0.2%. It was found, experimentally, that pulse-to-pulse consistency (required for dye laser pumping) was improved greatly by lowering the operating pressure of the laser to 17-20 psia (where 15psia is atmosphere).

A few key references for TEA lasers:

[1] Compact high-power TEA N2 laser B.S. Patel
Review Of Scientific Instruments, 49(9), Sept 1978
A compact TEA laser which required operates at atmospheric pressures and hence does not require a vacuum pump. Uses multi-segmented electrodes.

[2] A simple, high power nitrogen laser
D. Basting, et al.
Opto-electronics, 4 (1972) 43-44
An optimized and compact high powered nitrogen laser (low-pressure) which uses two flat-plate capacitors (top and bottom). Reported powers are 1.2MW for a 30cm tube. Also reported is the use of air instead of nitrogen gas. Not a TEA laser but a good analysis of design principles provided

[3] Compact TEA N2 Laser E.Bergmann
Review Of Scientific Instruments, 48(5), May 1977, pp. 545
A simple TEA laser using two aluminum electrodes and 0.4mm epoxy-glass PCB for the capacitors. Includes a mechanism for keeping the electrodes aligned.

[4] Compact high power subnanosecond nitrogen and 'open air' lasers at 760 torr W. Herden
Physics Letters, Vol. 54A, No. 1, 11 August 1975
A multi-segmented Blumlein laser which works with open air as a laser gas.

[5] Private Communication with Milan Karakas. See the LINKS page for his page in which a TEA nitrogen laser is optimized

[6] Spark gap power switching circuit for ... plasma gun Glenn Kuswa & Charles Stallings
Review Of Scientific Instruments, Vol 41, No 10, 1970, pp. 1429
Details construction of low inductance (7nH) spark gap switch which may be used for N2 lasers

[7] Compact High-Power N2 Laser: Circuit Theory and Design Adolph Schwab & Fritz Hollinger
IEEE Journal of Quantum Electronics, QE-12, No. 3, March 1966, p.183
An excellent analysis of transmission circuit theory as applied to a low-pressure nitrogen laser (equally applicable, though, to a TEA design).

[8] Pulsed molecular nitrogen laser theory Edward T. Gerry
Applied Physics Letters, Vol. 7, No. 1, 1965, pp. 6
Examines excitation mechanisms & saturation power output of N2 lasers

[9] Investigation of the Paschen curve of nitrogen ... D.W. Scholfield, et al
Journal of Applied Physics, Vol. 76, No. 3, 1994, pp. 1469
The Paschen curve for nitrogen, and other gas parameters, are investigated from 0.336 to 685 torr.

See the LINKS section of this site for links to many homepages detailing construction of TEA as well as other lasers.