RU2110875C1 - Semiconductor optical amplifier - Google Patents

Abstract

FIELD: quantum electronics, high- power single-mode and/or single-frequency high-coherent radiation sources used in systems of transmission of energy and information over long distances, design of medical equipment, laser technological equipment. SUBSTANCE: in semiconductor optical amplifier lead-out is distributed over surface of amplification region in specified order and each lead-out means is made in the form of recess of definite depth with reflectors on its faces and region transparent for brought-out radiation that provide both for withdrawal of portion of amplified radiation and for passage of its other portion for further amplification in next amplification region ( cell ) pumped with injection current. This leads to increase of effective amplification length for output signal of semiconductor operational amplifier and as outcome of this power of output radiation is considerably raised, pattern of its radiation is narrowed and multiplicity of output amplified signals from single optical radiation across input of semiconductor optical amplifier is provided. EFFECT: increased functional characteristics of amplifier. 13 cl, 7 dwg

Description

 The invention relates to quantum electronic technology, namely to a high-power single-mode and / or single-frequency highly coherent radiation source, which are used to pump solid-state and fiber lasers, to create laser radiation sources in the visible region of the spectrum (red, green and blue radiation) by generating a second harmonics in nonlinear optical crystals, are used in systems for transmitting energy and information over long distances, as well as in the creation of medical equipment, laser techno logical equipment.

 Semiconductor optical amplifier (POU) is also one of the most important elements of modern fiber-optic communication systems and information transfer. The input of the POC are optical signals, most often laser radiation.

 Various types of optical amplifiers are known: fiber-optic amplifiers pumped by semiconductor lasers [1] and POEs, in which the amplification of the input optical signals occurs directly in the active region of laser heterostructures.

 Fiber-optic amplifiers have a number of significant advantages and are used today both throughout the world and in domestic fiber-optic communication systems. However, POC due to its features: small dimensions, potentially lower cost, the ability to integrate them into optoelectronic circuits, certainly have great prospects for their use in building complex, in particular, branched communication networks.

 Currently known POAs have designs containing one optical input and one optical output after amplification of a weak input signal [2]. At the same time, for simple and effective circuitry solutions for complex branched communication networks, it would be desirable to have a POE that would have several optical outputs of an amplified input signal at one optical input. However, the values of the output signals of the known POEs are not sufficient to solve such problems.

 However, POCs with increased output radiation power due to the formation of an expandable gain region are currently known [3, 4]. At the same time, this design of the POC has certain drawbacks - difficulties in introducing radiation simultaneously into several fibers to create branched communication networks and insufficient output power in the POC with a large number of splitters.

 The closest in technical essence to the invention is a semiconductor optical amplifier, comprising a multilayer heterostructured active layer placed on a semiconductor substrate, containing an active amplification region, ohmic contacts, radiation output means with coatings and means for suppressing spurious emissions.

 The main design features of the known POA can be explained using FIG. one.

The POC is made in a semiconductor multilayer heterostructure 1, placed on the substrate 2, and consists of a strip active amplification region (PAO) 3, ending with a means 4 - output of amplified radiation by the cleaved face of the heterostructure 1. For this purpose, multilayer dielectric antireflection coatings are applied for 5 s deep enlightenment R≈5 • 10 ^-4 . The output force aperture is 4-5 microns. In addition, the design of the POU provides means for suppressing spurious reflections and re-reflections of the output signal, which can disrupt the single-mode mode of its operation. They are in the well-known POU [4] are the same antireflection coatings 5.

 The technical task of the present invention is to increase the effective length of the radiation gain in the optical amplifier when the radiation exits through the surface, which provides a significant increase in the output radiation power and its density for various operating modes, as well as narrowing and the ability to control the radiation pattern of the total radiation in the far field due to phased addition diffraction-limited output single-mode and / or single-frequency radiation, reduction of astigmatism of the aggregate over POC emission areas and increase heat removal efficiency.

 A semiconductor optical amplifier is proposed in which the active gain region is formed of at least two gain cells constituting at least one gain line, the cells are connected by radiation output means made in the form of additionally introduced recesses with a reflector and a region transparent to the output radiation. Moreover, the recess is located on the surface side of the heterostructure, the reflector is placed on the inclined surface of the recess frontal with respect to the input of the optical amplifier. In this case, the angle ψ was introduced, formed by the direction of the notch reflector rib on the surface of the heterostructure with the direction of the lateral sides of the strip gain region, selected in the range.

π / 2 - arcsin 1 / n <ψ <π / 2 + arcsin 1 / n,
where n is the refractive index of the region transparent to the output radiation, and the angle φ is introduced, formed by the normal mentally drawn in the plane of the active layer to the line of intersection of its plane with the plane of the recess reflector with a normal to the surface of the recess reflector, selected in the range
1/2 arcsin 1 / n <ψ <π / 2 - arcsin 1 / n.
The bottom of the recess in relation to the surface of the heterostructure is placed at a distance specified by the energy flux P _{in of} amplified radiation propagated during operation of the device, which is defined in the cross section of the heterostructure normal to its layers, at the beginning of the gain cell, in the region transparent for outputting radiation limited to the bottom recess, as well as specified by the total gain in the specified cell, depending on the given pump current, the length of the specified cell and the design of the heterostructure. In this case, the energy flux P _in selected in the range of 0.95 ... 0.001 of the total energy flux of the amplified radiation at the end of the previous amplification region, and the total gain in this cell is inversely proportional to the energy flux P _in . Further, in the area transparent to the output radiation and located along the propagation during the operation of the device of the radiation extraction reflected from the reflector, radiation output surfaces are introduced at least on one side adjacent to the external output surface, and the means for suppressing spurious emissions are made in at least a heterostructure in the form of an area located after the final excavation.

 The active amplification region can be made in various configurations. For example, the active amplification region can be made of width f along its entire length.

 To maintain a single-mode and / or single-frequency mode of operation and to obtain diffraction limited radiation: the first cell from the input of the active gain region is made expandable from the input of width b to width f, and the remaining cells are made of width f; the initial part from the entrance of the first cell is made of width b, and the continued part of the first cell is made expandable from width b to width f, the remaining cells are made of width f.

 Various options have been proposed to reduce losses during radiation removal after reflection from the recess reflectors and to obtain different output directions.

 When the substrate is opaque to the output radiation in a multilayer heterostructure, the layer between the substrate and the adjacent emitter has a semiconductor layer having a band gap (eV) exceeding the ratio of 1.24 to the wavelength (μm) of the generation of laser radiation propagated during operation of the device, and a thickness in the range of 5 ... 100 μm, and the surface of the radiation output is placed in the led layer.

 When radiation is output normal to the layers of the heterostructure in the case of a flat radiation output surface parallel to the layers of the heterostructure, the angle ψ is chosen equal to π / 2, and the angle φ is chosen to be π / 4.

 To reduce losses in the previous case, in the multilayer heterostructure, the layer between the emitter and the semiconductor layer adjacent to it by the semiconductor layer external to the radiation output side has a semiconductor layer with an optical thickness equal to a quarter of the radiation wavelength of the master laser propagating during operation of the device and with a refractive index equal to the square root of the product of refractive indices for the emitter layers and the adjacent semiconductor layer.

To output radiation along the normal to the plane of the layers of the heterostructure, in the particular case along the normal to the surface of the substrate, and excluding re-reflections in the case of a flat surface of the radiation output, inclined at an angle ε to the external output surface and intersecting it along a line parallel to the notch edge, the angle ψ for which it is chosen equal to π / 2:
a) when choosing the angle φ less than π / 4, the angle ε is given by
η • sin {ε - [(π / 2) -2φ]} = sinε,
where n is the refractive index of the region transparent to the output radiation;
b) when choosing the angle φ large π / 4, the angle ε is given by
n • sin {ε- [2φ-π / 2)]} = sinε,
where n is the refractive index of the region transparent to the output radiation.

 In the following case, in order to reduce spurious reflections and rereflections, it is proposed along the perimeter of the optical amplifier to suppress spurious emissions in the form of grooves with a depth not less than the depth of the heterostructure layers, from the side boundaries of the active radiation regions. Moreover, the sides of the grooves closest to them are placed at angles of total internal reflection to the preferred direction of suppressed spurious emissions and at distances at which during the operation of the device a lateral decrease in the radiation intensity is ensured to a value of no more than 0.1 of its maximum value in the corresponding active cross section area. Moreover, radiation absorbing material was introduced into the grooves, and it was proposed to select a semiconductor material having a band gap (eV) of not more than 1.24 ratio to the wavelength (μm) of radiation introduced during operation of the device. In this case, the width of the grooves is chosen at least at least three times the inverse value of the absorption coefficient of the semiconductor material for the specified radiation wavelength.

 To achieve a higher density of the radiation regions in the optical amplifier, various options are formed with phased addition to achieve a significant narrowing of the radiation pattern of the total radiation and the power density (higher than the total) in the far field. In this case, the strip gain region is made of at least two successively arranged gain bars placed at a given angle to each other, and at the point of rotation, the rulers are bounded by a reflector plane that intersects at least the heterostructure layers along the normal and also bounds the outer sides of the strip region reinforcing said rulers in places formed by the intersection of mentally continued inner sides of the strip reinforcement region with its outer sides, and and suppression of spurious emissions is further arranged at least along a portion of the sides of amplification lines.

 Moreover, it is proposed for phasing radiation in any gain line, at least on one gain cell, to form an autonomous ohmic contact.

 The essence of the present invention is the creation in the integrated performance of a new and original design of an amplifier such as a traveling wave with surface radiation, in which areas with one spectral composition of the output radiation are distributed in a predetermined manner within a linear or two-dimensional area of the luminous body. For the first time in an unobvious way. a part of the amplified radiation was output from the optical amplifier using the proposed radiation output means, made in the form of recesses with frontally placed reflectors and regions transparent to the output radiation. Moreover, these means ensure the passage of the remaining part (usually smaller) of the same amplified radiation to the subsequent amplification section, for which this remaining part of the radiation is an input signal (analogue of the signal from the master external laser for the POC). The latter in this amplification section is again amplified, possibly to the level of saturation, and is again output through the next similar radiation output means. So the process of alternately amplifying and outputting radiation can be repeated the required number of times.

 Consequently, a significant increase in the effective length of the radiation gain in the COI has been realized by providing periodic discharge of a part of the radiation and amplification of the remaining part, the proposal being new, has an inventive step and is industrially applicable.

 It is clear that in order to ensure that the output of the diffraction-limited single-mode and / or single-frequency coherent radiation is output throughout the amplification region, it is necessary to exclude the possibility of self-excited generation. The design of the recesses proposed for this purpose, their location, means of suppressing spurious radiation, means of scattering radiation, and the implementation of areas transparent to the emitted radiation made it possible to create various modifications of the proposed POE that make up a single general inventive concept.

 All this provides a significant increase in the output power, its density, makes it possible to achieve a narrowing of the radiation pattern of the total radiation, as well as its control due to the phased addition of diffraction-limited single-mode and / or single-frequency output radiation and the reduction of astigmatism from the radiation region of the POC with the possibility of radiation output from practically all the surface of the optical amplifier, the removal of large amounts of heat from the active amplification region and the relative technological simplicity of its manufacture.

 This also makes it possible to significantly increase the number of optical outputs from the POC, and if the back reflections of optically amplified signals to the active amplification layers are reduced, when they are removed from the POC, they can significantly increase the gain of the POC, in addition, this makes it possible to simplify the input of output radiation into the optical fiber and significantly reduce costs when using the proposed POU in complex branched fiber-optic communication systems.

 In FIG. 1 schematically shows the proposed POU, top view; in FIG. 2 - known POU, top view; in FIG. 3 is a longitudinal section AA in FIG. 2; in FIG. 4-6 are fragments of a longitudinal section AA in FIG. 2; in FIG. 7 - the proposed POU with a two-dimensional integrated amplification region in the form of a "snake".

The proposed device POU (Fig. 2 and 3) consists of gain cells 6, the first of which is composed of a strip section, called by us the preamplifier 7, and an expandable section, called by us the expandable amplification region (ROW) 8. The gain cells 6 are interconnected by means 4 output radiation. The input of the first gain cell 6 is the input of the PAOU 3. The width of its strip section is b, DOC 8 is expanded from width b to width f. Subsequent cells 6 of amplification f comprise PAOU 3. The means 4 for outputting most of the amplified radiation include recesses 9, on the front faces of which, with respect to the input of the POU, reflectors 10 are placed, as well as regions 11 that are transparent to the output radiation (indicated in dashed lines in Fig. 3 ) The recesses 9 have a depth x ₀ , which is determined in the case under consideration from the surface of the heterostructure 1 (in the general case, from the surface opposite to the radiation output surface). The recesses 9 limit on each side each gain cell 6, the sequence of which forms a gain line 12. The regions 13 and 14, suppressing spurious radiation, absorbing or scattering with further absorption spontaneous spurious reflections and re-reflections are made. They are located immediately after the last (along the amplification) recess 9 (end suppression regions 13) and on both sides of the entire amplification region of the optical amplifier with preamplifier 7 (lateral suppression regions 14) at a distance at which the lateral decay of the radiation intensity from the active amplification regions reaches no more than 0.1 of its maximum value in the corresponding cross section of the active region.

 In FIG. Figure 3 shows that the heterostructure 1 consists of an active layer 15, two emitters 16 and 17 surrounding it, a contact semiconductor layer 18 placed on the emitter from the opposite side to the substrate 2, and ohmic contacts 19 and 20, where ohmic contact 19 is an ohmic contact to the entire active amplification region, including all its parts, namely, preamplifier 7, ROW 9 and PAOU 3. In some cases, it is possible to autonomously perform part of ohmic contacts. In the substrate 2, contact 20 is made.

 The active layer 15 in real heterostructures 1, in particular with strained quantum-well sublayers [5], can include several quantum-well active sublayers with barrier sublayers separating them and two waveguide sublayers bordering emitters 16 and 17, respectively, but this will not play for the present invention fundamental role.

 The edges of the recess 9 can be coated with reinforcing protective coatings 21, and the frontal one with respect to the propagated amplified radiation, with a highly reflective coating 22 (Figs. 3-6).

To carry out the output of radiation through regions 11 transparent to the output radiation, namely the heterostructure layers 1 and the substrate 2, located along the path of the reflected amplified radiation from the reflector 10 of the recess 9, the radiation output surfaces 23 are made in the substrate 2 and are free of ohmic contact 20 and coated with multilayer dielectric antireflection coatings 5
In FIG. 4-6, fragments of sections show various options for the placement of reflectors 10 of recesses 9 and output surfaces 23 made either on the surface of the substrate 2 (Figs. 4 and 5) or in the recesses in the substrate 2 (Fig. 6). Moreover, the output surfaces 23 can be placed at different angles ε to the planes of the layers of the heterostructure 1 depending on the inclination of the reflector 10, i.e. from the value of the angle ψ (formed by the direction of the edge of the reflector 10 of the recess 9 on the surface of the heterostructure 1 with the direction of the sides of the PAOU 3) and the angle φ (formed by the normal mentally drawn in the plane of the active layer 15 to the line of intersection of its plane with the plane of the reflector 10 of the recess 9 with the normal to the surface of the reflector 10 of the recess 9).

 In FIG. 5 shows an additionally introduced antireflection layer 24 placed between the emitter 17 and the substrate 2.

 In FIG. 7 depicts rotary reflectors 25 and rotary gain cells 26 used to create a two-dimensional integrated PAOU 3 in the proposed POC.

 The proposed device operates as follows.

The input radiation (Fig. 2) through the input of the first amplification cell is introduced along its strip part in ROW 8, in which diffractionly diverging radiation is amplified. The magnitude of the amplified radiation can be calculated using known methods for a specific heterostructure 1 in a known manner with knowledge of the magnitude of the current applied to the DOC 8, as well as the length of the DOC 8. The radiation amplified to a given value P ₀ , falling on the front inclined mirror face of the recess 9 - reflector 10, is reflected from it and removed from the active layer 15. (In this particular case, shown in Fig. 2 at angles ψ = π / 2 and φ = π / 4, the rays are reflected at right angles to the substrate 2). However, since the recess 9 of depth x _{0 is} specially designed and formed, not all radiation is removed from the active layer 15, but only its part P _o . The remaining, usually the smaller part, passes directly under the bottom of the recess 9 and the radiation P _in (usually not less than 10 ^-3 ) enters the second gain cell 6 of the PAOU 3. Here, due to the pump current amplification applied to this cell 6, the radiation at the end of it again can be amplified almost to saturation (depending on the length of the gain cell 6 and the strength of the applied current), is again reflected from the reflector 10 of the next recess 9 and is output from the active region of the gain cell 6, etc.

 In FIG. 3 (a longitudinal section perpendicular to the layers of the heterostructure 1) shows a schematic distribution of the intensity I (x) of the radiation propagating along the active layer 15. The maximum of this distribution in our case, as for the commonly used symmetric heterostructures, is located in the center of the active layer 15, along the optical axis of symmetry, and the falling tails of the distribution capture emitters 16 and 17.

The calculation of the distribution of the radiation intensity I (x) for a specific heterostructure 1 for given compositions and thicknesses of all its layers does not presently present difficulties [6]. For each specific dependence I (x) obtained, the radiation power flux in the heterostructure 1 passing under the bottom of the notch through a section normal to the plane of the layers of the heterostructure 1 is
kP _o = K∫ _o I (x) dx = ∫ _x I (x) dx,
Where
P ₀ is the total flux of radiation power through the heterostructure 1.

With a chosen value of k equal to, for example, 0.1, you can find x ₀ , that is, the distance by which the bottom of the notch should be distant from the optical axis of the amplifier. It was determined that, for practicing the invention, P _{in is} selected in the range (0.99 ... 0.001) P ₀ .

 The indicated range of variation k = 0.99 ... 0.001 can also be justified by the known data [7] and the estimates made.

An energy flux equal to (1-k) P ₀ = P will be removed from the heterostructure 1 in the direction of the substrate 2, minus the losses, mainly diffraction, associated with this conclusion. These losses will be negligibly small if the width of the waveguide mode, cut off by the notch reflector, which is determined by the given configuration of the heterostructure and is cut off by the recess reflector, exceeds the wavelength of the generated radiation propagated in the heterostructure. This means that for the heterostructures used, the above condition is satisfied.

The proportion of the total flux equal to kP _0, minus losses associated with the occurrence of that part of the radiation in the radiation mode (mode in the process of forming the next generation of cell) is input _Rin power flow P coming next generation cell. From the estimates made, it was determined that the main fraction is the diffraction loss associated with the mode formation process, which depends on the value of k. The magnitude of these losses varies from a few percent P ₀ to values not exceeding 0.25 P ₀ .

The fraction of amplified radiation that has passed into the next cell is captured by its waveguide and will be amplified along its entire length until the next reflector of the notch; moreover, the necessity of fulfilling the condition under which the total gain in the cell was inversely proportional to the value of P _in was confirmed experimentally.

 We draw attention to the fact that the reflected part of the output power stream should be removed from the heterostructure 1 with minimal back reflections to the active layer. Back-reflected spurious emissions can lead to self-excitation of the generation regions in amplifying regions, i.e. to violation of the single-mode (single-frequency) nature of amplified radiation. To solve this issue, various design measures can be provided, relating both to the location of the recesses, i.e. to the angle ψ, to the inclination of the reflector, i.e. to the angle φ, to the regions 11, transparent for the output radiation, and to the regions 13 and 14 of suppressing spurious emissions.

 Consider the influence of the position of the reflectors 10 and the surfaces of the radiation output 23, as well as the design of the regions 11 transparent to the output radiation, on the implementation of the amplification mode in PAOU 3 with an increased effective length of the radiation amplification, as well as on the reduction in the likelihood of self-excited generation regions in the amplifying regions.

It is determined that in order to implement the necessary conditions of reinforcement, it is necessary that the directions of the ribs of the recesses are directed either perpendicular to the direction of the side surfaces of the PAOU 3, or at a certain angle ψ ≠ 90 ^o , with a limited range of values
π / 2 - arcsin 1 / n <ψ <π / 2 + arcsin 1 / n,
Where
n is the refractive index of the region transparent to the output radiation,
and the slope of the reflectors is limited by the range of angles φ
1/2 arcsin 1 / n <ψ <π / 2 - 1 / 2arcsin 1 / n,
Where
n is the refractive index of the region transparent to the output radiation.

 It is empirically confirmed that for the range of inclination of the edges of the recess 9 and the inclination of the reflectors 10 we can realize the output of radiation from the optical amplifier, a significant reduction in the scattering of spuriously reflected radiation, which, combined with the location of the bottom of the recess 9, allows the most efficient implementation of a sufficiently effective output of a part of the radiation its introduction into the next cell 6 for further amplification.

For the case when the wavelength (μm) of the emitted radiation is λ <1.24 / E _g , where E _g (eV) is the band gap of the substrate 2 material in the heterostructure 1, the substrate 2 must be removed from at least the region transparent to the output radiation.

In this case, when growing the heterostructure 1 between the substrate 2 and the emitter 17, an additional transparent semiconductor layer with a thickness in the range of 5 ... 100 μm can be grown. It is known that the transparency condition meets the requirement
λ> 1.24 / (E _g + δ),
Where
E _g (eV) is the band gap of the substrate material 2 in the heterostructure 1, and δ should be selected at least 0.1 eV.

 The output radiation can be directed both obliquely and perpendicularly to the planes of the layers of the heterostructure 1.

 As mentioned earlier, a perpendicular emission output is provided in the case of a flat radiation output surface 23 located parallel to the layers of the heterostructure 1, and the choice of the angle ψ equal to π / 2, and the angle φ equal to π / 4. However, it is possible to increase reflections from the surfaces of the layers of the heterostructure 1 and the surface 23 of the radiation outlet, in the particular case, from the surface of the substrate 2, which is highly undesirable (Fig. 4).

где
n₁₅, n₄, n₂ - показатели преломления четверьволнового просветвляющего слоя 24, эмиттера 17 и подложки 2 соответственно.As experiments have shown, in order to minimize this effect, it is desirable to transfer the reflected radiation from the active layer 15 of the heterostructure 1, for example, between the emitter 17 and the substrate 2 in the multilayer heterostructure 1, it is desirable to incorporate an antireflection semiconductor layer 24, the optical thickness of which is equal to a quarter of the wavelength of the radiation emitted (Fig. . 5). The enlightening properties of layer 24 are determined by the fulfillment of the condition

Where
n ₁₅ , n ₄ , n ₂ are the refractive indices of the four-wave antireflection layer 24, emitter 17 and substrate 2, respectively.

 Similarly, such a layer can be introduced from the side of another emitter with the opposite output radiation.

 The decrease in the effect of rereflections and spurious emissions when the directions of the output radiation are normal to the planes of the layers of the heterostructure 1, in particular, to the substrate 2 at angles ψ = π / 2 and φ ≠ π / 4 allows the choice of a certain slope of the flat surface of the radiation output to the planes of the layers of the heterostructure 1 (Fig. 6).

It was found that for ψ <π / 4 the angle ε is given by
η • sin {ε- [π / 2) -2φ]} = sinε,
Where
η is the refractive index of the region transparent to the output radiation,
and for the angle φ> π / 4, the angle ε is given by
η • sin {ε- [2φ- (π / 2)]} = sinε,
Where
η is the refractive index of the region transparent to the output radiation.

 In the places where the rays exit the substrate 2, free of metal contact layers of the ohmic contact 20, it is desirable to apply multilayer dielectric antireflection coatings 5.

 For the case when the substrate 2 is opaque to the output radiation, the above structural solutions remain valid, but at the same time, the substrate 2 must be removed at the places where the radiation exits.

As mentioned earlier, to reduce the back-reflected spurious emissions, we performed regions 13 and 14 of their suppression. For this, at the end of the last gain cell of the PAOU 3, immediately after the reflector 10 of the final recess 9, an end region 13 for suppressing spurious emissions is created in the heterostructure 1, namely, either the absorption or scattering region, followed by the absorption of a small part of the radiation _Pin , passed under the recess 9, and also suppression of reflected and reflected radiation in order to prevent self-excitation of the amplifier.

 It is proposed from the sides and at the end of the PAOU 3, after the last recess 9 of the reinforcement cell 6, to groove to create the end and partially lateral suppression regions 13 and 14.

 It has been empirically ascertained and confirmed by calculation estimates that the suppression of spurious emissions with minimum insertion loss is achieved by arranging the grooves at distances at which the lateral decay of the intensity of the generated or amplified radiation decreases to no more than 0.1 of the maximum value of the radiation intensity in the corresponding cross section active area. Moreover, these grooves should be made at least to the depth of all layers of the heterostructure 1, up to the substrate 2. It is desirable to form the sides of the grooves adjacent to the sides of the active regions at angles of total internal reflection to predominantly spurious radiation. They can be calculated using known methods.

To increase the effect of suppressing spurious emissions, the grooves are filled with radiation-absorbing materials with certain characteristics. As determined, the best materials are semiconductor materials with a band gap E _g (eB) <1.24 / λ. In this case, a groove width of at least 3 / α is required, where α is the absorption coefficient of the selected semiconductor material for a given wavelength. Semiconductor materials can be germanium, silicon, indium arsenide, etc.

 It is known [8], [9], the introduction of areas that suppress spurious radiation: both the end region 13 and the side regions 14. However, for this design of the CSP, the location, the angles of the sides of the grooves adjacent to the active regions, and the absorbing materials used are selected. It was found that increasing the distances between the groove 13 and the end side of the PAOU 3, as well as between the grooves 14 and the sides of the PAOU 3 significantly reduces the effect of suppressing spurious emissions. The smallest distance is determined by the technological capabilities and the requirements for eliminating rereflections from the edges of the grooves into the active region. The angles of inclination of the grooves are chosen precisely so as to exclude reverse re-reflections. Previously, they can be calculated using known methods and refined during the experiment, since they depend on each specific design of the POE, i.e. specific heterostructure and its elements. From the foregoing, it follows that the entire set of essential features of the proposed POE, including these means of suppressing spurious emissions, has an inventive step and novelty.

 The modifications of the elements proposed by us and considered here allow us to optimize the design of the POC and obtain excellent parameters, both in terms of the output power and the quality of the received radiation.

 To further increase the output power, it is possible to use various variants of the topologies of the PAOU 3. In FIG. 7 shows a variant of the POC, characterized by a two-dimensional region of surface radiation, formed from gain lines 12, for example, in the form of a “snake”. There are also other options for creating topologies of two-dimensional regions that are not presented in the present description (for example, "rolled up spiral", "rolled up spiral", etc.).

In FIG. 7, the input radiation amplified in the first amplification cell 6 is alternately reflected for output from the active layer 15 and amplified again, as previously described, reaches the end of the amplification line 12. Here, due to the introduction of a rotary gain cell 26 with a rotary reflector 25 formed at an angle of 45 ^° with respect to the lateral surfaces of the gain line 12 and normal to intersecting all layers of the heterostructure 1 up to the substrate 2, the radiation changes direction due to total internal reflection on this mirror at an angle of 90 ^o in one direction or another, depending on the direction of the rotation angle of the gain line 12. Next, the process is repeated in the second gain line 12 located at right angles to the first. At the end of the second line 12, also on the rotary cell 26, a similar rotary reflector 25 is formed. At the end of the last line 12 immediately after the recess 9 for outputting radiation, for example, through the substrate 2, an end region suppressing spurious radiation is formed region 13.

 By varying the sequence and angles of rotation and the length of the gain lines 12, it is possible to form various variants of two-dimensional figures of PAOU 3 and, accordingly, surfaces with transverse radiation strips regularly located on them.

 So, for example, can be performed POW, including 41 line 12 gain, and each subsequent line 12 gain is located at right angles to the previous one. In this case, the first gain line 12 will consist of two gain cells 6; the second, fourth and all even gain bars 12, up to the fortieth, from one gain cell 6; the third and all odd gain lines 12, up to the last odd 41 gain lines 12 — out of 6 gain cells 6. In total, in addition to rotary devices, 142 gain cells 6 must be provided. At each bend, a pivot cell 26 is placed (40 in total), bounded by pivoting reflector 25. Each of these mirror reflectors 25 should be located at an angle (π / 4 + -0.01) rad with respect to the amplified beam incident on them so that the direction of the amplified radiation in each of the odd six-cell gain bars 12 from 3 to 41 changes to the opposite.

Thus, a two-dimensional close-packed luminescence body was formed for the output amplified radiation having the form of a “snake” type. The total size of the near-square glow body for this IPLU is 3.0 • 3.2 mm ² = 9.6 mm ² .

 To implement a certain architecture of PAOU and rotary reflectors, the well-developed methods of planar technology and photography, as well as recently developed technologies for manufacturing etched mirrors, should be used [10].

 Our explanations of the work and the evidence of the essentiality of the device’s distinctive features allow us to determine the main advantage of the proposed design of the POC - the possibility of obtaining relatively simple known technological methods of ultrahigh levels of output radiation power while maintaining its single-frequency and / or single-mode properties with all its unique qualities: a narrow radiation pattern determined by the diffraction divergence of individual radiating elements OS and their coherent addition in the far field, the narrow emission spectrum, the high temperature stability of the emission wavelength, the uniformity of near field radiation, high reliability.

 The high powers of the POC structures according to the invention are ensured by the fact that PAOU 3 is made in the optical amplifier with an increased effective radiation amplification length due to the introduction of separate gain cells, at the boundaries of which a part of the radiation is “dumped” from the active region and the remaining part is introduced into the next cell for further gain in it.

 For high-quality heterostructures 1 that are close to ideal, the output rays reflected from the front reflectors 10 of the recesses 9 will be phased with each other and their coherent addition in this case provides not only the high output radiation powers indicated above, but also further (relative to the divergence of each single beam) narrowing the radiation pattern of the total radiation in each of two directions (along the generation line and perpendicular to it) is proportional to the number of added rays in each from the indicated directions. However, the structures are usually not ideal and in these cases some adjustment of the working conditions of the gain cells 6 to each other is necessary. The simplest is current tuning. In this case, the desired phase change is achieved by a given introduction of the concentration of injected carriers that change the optical path length for the amplified radiation in such a section. In some cases, it is possible to adjust only one cell 6 in line 12, but it may be necessary to adjust each cell 6. In this case, to obtain phased radiation, autonomous execution of ohmic contact layers 19 to each of the gain cells 6 is required.

 In fact, in the proposed POC for a single-frequency radiation propagating in PAOU 3, the traveling wave mode is implemented and, as a result, the output rays reflected from each recess, for example, by adjusting the current, can be phased out, which cannot be obtained by simply summing the powers of a large the number of conventional amplifiers [3]. In the latter case, it is also possible to obtain high values of output powers, however, it is not possible to achieve the other above-mentioned characteristics of the output radiation of the proposed POC. Therefore, the proposed POC cannot be considered as the result of adding up well-known solutions.

 It should also be noted that since the radiation can be removed in the proposed POCs through the substrate 2, it is possible to remove large heat fluxes from the active layers 15 of the POCs, usually located only at distances of several microns from the outer surface of the heterostructure 1. This allows one to obtain large output power from the proposed POC not only in short pulses, but also large power levels in continuous and quasi-continuous modes of its operation.

 Maintaining single-frequency and / or single-mode operation modes of the proposed POC with modifications is due to the fact that all active elements of the proposed device, as described previously, operate in the amplification mode of single-mode and / or single-frequency introduced radiation, which is achieved by its distinctive features. First of all, by the specific design of the strip gain region, namely, the location with respect to the heterostructure layers of 1 recesses (as defined above), allowing not only to remove part of the radiation from the active layer, but also to introduce into each subsequent gain cell a part of amplified radiation playing such same role as input for the first cell. In addition, the amplification mode in the gain cells is supported by means of suppressing possible spurious emissions arising from reflections and re-reflections (described earlier) introduced in the design of the COD at the end of the optical amplifier, after the last gain cell, its extraction, and from the sides of the active region, and also when radiation is removed from the active layer 15 through an area 11 that is transparent to the output radiation (requirements for the tilt angles of the reflectors 10 and output surfaces 23, to the antireflection layers 24). Eliminating the penetration of reflected optical rays into the active layers of 15 gain cells 6 prevents self-excitation and the generation of multimode laser radiation in them. It should be noted that the entire combination of the above means of suppressing spurious reflections in the proposed POU designs is more effective, not only in comparison with that used in the prototype by applying a multilayer antireflective dielectric coating 5 to the output face of the cleaved mirror, but also in comparison with other known means [8 , 9].

 Currently, various devices for reflecting radiation from a notch reflector placed at an angle to the direction of propagation of amplifying radiation [11] and devices for transmitting radiation through a waveguide layer under a notch and transferring radiation to the active layer of a subsequent region are known individually [12]. However, the formal addition of known solutions does not allow to obtain the proposed invention and solve the technical problem. The achievement of the proposed became possible only with the original and non-obvious combination in one node inside the amplified medium of the proposed new means of output and input of radiation of a certain configuration and located in a certain way, which determines the inventive step of the invention.

 It should also be noted here that the manufacture of the proposed POU is based on a number of well-known and, in most cases, proven technologies. In addition to manufacturing techniques for strained quantum-dimensional heterostructures [13] and messtrip active regions [14], this also applies to the etching technology of recesses at various angles to the heterostructure [15], and the technology for manufacturing an etched mirror [10]. This all ensured the industrial applicability of the invention.

 We determined that only the entire non-obvious combination of these essential features, which has an inventive step, novelty, and industrial applicability, allowed us to solve the technical problem: increase the effective length of the radiation gain in an optical amplifier when outputting radiation through the surface, which ensured a significant increase in the output radiation power and its density for various operating modes, as well as narrowing and the ability to control the radiation pattern of the total radiation in It field due to phased addition diffraction limited output single-mode and / or single-frequency radiation, reducing astigmatism plurality of surface regions of the radiation and increasing the heat removal efficiency.

 Example 1. The proposed POC with surface radiation (Fig. 2 and 3) with a single line 12 gain was made as follows.

A heterostructure 1 based on InGaAs and AlGaAs compounds of the following composition was grown on a polished n-type GaAs substrate 2:
first emitter 17 of n-type Al _0.31 Ga _0.69 As, 2.0 microns thick;
active layer 15, consisting of the following sequence of sublayers: an undoped waveguide sublayer of Al _0.17 Ga _0.83 As, a thickness of 0.09 μm, a GaAs barrier sublayer with a thickness of 6.0 • 10 ^-3 μm, an active sublayer of In _0.2 Ga _{0, 8} As, with a thickness of 7.0 • 10 ^-3 microns, and then the barrier, active, newly barrier and undoped waveguide sublayers of the above thickness and composition;
the second emitter 16 p-type Al _0.31 Ga _0.69 As, a thickness of 1.5 μm;
The contact layer 18 of doped p-type GaAs, a thickness of 0.3 μm.

 Heterostructure 1 with layers of the specified composition provides effective amplification of the introduced laser radiation at a wavelength of λ ≈ 980 nm.

 In heterostructure 1, by the methods of planar technology and ion-chemical etching, PAOU 3 was simultaneously formed, consisting of the first cell with the initial stripe part and DOC 8, and subsequent cells of the stripe configuration in the form of a gain line 12 consisting of six gain cells 6 separated by recesses 9.

 The initial strip part of the first cell 6 was made 3.0 μm wide and 0.3 mm long, the entrance width of the DOC 8 was 3.0 μm, the expansion angle of the DOC 8 was 0.1 rad, the length of the expandable part was 1.0 mm, and maximum width of expansion is 100 microns. The gain line 12 is made of the same width of 100 μm, the length of each gain cell 6 was 0.5 mm.

 The total length of the formed PAOU 3 on a semiconductor heterostructure 1 with six gain cells 6 with one gain line 12 was 4.3 mm.

 Seven recesses 9 were etched by ion chemical etching. The first of them was placed at the end of ROW 8 and at the beginning of the second 6 cell, and the rest through 0.5 mm at the end of each gain cell 6. The recesses 9 were formed so that their edges on the surface of the heterostructure 1 were directed perpendicular to the lateral surfaces of the gain line 12 with high accuracy provided by photolithography methods. They were faceted by two faces intersecting in the depth of the heterostructure 1. The angle of inclination of the mirror-polished front faces - reflectors 10 of all the recesses 9 was maintained within π / 4 ± 0.01 rad with respect to the normal to the layers of the heterostructure 1. There are special requirements for the other opposite the faces of the recesses 9 were not presented.

To determine the depth of the grooves 9 from the solutions of the wave equations with the corresponding boundary conditions for the above thicknesses and compositions (namely, their refractive indices) of the layers of the heterostructure 1, we found the radiation intensity distribution for the zero-order mode I (x) in the direction perpendicular to the layers of the heterostructure 1 (Fig. 3). The depth x _{0 of the} location of the recesses 9 is found from the condition that the radiation flux for this mode, flowing under the recess 9, is 10% of the total radiation flux of the mode through the heterostructure 1.

From the calculations it was found that the bottom of the notch should be located at a depth x ₀ = 2.22 μm from the upper boundary of the contact layer of the heterostructure. Experimentally, the depth of the recesses 9 was obtained in the range of 2.20 ... 2.30 μm, while the bottom of the recess 9 was deepened into the emitter layer 17 adjacent to the substrate 2 by ≈ 0.33 μm from the active layer (from the calculations, the value 0, 34 μm).

 A fairly small width (≈ 1.5 μm) of the bottom of the excavation was obtained experimentally. Therefore, you can not take into account when calculating the loss of radiation when passing under the bottom of the recess 9.

At distances of 5 ± 2 μm along the entire perimeter from the formed active elements of the device, grooves 13 and 14 were etched to the depth of the entire set of heterostructure layers 1, up to substrate 2. The width of the etched grooves was 6 ± 1 μm. The angle of inclination of the faces of the etched grooves 13 and 14 adjacent to the integral elements of the device was made equal to (1.0 ± 0.1) rad. The etched grooves 13 and 14 were filled with germanium having a radiation absorption coefficient> 10 ⁴ cm ^-1 for wavelengths less than 980 ± 5 nm.

 Further, by known methods, ohmic contacts 22, 23 were created on the p- and n-sides of the semiconductor wafer of the device. The common ohmic contact 23 on the substrate 2 was deposited after it was thinned with the total thickness of the substrate 2 and heterostructure 1 equal to 100 μm.

 To output radiation through the substrate 2, the metal layers of the ohmic contact 20 located directly below the recesses 9 were removed, and instead, multilayer dielectric antireflection coatings 5 were applied, the reflection coefficient of which did not exceed 0.1%.

 Then, by scribing, the plates were separated into crystals. The overall crystal size was made equal to 1.0 • 4.5 mm.

 Next, the crystals were p-side down soldered to a metallized synthetic diamond plate with high thermal conductivity. On the n-type side of substrate 2, a thin metal frame was soldered with slots in the places of radiation output.

 A diamond plate with a crystal was mounted on a heat-cooling device. To ensure the operation of the device, the metal output from the substrate 2 was connected to the minus of the power source, and the plus of the power source was connected to the strip contacts of the optical amplifier. The power supply provided a potential difference on the heterostructure 1 in the range of 1.5-2.2 V. At the same time, the current flow through the optical amplifier was controlled in the range of 1. ..10 A. The accuracy of setting the currents here and below was no worse than ± 5%.

 The low-power input single-frequency radiation of up to 50 mW was introduced into the POC. A current of 6 A was passed through the POU.

 The values of the output power densities were measured under each recess 9 separately. The average value of the output power per one cell 6 gain, amounted to about 0.33 watts. The total output power from PAOU 3 was obtained about 2.3-2.4 watts. For the radiation power measurements indicated here and below, the measurement accuracy was no worse than 20%.

 The measured wavelength of the POU radiation was 985.4 ± 0.1 nm. The width of the spectral line of radiation in this case was less than 0.1 nm. This value was limited by the resolution of the spectrometer used for measurements. The results of these measurements testified to a single-frequency operating mode of the POC.

 The divergence of the output radiation in the direction along the length of the gain line in each beam of six recesses near the substrate 2 was in the range 0.42-0.43 rad, and in the direction transverse to the length of the gain line for each of the rays within 10-15 mrad, which indicates a single-mode operation. Here and below, the angles of radiation divergence were measured at the level of 0.5 of the maximum power value. In the divergences given here and below, the accuracy of their measurements was no better than 15%.

 Example 1.1 The design basis for the POC is described in Example 1. The difference was in the autonomous performance of ohmic contacts 19 to each of the six cells and when adjusting the current through each cell to within 30 mA. The power characteristics are similar to Example 1. The divergence of the output radiation in the direction along the length of the gain line 15 in each beam of six recesses 12 near the substrate 2 was within 60-65 mrad, and in the directions transverse to the length of the gain line 15 for each of the rays within 11 -12 mrad.

 Example 2. Having chosen the design of the POA of example 1 as the basis, we tested other options for the manufacture and placement of areas transparent to the output radiation 11, recesses 9 with reflectors 10, and output surfaces 23 of the radiation.

Example 2.1 The angle ψ of the inclination of the recess 16 was chosen equal to 3 ± 0.10 ^o , while the angle φ of the inclination of the reflectors was chosen equal to 42 ± 0.05 ^o .

Example 2.2 In the multilayer heterostructure, an antireflection semiconductor layer 30 of composition Al _0.15 Ga _0.85 As, 0.072 μm thick, was grown (Fig. 5). The refractive index of this layer is 3.486, and the angles ψ and φ correspond to the indicated angles in Example 1.

 Conclusions from examples 2.1 - 2.2.

 The results of the measured energy and spatial spectral characteristics of the POC for options 2.1 - 2.2 were very close to each other and the results shown in example 1. The main difference was in the difference in the levels of the maximum output power of the generation, at which the single-mode generation is still maintained and the diffraction divergence of the output radiation is maintained each output beam. Accordingly, for the POC of examples 2.1 and 2.2 (in accordance with Fig. 4 and 5), the measured maximum achievable powers while maintaining the spatial spectral characteristics were 3.6; 2.8 watts

In addition, if for the variant of example 2.2, in accordance with FIG. 5, the output radiation was directed perpendicular to the surface of the substrate, then for the variant of example 2.1, in accordance with FIG. 4, the output radiation was directed at an angle ε = 22 ^o . It was found that the radiation output was in the region offset from the plane of the normal cross section passing through the reflector 10.

 If possible, implement the POU depicted in FIG. 7, on the basis of the design and with the operating modes described in Example 1.1, the following results could be obtained: output power values can be of the order of 100 W with the expected divergence of radiation in the far field for the direction along the gain lines 15 of the order of 60 ... 65 mrad , and in the direction perpendicular to it - about 1.0 mrad. This would be evidence of the single-mode nature of the individual output radiation from the gain cells 6 and their phased addition in the far field.

 A comparison of the characteristics of the well-known POC [4] and the proposed one showed that the proposed POC has a number of undeniable and significant advantages.

 Currently, it is not known that the combination of radiation output and input means in one node, placed in a certain way inside the amplified medium and made in the form of a specific configuration, which would allow to realize the task in a very small-sized device.

 As a result of this, it became possible to transfer radiation through the end face (3) to a multiple number of radiation leads through the surface. At the same time, the output power of the POU radiation was achieved approximately n-times greater than in the prototype, where n is the number of gain cells of the POU; the examples show that n ≈ 100. In reality, for large sizes of IPLU it is possible to obtain n ≈ 1000 or more.

 Important and new is that since the invention implements the regime of an amplified traveling wave of single-frequency radiation when it is regularly “discharged” through the surface of POU 15, the output rays reflected from the front reflectors of the recesses, for example, when the gain cells are tuned in current, are phased between themselves and their coherent addition in the far field provides not only high output radiation powers, but also a further (relative to the divergence of each individual beam) narrowing of the radiation pattern of the total treatment in each of the two directions (along the gain line and perpendicular to it) is proportional to the number of rays added in each of these directions.

 Consequently, the entire set of features of the invention is new, having an inventive step and is industrially applicable, and makes it possible to solve the technical problem: the number of optical outputs from the POC is significantly increased and with a decrease in the back reflections of optically amplified signals into active gain layers, when they are output from the POC, a significant increasing the gain of the POU, simplified input of the output radiation into the optical fiber and significantly reduced costs when using the proposed POW in complex branched fiber optic communication systems.

Sources of information
1. Optics Communic, 1992, Vol. 87, N 1-2, pp. 15-18.

 2. US patent 5260822, cl. H 01 S 3/19, 1993.

 3. IEEE J.of Quant. Electr., Vol. 29, N 6, June 1993, pp. 2028-2032.

 4. IEEE J.of Quant. Electr., Vol. 29, N 6, June 1993, pp. 2037-2042.

 5. US patent N 4845724, CL. H 01 S 3/19, 1989.

 6. Casey H., Panish M. Lasers on heterostructures. T. 1, 1981.

 7. Weinstein A.A. Electromagnetic waves. M. - Owls. radio, 1988.

 8. IEEE J. of Quantum Electronics, 1993, Vol. 29, N 6, pp. 2052-2057.

 9. IEEE Photonics Technology Letters, August 1993, Vol. 7, N 8, pp. 899-901.

 10. Osinski J.S., Mehuys D. et al., IEEE Photonics Technology Letters, 1994, Vol. 6, N 10, pp. 1185-1187.

 11. Ellectron. Lett., 1995, Vol. 31, N 13, pp. 1056-1057.

 12. European patent 0411145. A1, H 01 S 3/18, 1990.

 13. IEEE J. of Quantum Electronics, 1993, Vol. 29, N 6, pp. 1889-1894.

 RF patent 1831213, cl. H 01 S 3/19, 1990.

 J. Electr.Mater., 1990, Vol. 19, N 5, pp. 463-469.

Claims (13)

1. A semiconductor optical amplifier, comprising an active layer multilayer heterostructure placed on a semiconductor substrate, comprising an active amplification region, ohmic contacts, coated radiation output means and spurious radiation suppressors, characterized in that the active amplification region is formed of at least two cells amplification constituting at least one gain line, the cells are connected by radiation output means, made in the form of additionally inserted recesses from and a region that is transparent to the emitted radiation, the recess being located on the side of the heterostructure surface, the reflector is placed on the oblique surface of the recess frontal with respect to the input of the optical amplifier, and the angle ψ formed by the direction of the recess reflector edge on the heterostructure surface with the lateral sides of the strip gain areas selectable in the range
π / 2 - arcsin 1 / n <ψ <π / 2 + arcsin 1 / n,
where n is the refractive index of the region transparent to the output radiation,
and also introduced the angle φ formed by the normal, mentally drawn in the plane of the active layer to the line of intersection of its plane with the plane of the recess reflector with the normal to the surface of the recess reflector, selected in the range
1 / 2arcsin 1 / n <φ <π / 2 - 1 / 2arcsin 1 / n,
the bottom of the recess in relation to the surface of the heterostructure is placed at a distance specified by the energy flux P _{in x of the} amplified radiation propagated during operation of the device, which is determined in the cross section of the heterostructure normal to its layers at the beginning of the gain cell, and also specified by the total gain in the specified cell, depending on the given pump current, the length of the indicated cell and on the design of the heterostructure, while the energy flux P _{in x is} selected in the range of 0.95 - 0.001 on the total energy flux of the amplified radiation at the end of the previous of the existing amplification region, and the total amplification in the indicated cell is chosen inversely proportional to the energy flux P _{in x} , then to the radiation output surface, at least, are introduced at least to the region transparent to the output radiation and located along the propagation of the radiation reflected from the reflector during operation of the device one side adjacent to the external output surface, and the means of suppressing spurious emissions are made in the heterostructure at least in the form of a region located after the final excavation.  2. The device according to claim 1, characterized in that the active amplification region is made of width b along its entire length.  3. The device according to claim 1, characterized in that the first cell from the input of the active gain region is made expandable from the input of width b to width f, and the remaining cells are made of width f.  4. The device according to claim 1, characterized in that the part of the amplification region starting from the input of the first cell is made of width b, and the continued part of the first cell is made expandable from width b to width f, the remaining cells are made of width f.  5. The device according to one of claims 1 to 4, characterized in that in the multilayer heterostructure between the substrate and the adjacent emitter there is a semiconductor layer having a band gap, eV, exceeding the ratio of 1.24 to the wavelength, μm, of laser generation radiation propagated during operation of the device, and a thickness in the range of 5 to 100 μm, and the surface of the radiation output is placed in the introduced layer. 6. The device according to one of claims 1 to 5, characterized in that in the case of a flat radiation output surface located parallel to the heterostructure layers, the angle ψ is chosen equal to π / 2, and the angle φ is chosen equal to π / 4.
7. The device according to one of claims 1 to 6, characterized in that in the multilayer heterostructure the layer between the emitter and the adjacent semiconductor layer external to the radiation output side is a semiconductor layer with an optical thickness equal to a quarter of the radiation wavelength introduced during operation of the device , and with a refractive index equal to the square root of the product of the refractive indices for the emitter layer and the substrate. 8. The device according to one of claims 1 to 5, characterized in that in the case of a flat radiation output surface inclined at an angle ε to the external output surface and intersecting it along a line parallel to the notch edge, the angle ψ for which is chosen to be π / 2 , and when choosing the angle φ less than π / 4, the angle ε is given by
n • sin {ε - [(π / 2) -2φ]} = sinε,
where n is the refractive index of the region transparent to the output radiation. 9. The device according to one of claims 1 to 5, characterized in that in the case of a flat radiation output surface inclined at an angle ε to the external output surface and intersecting it along a line parallel to the notch edge, the angle φ for which is chosen to be π / 2 , and when choosing the angle φ large π / 4, the angle ε is given by
n • sin {ε- [2φ-π / 2)]} = sinε,
where n is the refractive index of the region transparent to the output radiation.  10. The device according to one of claims 1 to 9, characterized in that the means for suppressing spurious radiation is made in the form of grooves with a depth not less than the depth of the heterostructure layers on the side of the lateral boundaries of the active radiation regions, while the sides of the grooves closest to them are placed at full angles internal reflection to the predominant direction of suppressed spurious emissions and at distances at which, during operation of the device, a lateral drop in radiation intensity is ensured to values of no more than 0.1, its maximum Nogo value in the corresponding cross-section of the active region.  11. The device according to claim 10, characterized in that the material absorbing radiation is introduced into the grooves.  12. The device according to claim 11, characterized in that the semiconductor material selected by the radiation-absorbing material having a band gap, eV, is not more than 1.24 to the wavelength, μm, radiation introduced during operation of the device, and the width, μm, at least three times the reciprocal of the absorption coefficient of the semiconductor material for the indicated radiation wavelength is selected.  13. The device according to one of claims 1 to 12, characterized in that the strip gain region is made of at least two successive gain bars placed at a predetermined angle to each other, and at the turning point, the bars are bounded by a reflector plane that intersects normally at least layers of heterostructure, as well as bounding external lateral sides of the strip region of amplification of said lines at places formed by the intersection of mentally extended inner sides of the strip region of force Ia with its external sides, and a means of suppressing spurious radiations further arranged at least along a portion of the sides of amplification lines.  14. The device according to one of claims 1 to 13, characterized in that an autonomous ohmic contact is formed on at least one gain cell.

Images