MOPA lasers are pulsed lasers with very high power outputs. This discussion is limited to lasers for holography. MOPA stands for Master Oscillator with a Power Amplifier. The master oscillator power is independent of the final output power allowing for it to be designed with a Single Longitudinal Mode. After the Master Oscillator there is a Power Amplifier consisting of at least one stage but usually many stages.
Warning:MOPA lasers are often Class IV LASERS! Do not undertake the design and construction of one until you understand the electrical and electromagnetic safety issues! See Laser Safety for some introductory concepts.
- 1 Master Oscillator Design
- 1.1 Resonator Configurations
- 1.2 Aperture
- 1.3 Lasing Medium
- 1.4 Flash Lamps
- 1.5 Cavity Design
- 1.6 Polarizing Elements
- 1.7 Q-Switching
- 1.8 Output Couplers
- 1.9 Factors Effecting Spatial Mode
- 1.10 Factors Effecting Longitudinal Mode
- 2 Power Amplifier Design
- 3 Frequency Doubling
- 4 Filtering Out Un-Wanted Frequencies
- 5 Safety
- 6 Further Reading
- 7 Supliers
Master Oscillator Design
The master oscillator, as a minimum, consists of a lasing medium, a Q-switch, a highly reflecting mirror (HR mirror) and an output coupler (OC mirror). A master oscillator for holography will also have a polarizing element, Etalons, resonant reflectors and an aperture. The design of the Master Oscillator is of primary importance as is defines the limit of the properties of the final output beam.
For holography we want the Master Oscillator to have Single Longitudinal Mode and TEM00 temporal mode. We want a short pulse but if it is two short it will limit our coherence length. In considering different resonator designs it is useful to make a spread sheet of the different parameters. Ron Michael has created a remarkable one. MOPA Spread Sheet.
Insert equations to calculate limiting coherence length and maximium movement for a given pulse width.
Deisgning the resonator configuration controls a great deal of the parameters of the laser system. Not the least important of which is the reliability of the alignment. Since holographers are unlikely to have an unlimited budget the resonator design will become a compromise between available surplus parts and a stable design.
Important paramters controlled by the resonator:
- TEM mode (Spatial Mode)
- Coherence Length (Logitudinal Mode)
- Input power to the Amplifier section
- Alignment stability
It is important to remember that a solid laser rod acts as a negative lens during the intial flashlamp pulse due to thermal expansion where the outer radius absorbs more light intially. After repeated firings (faster than the thermal relaxation time) the outer radius of the rod will begin to cool faster than the center and therefore develop a positive thermal lensing effect. Some rods like ruby also will under go index of refraction changes due to inversion levels and therefore also have a lensing effect due to pump non-uniformity. In ruby dual flashlamp arrangements for example this effect can produce greater divergence in the flashlamp axis.
This is a popular way of enhancing spatial mode discrimation by the use of two spherical mirror configurations or 1/2 of that which is a plano concave arrangement. TEM mode discrimation is greatest for the confocal arrangement and least for the plano/plano arrangement. The Plano Concave is a good intermediate with typical g values of 0.8 to 0.9. g=1-L/R where L is resonator length and R is mirror radius. Typically the aperture is inserted next to the HR mirror radius and the plano OC can be etalons etc for longitudinal mode control.
This is an example of a un-stable design.
When an Etalon is used as an output coupler in a resonator configuraton, then it is generally called a resonant reflector due to the Etalon/multiple reflections that produce resonance with some modes enhanced and others diminished based on the Etalon thickness and index of refraction both of which can be controlled by temperature tuning. These RR are uncoated because of the high internal power density within the resonator due to the Fabry Perot reflections. Most RR are either single plate or multiple plate with air spacing to provide mode discrimination within the gain profile of the laser. Thin plates provide wide peak separations and wide separations provides thin peaks.
Inserting a aperture into the Oscillator cavity will prefer TEM00 mode by a process of depleting higher modes through preferential Fresnel Diffraction of the higher modes since higher order spatial modes also take up larger spatial cross-section.
The active material for the lasing medium needs to be carefully chosen for narrow line width, high gain and a useful transition frequency. The most common lines used are summarized below:
The ends of the laser rod can be finnished in a number of ways. They can have Brewster's Angled ends which polarizes the beam. They can have AR coated ends or if the ends are ground perfectly parallel, the rod itself can be an additional resonator in the cavity helping to creat a Single Longitudinal Mode.
- Note: Nd:Glass has very interesting properties but the broad wavelength and low gain make holographic resonator use difficult. It is an interesting choice for an amplifier however.
The medium's stimulated emission cross section Definition: from Encylopedia of Lasers Some emission cross-sections:
- Ruby 2.5E-20 cm^2
- Q-246Nd:silicate 2.9E-20 cm^2
- Yb:YAG 2.1E-20 cm^2
- Alexandrite 1.0E-20 cm^2
- Nd:YAG 65E-20 cm^2 1.06414 R2->Y3 transition
- Nd:YLF 18E-20 cm^2
- Nd:YV04 25E-20 cm^2
- Ti:Sapphire 41E-20 cm^2
Above was grouped into two catagories.
Group A with smaller cross-sections are able to store more extractable energy due to lower gain allowing a higher population inversion (more pumping) before onset of ASE depopulation and spontanteous emission (lifetime)losses. These are best for q switch operation yielding higher output.
Group B with larger cross-sections offer higher gains and are best used in CW or Quasi-CW operations yielding more efficiency and higher average power outputs in these modes.
Of course q switch operation can be achieved in either group with passive q switching more dependent on smaller cross-sections for higher output than active q switches.
Once gain reaches the physical conditions ripe for ASE, the depopulation increases and the gain levels off. Gain can be modulated by temperature for example, cooling ruby and heating nd:yag will allow more output extraction.
Absorption cross sections, absorption spectral range, and spontaneous fluorescence lifetimes are good indicators for pump rates (pulsed etc) and absorption overlaps with regard to efficient pumping and eventual laser gain.
It is the goal of a flash lamp to provide energy at the absorption bands of the active material. For Nd:YAG the important bands are 730nm to 760nm and 790nm to 820nm. (Note: There is a strong peak at 808nm making 808nm laser diodes extremely efficient as pump sources.) For Ruby the desired pump regions are 370nm to 420nm and 520nm to 690nm.
Since the cost the the active medium is an order of magnitude higher than the flash lamp the frugal holographer will pick the active material first. Once the laser rod has been selected the length of the flash lamp should approximately match the laser rod and the flash lamp bore diameter should approximately match the laser rod for good efficiency.
Flash Lamps are defined by a few simple parameters.
- Bore ID
- Flash Length
- Gas Fill (Argon or Krypton)
- Fill Pressure (linears 400 to 2000 Torr. Helicals are typically 300 Torr for easy firing due to longer arc lengths.)
One you know these 4 parameters the electrical behavior for a flash lamp is defined. Additional parameters defining the pump efficiency are the media in between the flash lamp and the active medium. Different envelopes with different properties have been developed. Also any cooling medium will also effect the pump radiation. (Envelopes and cooling materials have been designed to adsorb unwanted frequencies and to transmit desired frequencies.)
Krypton lamps are preferred for Nd:YAG for low pump energies and Xenon is preferred for higher pump energies. (Above 2x10^5 W/cm^3)
Generally surplus lamps found with an electrode that is a sharp point(cathode) are typical CW arc lamps like Krypton and have very thin 0.5mm walls to help in thermal conduction. These lamps will easily explode if you try and use them as flashlamps. Xenon/Krypton flashlamps generally have 1mm or better fused quartz walls and have rounded or blunt style electrodes. Both Xenon and Krypton lamps can be made into arc lamps for C/W use or the pulse style and the electrode design generally gives away the designed use but not the gas used or it's Torr/Atm fill pressure.
The energy that makes a flash lamp explode is known as the explosion energy. It is important to operate at a fraction of this energy to increase lamp life. At 60% of the explosion energy the lamp will fail in about 10^2 pulses. Additionally based on cavity use etc it best to derate the lamp since some energy is re-absorbed by the cavity etc. Most load calculations are done for free air thermal conduction etc.
Ohm's Law is V=IR. Voltage equals current times resistance.
Voltage control allows charging control of the storage capacitors and therefore vary the amount of pump energy desired. Two typical methods are:
- Variac control - Main power transformer is capable with rectifying circuits of fully charging the storage capacitors. The variac autotransformer allows control and varies the input voltage to the main power transformer.
- Solid state relay control - Here the SSR can cutoff the charging once a pre-determined voltage is reached. SSR are placed on the input primary of the main power transformer.SSR leak current and therefore some charging is un-avoidable in this design. Danger exists unless extra steps are taken to ensure capacitors are not charged.
All parts of the circuit must be able to handle current requirements and circuit breakers should be used to protect device and personnel. Proper grounding of the equipment chassis, laser head, and parts should be in place. Personnel should be kept away from all circuits by the use of the double protection of insulations and the use of chassis enclosures as a secondary isolation technique. See other safety requirements concerning procedures and testing for electrical device safety standards. Additionally for operators and the use of interlocks, warning labels, etc, and the need for a designated LSO:Laser safety officer in the use of lasers.
Oil caps designed for pulse use and some SCR communtation types with high Dv/Dt allow use in pulse storage applications. Oil capacitors that have lost more than 10 percent of it's rated capacitance value should not be used. Lifetime is limited due to stress near full voltage rating.
Electrolytic caps have higher energy density than others and allow for compact designs but due to electrolyte heating and drying, they are more prone to failures. This rating is also found in it's temperature ratings and it's ripple current ratings. And generally using the highest ripple rated caps in parallel to increase the overall ripple current rating helps in reducing the charging/discharging heat and prolonging the capacitor's life. Additionally low ESR values aid in lower internal heating. If caps are also added in series, the fewer number of series caps the better and they must be equalized with enough current flow from a voltage divider.4 caps in series should be about the reasonable maximum. Also care should be taken to avoid shorting through their common exterior metal cans if clamped and are used in series. Best to have them insluated mounted. Gas/liquid venting is possible and best used in proper orientation Used caps should be avoided since shelf life and ripple use may be unknown with regard to electrolyte drying. Best to use new recent manufactured capacitors with known specification.
All caps have the ability to explode and housings should be used. Additional fire protection and extinguishing should be available.
Resistors have voltage ratings that should not be ignored when used in HV circuits and across capacitors for discharging or voltage equalization. HVX, HVW resistors can have high ratings for voltage and some of the specialty resistors power resistors can have 64kv ratings, but typical wire-wound 225watt power resistors may be limited to 4kv or less and common 1/4 watt carbon comp generally are around 250v rated. Verfiy with mfg ratings before using. One reference: 
Trigger Coils circuits
Series injection is having a low inductance secondary coil of the transformer in series with the flashlamp.
External trigger is using a typical high inductance secondary coil's one lead wrapped around the flashlamp and the flashlamp and other coil lead referenced to ground.
Trigger coils are pulse transformers and made of ferrite RF materials which has a fast response and low staturation inductance level. Additionally potted to prevent arc-overs on coil windings. Since the main storage capacitor goes through in series injection type pulse transformers, the max peak current ratings of the transformer should be observed or destruction of the pulse transformer with a loud bang and flying potting material can be injurious
The cavity must be designed to reflect the light from the flash lamp to the active material (laser rod). The more evenly the active material is illuminated the better the beam profile will be. Since holography requires a very clean Gaussian beam careful attention needs to be paid to the cavity design.
Cavities can either be highly reflecting or diffuse reflecting. Highly reflecting designs are preferred. If only one linear flash lamp is being used for the master oscillator then the cavity design should be elliptical with the lamp and the rod at the foci of the ellipse. It can be shown that the efficiency is increased by making the ratio of the major axis to the minor axis as small as possible. Just enough room for the mounting and electrical connections is used.
The cavity can be made from any heat resistant material. Aluminum, copper and stainless steel are used. Highly polished aluminum can provide a sufficiently reflecting surface for a Ruby laser but aluminum is not as efficient as a silver plated cavity for Nd:YAG.
Cavities can be polished metals but better is to coat soft metals with nickle and polish it then deposit silver or gold. In the case of silver it must be coated with an overlayer to protect the silver from the atmosphere. SiO is usedfor telescopes and works wll for cavities.
It is important that the reflective frequencies match the absorbtion frequencies of the active material.
An overview of reflectivity:
- Evaporated Al is about 90% reflective from 200nm to 1000nm.
- Polished Al is about 10% less reflective in the UV and about 3% less reflective in the near IR.
- Silver reflectivity starts at about 350nm and reaches a high reflectivity at 420nm. Evaporated is about 5% better.
- Golds reflectivity starts at 500nm reaching full reflectivity at 620nm.
- From 500nm to 800nm evaporated gold is much more reflective than polished gold.
Any element that tends to polarize the beam inside the master oscillator cavity will help to ensure that all of the cavity's energy will go to that polarization. This can be Brewster's Angle rod ends on the active medium. Placing a window at Brewster's Angle. Or placing one of the optical elements at Brewster's Angle. IE. the Q-Switch.
A Q-Switch keeps the lasing medium from lasing until it is at a population inversion by blocking one of the mirrors. The Q-Switch design sets the pulse width. Q-switches can either be passive or active.
Passive Q-Switches work by using a saturable dye. Until there is enough light to bleach the dye it remains opaque. This allows the flash lamp to store energy in the laser rod until the stored energy reaches a threshold. A passive Q-Switch represents a large insertion loss (Even when bleached clear it still adsorbs a significant portion of the light).
If the laser starts to lase in TEM00 mode it will tend to bleach out the center of the Q-Switch first. This tends to reinforce the TEM00 mode.
Cr4:YAG is the most common Passive Q-Switch used.
There have been many active Q-Switch designs. Since we are only looking for a single pulse in holography they benefits are usually not worth the extra cost. Because they are complicated and expensive we will just list the types.
- Rotating Prism
- Translational Optic
- Rotating Disk with a hole
- Pockels Cell
- Kerr Cell
The simplest output coupler is a partially reflecting mirror. The optimum reflectivity can be calculated. A mirror can be designed with any figure or reflectivity profile. A mirror design with a radial reflectivity can help to insure TEM00 mode.
A resonant reflector is made with two or more parallel plates. It only allows a flat configuration but often the additional mode selectivity is desirable.
Factors Effecting Spatial Mode
High diffraction losses caused by a aperture and small volume of the gain medium have caused researchers to look for other alternatives to allow spatial mode control without the small mode volume of a typical stable resonator with just an aperture for mode control.
Cat's eye Resonator
Published in "Proceedings of the IEEE" Apr 1965 and again in April 1972 both papers by P. W. Smith indicated the use of a cat's eye resonator which is two flat mirrors with a convex lens whose focal length is 1/2 the distance between the mirrors and is placed at the half way point with an aperture at one mirror and the laser medium at the other mirror. As you close the aperture at one mirror, it forces the mode volume to become larger through the laser medium. This arrangement basically forms a confocal resonator (actually 1/2 of one) with large mode volume in the medium. According to Li and Smith they reported 2.5 times the output power than just from an aperture alone. Back in 1965 this was done with a He-Ne laser.
Of course you must consider the thermal lensing that can happen from a solid rod. During Q-switch operation this lens can be negative during this initial pumping. Depending on pump levels this lensing can become pronounced. So the use of telescopes or convex lenses must be calculated with the lensing of the rod.
To approximate the lens value pass a He-Ne beam down your amplifier rod and measure the beam spread before and after pumping to calculate your negative lensing at the power levels you want to use before designing the resonator and optical elements.
Telescopic Mode Control
Since the diffraction losses required for a TEM00 beam require a small aperture, inventive laser designers have sought methods to increase the utilization of the active material. One way to do this is to place a telescope into the resonator. The proper choice of lenses and spacing allows one to compensate for the thermal induced lensing of the Nd:YAG rod.
Factors Effecting Longitudinal Mode
The only way to ensure that successive shots are the exact same frequency is to temperature control the YAG and any resonant reflectors as well as any etalons. This is not a problem unless you are using multiple flashes for holographic measurements like double exposure holographic interferometry.
Power Amplifier Design
General Considerations on Setup Configurations
- Amplifier isolation between stages either by using spatial filters or wide separations and tilting the rod relative to each other help reduce feedback that leads to parasitic oscillation due to ASE.
- Isolators consisting of Faraday Rotators and waveplates can also be used to provide the best isolation but are quite expensive.
- Amplifier rods also can have face tilts of the arctan(Rod_diameter/Rod_length) as a minimum recommended tilt to reduce this problem.
- Additionally consideration should be given to ground rough barrels on the rod and possible water cooling to help reduce TIR (Total Internal Reflection) both which help reduce the parasitic oscillations that can occur in ring modes around the circumference of the rod.
Single Pass Configurations
Multiple Pass Configurations
There are many crystals that exhibit non-linear properties and can be used for frequency doubling. The most common doubling scheme for holographers is 1064nm to 532nm.
The choice of crystal has to do with damage thresholds, conversion efficiency and cost. The correct choice changes with each new price quote. :-)
Filtering Out Un-Wanted Frequencies
Since the frequency doubling efficiency is not 100%, there is always some of the original laser frequency in the output. This is customarily divided off in order to get accurate power readings and to keep from fogging the film.
For small frequency doubled lasers absorption filters can be employed that absorb 1064nm and transmit 532nm. The have very low damage thresholds and are not recommended for pulse lasers.
A mirror coated to reflect 532nm at 45 degrees and coated to transmit 1064 is another way to separate the unconverted light from a frequency doubled system. Note: It is equivalent to reflect 1064nm and transmit 532nm.
It is very important to dump the un-wanted energy into a beam ump as it can be quite dangerous. During alignment of the frequency doubling crystal it can represent more than 90% of the energy!
Pulsed lasers are not to be trifled with! The beam can not only do damage to eyes and skin but can burn holes and start fires! Make sure to send all beams to a Beam Dump.
The amount of energy stored in the power supply can kill. The voltages involved can jump large gaps. The capacitors can hold energy even when the power supply is unplugged. Never operate a laser with the cover removed. The cover is an important piece of safety equipment to protect from stray light and from high voltage.
Power supplies should be designed to bleed off capacitors when unused.
Stories of Failed Safety Programs
but hopefully will serve as a reminder...
Death by HV:
Lucky with HV:
But I had eyewear:
It also happens to companies:
- Solid State Laser Testing from SAM's Laser FAQ
- Ron Michael's Archive
- Solid-State Laser Engineering by W. Koechner ISBN 3-540-65064-4