Abstract
A wavelength conversion device includes a second-order nonlinear optical medium with a polarization inversion structure, wherein the wavelength conversion device performs wavelength conversion between three wavelengths according to a relationship of 1/λ.sub.1=1/λ.sub.2+1/λ.sub.3, a polarization inversion period Λ of the polarization inversion structure is divided into 2a regions, and when the 2a regions divided from the polarization inversion period Λ each has a width ratio of an inverted region and a non-inverted region of r to 1−r (where 0≤r≤1), a ratio value r is set such that, when one period in phase of a sine function from 0 to 2π is divided into 2a regions, a value of the sine function in a center of each divided region is (1−2r)±0.1.
Claims
1. A wavelength conversion device comprising a second-order nonlinear optical medium with a polarization inversion structure, wherein in a nonlinear optical process in which wavelength conversion is performed through an interaction in which three wavelengths of λ.sub.1, λ.sub.2, and λ.sub.3 have a relationship of 1/λ.sub.1=1/λ.sub.2+1/λ.sub.3, the polarization inversion structure has a polarization inversion period A which satisfies n.sub.1/λ.sub.1−n.sub.2/λ.sub.2−n.sub.3/λ.sub.3−1/Λ=0 where n.sub.1 is a refractive index at μ.sub.1, n.sub.2 is a refractive index at λ.sub.2, and n.sub.3 is a refractive index at λ.sub.3 in the second-order nonlinear optical medium, the polarization inversion period Λ is divided into 2a regions (where a is an integer of 2 or more), and when the 2a regions divided from the polarization inversion period A each has a width ratio of an inverted region and a non-inverted region of r to 1−r (where 0≤r≤1), a width ratio value r is set such that, when one period in phase of a sine function from 0 to 2π is divided into 2a regions in correspondence with the division of A into 2a regions, a value of the sine function in a center of each divided region is (1−2r)±0.1.
2. The wavelength conversion device according to claim 1, wherein the second-order nonlinear optical medium has a waveguide structure.
3. The wavelength conversion device according to claim 2, wherein the second-order nonlinear optical medium is LiNbO.sub.3, LiTaO.sub.3, or a mixed crystal thereof, and contains at least one of Mg, Zn, Sc, and In as an additive.
4. The wavelength conversion device according to claim 1, wherein the second-order nonlinear optical medium is LiNbO.sub.3, LiTaO.sub.3, or a mixed crystal thereof, and contains at least one of Mg, Zn, Sc, and In as an additive.
Description
BRIEF DESCRIPTION OF DRAWINGS
(1) FIG. 1 is a schematic diagram illustrating a polarization inversion structure of a wavelength conversion device of the present invention.
(2) FIG. 2 is a plan view of a polarization inversion structure for one period when the division number of the wavelength conversion device of the present invention is a=9.
(3) FIG. 3 is a diagram showing the normalized conversion efficiency with respect to the wavenumber of the period Λ of the polarization inversion structure in contrast to the two cases of division number a=1 and 9.
(4) FIG. 4 is a diagram in which calculation results for first and third orders of quasi-phase matching are plotted with the division number a on the horizontal axis and the normalized conversion efficiency on the vertical axis.
(5) FIGS. 5(a), 5(b), and 5(c) are plan views illustrating a comparative example, the first example of the present invention, and an enlarged view of the first example, respectively.
DESCRIPTION OF EMBODIMENTS
(6) FIG. 1 illustrates a schematic diagram for explaining a polarization inversion structure of a wavelength conversion device of the present invention. The description of FIG. 1 will be given of the case where division number a=3, that is, one period Λ of the polarization inversion structure is divided into 2a=6 regions. The division number a is an integer of 2 or more.
(7) As illustrated in FIG. 1(a), an area up to half of one period of a conventional ordinary polarization inversion structure as seen in the light propagation direction (for example, in the rightward direction) has upward polarization (indicated by an upward arrow in each gray region of FIG. 1(a)) and the latter half of one period has downward polarization (indicated by a downward arrow in each white region of FIG. 1(a)) paired with the upward polarization, thus producing an element with a periodic structure composed of repetitions of a structure with a polarization inversion pair corresponding to a pair of upward and downward arrows as in FIG. 1(a).
(8) FIG. 1(b) is a partially enlarged view of one period from 0 to Λ of the polarization inversion pairs of FIG. 1(a). At the same time, one period of a waveform of the trigonometric function sin (sine wave) is shown superimposed in FIG. 1(b). That is, the phase of the sine wave is represented by 0 to 2π on the horizontal axis at the upper end of FIG. 1(b). This structure of one period is divided into 6 regions (2a regions with a=3) by dotted lines, and vertical dashed-dotted lines are shown at the centers in the width direction of the 6 divided regions (where the width of each region is π/3).
(9) The values (representative values) of the sine wave at positions indicated by the dashed-dotted lines are +0.5, +1.0, +0.5, −0.5, −1.0, and −0.5 from the left.
(10) The representative value of each region, which is an (average) polarization value p of the region, can be considered to be represented by a ratio value r (0≤r≤1) of polarization inversion widths in each of the six divided regions.
(11) The ratio of the widths of a pair of upward and downward polarizations in one region with a width of π/3 is 1−r:r, where r is the width of the downward polarization, and when the value of the upward polarization is +1 and the value of the downward polarization is −1, the average value of the polarizations of the region is (+1)×(1−r)+(−1)×r=1−2r.
(12) That is, the average value p of polarizations in each region is in the range of 1≤1−2r≥−1 for the width ratio value r of 0≤r≤1 and the representative values of the trigonometric function sin can be fitted to 1−2r.
(13) FIG. 1(c) illustrates one period (corresponding to 0 to Λ in actual size and 0 to 2π in FIG. 1(b)) of the polarization inversion structure of the wavelength conversion device of the present invention obtained by constructing polarization structures of the divided regions with such ratio values r.
(14) A width ratio value r of the leftmost divided region in FIG. 1(c) (corresponding to 0 to π/3 in FIG. 1(b)) which satisfies p=1−2r for the representative value p=0.5 of the sine wave value is 0.25. That is, the ratio 1−r:r of the widths of the polarization inversion pair in the leftmost divided region in the bottom view, FIG. 1(c), is 0.75:0.25, that is, 3:1.
(15) A width ratio value r of the second divided region from the left in FIG. 1(c) (corresponding to π/3 to 2π/3 in FIG. 1(b)) which satisfies p=1−2r for the representative value p=1.0 of the sine wave value is 0 and the ratio of polarization inversion widths is 1:0. That is, the second divided region from the left in FIG. 1(c) is entirely composed of upward polarizations.
(16) The width ratio values of the polarization inversions of the divided regions are calculated in the same way to provide a plurality of pairs of inversion structures in a polarization inversion structure of one period (0 to Λ in actual size), thus forming a polarization inversion structure of the wavelength conversion device of the present invention illustrated in the bottom view, FIG. 1(c). In the wavelength conversion device of the present invention, such polarization inversion structures are formed over a plurality of periods. Here, a downward region is placed on the right side of each divided region, but each downward region may be placed at any position of the region, such as on the left side or in the center of the region, as long as the positions (phases) of the downward regions in the regions are aligned.
(17) In the wavelength conversion device of the present invention, polarization structures of such patterns to which one period of the sine wave is fitted are formed over a plurality of periods.
(18) FIG. 2 is an enlarged plan view of a polarization inversion structure of one period of a wavelength conversion device of the present invention when division number a=9. A polarization structure, whose width ratio is adjusted such that 18 (=2a) polarization values of one period of a sine wave are generated in 18 divided regions, is formed.
(19) FIG. 3 is a graph with the horizontal axis representing an amount obtained by normalizing the amount of phase mismatch Δk shown at the top of FIG. 3 by 2π/Λ and the vertical axis representing a normalized conversion efficiency (the energy conversion efficiency of the optical conversion device) corresponding to the corresponding normalized amount of phase mismatch. The polarization inversion regions are those of 1000 periods.
(20) The horizontal axis of FIG. 3 matches the order of quasi-phase matching, 1 on the horizontal axis corresponds to the efficiency of conversion into the original target conversion wavelength, and the other integer values correspond to the conversion efficiencies of parasitic wavelengths. The case of 2×1 divisions with division number a=1 is shown by a solid line graph and the case of 2×9 divisions with division number a=9 is shown by a dotted line graph superimposed thereon.
(21) In the case of 2×1 divisions (of the solid line graph), the normalized conversion efficiency at the target conversion wavelength of 1 on the horizontal axis is 1, while significant parasitic light is generated at 3 and 5 on the horizontal axis. On the other hand, in the case of 2×9 divisions (of the dotted line graph), it can be seen that the normalized conversion efficiency at the target conversion wavelength of 1 on the horizontal axis drops to about 0.6, while no significant parasitic light is generated at 3 and 5 on the horizontal axis.
(22) In consideration of these, the ordinary polarization inversion structure is considered to be that of the case of division number a=1.
(23) FIG. 4 is a diagram in which calculation results for the first and third orders are plotted with the division number a on the horizontal axis and the normalized conversion efficiency on the vertical axis. The calculation results are those when 1 to 9 are taken as the division number a on the horizontal axis in FIG. 4 and the polarization inversion period Λ is repeated 1000 periods. Specifically, FIG. 4 is a diagram in which normalized conversion efficiencies (.diamond-solid. and .square-solid.) of the first and third orders for a=1 to 9 and calculation results for the case of a=9 with errors of −0.1 or +0 allowed for the third order (.box-tangle-solidup. and .circle-solid.) are plotted as normalized conversion efficiencies on the vertical axis. (Note that the vertical axis in FIG. 4 is a logarithmic scale).
(24) Calculations were performed for the five cases of division number a on the horizontal axis in FIG. 4, a=1, 3, 4, 5, and 9. The case of a=1 is the same as that of an ordinary polarization inversion structure as described above. It can be seen that the first-order normalized conversion efficiency (plotted with .diamond-solid.) decreases to about 60% as the division number a increases as is also shown in FIG. 3, while the conversion efficiency of the third order (plotted with .square-solid.) further decreases exponentially as the division number a increases. (Note again that the vertical axis in FIG. 4 is a logarithmic scale).
(25) When a=3 or more in FIG. 4, the third-order conversion efficiency (.square-solid.) is one digit or more lower than the first-order normalized conversion efficiency (.diamond-solid.) and it is considered that wavelength conversion due to the third-order quasi-phase matching does not affect the decrease in energy of input light. From the viewpoint of safety, it is recommended to take a=4 or more.
(26) Incidentally, in the case of 18 divisions with a=9, the representative values of the sine wave of the 4th and 6th divided regions are about +0.94 (that is, at division locations before and after the value of 1 of the sine wave), and those of the 13th and 15th divided regions are about −0.94, while the third-order value represented by a point .square-solid. for a=9 in FIG. 4 rises a little as a result of calculation performed with errors allowed with the value rounded to ±0.9. The increase in efficiency at a=9 is considered to be due to the rounding error of the value.
(27) To obtain a reference for how much error is allowed, results calculated for sine values of 0.8 and 1.0 obtained by intentionally adding errors to the sine value 0.94 in the case of a=9 are shown by points .box-tangle-solidup. and .circle-solid. for a=9 in FIG. 4 as “THIRD ORDER−0.1” and “THIRD ORDER+0.1,” respectively. It can be seen that the third-order conversion efficiency is sufficiently low even if the error value changes by about ±0.1.
First Example
(28) FIG. 5 is a plan view of wavelength conversion devices formed of lithium niobate crystals having two types of polarization inversion structures prepared in connection with the first example of the present invention.
(29) FIG. 5(a) is a plan view of a wavelength conversion device with an ordinary periodic polarization inversion structure, in which the period of a single pattern polarization inversion structure is Λ==28.48 μm, according to a conventional example as a comparative example.
(30) FIG. 5(b) is a plan view of a wavelength conversion device with a periodic polarization inversion structure of a lithium niobate crystal having an inversion pattern, in which the period Λ=28.48 μm of the polarization inversion structure is divided into 18 regions (with division number a=9), according to the first example of the present invention. The method of the present invention involves an inversion pattern in which the polarization ratio value r of each divided region is adjusted such that the average polarization values of the divided regions become numerical values to which the sine function is fitted. FIG. 5(c) is an enlarged view of FIG. 5(b) for one period (similar to FIG. 2).
(31) The lengths of the two elements, that is, the wavelength conversion device with the single pattern polarization inversion structure according to the comparative example of FIG. 5(a) and the wavelength conversion device with the polarization inversion structure having patterns to which the sine wave is fitted according to the first example of FIG. 5(b), were 10 mm and 12 mm, respectively.
(32) To each element, 0.98 μm light was collimated with 200 mW and 1.47 μm light was collimated with 20 mW, each with a lens, and the collimated light beams were combined with a dichroic mirror and focused again on the center of each element with the lens, such that light of the two wavelengths was passed through each element under the same conditions.
(33) When optical power emitted from the output side was examined, 2.94 μm light (converted light) of 3 μW was confirmed for the single pattern element of FIG. 5(a) according to the comparative example. At the same time, 0.59 μm light (parasitic light) of 5 mW was observed.
(34) On the other hand, 2.94 μm light (converted light) of 3 μW was confirmed for the element with patterns to which a sine wave was fitted according to the first example of the present invention of in FIG. 5(b) as in the comparative example in FIG. 5(a). At the same time, the intensity of 0.59 μm light (parasitic light) was also examined, showing a reduction to a low value of 0.5 mW, which was 1/10 of that of the comparative example.
(35) Theoretically, the conversion efficiency of the wavelength conversion device having a polarization inversion structure with patterns to which a sine wave is fitted according to the first example drops to about 60% of the conversion efficiency of the ordinary single period element of FIG. 5(a) as shown in FIGS. 3 and 4. However, the length of the wavelength conversion device of FIG. 5(b) is 12 mm, which is 1.2 times that of the element of the comparative example of FIG. 5(a), and thus the conversion efficiency is 1.44 times, the square of 1.2 times. That is, the efficiency of the wavelength conversion device of the first example in FIG. 5(b) which imitates a sine wave is 0.9 times (≈0.6×1.44) that of the conventional example having a single pattern period and it is still expected that the output of the converted light is slightly smaller.
(36) Moreover, because the source light is focused by a lens, it is expected that as the length increases, the density of light will decrease at both ends and the efficiency will decrease. However, it is considered that the reason why the first example of FIG. 5(b) here has acquired mid-infrared converted light similar in intensity to that of the comparative example of FIG. 5(a) is that it can limit a reduction in optical power of input light due to generation of parasitic light of 0.59 μm and can perform wavelength conversion without weakening the source light within a practical range.
(37) Although an example of wavelength conversion using a bulk crystal of lithium niobate was shown in the first example above, the second-order nonlinear optical crystal material is not limited to lithium niobate (LiNbO.sub.3) and may be lithium tantalate (LiTaO.sub.3), a mixed crystal thereof, or other second-order nonlinear optical crystal materials. A small amount of additive selected from Mg, Zn, Sc, and In may also be added. In addition, although DFG generation of bulk crystals was shown in the first example, the same effects can be achieved using a waveguide type element.
INDUSTRIAL APPLICABILITY
(38) As described above, the present invention can provide a wavelength device element which can limit unintended wavelength conversion due to high-order quasi-phase matching and performs wavelength conversion without weakening source light within a practical range for originally intended wavelength conversion.
