COPPER VAPOR LASER CONSTRUCTION
Chemical and physical consideration
and introduction into the
Cu(II)Br laser construction
by Gruber B.
updated: Jan 18. 2010
| visits |
In
all pure copper vapor lasers, a heating system that provide
temperatures around 1400°C are needed in order to get a proper laser
output. The output wavelengths are 510.6 nm and 578.2 nm in pulsed or
quasi cw operating mode at higher repetition rates. By the use of
copper halide instead of pure copper as the active medium reduces the
requirement for the heating down to temperatures around 400 °C. This
lower temperatures are much easier to handle in terms of materials
involved for the resonator and cooling requirements for electrodes and
windows or mirrors. The benefit of the copper halide however has the
disadvantage, of a more complex power supply. To
dissociate the copper from its halogen atoms, a double pulse power
supply is needed that is capable of firing two HV pulses within the
recombination time of the copper halide molecule. If the power supply
can handle several kHz, a delay unit firing the second pulse is not
needed. HV pulse power supplies operating in the kHz range require fast
switching devices such as hydrogen thyratrons. The article describes
two topologies of pulse power supplies including two different copper
halide laser constructions.
Pure
copper vapor lasers need high voltage pulses for operation. in a very
simple construction the pulses can be produced mechanically via a motor
driven chopper. Even if this may sound simple it really is not. The
chopper construction has to widthstand the hot arcs during switching as
well as the high speed rotation and maintain proper isolation during
operation. The latter is mainly limited by sputtering effects -
depending on the athmosphere surrounding the electrodes. A much
better way is to use gas filled switching devices like ignitrons or thyratrons,
but this are very special devices and therefore they are
expansive too. A cheaper
way is to use spark gaps as switches. Compared
to the pure copper vapor laser (CVL) that require one HV pulse for
stimulated emmission, the copper halide laser (CHL) needs two.
Therefore the power supply design is complex. The delay time of the two
HV pulses is in the range of 100 µs to 280 µs, depending on the halide
used.
In any CHL design the topology of the powersuply is determined by the type of HV source defined by its voltage and power rating. One way is the use of neon sign transformers (NST), that deliver voltages in the 12-15 kV range at ~50 mA. Some types can be switched in parallel to increase the total output current. Another possibility is the use of Microwave oven transformers (MOT) rated at ~ 2 kV and 500 mA (~1kW). To increase the voltage of MOT, a voltage doubler, that is made of an series of parallel connected diodes, can be used. To generate high voltages at very low rise times a Marx Generator (MG) can be used. The principle of a Marx Generator is to charge a stack of capacitors in parallel and switch them in series if the first SG breaks down because of overvoltage or triggering.
The below circuit use one MOT (Ratio: 1:12, 1kW) and a voltage doubler that charge two serial connected high voltage capacitors (BOSCH: 1 µF / 6 kV) to about 6.5 kVDC. To further increase the voltage and getting the ability to trigger the pulses, a triggered Marx Generator is connected to the HV DC stage via current limiting resistors R10 and R11. They are needed to prevent to much stress on the pulse capacitors (in this simple design C1 ... C3 are made of a ceramic capacitor array). This resistor values can be optimized in order to match the circuit towards higher repetition rates. The inductors L1 ... L4 consists of two 20 µH / 8 A inductors in serial connection.
One of the most important component of this Marx Generator is the Triggered Spark Gap TSG1. The construction of the triggered and not triggered spark gaps (SG) is based on a glass tube with two flat brass electrodes on each side. One electrode has an isolated trigger electrode, whereas the other electrode is soldered to a copper pipe for the nitrogen gas support. The glass used for this spark gaps should have high transmission in the UV spectrum, for better synchronisation of the individual, optically coupled spark gaps that are fired by voltage breakdown.
The figure shows the whole Marx generator without the trigger unit.
Technical Realisation Of The Marx Generator:
The main part of the generator is the switching device. In this first project adjustable rod spark gaps were used instead of thyratrons due to cost constraints.
Spark Gap Design
Nitrogen filled, triggered spark gaps with adjustable main- and trigger electrodes.
Each spark gap consists of two brass main electrodes Œ with hard soldered copper pipes � , which are working as electrical connection and to ensure the gas support. To prevent short circuits, comming from electrode sputtering near the center of the inner glass surface, the main electrodes have a smaller top electrode area Ž . The operation with nitrogen improves quenching and thus allows higher repetition rates. Therefore the electrodes are sealed via O-rings � . On the outer side each electrode is mounted on a flange � , including three holes for the adjustment screws ‘ . The triggered spark gap has an additional tungsten trigger electrode ’ , which is isolated by the use of a ceramic pipe “ , which itself is mounted into a hard paper rod ” .
The following image shows the different assembled spark gap models, which are described above. The upper spark gap is a triggered model, the others are overvoltaged types. After the assembly leakage was tested with 3 bar air pressure under water, the spark gaps have been adjusted carefully using a caliper for a mechanical adjustment of a breakdown voltage of ~10 kV, depending on the pressure inside the sparkgap.
Different types of spark gaps
On top the triggered type, the others are simple non triggerable types
For the complete double pulse power supply, a second Marx Generator of the same type is needed. Both are connected via a triode spark gap (SG5) needed for AC coupling of the two generators to the resonator tube of the laser.
The figure shows the principal arrangement of the three decoupling electrodes of SG5.
There is also a gas flange for the nitrogen gas support, to improve quenching.
No synchronisation is necessary for this triode so the whole package can be made of ceramic.
Triggerelectronic
The trigger electronic itself consists of three 555 timer IC´s. The first timer works as oscillator up to 6 kHz. The following figure illustrates the delay unit, which is realized with U1 working as the delay pulse generator and U2 for the pulse width regeneration.
The delay time can be varied roughly by switching CDELAY and fine by trimming RDELAY between 1 µs and 100 ms. The output pulse width of the second ignition coil can be adjusted with CPW and RPW. Normally the pulse width for both trigger pulses is 20 µs, but it can be necessary to adapt it due to different ignition coil models or different firing preferences. The pulse width is 20 µs and the delay between the two pulses is 150 µs. Due to different dissociation times of different copper halides, delay times from 100 µs to 280 µs can be set. The power driver part of the trigger unit (not shown) use two 2N6547 fast switching transistors in darlington arrangement with a BD249. The ignition coils are electrically connected via a HV diode to the trigger electrode. To isolate the power stage from the trigger electronic a fast HCPL2530 optical coupler is used.
The figure shows the delay unit for the two trigger pulses
Spark Gap Arrangement In The Marx Generator
After evaluating and adjusting all spark gaps, they can be mounted in the Marx Generator. In order to keep the impedance of the Marx Generator as low as possible and the design compact and flexible the below design is used to form multiple stacks of Marx Generator stages. The relatively large size of the spark gaps determine most of the shape of the whole device. Using strong PCB frames, working once as mechanical basis for the spark gaps and as carrier for the capacitor array is a solution for both problems. Each frame consists of two clips, that electrically connect each spark gap electrode with the transmission line of the PCB. The frames also provide a short electrical path to the capacitor array and the inductors. Modular expansions of the whole system can easily be done by adding more frames.
Fully assembled frames with spark gap, capacitors and inductors therefore form one generator module. The following figures sow the arrangement of the spark gaps in the Marx generator circuit.
The capacitor array is expandable by adding more frames
Optimization of the Marx Generators is done by decreasing the current limiting resistors R10 and R11 or replace them with large inductors (about 10 mH) which are able to handle the charging current at the repetition rate. Increasing the pulse repetition time result into an reduced lifetime of the standard ceramic capacitors used in this project. To improve the time constant of each Marx Generator stage, the coupling inductors L can be changed. Using resistors instead of inductors might also be an option, depending on the impedance of the final design. Getting lower rise- and fall- times, the SG´s are filled with nitrogen at ~1 bar. This provides a better quenching and also cooling if gas circulation with gas cooling exists (not used in this design).
Using three separate microwave transformers with voltage doublers, connected to 3-phase mains in star arrangement, gives much better DC level and lowers the output ripple at high repetition rates.
Spark gap tests:
High voltage pulse capacitors are mostly large devices with about 2-5 cm diameter with wired (Stripline) or screw connectors. In the case of this Marx Generator, which should be as compact as possible, small wired ceramic capacitors are used.
The usage of many small capacitors instead a large one has some advantages if power is not the main reqiorement of your project. The small ceramic capacitors are less expansive compared to special HV pulse capacitors, if one capacitor of the array fails, it can be easily replaced by another one. Also the heat distribution in the capacitor array is much better than in one big capacitor. Therefore air cooling at high repetition rates will be sufficient. The most important fact is that the high peak current is distributed over the capacitor array. Due to the skin effect, which becomes important at those rise- and fall times, a greater surface of the connectors, result in lower impedance and therefore in higher peak currents.
After the successful construction of all pulse capacitor arrays, the developed spark gaps were tested, by switching the energy of the capacitor array into a short circuit. That seems to be the maximum stress condition for all components. Once the overvoltaged spark gaps succeeded, the triggered model was tested with the same result. Careful adjustment of the trigger electrode was necessary, to minimize the misfiring of the switch.
The Laser head consists of the resonator
rod mounted between the two brass electrodes that have itself one copper pipe
for the gas support or the vacuum and another bigger copper pipe which goes to
the output windows. This copper pipes works also as cooling area for the copper
halide vapor so that it condenses and don´t reach the output windows.
Output windows, which are adjustable mounted, should be high optical quality types with an transmission maximum between 500 and 580 nm, which includes the two lasing wavelengths of the CHL. An aluminum coated mirror at one end reflecting the beam once so that the beam leaves the laser head at the output window on the other side. This HQ window has a reflection of about 2% thus forming a resonator. CVL and CHL have high gains therefore no resonator is needed to operate, because of the high photonic emission. Both out coupling windows are mounted at a tilt angle of 2° on a brass disk with a center hole.
Version 1:
Copper halide lasers need about 400 °C to operate. This is far below the 1400° C which pure copper vapor laser need and therefore THE reason for CHL. In order to build a compact resonator (version 1) a heating spiral was used. This spiral was in direct contact with a second quartz tube, which covered the resonator rod. Each quartz tube had a 2 mm wall and the whole construction isolated with 4mm quartz glass and a 1 mm air gap between the two tubes, kept in place via mica. The following image show the heating stress test. To compensate the length expansion of the heating wire during operation, springs are mounted on both ends of the coil.
This design had the disadvantage, that the heat ionizes the quartz tube and therefore leads into higher conductivity of the glass wall. Since the high voltage power supply is switched on, the sparks touches the inner wall if there is any turbulence or instability of the plasma. If this happens, tiny current flows through the wall into the heating wire, which has near earth potential, compared to the high voltage of the plasma. The plasma channel collides with the amorphous quartz molecules and heats it up to the melting point of the glass. After the melting point is reached, a tiny channel develops which leads to a leakage of the system.
This problem was solved in the new resonator model, which uses infrared heating rods in 25 mm distance from the outer surface of the laser rod.
Version 2:
The heating is done by three 600 W IR heating elements, which are arranged in 120° angle around the tube. Around the heating elements a ceramic housing forms the constant temperature oven. The oven contains also a temperature sensor and insulating material.
The most favorable conditions for obtaining output in the heated regime were realized in a gas discharge tube with an internal diameter of 2 cm and an active length of 60 cm. In this tube, lasing occurred in a mixture of copper bromide vapor and helium. The following figure shows the plasma test of the first resonator model. The color of the plasma tended a little bit into violet due to the higher vapor pressure of 30 torr.
Version 3:
Building of another resonator model improved the tightness and isolation of the infrared heating elements. The reachable pressure was about 1 torr, using a two stage mechanical rotary pump. After the system was flushed with helium and evacuated below 20 torr, gray-white sparks could be seen therefore leading to the assummtion that no air was in the system. The heating isolation required a resonator whith more free space around the plasma tube for the infrared heating elements. The new design had also expanded cooling zones for the condensation of the copper halide vapor, before it reaches the surface of the mirrors.
Complete System :
After all described components are working properly, a complete copper vapor laser system can be build. But even if a new clean system is used, there are many steps necessary to get this laser system running. The following explanations describe the startup and shutdown procedures.
In the following figure all parts of the whole CHL system are shown. The three main blocs around the laser resonator itself are the pulse generation-, temperature control- and the vacuum module. The two Marx generators MG1 and MG2 are DC sourced via two separate high power transformers with voltage doublers, which are itself supplied from 400V/3 phase mains. The digital part of the optically isolated trigger electronic TE switch the darlington stage of the power unit PU. Triggering of the first stage of the Marx generators MG1 and MG2 is initialized by the ignition coils IC1 and IC2. Both Marx generators are connected to the laser head through a three-electrode spark gap SG5. A standard PID controller for temperature regulation by switching the infrared heating elements on and off. The vacuum system consists of all magnetic valves, flow limiters and gauges necessary for the helium and nitrogen gas support for the laser head and the spark gaps.
Preparations:
The main part of the copper halide laser is the active medium, the halide itself. Due to the nature of halides, they are extremely hydrophilic. Therefore it is necessary to dehumidify the halide powder in a special way. First the wet halide powder must be heated in a scientific glass up to 120° C to get most of the water out of it. After this the powder is placed in the middle of a glass tube, which has a nitrogen support connected to a two-stage filter on one side followed by silica gel and is connected on the other side to a vacuum pump. By heating the glass tube up to 120° C the rest of the water vaporizes and will be carried out by the nitrogen through the vacuum pump. A duration of 15 minutes for this procedure seems to be enough and the now dry halide powder can be filled into the resonator directly out from the drying apparatus into the also 120° C heated resonator. To fill the dry halide into the resonator, the whole device must be raised up until the powder slips into. The following figure illustrates the drying apparatus.
In order to reach the low pressure, which is necessary for the
lasing action to take place, a quite complex vacuum system is necessary. The
functionality of this system is the support of the helium buffer gas and the
nitrogen for flushing. Using two stage vacuum magnetic valves, it is possible to
obtain a full automatic startup and refill procedure. The next figure
illustrates the whole vacuum system, which was used for both resonator models.
Magnetic valve Y1 controls the path to the 2-stage vacuum pump. The vacuum is
measured with an ordinary vacuum gauge. Y2 switches the helium gas flow and Y3
the nitrogen, which is used for flushing the system with dry nitrogen to prevent
humidity distortion of the halides. The nitrogen passes a filter and a dryer,
which heats it up to 120 °C. To minimize the turbulence during the filling
phase, restrictor valves are inserted. For the helium restrictor valve a
rotameter indicates the gas flow.
The vacuum system is realized by using a 2 stage mechanical vacuum pump (Leybold), which is able to reach 1 torr.
Leybold Trivac D 1.6B
To compensate the pressure increasing after heating the copper halide to 400° C, the vacuum pump should be able to reach less than 1 torr. A better vacuum equipment including a 2 stage mechanical pump and a turbo molecular pump solve this problem. Those low pressures are measured by convectrons (Granville-Phillips).
Pfeiffer Turbo molecular pump (TPH055) and vacuum gauge (Series 275 Mini-Convectron Gauge 275876)
The next project target is a more powerful system that operates reliable for many hours. The topology of the new HV power supply is based on a resonant charginge circuit with a thyratron switching element sourced by a professional HV lab power supply (FUG 0-35kV, 7000W)
FUG 7kW HV supply and arc test at maximum load
Different Hydrogen Thyratrons used in this project
Details on the new project will follow soon. In the meanwhile some impressions of other projects are shown below.
1000 mW DPSS 532 nm, 200 mW Diode lasers 650 nm in operation
Er:YAG laser rod and 400W cw Er:YAG laser resonator with water cooling (double elliptical pump chamber)
Nd:YAG laser 50W CW @ 1064nm, Q-switched (max: 20kHz), SHG: 532nm & THG 266nm @ 3,3kW/60ns. Resonator length 1m THG BBO outside resonator
Nd:YAG - SHG 532nm @1kHz, 20A lamp current cutting a hole in 0.7mm stainless steel
Dye laser flow through cuvette with Rhodamine B, longitudinal pumped with 532nm
Video of the new CVL resonator tube - first tests with air (m4v format, 15MB)
