MIROPA is not an OPO!! It is instead an OPA (A for amplifier as opposed to O for oscillator), it has no cavity and operates with very high efficiency for a single pass of the nonlinear crystal leading to unsurpassed robustness and simplicity. The lack of cavity means that MIROPA can very easily be used with many different pump lasers or even single shot as it does not have to be synchronised to any a particular pump repetition frequency. The pump pulses are also not required to be in a regular repetitive train, so pulse bursts, varying repetition rates, etc can be used. The parametric gain is very high and the output beam and pulse parameters are directly determined by the properties of the pump. We seed the OPA in the NIR with a Gaussian mode from a tunable CW laser and using a diffraction limited pump beam results in high quality signal and idler beams/pulses. MIROPA uses a patented operating regime that enables the use of long nonlinear crystals where pulse walk-off does not limit the interaction. By this means we obtain high conversion efficiency and a wide tuning range for pump pulses from ~50fs to a few hundred fs. MIROPA can also operate with picosecond pump pulses although the conversion efficiency is reduced.
They are nice and very well known product families (we have one of each in our ANU lab!), but they operate in a very different parameter space from MIROPA. The key issue with Topas and related OPAs such as Palitra, is that they are designed for very high peak power pump pulses, typically what you get from a regenerative Ti:sapphire system: around 1mJ in 100fs but at low repetition rates (1kHz or 10s kHz). Of the available units, the performance is:
MIROPA is designed for high repetition rate high average power lasers say 1-10W average power in the MHz pulse repetition rate range (ie 10’s of nJ pulses). MIROPA achieves typically >50% conversion efficiency at the peak of its tuning response in the 3000-4000nm region.
Another issue with the kHz high power systems is that inevitably you end up using large beams and short crystals. Because the power is so high the gain is huge and any hotspots in the beam cause preferential amplification in those hot spots. Tis is exacerbated because the Rayleigh length of the beam is >>> the crystal length. This means that output beam can be badly distorted.
In MIROPA we seed with a tunable CW pump and the pump geometry is such that the crystal length ≈ Rayleigh length of the focussed pump beam. This means that the pump is a diffraction-limited Gaussian that cannot contain hot spots and hence the output beam quality can be high.
The usual “it depends” answer applies, cost and choice of pump laser being the determinants. As it uses PPLN, the upper wavelength limits is somewhere in the 4.5-5um range. In the lab with a 1041nm pump we have demonstrated 2800nm to 4650nm. The model displayed at Photonics West used a single low cost tunable seed laser which provided tuning from 3300 to 4400nm. Adding another tunable seed laser would enable access down to around 2800nm. Currently seed laser sources are not readily available to move to wavelengths below 2800nm although we are working on that issue. Conversion efficiency is also untested below 2800nm and remains to be verified.
We are also planning to test alternative pump lasers for example Ti:sapphire systems operating in the 900-950nm range such as the Coherent Chameleon II. This allows some adjustment of the tuning range because a better range of seed lasers is available. We will be developing a specification for a Ti:sapphire pumped MIROPA during 2016.
Essentially that depends on the pump laser. The Spectra Physics HighQ-2 system demonstrated at Photonics West provided about 1.5W average power, and we obtained about 50% total conversion efficiency (signal and idler combined), although some power is lost in the filters. Thus we can output a maximum average of >200mW in the MIR idler at the peak of the tuning curve (3750nm) with a peak pulse power of >14KW. Signal; output can also be provided at around 500mW of average power. By increasing the pump power more output can be obtained providing the pump pulse duration remains the same. We expect to be able to scale the power up about a decade more power and down a factor of 2 without encountering any serious issues. In this context we use a system based on a Spectra Physics Femtotrain pump in one of our laboratories. This delivers about 2.7W at 21MHz and as a result its energy per pulse is about 5x higher than the HighQ-2. This generates 300mW of average power (the conversion efficiency is not as high as HighQ-2 because Femtotrain produces 500fs rather than 200fs pulses). Nevertheless, this demonstrates that higher pulse energy is possible (we obtain 4.5x the pulse energy from the Femtotrain compared with HighQ-2).
Below is the power tuning curve for the device displayed at Photonics West. At the long wavelength end the power rolls off for two reasons: due to reducing seed power and due to absorption by CO2 in the atmosphere. At the short wavelength end there is again a reduction in seed power, but also the drop of is also characteristic of the operating principle that becomes less efficient as you move towards the degeneracy point.
If we supply MIROPA with a HighQ-2 pump in the package demonstrated at Photonics West the cost is US$160-170k depending on destination and whether the price includes distributor support for installation and commissioning by our agent. A stand-alone MIROPA costs around US$95k. Other options are under active development (e.g. built in spectrometer, power monitor, NIR output, extended tuning range, Second Harmonic, broadband output, SC output, etc) costs are yet to be determined for these.
This depends on the pump laser. With the HighQ-2, the pulses are close to transform limited and about equal to the pump pulse in duration at ≈150-200fs. We have briefly tested MIROPA with a Coherent Fidelity II (50fs pulses) and it worked well but we did not have time to FROG the pulses. Simulations indicate we should obtain around 120fs. We run a system in our a Laboratory which uses a 500fs pump pulse which generates near transform-limited pulses around 300fs in duration.
Recently we demonstrated a system for post-chirping and compressing the output pulses (see application note 2) and achieved pulses as short as 55fs or about 4 optical cycles with 80-100mW of average power using the HighQ-2 pump laser (200fs). The module for chirping and compressing the MIROPA pulses has been designed to fit within the current compact MIROPA footprint and is expected to be formally released by the end of 2016.
In principle any ultrafast laser operating close to 1.0µm (Yb or Nd laser) with near transform limited pulses and energies in the 10nJ to ≈1µJ should work. We chose the Spectra-Physics HighQ-2 as the pump for our standard package because it is a robust turnkey industrial style laser although there are several alternatives from other manufacturers which can be considered. An important point is the absence of a cavity means that there is no need to choose a pump laser that has a specific repetition rate.
Other pump laser wavelengths are possible, for example Ti:sapphire systems capable of tuning beyond 900nm should be practical and solutions for Ti:sapphire pump are under active development.
Currently it takes about 15s to tune from 3300nm to 4400nm and this is determined by the piezo-positioning stage used to move the nonlinear crystal.
Power stability is largely determined by the pump laser. With the HighQ-2 logged a 50 hour stability at <1% RMS. Recent tests show <0.1% over 12 hours. Due to the cavity-less short optical path architecture, MIROPA does not routinely need any realignment.
The control software is custom written in Labview by Hotlight Systems and the instrument is addressed via a single USB port using virtual COM ports. A self standing Labview VI that tunes the OPA can be provided upon request.