Patent Application
20120057608
The present invention relates in general to intra-cavity frequency converted lasers. The invention relates in particular to intra-cavity frequency conversion by sum frequency mixing in a common leg of a branched laser-resonator.
Intra-cavity sum-frequency mixing in a branched laser-resonator has been proposed as a means of generating wavelengths in the visible spectrum that are not available using a simple intra-cavity frequency doubling approach with common solid-state gain-media. In such a branched resonator arrangement, one solid-state gain-medium is included in one of two separate branches for generating one wavelength of laser radiation and another solid-state gain-medium is included in the other separate branch for generating a different wavelength of laser radiation. An optically nonlinear crystal is included in a common branch for sum-frequency mixing the two different wavelengths. The arrangement has the advantage that the full circulating power of each branch is available for the sum-frequency conversion process, with power being extracted from the resonator combination only as sum-frequency radiation.
By way of example, in U.S. Pat. No. 5,345,457, a branched resonator is described wherein a neodymium-doped yttrium aluminum garnet (Nd:YAG) gain-medium is included in each of the separate branches, with end-mirrors of the branches arranged such that radiation having a wavelength of 1064 nanometers (nm) is generated in one branch and radiation having a wavelength of 1318 nm is generated in the other branch, both wavelengths, of course, being characteristic of the Nd:YAG gain-medium. Sum-frequency mixing in the common branch generates radiation having a wavelength of 589 nm. Sum-frequency mixing in the common branch provides that high circulating power of both wavelengths is available for sum-frequency mixing.
A potential problem with the sum-frequency generating arrangement of the '457 patent is that the generated sum-frequency radiation can be noisy. This is because solid-sate gain-media doped with rare earth or transition metals such as neodymium (Nd), thulium (Tm), holmium (Ho), erbium (Er), ytterbium (Yb), chromium (Cr), and praseodymium (Pr) all have long excited-state lifetimes ranging from several microseconds (μs) to a few milliseconds (ms).
It was recognized in a paper “Large-amplitude Fluctuations Due to Longitudinal Mode Coupling in Diode-Pumped Intracavity-Doubled Nd:YAG Lasers” T. Baer, J. Opt. Soc. Am., 3, 9, (1175-1179), September 1986, that when doing intra-cavity frequency-conversion in lasers with such gain-media, the long excited-state lifetimes gave rise to chaotic noise fluctuations and instability in the frequency converted output because of mode-coupling effects. These chaotic fluctuations became known to practitioners of the art as “the green problem” having been described in terms of frequency-doubling 1064-nm (Near-IR) radiation to provide 533-nm (green) radiation. Frequency doubling can be considered a special case of sum-frequency mixing wherein the wavelengths of radiation being mixed are equal.
In U.S. Pat. No. 5,446,749 (with the above-mentioned T. Baer as one of three inventors), a solution to the “green problem” is described. The solution involves increasing the number of fundamental longitudinal oscillating modes to a point where there are so many that the above discussed chaotic fluctuations “average out”. Generation of at least a few tens of modes is described as being necessary for the solution to be effective. This became, and still is, a commercially successful solution. Frequency-doubled output noise than 3% RMS is routinely achieved. A drawback, however, is that in order to generate the required number of modes a long resonator (with length well in excess of 1 meter) is required. This would provide problems in a branched resonator arrangement, and move away from a current trend to make commercial lasers with a compact “footprint”.
Another solution to the green-problem that has enjoyed equal commercial success is to perform intra-cavity frequency-doubling in a traveling-wave ring-resonator operating in a single longitudinal mode. Similar if not superior output noise reduction is achieved. It is difficult however to adapt a ring-resonator to a branched operation for sum-frequency mixing different wavelengths. Further, there is a practical long-wavelength operating limit for traveling-wave ring-resonators. This is due to practical long wavelength limits of optical diodes (absorption of which increases with increasing wavelength) needed to achieve unidirectional circulation in the resonator.
In U.S. Pat. No. 7,362,783 a sum-frequency mixing approach is claimed wherein one solid state gain-medium is a four-level gain-medium and the other is a three-level or quasi three-level gain-medium. It is argued that the two different solid-state gain-media allow more efficient generation of a wider range of wavelengths than would be possible with two of one or the other type of solid-state gain-medium. Sum-frequency-mixing arrangements described and claimed include the branched-resonator intra-cavity sum-frequency mixing arrangement of the above discussed '457 patent, and other arrangements, including extra-cavity sum-frequency mixing.
It is taught in the '783 patent that the sum-frequency mixing process is less noisy than frequency-doubling but that the system may still tend to be unstable, since the two lasers are subjected to a non-linear coupling by the frequency-converting mechanism. It is taught that the nonlinear coupling effect can be reduced by decoupling the two lasers by inserting the non-linear crystal in only one of the two laser cavities while the other one is isolated from the nonlinear sum-frequency mixing crystal. This of course requires giving up the more efficient coupled branched resonator arrangement. As the efficiency of the sum-frequency depends on the power of both radiations being mixed this means that whatever nose reduction is obtained is obtained at the expense of efficiency.
It is also taught in the '783 patent that the non-linear coupling can be avoided altogether by placing the nonlinear crystal outside of both resonators. This is referred to by practitioners of the art as “extra-cavity frequency mixing”, and is an even lower efficiency arrangement, suitable only for sum-frequency mixing high-intensity pulsed radiations, which in turn requires phase control for the two resonators to ensure temporal pulse overlap in the nonlinear crystal.
There is a need for a low-noise branched-resonator arrangement for intra-cavity sum-frequency mixing that requires neither long resonators nor ring-resonators. Preferably, the approach should be compatible with compact resonators, and should be suitable for continuous-wave (CW) or pulsed, Q-switched operation.
In one aspect of the present invention, optical apparatus comprises an optically nonlinear element and first and second laser-resonators having first and second branches. The first and second laser-resonators are optically coupled such that the first branches thereof are coaxial with each other and the second branches thereof are separate from each other. The first laser-resonator includes a first gain-medium located in the second branch thereof, and the second laser-resonator includes a second gain-medium located in the second branch thereof. The first gain-medium has an excited-state lifetime greater than about 10 microseconds, and the second gain-medium has an excited-state lifetime less than about 100 nanoseconds. Means are provided for energizing the first and second gain-media such that radiation having a first wavelength circulates in the first laser-resonator and radiation having a second wavelength circulates in the second laser-resonator. The optically nonlinear element is located in the coaxial first branches of the first and second laser-resonators and arranged to sum-frequency mix the circulating first and second wavelength radiations to generate radiation having a third wavelength shorter than that of the first and second wavelengths.
In a preferred embodiment of the apparatus, the first gain-medium is a solid-state gain-medium and the second gain-medium is a surface-emitting semiconductor gain-medium. The short excited-state lifetime of the semiconductor gain-medium substantially reduces the above discussed noise and instability associated with prior-art intra-cavity frequency-converted lasers using only the longer lifetime solid-state gain-media.
In an example of the inventive laser wherein the solid-state gain-medium is Nd:YVO4 generating radiation having a wavelength of 1342 nm, and the semiconductor gain-medium generates radiation having a wavelength of 1064 nm, 593 nm CW output stabilizes to less than 1% RMS noise about 50 microseconds after turn-on. Retaining one solid-state gain-medium in the inventive apparatus provides that the inventive apparatus can operated in a Q-switched mode in addition to being operable in the CW mode.
The accompanying drawings, which are incorporated in and constitute a part of the specification, schematically illustrate a preferred embodiment of the present invention, and together with the general description given above and the detailed description of the preferred embodiment given below, serve to explain principles of the present invention.
Referring now to the drawings, wherein like components are designated by like reference numerals,
Resonators 12 and 14 are co-axial between mirror 18 and reflective coating 22 on birefringent filter 24. The two resonators can be considered as a single, compound resonator comprising the common coaxial portion with two branches. One branch 12A is between coating 22 and mirror 16 and the other branch 14A is between coating 22 and mirror structure 28.
Resonator 12, in branch 12A thereof, includes a solid-state gain-medium 32, assumed here for example, to be neodymium-doped yttrium vanadate (Nd:YVO), and optionally, a Q-switch 40. Gain-medium 32 is end-pumped by diode-laser radiation delivered through mirror 16. The diode-laser radiation preferably has a wavelength of about 808 nm. Mirror 16 is highly reflective for the 1342 nm fundamental wavelength of the Nd:YVO4 gain medium. Mirror 16 is coated to be highly reflective for 1342 nm and highly transmissive for the 808 nm.
As a result of the optical pumping of gain-medium 32, fundamental-wavelength radiation having a wavelength of about 1342 nm circulates in resonator 12 as indicated by arrowheads F1. Coating 22 is designed for maximum reflectivity at 1342 nm for radiation s-polarized with respect to the filter, i.e., perpendicular to the plane of incidence. Accordingly, the circulating 1342-nm radiation is plane-polarized in a plane perpendicular to the drawing as indicated by arrowhead PF1.
Resonator 14, in branch 14A thereof, includes a multilayer, surface-emitting, OPS gain-structure 30 of OPS-structure 26, surmounting mirror structure 28 of the OPS-structure. By way of example, gain-structure 30 can be an InGaAs/GaAs (active layers/substrate) structure having a peak-emission wavelength of about 1064 nm. Gain-structure 30 is optically pumped by diode-laser radiation preferably having a wavelength of about 830 nm, although pumping with 808-nm radiation is possible. Birefringent filter 24 is configured to select and fix a fundamental lasing wavelength from the gain-bandwidth of the OPS gain-structure. Coating 22 on birefringent filter 24 is designed to be maximally transmissive for p-polarized radiation (plane-polarized parallel to the plane of incidence of the filter) at the selected wavelength. Accordingly, when the OPS-structure is optically pumped, fundamental radiation having the fixed wavelength circulates in resonator 14 as indicated by arrowheads F2. Because of the coating design and a Brewster angle inclination of the birefringent filter, radiation F2 is plane-polarized in the plane of the drawing as indicated by arrowheads PF2.
In the coaxial portions of resonators 12 and 14, adjacent mirror 18, is an optically nonlinear crystal 34 arranged for type-2 sum-frequency mixing of 1342-nm and 1064-nm radiation to provide radiation having a wavelength of about 593 nm, which is a useful wavelength for medical laser applications. Mirror 18 is coated for maximum reflectivity at both fundamental wavelengths and the sum-frequency wavelength. Sum-frequency radiation indicated by arrowheads S is generated in a double pass of the fundamental wavelength radiations in crystal 34, but once having been generated, propagates in only one direction away from the crystal. Mirror 20 is coated for maximum reflectivity at both fundamental wavelengths and for maximum transmission at the sum-frequency wavelength. Accordingly, the sum-frequency radiation exits the coaxial portion of resonators 12 and 14, via mirror 20, as output radiation.
In an experimental evaluation of the apparatus of
The noise and instability experienced in prior-art intra-cavity sum-frequency conversion arrangements discussed above results from a strong interaction between gain (G), linear cavity (resonator losses) (α), intra-cavity (intra-resonator) intensity (I), nonlinear coupling (ε) mode-overlap (β). The interaction depends on the excited-state (fluorescence) lifetimes (τf1 and τf2) of the individual gain-media.
The nonlinear frequency conversion process is quasi instantaneous and depletes gain. It the excited-state lifetime of a gain-medium is long the (build-up) time for the gain to be replenished will be correspondingly long. If the mixed-mode gain is below threshold, the gain for an individual mode can be driven above threshold, which results in gain-switching from one mode to the other. This gain-switching may be permanently excited and may never be damped out. If the excited-state lifetime is sufficiently short, however, the buildup will be relatively rapid, and any initial gain-switching will be rapidly damped out.
There is always an interaction between the cavity lifetime (τc) of a laser-resonator (which depends on the resonator length and output coupling) and the excited-state lifetime if. This requires that for gain-media with a long τf, for example, greater than about 10 microseconds, a resonator length in excess of 1 meter is can provide for stable multi-mode operation. In the event that a compact resonator is required, for example, having length of about 50 cm or less, only a gain-medium with a short τf, for example, less than 100 nanoseconds (ns), will provide stable operation in an arbitrary number of longitudinal modes.
It would seem from the above discussed considerations that a branched resonator with two different short-lifetime gain-media, such as OPS gain-media would provide a solution for a compact coupled branched-resonator sum-frequency mixing arrangement with stable output. There are distinct disadvantages and shortcomings, however, in using two OPS gain-media to provide for the sum-frequency mixing. One disadvantage is that at wavelengths longer than about 1100 nm OPS gain-structures become less efficient and the maximum power limitation resulting from thermal roll-off is much less than at the shorter wavelengths. Another disadvantage is that Q-switched operation is not possible because of the same short excited-state lifetime that provides for stable operation. It is very important to retain a solid-state gain-medium as one of the gain-media at least because of long-wavelength efficiency.
On its face then, and assuming a system would only be as good as its weakest link, it would seem that the long-lifetime related noise problems of the solid-state gain-medium would mean that stability could only be achieved by taking one of the above-discussed prior-art measures to deal with those problems. The above discussed experimental results, however, indicated that this was not necessary. In order to understand why this occurred, it is necessary to investigate further the interactions between elements of the inventive sum-frequency arrangement. A discussion of such an investigation is set forth below.
The sum-frequency ISFG(n), where n is some arbitrary time, in a coupled two-branch resonator (where the nonlinear crystal interacts with both resonator branches) with two modes oscillating in each branch can be approximated by an equation:
ISFG(n)=I1(n)·I3(n)+I1(n)·I4(n)+I2(n)·I3(n)+I2(n)·I4(n) (1)
where I1, I2, are respectively the instantaneous intensities of the first and second modes of the first gain-medium and I3, and I4 are respectively the instantaneous values of the first and second modes of the second gain-medium. In order to compute ISFG(n) and gain as a function of time it is necessary to solve eight differential equations, more specifically, four pairs of differential equations, each pair having one element representing change in intensity with time and the other representing change of gain with time. This can be done numerically by computer, using a fourth order Runge-Kutta method. The vector of derivatives with respect to time is represented below by equation (2).
G02−(β2·y4+β34·y6+1)·y5 (2)
In the vector-elements: y0 and y2 represent I1 and I2, the time-dependent intensities of the first and second modes of the first gain-medium; y1 and y3 represent G1 and G2, the time-dependent gains of the first and second modes of the first gain-medium; y4 and y6 represent I3 and I4, the time-dependent intensities of the first and second modes of the first gain-medium; y5 and y7 represent G3 and G4, the time-dependent gain of the first and second modes of the second gain-medium. For further clarification of the vector terms consider the first and second elements. The first element represents the intensity change with time of the intensity of the first mode of the first gain-medium, i.e.,
Here, τc is the resonator round trip time (determined by the resonator optical length), which, for convenience of calculation is assumed to be the same for both resonators; α1 is the linear loss for the first resonator, essentially the same for both modes of the resonator; and ε is the coupling coefficient for the sum-frequency generation and is applied to the sum of y4 and y6 (the intensities of the two modes from the other gain-medium).
The second element of vector (2) represents the gain-change with time of the first mode of the first gain-medium, i.e.,
Here, tf1 is the excited-state lifetime of the first gain-medium (the same for each mode); G01 is the small-signal gain for that gain-medium (gain the same for each mode); the product β1·y0 is the gain saturation for the first mode of the first gain-medium; and the product β12·y2 is the cross-saturation from the second mode of the first gain-medium. It should be noted that while only the first pair of terms of vector (2) have been explained, the other pairs of terms follow a similar pattern and the explanation of those will be evident from the foregoing explanation of the first pair of terms.
Now, despite the complexity of the mathematical model presented above it can be recognized that it is the gain-change elements of matrix (2) that depend on the excited-state lifetimes on the gain-media, and that those elements actually depend on the reciprocal of the excited state lifetime. This suggests, without being limited to a particular hypothesis, that in the case of the present invention, where the OPS gain-structure has a much shorter excited-state lifetime than that of the solid-state gain-medium it will be the OPS-resonator that dominates the stability of the sum-frequency generation arrangement and not (as would be expected on general consideration) the otherwise-noisy, solid-state resonator. This is confirmed by calculations and experimental results for an example of the branched resonator of
Calculations were made by computer using a fourth order Runge-Kutta routine in MATHCAD software available from PTC Corporation of Needham, Mass. Assumptions made in the calculations are as follows. Graphs were provided by plotting 40000 equally spaced points within a time period from t1=0 to t2=90 μs. Two modes in each resonator branch are assumed to oscillate. Each resonator branch has a length of 500 millimeters (mm) giving a value for τc of 3.33 ns. Each resonator is assumed to have the same linear loss α1=α2=0.01.
Solid-sate gain-medium 34 is Nd:YVO4 (generating 1342-nm radiation), assumed to have a small signal gain (G01) of 0.12, an excited state lifetime τf1 of 90 μs, and saturation β1=1.0. The OPS gain-structure 30 (generating 1064 nm radiation) is assumed to have a small signal gain (G02) of 0.05, an excited state lifetime τf1 of 0.01 μs (10.0 ns), and saturation β2=1.0. Cross-saturation values were assumed as follows: β12=β21=0.5; and β34=β43=0.9.
The coupling coefficient (ε) for optically nonlinear crystal 34 was assumed to be 5 10−5. It was assumed that the optically nonlinear crystal was phase-matched only for sum-frequency generation between modes of the two resonator branches and not for generation of second-harmonics of individual modes or sum-frequency mixing between modes of the same resonator branch.
This unstable behavior is to be expected from the above-discussed teachings of the prior-art, as the resonator branches are too short to provide the noise mitigation proposed in the U.S. Pat. No. 5,446,749. The U.S. Pat. No. 7,362,783 teaches sum-frequency generation in a coupled branched resonator (with two solid-state gain-media) may tend to be unstable and recommends achieving stability by inserting the non-linear crystal in only one of the laser-resonators, with the other resonator being isolated from the nonlinear crystal.
Note that the initial sum-frequency spike of the calculated performance depicted in
Recapitulating here, the present invention solves the problem of noise and in the output of an intracavity sum-frequency mixed branched coupled resonator laser having two solid-state gain-media by replacing one of the solid-state gain-media with a gain-medium having a very short excited-state lifetime. An optically pumped semiconductor gain-medium is one such gain-medium which is particularly suitable. The low-noise performance of the inventive laser is independent of the resonator length and can be achieved by a compact arrangements with resonator lengths less than about 0.5 meters. Gas laser gain-media have comparably short excited-state lifetimes but have low gain per unit length and accordingly require a long resonator to provide adequate power.
The short excited-state lifetime gain-medium is so effective in reducing above-discussed noise problems experienced in prior-art intra-cavity sum-frequency mixed solid-state lasers (which problems are due to the long excited-state lifetime characteristic of all solid-state gain-media), that one-solid state gain-medium can be retained in the inventive laser. Retaining one solid-state gain-medium is particularly important as that gain-medium can be used to generate wavelengths longer than about 1100 nm up to about 2000 nm which cannot be easily generated at the same power or efficiency with a OPS-gain structure. Solid state gain-media suitable for use in the present invention include any rare earth or transition metal doped host.
A wide-range of wavelengths shorter than 1100 nm can be generated using OPS structures. Suitable structures include, but are not limited to, InGaAsP/InP InGaAs/GaAs, AlGaAs/GaAs, InGaAsP/GaAs and InGaN/Al2O3 (active layer/substrate), which provide relatively-broad spectra of fundamental-wavelengths in ranges, respectively, of about 850 to 1100 nm; 700 to 850 nm; 620 to 700 nm; and 425 to 550 nm. There is, of course, some overlap in the ranges. This means that the inventive sum frequency laser can be configured to generate stable low-noise output at wavelengths from about 300 nm or less up to about 830 nm.
Another advantage of retaining a solid-state gain-medium is that it allows the inventive laser to be operated in a Q-switched pulse mode, by locating a Q-switch in the separate solid-state branch of the laser, as indicated by Q-switch 40 in resonator branch 12A of
In summary, the present invention is described in terms of a preferred and other embodiments. The invention is not limited, however to the embodiments described and depicted herein. Rather the invention is limited only to the claims appended hereto.
